Proefschrift Koorman by Nicole Nijhuis

Networking for proteins

Uitnodiging

Networking for proteins A yeast two-hybrid and RNAi profiling approach to uncover C. elegans cell polarity regulators

Thijs Koorman

Voor het bijwonen van de openbare verdediging van het proefschrift

Networking for proteins van Thijs Koorman op woensdag 20 januari 2016 om 14:30 uur in het Academiegebouw Domplein 29 te Utrecht Na afloop van de verdediging is er een receptie bij Madeleine, Wed 3a te Utrecht

2016

Heeft u vragen? Neem a.u.b. contact op met mijn paranimfen: Vincent Portegijs v.c.portegijs@uu.nl Martijn Gloerich gloerich@stanford.edu

Networking for proteins A yeast two-hybrid and RNAi profiling approach to uncover C. elegans cell polarity regulators

Thijs Koorman

ISBN: 978-90-393-6423-9 Copyright ÂŠ by Thijs Koorman. All rights reserved. No part of this book may be reproduced in any form, by any print, microfilm or transmitted in any other means, without prior permission of the author. Picture cover by: Cover design by: Layout by: Printed by:

Shutterstock Thijs Koorman and Gildeprint Gildeprint Gildeprint

Thesis financially supported by: Utrecht University The research described in this thesis was performed at the Utrecht University within the framework of the Cancer, Stem Cells and Developmental Biology graduate school. Dit onderzoek is gesubsidieerd door NWO (Vernieuwingsimpuls Vidi)

Networking for proteins A yeast two-hybrid and RNAi profiling approach to uncover C. elegans cell polarity regulators Netwerken voor eiwitten Een yeast two-hybrid en RNAi methode om C. elegans cel polariteit regulatoren te ontdekken (met een samenvatting in het Nederlands)

Proefschrift ter verkrijging van de graad van doctor aan de universiteit Utrecht op gezag van de rector magnificus, prof.dr. G.J. van der Zwaan, ingevolge het besluit van het college van promoties in het openbaar te verdedigen op woensdag 20 januari 2016 des middags te 2.30 uur door

Thijs Koorman geboren op 29 juni 1986 te Avereest

Promotor:

Prof.dr. S.J.L. van den Heuvel

Copromotor: Dr. M. Boxem

“Nothing in life is to be feared, it is only to be understood. Now is the time to understand more, so that we may fear less”

“Eigenlijk weten wij heel veel over heel weinig”

Table of contents Chapter 1: Introduction and scope 9 Proteomics approaches in the study of cell polarity in the nematode Caenorhabditis elegans Chapter 2: A combined binary interaction and phenotypic map of C. elegans polarity proteins

Chapter 3: Identification of human protein interaction domains using an ORFeome-based yeast two-hybrid fragment library

Chapter 4: Controlled sumoylation of the Mevalonate pathway enzyme HMGS-1 regulates metabolism during aging

129

Chapter 5: General discussion and concluding remarks 165 Addendum Summary Nederlandse samenvatting Curriculum vitae and publications Dankwoord

177 179 183 188

Chapter 1

Proteomics approaches in the study of cell polarity in the nematode Caenorhabditis elegans

Thijs Koorman

Utrecht University, Department of Biology, Padualaan 8, 3584 CH Utrecht, the Netherlands

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39 10

Proteomics approaches in the study of cell polarity in the nematode Caenorhabditis elegans

Large scale biology to study cellular processes

hysical interactions between macromolecules form the basis for all processes that are essential in living organisms. Understanding which proteins interact may clarify processes in both health and pathological conditions. Hence, much effort has been put into elucidating interactions between proteins. Large scale protein-protein interaction (PPI) networks, or interactomes, provide a comprehensive view of the interactions that can take place in a cell1 (Fig 1). When combined with other â&#x20AC;&#x2DC;omicsâ&#x20AC;&#x2122; data sources, for example on gene regulation or phenotypes, interaction networks may provide a powerful scaffold for understanding biological processes and complex diseases like cancer1,2. Affinity purification followed by mass spectrometry (AP/MS) and the yeast 2-hybrid system (Y2H) are the two most commonly used techniques to identify PPIs on a large scale. Where mass spectrometry is used to capture interactions in complexes, which can either be direct or indirect, the Y2H system is an ideal system to capture direct physical interactions between proteins in both small and large scale screens3. Of the two techniques, the Y2H system is the most cost- and time-effective to execute on a large scale, as it requires only manipulation of plasmids and the growth of yeast strains. To date, Y2H-based interactome maps are available for many organisms, including yeast, E. coli, Drosophila, C. elegans, Arabidopsis, and humans4-9. AP/MS is more laborious and costly to operate on a large scale, and in particular the generation of large numbers of proteins tagged endogenously with an affinity tag is often difficult10. This technique has mostly been used on smaller and medium scale studies to find new players in certain pathways. For example, a PPI network was generated of the Hippo growth control pathway in Drosophila11. Larger scale studies with AP/MS were used to generate protein complex maps for yeast and Drosophila S2 cells12,13. Recently a consortium of research groups attempted to profile the human proteome with AP/MS which resulted in an interactome map containing 84% of all protein-coding genes14. In this section of chapter 1, I will provide an overview of high-throughput interaction mapping studies in C. elegans with a focus on the Y2H system.

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Node

Edge

Individual proteins

Biological processes

Genome Wide

Figure 1 visual representation of networks. Networks are usually visualized as a graph of nodes connected by edges. In protein networks, nodes represent proteins and edges represent physical interactions between them. In such a map, edges are non-directed, since it cannot be inferred which of the two binding partners acts upon the other. Networks can be produced on three scales, on individual genes to decipher functional interactions (left figure), on biological processes (middle) to clarify pathways and processes (a box resembles a process) or on a genome wide scale (right).

Genomic resources in C. elegans Sydney Brenner selected C. elegans as a model organism to study neuronal development in 1965[ref. 15]. Sulston and Horvitz deciphered the complete cell lineage of the 959 somatic nuclei present in hermaphroditic adults, and described that C. elegans follows a fixed and reproducible pattern of development15,16. In the years to follow, a lot of effort was put into understanding genetic traits and mutations by mapping mutant loci to genes, which required intimate knowledge of the C. elegans genome. One of the earliest genome projects for C. elegans was started in the 1980â&#x20AC;&#x2122;s in which a strategy was undertaken to generate a resource of overlapping cosmids and yeast artificial chromosomes covering the entire genome17,18. This map was nearly completed in 1989, and at that point it was decided that the entire genome of C. elegans should be sequenced19. This effort led to the first draft of the C. elegans genome sequence in 1998[ref. 20]. After this achievement the post-genomic era for C. elegans began. To date, the C. elegans genome is one of the most complete genomic sequences and annotated with an extremely high degree of accuracy, which has not been achieved for any other metazoan. All systematic and functional datasets are stored in a centrally managed C. elegans database known as Wormbase21. The next discovery that would change the scientific landscape in studying gene regulation and protein function was the finding that genes could be inhibited by injecting dsRNA, a process known as RNA-mediated interference (RNAi)22. It did not take long to discover that soaking worms in dsRNA, or feeding C. elegans bacteria that

Proteomics approaches in the study of cell polarity in the nematode Caenorhabditis elegans

produce dsRNA expressed from plasmid, led to a quick and stable gene silencing response23. The availability of the genome sequence and the advances in RNAi led to the generation of genome wide RNAi bacterial feeding libraries, which are excellent resources for large-scale functional studies24,25. Another valuable genomic resource was generated by the cloning of predicted protein-coding open reading frames (ORFs) from C. elegans cDNAs into a Gatewaysystem entry vector26. This ORFeome library allows rapid cloning of ORFs into different expression systems. For example, the ORFeome library was used to generate a full-length Y2H prey-library used for various high-throughput studies. To date approximately 12,500 C. elegans ORFs are available in the ORFeome 3.1 dataset27. The Y2H system as a method to map protein interactions The Y2H system was originally described in 1989[ref. 28]. Since then, the system has been optimized in such a way that it is applicable on a high-throughput scale29. The Y2H-system is based on the reconstitution of a modified transcription factor, such as Gal4p, in the nucleus of yeast, which results in the activation of reporter genes required for growth (Fig 2). In the case of Gal4p, a bait-protein is fused to the Gal4p DNA-binding domain (DB) and a prey-protein is fused to the transcriptional activation domain (AD). If the bait and prey-proteins physically interact, the two domains are brought together with the result that yeast is able to survive and Figure 2 the Y2H system. A physical interaction between proteins 2 the Y2H the system. physical interaction between grow on selective plates. XFigure and Y reconstitutes Gal4pA transcription factor which results X and of Y reporter reconstitutes Gal4p for transcription the activation genesthe required growth. DB:factor DNAA first draft of the inproteins which results in AD: the activation activation domain. of reporter genes required for binding domain. C. elegans interactome was growth. DB: DNA-‐binding domain. AD: activation domain. created by using 3,024 predicted worm ORFs as baits which were screened against two Gal4 AD-libraries: the ORFeome-based library mentioned above and an AD cDNA library7. This resulted in more than 4,000 protein-protein interactions of which a subset was validated with affinity purification. Mapping specific signaling pathways, such as C. elegans TGFß signaling, has shown that in-depth systematic analysis of all the genes involved in a signaling route may reveal novel proteins or new links between known regulators30. Furthermore, complementation of interactomes with forward or reverse genetics adds functional information to protein pairs, which adds an additional level of detail to interactome maps31,32.

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

For all these studies, full-length bait clones were tested. However, full-length proteins are sometimes incapable of interacting in the Y2H system. Full-length proteins can be folded in such a way that the interaction domain is masked or folded incorrectly. The use of protein fragments would partially bypass these issues and thereby increase the number of interactions identified in Y2H-screens. Furthermore, fragment-based screens provide information on the specific domains that facilitate the interaction, especially when applied for both bait and prey libraries (Fig 3). This fragment-based Y2H approach was tested by generating an AD-Fragment library for 750 early embryogenesis C. elegans genes. The approach increased the detectability of interaction in the Y2H-screens, and led to the identification of interaction domains for over 200 proteins33. The same strategy was applied to human genes in which the human ORFeome library was randomly fragmented34. In our research we used a fragment-based high-throughput Y2H approach to identify PPI of polarity regulators. In the years to come, all the advances of the last decade and the current high standards of high-throughput protein interaction mapping may address fundamental topics with high precision. A

Protein-Y

Stop

III

Minimal region of interaction

Start

III

Figure 3A detection of the minimal interacting region of a protein. Protein interaction domains can be identified when a interaction region is shared, the minimal region of interaction (yellow region) is shared by three fragments. Figure 3B domain interaction mapping. Most proteins combine modular domains, such as catalytic and scaffold interaction domains that results, when combined, in more detailed interaction maps because of the identified minimal region of interaction.

Three conserved protein complexes that control cell polarity Polarity is a fundamental property that is essential for a variety of biological and cellular processes. Many cell types need to organize their membrane in molecular compartments to spatially restrict signaling, organize directional trafficking, or provide a cue for cells to respond to their environment (Fig 4). One example of the importance of cell polarity is the ability of cells to divide asymmetrically. During the development of multicellular organisms, cell specification by asymmetric division allows the generation of daughter cells with different developmental potential. Pluripotent cells generate diversity that can be achieved by either intrinsic or extrinsic cues. During intrinsic asymmetric division, cell fate determinants

Proteomics approaches in the study of cell polarity in the nematode Caenorhabditis elegans

are unequally distributed, and the mitotic spindle is positioned such that two daughter cells with different developmental potential are generated. Cell polarity is also essential for the functioning of epithelial cells, which are the most common type of cells in mammals35. Epithelial cells are polarized along a vertical axis into apical and basolateral domains. Polarity allows epithelial cells to build spatially restricted structures such as microvilli or cilia on the apical border, and junctions and desmosomes on lateral regions, and allows these cells to attach to stroma on the basal side. Aberrant polarity or loss of cell polarity is considered as one of the hallmarks of cancer, affecting oncogenic signaling pathways in tumor cells, causing deregulation of proliferation, and promoting tissue invasion and metastasis36. Three conserved groups of proteins have been identified that control cell polarity. The PAR complex, the Crumbs complex, and the Scribble group of proteins control the overall architecture of a diverse set of cells, tissue types and organs37. In this section of chapter one, I will provide an overview of the proteins that affect the polarization machinery in a variety of cell-types in C. elegans. In the final section of this chapter I will elaborate on how the implementation of other mechanisms and cellular processes relate to cellular polarization. Unequal distribution of: Proteins

Apical Anterior

Posterior

Junctions Lipids Cytoskeleton Endosomal structures

Basal

Figure 4 schematic representation of cell polarity. Cell polarity is the unequal, or asymmetric, distribution of molecular and cellular components along a vectorial axis. The asymmetric distribution of anterior and posterior determinants allows the C. elegans zygote to divide asymmetrically (middle figure). Epithelial cells are apical-basal polarized, which allows the generation of spatial restricted structures (right).

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

The PAR-complex Pioneering work by Kemphues and colleagueâ&#x20AC;&#x2122;s resulted in identification of the PAR proteins in a screen for Partitioning defective mutants in the C. elegans zygote38. Embryos mutant for any one of six par genes partition cytoplasmic components symmetrically, instead of asymmetrically, which results in two daughter cells both equal in size and composition. Polarization of the embryo is triggered when the mature centrosome contacts the cortex upon sperm entry, which locally weakens the contractility of the actomyosin cytoskeleton, resulting in an actomyosin cytoskeletal flow toward the anterior-posterior ends39. This process appears to be mediated by local down regulation of the RhoGEF ECT-2, which leads to inactivation of Rho, and inactivation of the regulatory myosin light chain NMY-2[ref. 40]. Before fertilization, the PDZ-domain scaffolding proteins PAR-3 and PAR-6 and the atypical protein kinase C aPKC/PKC-3 are uniformly distributed along the cortex. As a consequence of the cortical flow, PAR-3/PAR-6/PKC-3 are transported anteriorly (Fig 5), and the serine-threonine kinase PAR-1 together with the Ring-finger protein PAR-2 relocalize from the cytoplasm to the posterior cortex41,42,43. Antagonistic interactions between these two groups of PAR complexes then establish two different cortical domains42. Embryos mutant for par-2 exhibit an extension of the anterior PAR-3/PAR-6 cortical domain while worms mutant for either par-3 or par-6 show the opposite phenotype44. The opposing localization of the two PAR-complexes is achieved by the antagonistic phosphorylation of PAR-2 by aPKC, and of PAR-3 by PAR-1. The Rho GTPase CDC-42 plays a role in the stable maintenance of the polarized domains45. The semiCrib and PDZ domains of PAR-6 physically interact with active CDC-42, which contributes to the localization of PAR-6[46,47]. During polarization, the serine threonine kinase PAR-4 and the 14-3-3 protein PAR-5 are uniformly distributed throughout the cytoplasm. The phenotypes of par-4 or par-5 mutant embryos are weak, but polarity is still disturbed in these mutants38,48,49. C. elegans PAR-4 is orthologous to human liver kinase B 1 (LKB1), that activates various kinases including PAR-1 in humans, but this activity has not been shown for C. elegans PAR-1[ref. 50]. PAR-5 binds phosphorylated proteins such as PAR-1, PAR-3, Lethal giant larvae (LGL) and PAR-5 was shown to bind to Drosophila PAR-3/BAZOOKA upon PAR-1 phosphorylation in Drosophila51. In short, the combination of actomyosin contractility, anterior cortical flow, and antagonizing activities of the PAR complex ensures stable polarization along an anterior to posterior axis.

Proteomics approaches in the study of cell polarity in the nematode Caenorhabditis elegans

PAR-polarity after asymmetric division PAR polarity after fertilization

PAR-polarity before asymmetric division

PAR-3 / PAR-6 / PKC-3 PAR-1 / PAR-2

A P

Anterior Posterior

Figure 5 PAR-polarity in the C. elegans embryo. After fertilization, the embryo is polarized in an anterior to posterior axis. The two PAR-complexes are asymmetrically distributed along the cortex in which PAR3/PAR-6 and PKC-3 (green) localize to the anterior and PAR-1 and PAR-2 (red) to the posterior daughter cell.

After the second round of zygotic divisions, polarity is newly defined as radial or apical to basal, instead of anterior to posterior. At this stage PAR-3, PAR-6 and PKC3 localize to outer cell surfaces, and are excluded from the sites of cell-cell contact. A screen for mutants in which the cortical distribution of PAR-6 was uniform at the four cell stage identified the gene pac-1. PAC-1 is a GTPase activating protein (GAP) that promotes formation of the inactive GDP-bound form of CDC-42. PAC-1 localization is confined to cell-cell contacts, limiting the recruitment and restriction of PAR-6 to contact free cell surfaces52,53. With the exception of PAR-2, the PAR proteins are highly conserved throughout the animal kingdom and are essential for polarity establishment in a number of different cell types. In cells responding to a chemo-attractant for example, the PAR proteins localize to the leading edge, modulating the cytoskeleton and providing a migration cue54. In epithelia, PAR-1 defines the basolateral domain while PAR-3/ PAR-6 and aPKC localize apically55. The PAR-complex can be thus considered as one of the core polarity components, functioning in various different cell types. The Crumbs-complex The Crumbs (Crb) protein complex was originally identified in Drosophila as a regulator of epithelial polarity in the embryo, controlling the establishment of apical membrane identity56. The complex consists of Crb, PALS1 (protein associated with LIN-7, Stardust in Drosophila), LIN-7 and PATJ (PALS1 associated tight junction

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

protein) and interacts with the PAR complex to stabilize the junctional Cadherin complex56 (Fig 6). Drosophila mutants lacking either Crb or Sdt show similar defects and die early in development while PATJ and DLin-7 mutants are viable57,58,59. Crb is a transmembrane protein. The large extracellular domain is composed of EGF and Laminin motifs in the extracellular domain. The short intracellular domain consists of a Carboxyl-terminal PDZ-binding motif that binds to the PDZ-domain of PALS1. PALS1 is a scaffolding protein that binds to LIN-7 and PATJ via its L27 domain58,59,60. In epithelial cells, PATJ is essential for epithelial polarization, positively promoting the Crumbs complex, and stabilizing tight junctions61. In Drosophila photoreceptor cells, loss of LIN-7 causes light-induced retinal degradation, and PATJ plays a role in proper localization of adherens junctions62,63. The CRB-complex is conserved in C. elegans and mammals. C. elegans has three members of the Crumbs family, eat-20, crb-1 and the recently identified crb-3, three candidates for the MAGUK domain containing protein PALS1 magu-1, magu-2 and magu-3, one member of PATJ and one member of LIN-7[ref 64]. In mammals, the Crbfamily of proteins contains three members; CRB1, CRB2 and CRB3. CRB1 and 2 have a similar protein domain composition as Drosophila Crb, while Crb3 lacks the large extracellular domain. All three Crumbs proteins are characterized as polarizing factors in all model organisms except for C. elegans. In mammals the Crb3 protein is the most frequent Crumbs protein associated with cancer. Crb3 is expressed in epithelial cells and functionally important, a Crb3 knockout mice show extensive epithelial morphogenesis defects and die shortly after birth65,66. The function of the Crumbs proteins in C. elegans is unexpected because worms lacking all three Crumbs genes have no obvious epithelial defects and the Crumbs proteins are not essential for viability67. eat-20 mutant worms exhibit defects in pharyngeal pumping68. LIN-7 was originally identified in C. elegans as a cell junction associated protein required for the basal localization of the EGF-receptor kinase LET-23 in vulva cells69. Worms mutant for lin-7 show lethality but exhibit no obvious epithelial polarity defects70. Mammals have a LIN-7 homologue known as Velis/ MALS59. It was shown that loss of LIN-7 in polarizing MDCK cells functions in the basal sorting machinery and that apical localization of LIN-7 is dependent on Î˛-catenin71. In neurons, LIN-7 collaborates with LIN-2 (CASK) and LIN-10 (MINT) regulating exocytosis in the establishment of neuronal cell adhesion72. Further studies in C. elegans may elucidate the molecular details of how the Crumbs complex functions in cellular polarization.

Figure 6 overview of the polarity complexes and their interactions. Left: cellular localization of polarity proteins. Right: a schematic overview of the protein domains of the proteins in the three polarity complexes. Symbols are given for the protein domains. Solid line arrows indicate direct interactions and the dashed arrows phosphorylations. In blue the apical region and the Crumbs complex, in orange and green the sub-apical domain and the PARcomplex, and in red the basolateral region with the Scribble-group of polarity proteins. Lipid modifiers are shown in yellow.

Proteomics approaches in the study of cell polarity in the nematode Caenorhabditis elegans

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

The Scribble group of proteins Besides the PAR and the Crumbs protein complexes, epithelial polarity depends on a third conserved polarity group, the Scribble group of proteins. This group consists of three proteins namely Scribble (scrib), Discs large (dlg) and Lethal Giant Larvae (lgl) (Fig 6). Scribble, encoding a leucine-rich repeats and PDZ-domain (LAP) protein, was originally identified in a Drosophila screen for maternal effect mutations that disrupt aspects of epithelial morphogenesis73,74. Dlg and lgl were identified in a screen for tumor suppressors in Drosophila epithelial cells. Mutant clones exhibited overproliferation and miss-localization of apically restricted proteins74. These proteins act in concert to maintain the identity of the basolateral domain. Drosophila epithelial cells mutant for any of the genes encoding Scribble complex proteins show mislocalization of apical and junctional proteins to the basolateral domain75. Thus, the main role of the Scribble group appears to be exclusion of apical proteins from the basolateral domain. Physical interactions between the proteins in the complex have not been shown. One homologue of Scribble has been described in the C. elegans and mammal genomes76,77. C. elegans Scribble is encoded by the let-413 gene and was identified in an adherens junction deficiency screen76. Worms mutant for let-413 exhibit elongation defects in the early embryo which results in lethality78. In C. elegans epithelial cells, LET-413 functions in concert with DLG-1 to form adherens junctions and restrict the junctions to subapical domains79. Human Scribble was isolated in a screen designed to identify targets of the E6 onco-protein of the Human Papilloma virus, which is the major causative agent for cervical cancer77,80. Scribble genes in humans encode for membrane associated proteins, which localize basolaterally in epithelial cells81. How mammalian Scribble may function as tumor suppressor depends on its cellular context. Loss of Scribble may induce invasive behavior of mammary epithelial cells by activation of Ras and Raf, modulating MAPK-signaling82. It remains to be elucidated whether Dlg and Lgl can regulate Ras signaling. Furthermore, in Drosophila, clones of cells with loss of Scribble combined with an activated RasV12 mutants exhibit massive overproliferation and activate JNK and JAK/STAT signaling83. A mouse mutation allele model of Scrib (circletail) showed severe neural tube defects, and mouse scrib genetically interacts with looptail, a homolog of the planar polarity protein Van Gogh/Strabismus84. Thus, Scribble appears to play a role in planar cell polarity as well. Dlg is a scaffolding protein and has three PDZ (PSD95/DLG/ZO1) domains, a Src homology (SH3) domain and a GUK domain85. In spite of the clear tumor suppressor role described in Drosophila, the function of dlg in C. elegans and mammalian epithelia is less clear86. C. elegans has a sole dlg homologue, dlg-1, which is required for proper

Proteomics approaches in the study of cell polarity in the nematode Caenorhabditis elegans

adherens junction formation. In a dlg-1 mutant background, proteins normally restricted to apical regions within epithelial cells, such as the PAR and Crumbs complex proteins, are not properly localized79. Immuno-precipitation studies with Drosophila Dlg revealed a large complex containing e.g Scrib, but also the scaffold protein LIN-2, and the PDZ-containing proteins LIN-7 and LIN-10. The above mentioned C. elegans LIN-2/LIN-7/LIN-10 complex has a distinct role in vulva cells, mediating correct basal localization of the LET-23 EGF receptor69. Four mammalian dlg genes (DLG1-4) and an additional hDlg5 gene in humans are described, which makes it complicated to analyze because of possible functional redundancy87. Studies in mammalian cells showed that all paralogs show different epithelial localization patterns, and that only the loss of DLG3 is associated with defects in cell polarity and apical junction formation88. DLG4 (also known as postsynaptic protein PSD95) has an important role in neurons where it functions in synapse development89. Partial functional conservation was shown by the ability of rat DLG1 and DLG3 to rescue the Drosophila dlg overproliferation phenotype90. Lgl mutant fly larvae exhibit overproliferation of neuronal cells in the brain, and of epithelial cells in the imaginal discs. Lgl localizes together with Scibble and Dlg at basolateral membranes but can be released to the cytoplasm upon phosphorylation74. It is postulated that LGL can associate with the cytoskeleton in the cytoplasm. C. elegans has one sole homologue of the lgl gene, termed lgl-1, but an epithelial role of the protein has not been described. However, in the early embryo, LGL-1 has been shown to act redundantly with PAR-2 and localizes to the posterior cortical domain. This is achieved by mutual elimination induced by phosphorylation of conserved residues on LGL-1 by aPKC from the anterior PAR-complex42,91,92. In mammals two proteins are related to Drosophila lgl. Both proteins are expressed in a wide variety of cell types. Upon epithelial polarization, human lgl (Hugl) is released from the cortex and localizes together with NMY-II heavy chain at the cytoskeleton and is postulated to play a role in vesicle exocytosis93.

The spatiotemporal control of cooperative crosstalk in cell polarity For a cell to be properly polarized, it is essential that the cortical domains are well defined and proteins are correctly localized. Through either cooperative or antagonistic mechanisms, the polarity complexes adjust local signaling and function to achieve cellular asymmetry and vectorial transport. These mechanisms act differently depending on the cellular context and cell type. How this is achieved and by which underlying mechanisms is not well understood. One reoccurring

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

mechanism of polarity proteins is the ability to exclude proteins from cortical regions such as the antagonistically functioning PAR-protein complexes in the C. elegans embryo42. It recently became clear that cellular processes, such as trafficking of proteins, are functionally linked to epithelial cell polarity. A C. elegans genome-wide image-based RNAi screen to identify genes involved in correct localization of apical RAB-11 positive recycling endosomes and lumen morphology identified several polarity proteins. In this screen, PAR-5 was identified as a critical regulator of recycling endosomes. PAR5 depletion showed patches of basolateral localization of RAB-11 positive recycling endosomes94. In addition, an RNAi based screen for defects in the endocytosis and secretion of Yolk protein YP170 showed that efficient endocytosis is dependent on PAR-3/PAR-6/aPKC and CDC-4295. Furthermore, loss of the clathrin adaptor AP-1 did not alter epithelial polarity establishment or the formation of villi. However, AP-1 loss ectopically induces lumen formation inside the normal C. elegans intestine and loss of the polarized distribution of apical and basal lateral transmembrane proteins96. These examples show that correct organization of different aspects of the trafficking machinery is dependent on apical basal distribution of plasma membrane proteins. Spatially restricted cytoskeletal adjustments, such as apical actin rearrangements, influence the formation of microvilli on epithelial cells. In this process, the local activation of the Rho-GTPase CDC-42 has been shown to be an important event40,96. Cytoskeletal linkers are known to be essential in the formation of microvilli, activation of ERM-proteins in apical regions allows the attachment of the actin cytoskeleton to the cell membrane97. The MDCK cell line is frequently used and provides an easy and traceable 3D-culture model for studying polarity98. In MDCK cysts, correct apical localization of ERM-proteins and villi formation during polarization requires an intracellular trans-cytosis event in which a signaling platform of Podocalyxin/ NHERF/Ezrin is recruited from the basal lamina upon inactivation of RhoA99. This illustrates the combined involvement of several basic cellular processes to achieve a correct polarized state. Epithelial polarity establishment is not exclusively regulated by intrinsic regulators. The epithelial microenvironment, the extracellular matrix (ECM) and membrane lipids known as phosphatidylinositol (PIs), determine the axis on which cells polarize (Fig 6). The ECM is able to provide a scaffolding landmark that consists of fibrous structural proteins, sugars, fluid and signaling molecules that define positional information and differentiation cues100. How the ECM contributes to polarization and the molecular nature of molecular specification remains unclear. However, it appears that ECM stiffness mediated by collagens, and

Proteomics approaches in the study of cell polarity in the nematode Caenorhabditis elegans

cell confinement may alter correct spindle positioning during the first division of polarizing MDCK cells in 3D cell culture101. Furthermore, in the same cell system, signals originating from the ECM effect intracellular signaling, as is shown for Î˛1-integrins. Activated integrins are able to suppresses RhoA activity which in turn leads to less ROCK activity and stable microtubules at the basal region of an epithelial cell101. Spatiotemporal regulated production of membrane lipids by kinases such as PI-3K and phosphatases such as PTEN are essential for signaling and function. The lipid PtdIns(3,4,5)P3 (PIP3) is enriched at basolateral membranes whereas PtdIns(4,5)P2 (PIP2) is apically localized. How membrane lipid asymmetry is initially coordinated during polarization is unclear. It has been shown that for a cell to be polarized correctly, a balance of specific lipids is crucial for homeostasis102. The kinase PI-3K and the phosphatase PTEN play essential roles in the maintenance of this phospholipid distribution. PTEN is activated by CDC-42 and localizes to apical and subapical regions together with PAR-3, removing a phosphate from PIP3 to maintain PIP2103. PTEN functionally antagonizes PI-3K and thereby excludes it from apical regions restricting it to basolateral domains. PI-3K phosphorylates PIP2 to PIP3 which is postulated to be regulated through the GTPase Rac1 activated via integrin dimerization, and E-cadherin104 (Fig 6). C. elegans lipid biogenesis in the intestine during tubulogenesis converts apical-basal polarity to lumen formation by a yet unknown mechanism105.

Cell Polarity regulation in C. elegans To study cell polarity we use the nematode Caenorhabditis elegans. C. elegans has a number of polarized tissues, which we have used to identify the functional importance of candidate polarity-related protein interactions in chapter 2. Here, I provide an overview of the various tissues examined in this thesis. Epithelial tissues C. elegans has three types of epithelial tissues. Classic epithelium, which forms a sheet or a tube as for example found in the hypodermal seam cells or the intestine, contractile epithelial cells like the spermatheca, and epithelial cells functioning as a single unit such as the excretory system. As in other animals, C. elegans epithelial cells display apical-basal polarity106. In the studies described in this thesis I have examined the hypodermal seam cells, the intestinal epithelium and the excretory system.

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

The hypodermal seam cells Seam cells are hypodermal epithelial cells that span the complete length of the worm and are embedded in the hypodermal syncytium (Fig 7). The cells apically secrete collagens and other stromal proteins to form a subcutaneous layer contributing to the cuticle. The first hypodermal cells arise during embryonic development and larvae hatch with ten seam cells (H0/H1/H2, V1-V6 and T) on each lateral side of the animal. An important characteristic of the seam cells is their ability H1 H2 V1 V2 H0 L2/L3 V3 to divide in an asymmetric T V4 V5 V6 stem-cell like pattern in each larval stage. In the second larval stage, the asymmetric apical/basal division is preceded by one extra symmetric division anterior/posterior (except in H0 and V5)16. After Figure 7 schematic representations of the hypodermal Seam each asymmetric division, the cells. The 16 seam cells (red) are polarized in an apical to basal smaller anterior cell migrates and anterior to posterior axis. into the cuticle and fuses with the hypodermis while the larger posterior cell will retain the seam cell fate. The H0 and T cell remain as single seam cells, and V5 will differentiate to produce posterior deirid cells and a single neuron. In the final stage of larval development, C. elegans has 16 seam cells. In the L4 stage, the seam cells terminally differentiate, fuse into a single syncytial cell, and produce the cuticular structure known as adult alae. The seam cells are attached to each other and to the surrounding hyp7 syncytial cell through apical junctions. The presence of apical junctions and the localization pattern of the polarity complexes suggest that the seam cells are polarized in an apical-basal manner. Loss of either PAR-3 or CDC-42 results in gaps between the seam cells107. A candidate-based RNAi screen that focused on identifying genes involved in the apical localization of the adherens junction E-cadherin in the seam cells showed that a newly defined protein SOAP-1 (sorting of apical proteins 1) acts in concert with the clathrin adaptor AP-1 in restricting E-cadherin apically in the seam cells108. Seam cells are also responsive to anterior to posterior Wnt-signaling. Non-canonical Wntsignaling is required for the asymmetric division of the V1-V4 and V6 lineage109. Loss of the C. elegans Wnt target POP-1, a homologue of the TCF-LEF transcription factor, resulted in overproliferation of seam cells, indicating an essential role in seam cell-fate specification for Wnt signaling110.

Proteomics approaches in the study of cell polarity in the nematode Caenorhabditis elegans

The intestinal epithelium The pharynx and the intestine together comprise the digestive tract of the worm. The pharynx is responsible for bacteria uptake and for grinding the bacteria before they enter the intestine. The pharynx is surrounded by various cell types, such as neurons and muscle cells. The internal cell type of the bulb consists of epithelial cells that display apical-basal polarity with the apical surfaces lining the pharynx cavity. The ground bacteria enter the intestine via the pharyngeal-intestinal valve. The intestineâ&#x20AC;&#x2122;s primary function is the uptake of nutrients and fluids from the lumen. The organ develops entirely from the E blastomere. After the first embryonic cell division, upon gastrulation, the E daughter cells differentiate into endoderm precursor cells which apically ingress by constriction. The C. elegans intestine consists of 22 large epithelial cells, four cells next to the pharynx known as the foregut and nine rings of two cells, with distinct apical-basal and lateral regions16,111 (Fig. 8). The apical surfaces are lined by microvilli brush borders. These specialized structures are composed of actin filled protrusions that extend into the lumen from the apical surface and are attached apical/basal to the sub-apical terminal web, Figure 8 schematic representation of the intestinal a cytoskeletal network of actin epithelium. The intestinal epithelium is a tube like structure and intermediate filaments. with an apical lumen (red) and a basolateral domain (blue). Two opposite cells are attached by junctional proteins (green) The microvilli expand the that form a junctional belt. apical surface and the capacity of the intestine to take up fluids and nutrients from the lumen112. Most of the tube is composed of only two opposed intestinal cells. Apical-basal polarity is established during intestinal morphogenesis. The cues that trigger apical/basal polarization of the intestinal cells are not fully understood. One signal may involve a re-localization of the Microtubule Organizing Centre (MTOC) from the centrosome to the apical surface. This process coincides with the appearance of apical foci containing PAR-3, PAR-6, and microtubule nucleators. When the MTOC was ablated at the onset of polarization, the apical surface did not develop113. Furthermore, par-3 mutant worms developed non-apical foci of cytoskeletal clusters containing PAR-6, showing that PAR-3 is required for correct localization of PAR-6 and apical microtubules113. A parallel study showed that C. elegans endoderm polarization involves similar mechanisms as in Drosophila, in which the same foci that were enriched for PAR-3, PAR-6, and aPKC also contain junctional proteins such as D-Ecad114. Surprisingly, the C. elegans CRUMBS proteins

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

have no role in intestinal epithelial organization or maintenance and animals lacking all three Crumbs homologs are viable67. In addition, inhibition of the basolateral polarity complexes or the junctional protein AJM-1 resulted in disorganized apical junctions, cytoplasmic localization of baso-lateral components and extension of the apical domain115. Depletion of the apical cytoskeletal linker ERM-1 showed no defect in early polarization of the C. elegans embryo, but during endoderm specification, the localization of apical junction proteins failed. This resulted in lethality which is due to a loss of apical actin and reduced villi formation111,116. Surprisingly, rap-2 mutant worms, which showed a reduction in the levels of the phosphorylated and activated form of ERM-1 in the apical brush border, are viable117. This would mean that the phosphorylation of ERM has a non-essential role in C. elegans intestinal villi. Notably, the separation of the apical membranes which initiates lumen formation is linked to membrane lipid composition and the appearance of apical vesicles, identified to be RAB-11 positive recycling endosomes94. A screen to identify defects in tubulogenesis revealed genes encoding for lipid-biosynthetic enzymes that converts apical-basal polarity to lateral lumen formation. It is thus likely that the apical membrane lipid composition of C. elegans may function as a cortical landmark directing positional trafficking105,96. The excretory canal The excretory system is comprised of four different types of cells: a duct cell, a pore cell, a fused pair of gland cells and a canal cell. The function of the excretory system is not completely understood, but is presumed to include maintaining osmotic homeostasis and waste elimination, equivalent to mammal kidneys118,119. Furthermore, the excretory system may have a role in larval molting120. The canal cell apical/basal is the largest cell in C. elegans, Figure 9 schematic representation of the excretory canal. The and forms an H-shaped tube excretory canal is a single polarized cell spanning the nearly spanning the nearly complete complete length of the worm (green). The canal is polarized in length of the animal. The both anterior to posterior and an apical basal axis and highly enriched for vesicles (grey). canal cell displays apical basal polarity and develops differently from most epithelial tissues. The canals are made up by a luminal cavity, filled with fluid. The cytoplasm contains many vesicles and canaliculi, and is lined by an apical and a basolateral membrane121 (Fig 9).

Proteomics approaches in the study of cell polarity in the nematode Caenorhabditis elegans

The molecular mechanisms involved in the development and polarization of the canal are mostly unclear. Early genetic studies identified twelve exc genes. Mutation of the exc genes causes cysts to develop in the canal122. One geneâ&#x20AC;&#x2122;s function could be linked to polarity namely fgd-1. FGD-1 is a guanine nucleotide exchange factor for CDC-42 and is required for proper localization of cytoskeletal elements in the apical surface. In humans, mutations in FGD-1 cause cranial disorders123. C. elegans mutant for fgd-1 are depleted of apical recycling endosomes whereas overexpression leads to enrichment of endosomes and failing of apical lumen formation124,125. Recent studies have yielded new insights into the process of lumenogenesis. Loss of erm-1 was shown to result in the formation of cyst-like structures in the canal. ERM-1 is the single C. elegans ortholog of the Ezrin-Radixin-Moesin (ERM) family of membranecytoskeleton linking proteins. ERM-1 is enriched at luminal borders and required for tubular organ lumen morphogenesis. For the excretory canal, ERM-1 recruits vesicles and stabilizes actin which results in vesicle fusion in the lumen in such a way that it extends. This process is mediated by the STE20 like kinase GCK-3, which is postulated to phosphorylate ERM-1 in humans, changing it to an active conformation and linking the actin cytoskeleton to the membrane126,127. The GTPase RAL-1 controls apical vesicle fusion and acts in concert with the vesicle tethering exocyst complex. C. elegans mutant for either sec-8, ral-1 or sec-5, showed defects in the extension of the excretory canal. The same study showed that the PAR-complex localizes apically in the excretory system128. In our RNAi screens, knockdown of erm-1 exhibits a severe canal defect phenotype. Depletion of either par-2 or par-4, and the planar polarity protein vang-1 resulted in a mild phenotype. This result opens up new questions such as how do polarity proteins control the development of this system and why does inactivation of the PAR-3/PAR-6/aPKC not cause defects in this tissue? Neuronal polarity den The worm consists of 958 drite s somatic nuclei of which 302 are neurons. C. elegans neurons can axo n be divided into two distinct and independent neuronal systems, Figure 10 schematic representation of the C. elegans PVD neuron. Neurons are highly polarized structures with a somatic and a pharyngeal distinct polarized compartments such as an axon (red) or nervous system. The somatic dendrites (green) originating from the cell body (grey). nervous system is primarily localized in the hypodermis and the body wall muscle cells. There are various types of neuronal cells such as motor and sensory neurons. Neurons are complex

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

and highly polarized cells with multiple specialized functional compartments, known as dendrites and axons129 (Fig 10). Where dendrites function in detecting mechanical stimuli, the axons deliver chemical signals to target cells via synapses. Three conceptual steps are described in how neuronal polarity can be achieved130. Initially, after mitotic division, polarity needs to be established. Cortical landmarks or polarity proteins are created and molecules localize asymmetrically in such a way that the axon and dendrite location is specified. The next step in polarization is the outgrowth of dendrites and axons in which structural and cytoskeletal proteins are targeted to either dendrites or the axon. Finally, the polarized morphology needs to be maintained to ensure specification by directed transport of cargo and specific molecular composition in each compartment. In Drosophila and mammals, members of the polarity protein complexes are involved in establishing neuronal polarity. How the three polarity complexes function in C. elegans neurons is not yet clear. The apical PAR-complex has distinct functions in a variety of C. elegans tissues, but not in neurons. In contrast, PAR-4 facilitates the specification of the axon and the dendrites131,132. A similar role was found for the PAR-1 like kinase SAD-1[ref. 133]. Although no clear roles for the other members of the PAR-complex proteins are known, in mammals the PAR-complex functions in axon guidance during neuronal development134. As for the Crumbs complex, there is no evidence that the C. elegans PAR complex plays a role in neuronal polarity. EAT-20 is expressed in pharyngeal neurons and muscle cells. C. elegans mutant for eat-20 exhibit a reduced pharyngeal pumping rate but it is not known if this is due to a neurological defect68. For the Scribble-complex, no neuronal function is described for C. elegans. Our screens showed that depletion of lgl-1 resulted in the appearance of axon specific vesicles in dendrites. The Drosophila Scribble-complex plays an essential role in the polarization of neuroblasts before asymmetric division135. Later functions in the neuron have not been described. In a mouse model for ocular cancer the scribble-complex is miss localized in retinal neurons and postulated as a marker for late tumor progression136. In humans, HDlg localizes in synapses and functions as a scaffold protein mediating vesicle release and receptor localization137. Two separate functions for human Scribble in neuronal cells are described. Neurons in culture respond to neuronal growth factor, which downstream acts on Scribble. At synapses, Scribble interacts with Î˛-catenin, orchestrating a downstream function to cluster vesicles in synapses135. A surprising observation in C. elegans was that the planar polarity proteins VANG-1, PRKL-1 and DSH-1, which have no clear planar polarity function in the worm, cause neurodevelopmental defects of motor neurons138. This result shows a new unexpected role for the planar polarity proteins. A clear role has been described for Wnt proteins in neurons and has been implicated in multiple aspects of C. elegans and mammals neuronal polarity as e.g migration of Q-cells along the 28

Proteomics approaches in the study of cell polarity in the nematode Caenorhabditis elegans

anterior-posterior axis139. Whether the polarity proteins exhibit similar mechanisms in neuronal development and cellular processes remains to be established.

Summary and scope of this thesis In this chapter, I discussed how C. elegans research has evolved through the omics era and introduced the polarity proteins that are involved in the development of polarized tissues in C. elegans. Over the years, a number of polarity regulators have been identified. Perhaps the best understood polarizing system is the establishment of apical basal polarity in epithelia, which is mediated by the opposing actions of the apical PAR and Crumbs complexes, versus the basolateral Scribble group of proteins (Chapter 1). Even in epithelia, we only partially understand the mechanisms involved, and they vary between tissues and organisms. In addition, many of the mechanistic details of how polarity integrates with other aspects of cell polarity, such as establishment of a polarized trafficking machinery, cytoskeletal rearrangement, or cell morphogenesis, remain to be discovered. Ultimately, for a complete understanding of cell polarity, we need to obtain a comprehensive overview of not only the proteins involved, but also the molecular interactions between them. Commonly used protein interaction identification techniques, such as the yeast twohybrid system applied in this thesis, have been optimized over the years such that they can be used on a large scale with high precision. In Chapter 2 we generated a map of candidate interactions that may contribute to cell polarity. We can now begin to investigate how these interactions contribute to cell polarity. Ultimately, a detailed understanding of the molecular mechanisms that control cell polarity will allow us to better understand how alterations in gene function contribute to the development of pathological conditions such as cancer. Although interaction maps provide a wealth of information, detailed descriptions of how the protein domains that mediate these interactions remain elusive since most libraries contain only full-length genes. In Chapter 3 we expanded on a C. elegans fragment-based Y2H library strategy by generating a genome wide human fragment library. To demonstrate the quality of the library, we screened several bait proteins required for cell division and polarity and tested the obtained interactions experimentally. The Y2H system has proven to be a strong and versatile technique to identify binary protein interactions on a large scale. However, the system also still proves to be extremely valuable in identifying functional links on smaller scale for more specific research questions to unravel molecular mechanisms. In Chapter 4 we show that the sumo-protease ULP-4 interacts with HMGS-1 and that this interaction regulates mevalonate metabolism during C. elegans aging. ď&#x201A;ž 29

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26.

Vidal, M., Cusick, M. E. & Barabási, A.-L. Interactome networks and human disease. Cell 144, 986– 98 (2011). Lage, K. Protein-protein interactions and genetic diseases: The interactome. Biochim. Biophys. Acta 1–10 (2014). doi:10.1016/j.bbadis.2014.05.028 Walhout, A. J., Boulton, S. J. & Vidal, M. Yeast two-hybrid systems and protein interaction mapping projects for yeast and worm. Yeast Chichester Engl. 17, 88–94 (2000). Yu, H. et al. High-quality binary protein interaction map of the yeast interactome network. Science 322, 104–10 (2008). Rajagopala, S. V. et al. The binary protein-protein interaction landscape of Escherichia coli. Nat. Biotechnol. 32, 285–90 (2014). Guruharsha, K. G. et al. A protein complex network of Drosophila melanogaster. Cell 147, 690–703 (2011). Li, S. et al. A map of the interactome network of the metazoan C. elegans. Science 303, 540–543 (2004). Arabidopsis Interactome Mapping Consortium. Evidence for network evolution in an Arabidopsis interactome map. Science 333, 601–607 (2011). Rolland, T. et al. A Proteome-Scale Map of the Human Interactome Network. Cell 159, 1212–1226 (2014). Lambert, J.-P. et al. Mapping differential interactomes by affinity purification coupled with dataindependent mass spectrometry acquisition. Nat. Methods (2013). doi:10.1038/nmeth.2702 Kwon, Y. et al. The Hippo Signaling Pathway Interactome. Science (2013). doi:10.1126/ science.1243971 Babu, M. et al. Interaction landscape of membrane-protein complexes in Saccharomyces cerevisiae. Nature 489, 585–589 (2012). Brunner, E. et al. A high-quality catalog of the Drosophila melanogaster proteome. Nat. Biotechnol. 25, 576–583 (2007). Kim, M.S. et al. A draft map of the human proteome. Nature 509, 575–81 (2014). Brenner, S. The genetics of Caenorhabditis elegans. Genetics 77, 71–94 (1974). Sulston, J. E. & Horvitz, H. R. Post-embryonic cell lineages of the nematode, Caenorhabditis elegans. Dev. Biol. 56, 110–156 (1977). Coulson, A. et al. YACs and the C. elegans genome. BioEssays News Rev. Mol. Cell. Dev. Biol. 13, 413–417 (1991). Waterston, R. & Sulston, J. The genome of Caenorhabditis elegans. Proc. Natl. Acad. Sci. U. S. A. 92, 10836–10840 (1995). Watson, J. D. & Cook-Deegan, R. M. Origins of the Human Genome Project. FASEB J. Off. Publ. Fed. Am. Soc. Exp. Biol. 5, 8–11 (1991). C. elegans Sequencing Consortium. Genome sequence of the nematode C. elegans: a platform for investigating biology. Science 282, 2012–2018 (1998). Harris, T. W. et al. WormBase 2014: new views of curated biology. Nucleic Acids Res. 42, D789–793 (2014). Fire, A. et al. Potent and specific genetic interference by double-stranded RNA in Caenorhabditis elegans. Nature 391, 806–811 (1998). Tabara, H., Grishok, A. & Mello, C. C. RNAi in C. elegans: soaking in the genome sequence. Science 282, 430–431 (1998). Kamath, R. S. & Ahringer, J. Genome-wide RNAi screening in Caenorhabditis elegans. Methods San Diego Calif 30, 313–321 (2003). Poulin, G., Nandakumar, R. & Ahringer, J. Genome-wide RNAi screens in Caenorhabditis elegans: impact on cancer research. Oncogene 23, 8340–8345 (2004). Reboul, J. et al. C. elegans ORFeome version 1.1: experimental verification of the genome annotation and resource for proteome-scale protein expression. Nat. Genet. 34, 35–41 (2003).

Proteomics approaches in the study of cell polarity in the nematode Caenorhabditis elegans

27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50.

Lamesch, P. et al. C. elegans ORFeome version 3.1: Increasing the coverage of ORFeome resources with improved gene predictions. Genome Res. 14, 2064–2069 (2004). Fields, S. & Song, O. A novel genetic system to detect protein-protein interactions. Nature 340, 245–246 (1989). Caufield, J. H., Sakhawalkar, N. & Uetz, P. A comparison and optimization of yeast two-hybrid systems. Methods San Diego Calif 58, 317–24 (2012). Tewari, M. et al. Systematic interactome mapping and genetic perturbation analysis of a C. elegans TGF-beta signaling network. Mol. Cell 13, 469–482 (2004). Astin, J., Merry, A., Rajan, J. & Kuwabara, P. E. Caenorhabditis elegans functional genomics: omic resonance. Brief. Funct. Genomic. Proteomic. 3, 26–34 (2004). Colland, F. & Daviet, L. Integrating a functional proteomic approach into the target discovery process. Biochimie 86, 625–632 (2004). Boxem, M. et al. A protein domain-based interactome network for C. elegans early embryogenesis. Cell 134, 534–45 (2008). Waaijers, S., Koorman, T., Kerver, J. & Boxem, M. Identification of Human Protein Interaction Domains using an ORFeome-based Yeast Two-hybrid Fragment Library. J. Proteome Res. 12, 3181– 3192 (2013). Macara, I. G., Guyer, R., Richardson, G., Huo, Y. & Ahmed, S. M. Epithelial Homeostasis. Curr. Biol. 24, R815–R825 (2014). Halaoui, R. & McCaffrey, L. Rewiring cell polarity signaling in cancer. Oncogene 1–12 (2014). doi:10.1038/onc.2014.59 St Johnston, D. & Ahringer, J. Cell polarity in eggs and epithelia: Parallels and diversity. Cell 141, 757–774 (2010). Kemphues, K. J., Priess, J. R., Morton, D. G. & Cheng, N. S. Identification of genes required for cytoplasmic localization in early C. elegans embryos. Cell 52, 311–320 (1988). Munro, E., Nance, J. & Priess, J. R. Cortical flows powered by asymmetrical contraction transport PAR proteins to establish and maintain anterior-posterior polarity in the early C. elegans embryo. Dev. Cell 7, 413–424 (2004). Motegi, F. & Sugimoto, A. Sequential functioning of the ECT-2 RhoGEF, RHO-1 and CDC-42 establishes cell polarity in Caenorhabditis elegans embryos. Nat. Cell Biol. 8, 978–985 (2006). Zonies, S., Motegi, F., Hao, Y. & Seydoux, G. Symmetry breaking and polarization of the C. elegans zygote by the polarity protein PAR-2. Dev. Camb. Engl. 137, 1669–1677 (2010). Hoege, C. et al. LGL can partition the cortex of one-cell Caenorhabditis elegans embryos into two domains. Curr. Biol. 20, 1296–1303 (2010). Boyd, L., Guo, S., Levitan, D., Stinchcomb, D. T. & Kemphues, K. J. PAR-2 is asymmetrically distributed and promotes association of P granules and PAR-1 with the cortex in C. elegans embryos. Dev. Camb. Engl. 122, 3075–3084 (1996). Labbé, J. C., Pacquelet, A., Marty, T. & Gotta, M. A genomewide screen for suppressors of par2 uncovers potential regulators of PAR protein-dependent cell polarity in Caenorhabditis elegans. Genetics 174, 285–295 (2006). Gotta, M., Abraham, M. C. & Ahringer, J. CDC-42 controls early cell polarity and spindle orientation in C. elegans. Curr. Biol. CB 11, 482–488 (2001). Joberty, G., Petersen, C., Gao, L. & Macara, I. G. The cell-polarity protein Par6 links Par3 and atypical protein kinase C to Cdc42. Nat. Cell Biol. 2, 531–539 (2000). Aceto, D., Beers, M. & Kemphues, K. J. Interaction of PAR-6 with CDC-42 is required for maintenance but not establishment of PAR asymmetry in C. elegans. Dev. Biol. 299, 386–397 (2006). Morton, D. G. et al. The Caenorhabditis elegans par-5 gene encodes a 14-3-3 protein required for cellular asymmetry in the early embryo. Dev. Biol. 241, 47–58 (2002). Morton, D. G., Roos, J. M. & Kemphues, K. J. par-4, a gene required for cytoplasmic localization and determination of specific cell types in Caenorhabditis elegans embryogenesis. Genetics 130, 771–790 (1992). Baas, A. F. et al. Complete polarization of single intestinal epithelial cells upon activation of LKB1 by STRAD. Cell 116, 457–466 (2004).

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74.

Benton, R. & St Johnston, D. Drosophila PAR-1 and 14-3-3 inhibit Bazooka/PAR-3 to establish complementary cortical domains in polarized cells. Cell 115, 691–704 (2003). Anderson, D. C., Gill, J. S., Cinalli, R. M. & Nance, J. Polarization of the C. elegans embryo by RhoGAP-mediated exclusion of PAR-6 from cell contacts. Science 320, 1771–4 (2008). Chan, E. & Nance, J. Mechanisms of CDC-42 activation during contact-induced cell polarization. J. Cell Sci. 126, 1692–702 (2013). Wang, S. et al. Tiam1 interaction with the PAR complex promotes talin-mediated Rac1 activation during polarized cell migration. J. Cell Biol. 199, 331–345 (2012). Goldstein, B. & Macara, I. G. The PAR proteins: fundamental players in animal cell polarization. Dev. Cell 13, 609–622 (2007). Tepass, U. Epithelial Cells: Regulation of Polarity, Junctions, Morphogenesis, Cell Growth, and Survival. Annu. Rev. Cell Dev. Biol. 28, 655–685 (2012). Tepass, U., Theres, C. & Knust, E. crumbs encodes an EGF-like protein expressed on apical membranes of Drosophila epithelial cells and required for organization of epithelia. Cell 61, 787–799 (1990). Bachmann, A., Schneider, M., Theilenberg, E., Grawe, F. & Knust, E. Drosophila Stardust is a partner of Crumbs in the control of epithelial cell polarity. Nature 414, 638–643 (2001). Roh, M. H. et al. The Maguk protein, Pals1, functions as an adapter, linking mammalian homologues of Crumbs and Discs Lost. J. Cell Biol. 157, 161–172 (2002). Kamberov, E. et al. Molecular cloning and characterization of Pals, proteins associated with mLin-7. J. Biol. Chem. 275, 11425–11431 (2000). Lemmers, C. et al. hINADl/PATJ, a homolog of discs lost, interacts with crumbs and localizes to tight junctions in human epithelial cells. J. Biol. Chem. 277, 25408–25415 (2002). Bachmann, A., Grawe, F., Johnson, K. & Knust, E. Drosophila Lin-7 is a component of the Crumbs complex in epithelia and photoreceptor cells and prevents light-induced retinal degeneration. Eur. J. Cell Biol. 87, 123–136 (2008). Bulgakova, N. A., Kempkens, O. & Knust, E. Multiple domains of Stardust differentially mediate localisation of the Crumbs-Stardust complex during photoreceptor development in Drosophila. J. Cell Sci. 121, 2018–2026 (2008). Armenti, S. T. & Nance, J. Adherens junctions in C. elegans embryonic morphogenesis. Subcell. Biochem. 60, 279–299 (2012). Li, P., Mao, X., Ren, Y. & Liu, P. Epithelial cell polarity determinant CRB3 in cancer development. Int. J. Biol. Sci. 11, 31–37 (2015). Whiteman, E. L. et al. Crumbs3 is essential for proper epithelial development and viability. Mol. Cell. Biol. 34, 43–56 (2014). Waaijers, S., Ramalho, J. J., Koorman, T., Kruse, E. & Boxem, M. The C. elegans Crumbs family contains a CRB3 homolog and is not essential for viability. Biol. Open 4, 276–284 (2015). Shibata, Y., Fujii, T., Dent, J. A., Fujisawa, H. & Takagi, S. EAT-20, a novel transmembrane protein with EGF motifs, is required for efficient feeding in Caenorhabditis elegans. Genetics 154, 635–646 (2000). Kaech, S. M., Whitfield, C. W. & Kim, S. K. The LIN-2/LIN-7/LIN-10 complex mediates basolateral membrane localization of the C. elegans EGF receptor LET-23 in vulval epithelial cells. Cell 94, 761–771 (1998). Stetak, A. et al. Cell fate-specific regulation of EGF receptor trafficking during Caenorhabditis elegans vulval development. EMBO J. 25, 2347–2357 (2006). Perego, C., Vanoni, C., Massari, S., Longhi, R. & Pietrini, G. Mammalian LIN-7 PDZ proteins associate with beta-catenin at the cell-cell junctions of epithelia and neurons. EMBO J. 19, 3978– 3989 (2000). Butz, S., Okamoto, M. & Südhof, T. C. A tripartite protein complex with the potential to couple synaptic vesicle exocytosis to cell adhesion in brain. Cell 94, 773–782 (1998). Bilder, D. & Perrimon, N. Localization of apical epithelial determinants by the basolateral PDZ protein Scribble. Nature 403, 676–680 (2000). Bilder, D., Li, M. & Perrimon, N. Cooperative regulation of cell polarity and growth by Drosophila tumor suppressors. Science 289, 113–116 (2000).

Proteomics approaches in the study of cell polarity in the nematode Caenorhabditis elegans

75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88. 89. 90. 91. 92. 93. 94. 95. 96. 97.

Bilder, D., Schober, M. & Perrimon, N. Integrated activity of PDZ protein complexes regulates epithelial polarity. Nat. Cell Biol. 5, 53–58 (2003). Köppen, M. et al. Cooperative regulation of AJM-1 controls junctional integrity in Caenorhabditis elegans epithelia. Nat. Cell Biol. 3, 983–991 (2001). Nakagawa, S. & Huibregtse, J. M. Human scribble (Vartul) is targeted for ubiquitin-mediated degradation by the high-risk papillomavirus E6 proteins and the E6AP ubiquitin-protein ligase. Mol. Cell. Biol. 20, 8244–8253 (2000). Legouis, R. et al. LET-413 is a basolateral protein required for the assembly of adherens junctions in Caenorhabditis elegans. Nat. Cell Biol. 2, 415–422 (2000). Bossinger, O., Klebes, a, Segbert, C., Theres, C. & Knust, E. Zonula adherens formation in Caenorhabditis elegans requires dlg-1, the homologue of the Drosophila gene discs large. Dev. Biol. 230, 29–42 (2001). Takizawa, S. et al. Human scribble, a novel tumor suppressor identified as a target of high-risk HPV E6 for ubiquitin-mediated degradation, interacts with adenomatous polyposis coli. Genes Cells 11, 453–464 (2006). Dow, L. E. et al. hScrib is a functional homologue of the Drosophila tumour suppressor Scribble. Oncogene 22, 9225–9230 (2003). Dow, L. E. et al. Loss of human Scribble cooperates with H-Ras to promote cell invasion through deregulation of MAPK signalling. Oncogene 27, 5988–6001 (2008). Wu, M., Pastor-Pareja, J. C. & Xu, T. Interaction between Ras(V12) and scribbled clones induces tumour growth and invasion. Nature 463, 545–548 (2010). Murdoch, J. N. et al. Circletail, a new mouse mutant with severe neural tube defects: chromosomal localization and interaction with the loop-tail mutation. Genomics 78, 55–63 (2001). Humbert, P., Russell, S. & Richardson, H. Dlg, scribble and Lgl in cell polarity, cell proliferation and cancer. BioEssays 25, 542–553 (2003). Humbert, P. O. et al. Control of tumourigenesis by the Scribble/Dlg/Lgl polarity module. Oncogene 27, 6888–6907 (2008). Roberts, S., Delury, C. & Marsh, E. The PDZ protein discs-large (DLG): The ‘Jekyll and Hyde’ of the epithelial polarity proteins. FEBS J. 279, 3549–3558 (2012). Van Campenhout, C. A. et al. Dlg3 trafficking and apical tight junction formation is regulated by nedd4 and nedd4-2 e3 ubiquitin ligases. Dev. Cell 21, 479–491 (2011). Wu, Q., DiBona, V. L., Bernard, L. P. & Zhang, H. The polarity protein partitioning-defective 1 (PAR-1) regulates dendritic spine morphogenesis through phosphorylating postsynaptic density protein 95 (PSD-95). J. Biol. Chem. 287, 30781–30788 (2012). Thomas, U., Phannavong, B., Müller, B., Garner, C. C. & Gundelfinger, E. D. Functional expression of rat synapse-associated proteins SAP97 and SAP102 in Drosophila dlg-1 mutants: effects on tumor suppression and synaptic bouton structure. Mech. Dev. 62, 161–174 (1997). Beatty, A., Morton, D. & Kemphues, K. The C. elegans homolog of Drosophila Lethal giant larvae functions redundantly with PAR-2 to maintain polarity in the early embryo. Dev. Camb. Engl. 137, 3995–4004 (2010). Beatty, A., Morton, D. G. & Kemphues, K. PAR-2, LGL-1 and the CDC-42 GAP CHIN-1 act in distinct pathways to maintain polarity in the C. elegans embryo. Dev. Camb. Engl. 140, 2005–2014 (2013). Cao, F., Miao, Y., Xu, K. & Liu, P. Lethal (2) Giant Larvae: An Indispensable Regulator of Cell Polarity and Cancer Development. Int. J. Biol. Sci. 11, 380–389 (2015). Winter, J. F. et al. Caenorhabditis elegans screen reveals role of PAR-5 in RAB-11-recycling endosome positioning and apicobasal cell polarity. Nat. Cell Biol. 14, 666–76 (2012). Balklava, Z., Pant, S., Fares, H. & Grant, B. D. Genome-wide analysis identifies a general requirement for polarity proteins in endocytic traffic. Nat. Cell Biol. 9, 1066–1073 (2007). Shafaq-Zadah, M., Brocard, L., Solari, F. & Michaux, G. AP-1 is required for the maintenance of apico-basal polarity in the C. elegans intestine. Dev. Camb. Engl. 139, 2061–70 (2012). Saotome, I., Curto, M. & McClatchey, A. I. Ezrin is essential for epithelial organization and villus morphogenesis in the developing intestine. Dev. Cell 6, 855–864 (2004).

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

98. 99. 100. 101. 102. 103. 104. 105. 106. 107. 108. 109. 110. 111. 112. 113. 114. 115. 116. 117. 118. 119. 120. 121. 122.

Dukes, J. D., Whitley, P. & Chalmers, A. D. The MDCK variety pack: choosing the right strain. BMC Cell Biol. 12, 43 (2011). Bryant, D. M. et al. A Molecular Switch for the Orientation of Epithelial Cell Polarization. Dev. Cell 1–17 (2014). doi:10.1016/j.devcel.2014.08.027 Bryant, D. M. & Mostov, K. E. From cells to organs: building polarized tissue. Nat. Rev. Mol. Cell Biol. 9, 887–901 (2008). Rodríguez-Fraticelli, A. E., Auzan, M., Alonso, M. a., Bornens, M. & Martín-Belmonte, F. Cell confinement controls centrosome positioning and lumen initiation during epithelial morphogenesis. J. Cell Biol. 198, 1011–1023 (2012). Shewan, A., Eastburn, D. J. & Mostov, K. Phosphoinositides in cell architecture. Cold Spring Harb. Perspect. Biol. 3, 1–17 (2011). Von Stein, W., Ramrath, A., Grimm, A., Müller-Borg, M. & Wodarz, A. Direct association of Bazooka/PAR-3 with the lipid phosphatase PTEN reveals a link between the PAR/aPKC complex and phosphoinositide signaling. Dev. Camb. Engl. 132, 1675–1686 (2005). Kovacs, E. M., Ali, R. G., McCormack, A. J. & Yap, A. S. E-cadherin homophilic ligation directly signals through Rac and phosphatidylinositol 3-kinase to regulate adhesive contacts. J. Biol. Chem. 277, 6708–6718 (2002). Zhang, H. et al. Apicobasal domain identities of expanding tubular membranes depend on glycosphingolipid biosynthesis. Nat. Cell Biol. 13, 1189–201 (2011). Pasti, G. & Labouesse, M. Epithelial junctions, cytoskeleton, and polarity. WormBook 1–35 (2014). doi:10.1895/wormbook.1.56.2 Welchman, D. P., Mathies, L. D. & Ahringer, J. Similar requirements for CDC-42 and the PAR-3/ PAR-6/PKC-3 complex in diverse cell types. Dev. Biol. 305, 347–357 (2007). Gillard, G. et al. Control of E-cadherin apical localisation and morphogenesis by a SOAP-1/AP-1/ clathrin pathway in C. elegans epidermal cells. Dev. Camb. Engl. (2015). doi:10.1242/dev.118216 Gleason, J. E. & Eisenmann, D. M. Wnt signaling controls the stem cell-like asymmetric division of the epithelial seam cells during C. elegans larval development. Dev. Biol. 348, 58–66 (2010). Ren, H. & Zhang, H. Wnt signaling controls temporal identities of seam cells in Caenorhabditis elegans. Dev. Biol. 345, 144–155 (2010). Leung, B., Hermann, G. J. & Priess, J. R. Organogenesis of the Caenorhabditis elegans intestine. Dev. Biol. 216, 114–134 (1999). Macqueen, A. J. et al. ACT-5 Is an Essential Caenorhabditis elegans Actin Required for Intestinal Microvilli Formation ?. 16, 3247–3259 (2005). Feldman, J. L. & Priess, J. R. A role for the centrosome and PAR-3 in the hand-off of MTOC function during epithelial polarization. Curr. Biol. CB 22, 575–582 (2012). Goulas, S., Conder, R. & Knoblich, J. A. The Par complex and integrins direct asymmetric cell division in adult intestinal stem cells. Cell Stem Cell 11, 529–540 (2012). Segbert, C., Johnson, K., Theres, C., van Fürden, D. & Bossinger, O. Molecular and functional analysis of apical junction formation in the gut epithelium of Caenorhabditis elegans. Dev. Biol. 266, 17–26 (2004). Carberry, K. et al. The novel intestinal filament organizer IFO-1 contributes to epithelial integrity in concert with ERM-1 and DLG-1. Dev. Camb. Engl. 139, 1851–62 (2012). Gloerich, M. et al. Rap2A links intestinal cell polarity to brush border formation. Nat. Cell Biol. 14, 793–801 (2012). Sigurbjörnsdóttir, S., Mathew, R. & Leptin, M. Molecular mechanisms of de novo lumen formation. Nat. Rev. Mol. Cell Biol. 15, 665–676 (2014). Barr, M. M. Caenorhabditis elegans as a model to study renal development and disease: sexy cilia. J. Am. Soc. Nephrol. JASN 16, 305–312 (2005). Nelson, F. K. & Riddle, D. L. Functional study of the Caenorhabditis elegans secretory-excretory system using laser microsurgery. J. Exp. Zool. 231, 45–56 (1984). Liégeois, S., Benedetto, A., Michaux, G., Belliard, G. & Labouesse, M. Genes required for osmoregulation and apical secretion in Caenorhabditis elegans. Genetics 175, 709–724 (2007). Buechner, M., Hall, D. H., Bhatt, H. & Hedgecock, E. M. Cystic canal mutants in Caenorhabditis elegans are defective in the apical membrane domain of the renal (excretory) cell. Dev. Biol. 214, 227–241 (1999).

Proteomics approaches in the study of cell polarity in the nematode Caenorhabditis elegans

123. Gao, J., Estrada, L., Cho, S., Ellis, R. E. & Gorski, J. L. The Caenorhabditis elegans homolog of FGD1, the human Cdc42 GEF gene responsible for faciogenital dysplasia, is critical for excretory cell morphogenesis. Hum. Mol. Genet. 10, 3049–3062 (2001). 124. Suzuki, N. et al. A putative GDP-GTP exchange factor is required for development of the excretory cell in Caenorhabditis elegans. EMBO Rep. 2, 530–535 (2001). 125. Mattingly, B. C. & Buechner, M. The FGD homologue EXC-5 regulates apical trafficking in C. elegans tubules. Dev. Biol. 359, 59–72 (2011). 126. Khan, L. a et al. Intracellular lumen extension requires ERM-1-dependent apical membrane expansion and AQP-8-mediated flux. Nat. Cell Biol. 15, 143–56 (2013). 127. Kolotuev, I., Hyenne, V., Schwab, Y., Rodriguez, D. & Labouesse, M. A pathway for unicellular tube extension depending on the lymphatic vessel determinant Prox1 and on osmoregulation. Nat. Cell Biol. 15, 157–68 (2013). 128. Armenti, S. T., Chan, E. & Nance, J. Polarized exocyst-mediated vesicle fusion directs intracellular lumenogenesis within the C. elegans excretory cell. Dev. Biol. 394, 110–121 (2014). 129. White, J. G., Southgate, E., Thomson, J. N. & Brenner, S. The structure of the nervous system of the nematode Caenorhabditis elegans. Philos. Trans. R. Soc. Lond. B. Biol. Sci. 314, 1–340 (1986). 130. Ou, C. Y. & Shen, K. Neuronal polarity in C. elegans. Dev. Neurobiol. 71, 554–566 (2011). 131. Teichmann, H. M. & Shen, K. UNC-6 and UNC-40 promote dendritic growth through PAR-4 in Caenorhabditis elegans neurons. Nat. Neurosci. 14, 165–172 (2011). 132. Chien, S.-C., Brinkmann, E.-M., Teuliere, J. & Garriga, G. Caenorhabditis elegans PIG-1/MELK acts in a conserved PAR-4/LKB1 polarity pathway to promote asymmetric neuroblast divisions. Genetics 193, 897–909 (2013). 133. Shelly, M. & Poo, M. M. Role of LKB1-SAD/MARK pathway in neuronal polarization. Dev. Neurobiol. 71, 508–527 (2011). 134. Lee, S. et al. Atypical protein kinase C and Par3 are required for proteoglycan-induced axon growth inhibition. J. Neurosci. Off. J. Soc. Neurosci. 33, 2541–2554 (2013). 135. Albertson, R., Chabu, C., Sheehan, A. & Doe, C. Q. Scribble protein domain mapping reveals a multistep localization mechanism and domains necessary for establishing cortical polarity. J. Cell Sci. 117, 6061–6070 (2004). 136. Vieira, V. et al. Differential regulation of Dlg1, Scrib, and Lgl1 expression in a transgenic mouse model of ocular cancer. Mol. Vis. 14, 2390–2403 (2008). 137. Satoh, K. et al. DAP-1, a novel protein that interacts with the guanylate kinase-like domains of hDLG and PSD-95. Genes Cells Devoted Mol. Cell. Mech. 2, 415–424 (1997). 138. Sanchez-Alvarez, L. et al. VANG-1 and PRKL-1 cooperate to negatively regulate neurite formation in Caenorhabditis elegans. PLoS Genet. 7, (2011). 139. Sawa, H. & Korswagen, H. C. Wnt signaling in C. elegans. WormBook Online Rev. C Elegans Biol. 1–30 (2013). doi:10.1895/wormbook.1.7.2

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 2

A combined binary interaction and phenotypic map of C. elegans cell polarity proteins

T. Koorman1, M. van der Voet1,2, D. Klompstra3,4, J. J. Ramalho1, S. Nieuwenhuize1, S. J. L. van den Heuvel1, J. Nance3,4, and M. Boxem1,5

In revision for Nature Cell Biology

1. Division of Developmental Biology, Department of Biology, Faculty of Science, Utrecht University, The Netherlands.

2. Present address: Department of Human Genetics, Donders Institute for Brain, Cognition

and Behaviour, Radboud university medical center, Nijmegen, The Netherlands.

3. Helen L. and Martin S. Kimmel Center for Biology and Medicine at the Skirball Institute

of Biomolecular Medicine, NYU School of Medicine, New York, NY 10016, USA.

4. Department of Cell Biology, NYU School of Medicine, New York, NY 10016, USA.

5. Correspondence should be addressed to M.B. (e-mail: m.boxem@uu.nl)

Abstract

he establishment of cell polarity is an essential process for the development of multicellular organisms and the functioning of cells and tissues. Here, we combine large-scale protein interaction mapping with systematic phenotypic profiling, to study the network of physical interactions that underlies polarity establishment and maintenance in the nematode Caenorhabditis elegans. Using a fragment-based yeast two-hybrid strategy, we identified 439 interactions between 296 proteins, as well as the protein regions that mediate these interactions. Phenotypic profiling of the network resulted in the identification of 100 physically interacting protein pairs for which RNAi-mediated depletion caused a defect in the same polarity-related process. We demonstrate the predictive capabilities of the network by showing that a newly discovered physical interaction between the RhoGAP PAC1 and PAR-6 is required for radial polarization of the C. elegans embryo. Our network represents a valuable resource of candidate interactions that can be used to further our insight into cell polarization.

A combined binary interaction and phenotypic map of C. elegans cell polarity proteins

Introduction The ability to generate an axis of polarity is a fundamental property of cells. Polarity is a prerequisite for cell migration and asymmetric cell divisions, and many cell types need to establish functionally distinct domains along a polarity axis in order to perform their functions. The loss of epithelial polarity is associated with a number of diseases, including inflammatory bowel disease1 and retinal dystrophies2. Moreover, all invasive solid tumors have lost epithelial polarity and integrity, and it is becoming increasingly likely that polarity loss causally contributes to tumorigenesis3â&#x20AC;&#x201C;6. Studies in the nematode Caenorhabditis elegans and the fruit fly Drosophila melanogaster have identified a series of cortically localized proteins that can drive the establishment of polarity. The PAR3/PAR6/aPKC and Crumbs/SDT/PATJ complexes together promote apical domain identity7â&#x20AC;&#x201C;9. The LGL/SCRIB/DLG proteins, which were originally identified as tumor suppressors10â&#x20AC;&#x201C;12, and the Yurt/ Coracle/Neurexin IV/NaK-ATPase group, promote basolateral identity8,13. These cortical polarity complexes act together with a number of different components, such as Rho family GTPases, junctional components, and cytoskeletal linkers of the ERM family, in establishing polarity. Although it is clear that mutual exclusion is a key mechanism by which cortical polarity regulators establish polarity14,15, we still lack a detailed mechanistic understanding of how the cortical polarity regulators are segregated into distinct domains. Moreover, we know little of the mechanisms through which cortical polarity is integrated with cellular events such as cytoskeletal rearrangement, organization of a polarized trafficking machinery, and functional specialization of membrane domains. A full understanding of polarity establishment will require a comprehensive knowledge of the proteins involved in this process and the molecular interactions between them. Large-scale protein interaction mapping techniques excel at identifying potential functional interactions between proteins, and the proteins involved in these interactions can be functionally examined in detail in vivo. Here, we combine largescale protein interaction mapping with phenotypic profiling to study the network of physical interactions that underlies polarity establishment and maintenance in the nematode Caenorhabditis elegans. Starting with 69 evolutionarily conserved polarity regulators, we established a C. elegans polarity interaction network (CePIN) of 439 interactions between 296 proteins. To increase the completeness of the interaction map and to systematically identify the domains responsible for the interactions, we used multiple fragments for each bait protein and screened both a traditional cDNA library and a library designed specifically to map interaction domains16. We

combined the physical interaction map with phenotypic profiling of the proteins in the network by RNAi depletion in a series of polarity marker strains. This revealed 100 protein pairs that exhibited a phenotypic defect in the same polarity related process. These pairs are strong candidates for a functional interaction in vivo. To demonstrate the value of our approach, we studied the interaction identified between PAR-6 and PAC-1/ARHGAP21 in detail. The physical interaction proved important for the function of PAC-1 in radially polarizing PAR-6 in early embryonic cells. The protein interaction network we identified and its phenotypic characterisation provide a resource for future studies into cell polarity, and will contribute to our understanding of this essential process.

Materials and methods Y2H Vectors Two new Y2H vectors were generated: the Gal4-DB vector pMB28 and the Gal4AD vector pMB29, derived from parent vectors pPC97 and pPC86 respectively1,2. Both encode a flexible linker (GGGG) upstream of the cloned ORF, and facilitate cloning using AscI and NotI. pMB28 also contains the CAN1 gene to allow counter selection on plates containing canavinine, while pMB29 contains the CYH2 gene to allow counter selection on plates containing cycloheximide. An oligonucleotide linker containing the flexible linker and AscI and NotI restriction sites was inserted into the parent vectors digested with SmaI and SacI. To insert CAN1 into pMB28, a 2.3 kbp fragment containing the wild-type CAN1 gene flanked by ApaI restriction sites was PCR amplified from yeast genomic DNA, and cloned into the pPC97 ApaI site (primers used: 5’-aaGGGCCCgaagagtggttgcgaacagag and 5’-aaGGGCCCtggttctaggttcgggtgac). To insert CYH2 into pMB29, an EcoRV fragment containing CYH2 was cut from Clontech vector pAS2-1, and ligated into vector digested with Acc65I and treated with Klenow + dNTPs to generate blunt ends. Generating full-length DB-ORF clones For each of the 69 polarity genes we designed primer pairs to amplify the complete coding sequence of the longest splice variant predicted in Wormbase version WS224. Forward primers contained an AscI site immediately upstream of the ATG (5’-aGGCGCGCC-ORF), and reverse primers a NotI site immediately downstream of the stop codon (5’-aGCGGCCGC-Stop-ORF). Gene amplification was performed on 100 ng of a mixed stage C. elegans cDNA library using High Fidelity Hot Start KOD

A combined binary interaction and phenotypic map of C. elegans cell polarity proteins

polymerase (Novagen). We used a 50 µl PCR reaction mixture and the following amplification program: 95°C for 2:00 min, 35 × (95°C for 0:15 min, 58°C for 0:30 min, 68 °C for 1:00 min for each kbp to be amplified), 68°C for 8:00 min. PCR products of the correct size were purified from an agarose gel, digested with AscI/NotI, and ligated into pMB28 linearized with AscI and NotI. After transformation into E. coli strain DH5α, correct clones were identified by sequencing with internal gene specific primers and two external primers (DB: 5’-GGCTTCAGTGGAGACTGATATGCCTC and TERM: 3’-CGCCAGAGGTTTGGTCAAGTCTCC). For each gene, both a glycerol stock and miniprep of a fully sequence verified clone were stored. Generating fragment DB-ORF clones We manually designed primers to amplify fragments of each wild-type full-length polarity clone. Fragment design was based on the size of the gene and the protein domain architecture as predicted by Pfam (http://pfam.sanger.ac.uk/) and SMART (http://smart.embl-heidelberg.de/). PCR amplification was performed as above for full-length proteins, using 50 ng of a mix of 65 sequence-verified wild-type fulllength polarity clones as template in a 20 µl reaction volume. For 7 genes for which we were not able to obtain a full-length clone, 100 ng of the mixed stage C. elegans cDNA library was used as template. PCR products of the correct size were purified from an agarose-gel, digested with AscI/NotI, and ligated into pMB28 linearized with AscI and NotI. After transformation into E. coli strain DH5a, colony PCR using primers DB and TERM was used to identify clones carrying an insert of correct size. Colony PCRs were performed using TaqBlue Recombinant (Fermentas) according to the manufacturer’s instructions. Six positive clones of the same bait fragment were pooled, and cultured for a glycerol stock and miniprep. Yeast strains and transformation Y2H experiments were done with the S. cerevisiae strains Y8800 (MATa) and Y8930 (MATα)3. The strains contain three Gal4p inducible reporter genes: GAL1::HIS3, providing growth on plates lacking histidine, GAL7::LacZ, for colorimetric detection of Gal4P activity, and GAL-2::ADE2, providing growth on media lacking adenine, as well as a colorimetric indication of Gal4p activity. The genotype of the Y8800 and Y8930 strains is MAT(a or α) leu2-3,112 trp1-901 ura3-52 his3-200 gal4Δ gal80Δ cyh2R GAL1::HIS3@LYS2 GAL2::ADE2 GAL7::lacZ@met2. All yeast transformations were done using the Te/LiAc transformation method4.

Generation of bait strains for Y2H screens Full-length and fragment Gal4-DB::ORF fusions in vector pMB28 were transformed into yeast strain Y8930 and successful transformants were selected on SD –Leu plates. All bait strains were then plated on SD –Leu –His plates to test for reporter gene activation in the absence of AD-plasmid. Bait strains able to grow on SD –Leu –His plates were discarded as auto-activators. Generation of Y8800 library stocks Two Y2H AD libraries were screened: a previously generated library carrying fragments of 749 genes required for early embryogenesis5, and a mixed-stage cDNA library cloned in a vector carrying the wild-type CYH2 gene for counter selection (a gift from X. Xin and C. Boone, University of Toronto). Yeast strain Y8800[ref. 3] was transformed with 30 µg of each library, and was plated on SD –Trp plates to select transformants. After three days of growth at 30°C, all colonies were harvested in YEPD medium containing 20% glycerol, and frozen in aliquots of 1 ml with an estimated optical density at 600 nm (OD600) of 30. Library screening All library screens were done using the mating approach6. An aliquot of AD-library yeast was thawed and diluted to an OD600 of 3 (AD-Fragment library) or 6 (ADcDNA library). One ml of library yeast was mixed in a 2 ml Eppendorf tube with an equal volume of bait yeast grown overnight (O/N) in 10 ml YEPD in a 50 ml conical bottom tube in a 30°C shaking incubator, diluted to the same OD as the library prior to mixing. Mixed yeast were briefly spun down (30 sec at 2000 g), resuspended in 300 ml H2O and plated on a 10 cm YEPD-agar plate. Yeast were allowed to mate for 4 hours at 30°C. After mating, the yeast was scraped off the YEPD-agar plate in 2 ml of H2O, briefly spun down, resuspended in 600 ml H2O and plated in equal amounts on two 15 cm SD –Leu –His –Trp plates. To check mating efficiency, a 1:10,000 dilution was plated on an SD –Leu –Trp plate. After growth for 4 days on the SD –Leu –His –Trp plates, colonies were picked into a 96-well plate containing 25 ml of H2O (a maximum of 96 colonies were picked per bait), and 4 ml of yeast suspension was spotted on two fresh SD –Leu –His –Trp plates. Control yeasts containing protein pairs of known reporter activity strength were added to these plates. After two days of growth at 30°C, yeast from one of the plates was stored as a back-up in YEPD medium containing 20% glycerol. The other plate was replica plated to four different assay plates: an SD –Leu–Trp –His plate to validate activation of the HIS3 reporter, an SD –Leu–Trp –His + 2 mM 3-Amino-1,2,4-triazole (3-AT) plate to test activation of the HIS3 reporter gene under more stringent conditions (3-AT inhibits the product of

A combined binary interaction and phenotypic map of C. elegans cell polarity proteins

the HIS3 gene), an SD –Leu –Trp –Ade plate to gauge activation of the ADE2 reporter gene, and finally an SD –Leu –His + 1 mg/ml cycloheximide plate to identify de novo auto-activators arising during the screen by measuring the activation of the HIS3 reporter gene after elimination of the AD plasmid through counter selection7. All plates were kept at 30°C for two days before scoring of the phenotypes, with exception of the SD –Leu –Trp –Ade plates which were kept for 5 days at room temperature before scoring. Yeast lysis and PCR amplification of plasmid inserts A swath of yeast cells, approximately 5 µl in volume, was transferred from a fresh YEPD plate grown O/N at 30°C to 20 µl of yeast lysis buffer (0.1 M NaPO4 buffer pH 7.4, 2.5 mg/ml Zymolase 20T, Seikagaku Corporation) in a 96 well plate. Plates were incubated in a PCR machine for 10 min at 37°C followed by 10 min at 94°C. For PCR amplification of plasmid inserts, we used High Fidelity Hot Start KOD Polymerase (Novagen), and 2 µl of lysed yeast as template in a 20 µl reaction volume, running the following amplification program: 95°C for 2:00 min, 35 × (95°C for 0:15 min, 58°C for 0:30 min, 68°C for 5:00 min), 68°C for 8:00 min. Use of a 5 min elongation time was essential for efficient amplification. Primers used were DB: 5’-GGCTTCAGTGGAGACTGATATGCCTC as forward primer for pMB28, AD: 5’-CGCGTTTGGAATCACTACAGGG as forward primer for pMB29 and TERM: 5’-CGCCAGAGGTTTGGTCAAGTCTCC as reverse primer for both vectors. DNA sequencing All DNA sequencing was performed by Macrogen Europe (http://dna.macrogen. com). Sequence analysis All yeast colonies identified in the library screening that activate the HIS3 reporter and do not show auto-activation of the HIS3 reporter in the absence of the AD::ORF plasmid were selected for sequence analysis. From each yeast colony, the ORF sequence contained in the Gal4::AD plasmid was amplified by PCR using primers AD and TERM as described above, and sent for Sanger sequencing using primer AD. All sequences were analyzed by phred8,9, using the default settings and the – trim_alt flag. Bases with quality scores below the default threshold of 0.5 were eliminated from analysis. Next, vector ends were identified and removed from the traces by searching each trace for an exact match with the 10 bp sequences flanking the insertion site. Sequences where no 5’ vector end could be found were eliminated from further analysis. Next, the identity of the ORF was determined by DNA BLAST

analysis of the sequence using an in-house BLAST database containing all protein coding genes present in Wormbase release WS235. To determine whether the ORF was in frame with the Gal4::AD sequence, we determined the identity of the ORF by protein BLAST using the translated protein sequence. A trace was considered in frame if the DNA and protein blast identities match, or when no stop codon was encountered in the first 100 codons of the sequence trace. Out of frame sequences were eliminated from further analysis. For display of identified interactors in a genome browser style web interface, we determined the genome coordinates of each sequence. If the sequence exactly matched a predicted splice variant present in WS235, the coordinates of that splice variant were used, truncated on the 5’ or 3’ end if the sequence was not full length. If no exact match could be found, for example because we identified a novel splice variant, coordinates were generated using BLAT (http://www.kentinformatics.com/). Confirmation of bait identity For each interaction identified multiple times, we confirmed the bait protein identity by PCR amplifying and sequencing of the bait construct from a representative yeast colony. Retesting by plasmid isolation Protein pairs identified a single time from the AD-cDNA library were retested by isolating the Gal4-AD plasmid from the original yeast clone identified in the Y2H-screen. Plasmid DNA was isolated from yeast by using a modified bacterial miniprep kit procedure (Machery-Nagel Nucleospin) procedure. A patch of yeast was inoculated in 10 ml of YEPD medium and incubated overnight at 30°C in an incubator rotating at 220 rpm. Two times 2 ml of the culture was pelleted for 30 seconds at 2000 g in a 2 ml Eppendorf tube. After removing the supernatant, 250 ml A1 buffer was added together with 350 mg of 0.02 mm glass beads. Yeast were then mixed on an Eppendorf tube mixer for 5 minutes at 1400 rpm. For the rest of the procedure the manufacturers protocol was followed. The final purification contains a mixture of Gal4-AD and Gal4-DB plasmid at low concentration. To obtain suitable amounts of the Gal4-AD plasmid, 5 µl of the plasmid prep was transformed into E. coli strain DH5α and plated on LB plates containing 50 µg/ml Ampicillin. Colonies containing Gal4-AD plasmid were then identified by colony PCR using primers AD and TERM, and plasmid was purified using a standard miniprep procedure. Plasmid was transformed into yeast strain Y8800, mated with the corresponding Y8930 bait strain, and assayed for reporter activity as in the library screening procedure. Protein pairs that failed to activate Y2H reporters were not included in the dataset.

A combined binary interaction and phenotypic map of C. elegans cell polarity proteins

Network graphs Graphical network representations were plotted using Cytoscape version 2.8. Manual adjustments were made with Adobe Illustrator version Creative Suite 6. PubMed searches PubMed searches for interactions between C. elegans proteins were guided by automated searches of the Textpresso database (http://www.textpresso.org/ nematode/) for the co-occurrence within the same publication of interacting pairs where both bait and prey proteins have associated gene names in Wormbase (all published genes are assigned names). Publications in which both proteins of an interacting pair are mentioned were then manually read to determine if an interaction was published. We also manually searched PubMed for interactions between homologous proteins in other organisms, using candidate homologs identified as described above. Comparison with IntAct To identify interactions present in IntAct (https://www.ebi.ac.uk/intact/), we used a Perl script to query IntAct release 27 downloaded into a local MySQL table. To identify orthologous interactions, we compiled a list of potential homologs of the C. elegans proteins in our network from Homologene build 67 (http://www.ncbi. nlm.nih.gov/homologene), Inparanoid release 7 (http://inparanoid.sbc.su.se/cgibin/index.cgi), Orthomcl version 5 (http://orthomcl.org/orthomcl/), and Ensembl release 73 (http://www.ensembl.org/index.html). All candidate homologs were used to query the database, and hits were manually curated to identify interactions between likely true homologs (as opposed to, for example, an interaction involving a related protein in the same protein family). As the IntAct database identifies proteins mainly by their Uniprot IDs, all protein names were converted to Uniprot IDs using the ID mapping table of the October 2013 Uniprot release (http://www.uniprot. org/). Search spaces used for computational analysis Several of the analyses are based on a comparison between the protein interactions we experimentally identified, and all protein pairs between which an interaction could be identified (the search space). We screened two very different AD library types with correspondingly different search spaces: a cDNA library, which potentially contains all proteins encoded by the C. elegans genome, and the early embryogenesis fragment library generated previously5, which encodes fragments of 749 proteins. The proteins present in the AD-Fragment library were identified

based on their shared role in early embryonic development, and already have very similar expression profiles and gene ontology descriptions. For this reason, all of the analyses were performed separately for interactions identified in these two library types. The search spaces were defined as follows: for the AD-cDNA library screens, all possible pairs between each of the 69 bait proteins and all predicted C. elegans proteins; for the AD-Fragment library screens, all possible pairs between the 69 bait proteins and the 749 proteins present in the library. GO term analysis We used the Hybrid Relative Specificity Similarity based on Gene Ontology (HRSS) software package10 in combination with GO database release 201310 (http://www. geneontology.org/) to calculate semantic similarity of GO terms scores for protein pairs. We limited our analysis to the Biological Process domain of the ontology. To determine whether a set of protein interactions was enriched for similar GO terms, we compared the distribution of similarity scores for interacting protein pairs with the distribution of scores for the pairs present in the corresponding search space. A custom R script was used to plot the distribution of GO similarity scores, and to calculate the enrichment of pairs with a high (0.9 â&#x20AC;&#x201C; 1.0) similarity among interacting proteins compared to the search space. P-values were calculated using the two-sided Fisherâ&#x20AC;&#x2122;s exact test, and the enrichment values shown correspond to the relative risk. Gene expression profiling comparison For comparison of gene expression profiles, we calculated the pair-wise Pearson Correlation Coefficients between all gene pairs across a large compendium of gene expression microarray data. C. elegans gene expression microarray data collected in Wormbase release WS236 were downloaded from the C. elegans SPELL website (http://spell.caltech.edu:3000). Each dataset was processed by a Perl script to obtain uniformly formatted data files with a single row of data for each protein-coding gene present in the dataset. For each dataset, we generated normalized Fisher z transformed pair-wise correlation scores as described in Hibbs et al.11. Briefly, missing values were imputed using the KNN impute algorithm with K = 10 using Euclidean distance. Next, pair-wise Pearson Correlation Coefficients were calculated within the left singular vectors obtained by singular vector decomposition (SVD) of the imputed microarray data. After Fisher z-transformation, the correlation scores were normalized by subtracting the mean z correlation and dividing by the SD of the z correlations. From this analysis, we excluded datasets with fewer than 5 different conditions. We also excluded several datasets where many genes had exactly the same expression levels in all conditions by using only datasets where >90% of genes have

A combined binary interaction and phenotypic map of C. elegans cell polarity proteins

a unique sum total of expression values. To enable comparison of gene pairs across the compendium of expression data, the average of all z correlations was calculated, and converted back to PCC values. Gene pairs which had data in less than 50% of the datasets from the final matrix of PCC were excluded. The final compendium contains data for 17,287 protein-coding genes compiled from 110 datasets and 2394 conditions. All calculations and plotting of PCC value distributions were done in the R language12. Comparison with WormNet The WormNet v3 dataset13 was downloaded into a local MySQL table, and a Perl script was used to query this table for the presence or absence of gene pairs, and to extract the log-likelihood score (LLS) measuring the probability of a true functional linkage. Enrichment values correspond to the relative risk, derived by comparing the fraction of interactions identified by Y2H that are present in WormNet, with the fraction of the search space pairs present in WormNet. Corresponding p-values were calculated in the R language using the two-sided Fisher’s exact test. The average LLS for protein pairs present in WormNet was also calculated. Mammalian expression constructs Three mammalian cell culture expression vectors were used in this study: pCI-NEOBirA14 was used to express the BirA biotin ligase, Avi-mCherry-C1[ref. 15] to express proteins tagged with mCherry and the Avi tag (which is recognized and biotinylated by BirA), and pEGFP-C1 (Clontech) to express proteins tagged with EGFP. To clone ORFs into these vectors, sequences were PCR amplified from 100 ng of a mixed stage C. elegans cDNA library using High Fidelity Hot Start KOD polymerase (Novagen). PCRs were done in a 50 µl volume, using the following amplification program: 95°C for 2:00 min, 35 × (95°C for 0:15 min, 58°C for 0:30 min, 68°C for 1:00 min for each kbp to be amplified), 68°C for 8:00 min. PCR products were gel-purified, digested with appropriate enzymes, and cloned into linearized Avi-mCherry or pEGFP-C1 vectors. Enzyme combinations used were BglII/Acc65I, Xho/HindIII, SalI/Acc65I, SalI/BamHI and SmaI/XmaI. Restriction sites were added to the PCR primers. Co-affinity purification Co-affinity purifications were done as described16. Briefly, for every protein pair to be tested, 3 sets of plasmids were transfected into HEK293 cells. 1: Avi-mCherrybait (7 µg) + empty EGFP vector (7 µg) + BirA (5 µg), 2: empty Avi-mCherry vector (7 µg) + EGFP-prey (7 µg) + BirA (5 µg), 3: Avi-mCherry-bait (7 µg) + EGFP-prey (7 µg) + BirA (5 µg). After 18-24 h of growth, Avi-tagged proteins were purified

from cell lysates using 25 µl of Dynabeads M-280 streptavidin (Life technologies/ Invitrogen). Beads were washed three times in wash buffer (20 mM Tris-HCl, pH 8.0, 150 mM NaCl, 0.5% Triton X-100, and protease inhibitors; Roche), and bound proteins were eluted with 2x SDS sample buffer. Protein samples were separated on 10% acrylamide gels, and subjected to western blotting. EGFP was detected using rabbit polyclonal anti-GFP (Abcam ab6556, 1:1000), and biotinylated Avi-mCherry was detected using Streptavidin coupled to horseradish peroxidase (Pierce High Sensitivity Streptavidin HRP, 1:20,000). Defining Minimal regions of interaction Minimal regions of interaction (MRIs) for each interaction are defined as the region of the prey protein that is present in all of the fragments found to interact with the bait protein. For each interacting prey clone found, we determined the start and stop position relative to the full-length protein. When multiple splice variants were predicted for a gene, the protein encoded by the splice variant we identified most frequently was used. Maintenance of C. elegans C. elegans strains were cultured under standard conditions17. All RNAi experiments were performed at 20°C, except for experiments using strain JH2647, which were performed at 25°C. The following strains were used: N2: wild-type, BOX56: mibIs31[dlg-1::GFP-Avi + pMyo-3::mCherry]V, CX9797: kyIs445[des-2::mCherry::RAB-3; des-2::SAD-1::GFP; odr-1::DsRED], DH1033: sqt-1(sc103) II; bIs1[vit-2::GFP + rol6(su1006)] X, JH2647: axIs1928 [mCherry::par-6]; axIs1182 [GFP::par-2], ML846: vha-5(mc38) IV; mcEx337[vha-5(+)::GFP + rol-6(su1006)], MZE1: unc-119(ed3) III; cbgIs91[Ppept-1::pept-1::DsRed + unc-119(+)]; cbgIs98[Ppept-1::GFP::rab-11.1; unc119(+)], RT408: unc-119(ed3) III; pwIs116 [rme-2::GFP + unc-119(+)], STR62: hrtIs3 [des2::myristoyl::GFP, unc-122::DsRed], SV1009: heIs63[Pwrt-2::GFP::PH; Pwrt-2::GFP::H2B; Plin-48::mCherry], JJ1404: unc-32(e189) III, FT992: pac-1(xn6) unc-32 (e189) III, FT1456: xnIs503[Ppie-1::GFP:pac-1ΔPAR-6 binding domain]; pac-1(xn6) unc-32 (e189) III. C. elegans RNAi constructs RNAi clones were primarily derived from the genome wide Vidal full-length HT115 RNAi feeding library derived from the ORFeome 3.1 collection18. When not available, we used a clone from the genome wide Ahringer fragment HT115 RNAi feeding library19. RNAi clones for all genes used as baits in the Yeast 2-Hybrid screens were sequence verified. Some RNAi clones where not available in either library, and were constructed by subcloning either full-length or 1kb of the corresponding cDNA into a

A combined binary interaction and phenotypic map of C. elegans cell polarity proteins

modified L4440 RNAi feeding vector which has an oligonucleotide linker containing a flexible linker and AscI and NotI restriction sites. Genes were inserted using enzymes AscI/NotI. To confirm RNAi clones, L4440 plasmids from single colonies were isolated and sequenced with forward primer 5’-TAATACGACTCACTATAGGG, or reverse primer 3’-GTAATACGACTCACTATAGGG. A HT115 bacterial clone expressing the L4440 vector with no insert was used as a control in feeding experiments. C. elegans RNAi by feeding All RNAi experiments were performed with worm strains maintained at 20°C, except for strain JH2647, which was maintained at 25°C, since germline expression of the mCherry::PAR-6 transgene was not detectable at 20°C. Bacteria were cultured in Lysogeny Broth (LB) supplemented with 100 mg/ml ampicilin (Amp) and 2.5 mg/ml tetracyclin (Tet) at 37°C in an incubator rotating at 200 rpm. HT115 bacteria carrying the desired RNAi clones were cultured in 2 ml volume for 6-8 h in a 10 ml round bottom propylene tube, before being transferred to 10 ml medium in a 50 ml conical tube for overnight culturing. The next day, Isopropyl β-D-1-thiogalactopyranoside (IPTG) was added to a final concentration of 1 mM, and bacteria were cultured for 1 h before harvesting, to induce production of T7 polymerase. Bacterial cultures were pelleted by centrifugation at 3000 g for 10 min, and taken up in 2 ml of LB, resulting in a 5x concentrated bacterial suspension. Nematode Growth Media (NGM) agar plates supplemented with 100 μg/ml Amp, 2.5 μg/ml Tet, and 1 mM IPTG were seeded with 250 μl of bacterial suspension, and kept at room temperature (RT) for 48 hr before placing worm on them. L4 hermaphrodites were placed on agar plates seeded with bacteria expressing double-stranded RNA against a target gene. Phenotypes were then analyzed in the F1 generation at the L4 stage, with the exception of the mCherry::PAR-6 embryonic polarity marker line, which was screened at the embryonic stage, and the yolk trafficking strain, which was screened at the young adult stage. For genes that caused an embryonic lethal phenotype and could not be assayed in larvae, we repeated the inactivation, starting the feeding RNAi at either the L1 and examining the L4 or adult stages of the same generation. Similarly, several genes caused severe germline defects when inactivated, preventing an examination of yolk protein secretion and uptake. These genes were re-screened starting the feeding procedure in the L4 stage and analyzing adults of the same generation. For genes that caused larval arrest and needed to be assessed in either young adult or embryonic stage, we initiated the feeding RNAi at the L3 stage.

Fluorescence microscopy and live imaging Microscopy was performed at 20°C. Two microscope setups were used: a wide-field immunofluorescence microscope and a spinning disc confocal microscope. The widefield fluorescence microscope is a Zeis Axioplan2 upright microscope equipped with a HAL 100 halogen visible light source controlled by an internal shutter, an HXP 120 C metal halide fluorescence light source controlled by a Uniblitz VMM-D1 shutter, a 25x/0.8 NA imm Corr objective with DIC slider, 63x and 100x Plan-APOCHROMAT 1.4 NA objectives with DIC sliders, a DIC polarizer, and an Axiocam MRm CCD monochrome camera. The filter turret is equipped with the following Zeiss filter sets: filter set 34 for DAPI (excitation: 390/22 nm band pass, dichroic mirror: 420 nm, emission: 460/50 nm band pass), filter set 47 for CFP (excitation: 436/20 nm band pass, dichroic mirror: 455 nm, emission: 480/40 nm band pass), filter set 13 for GFP (excitation: 470/20 nm band pass, dichroic mirror: 495 nm, emission: 505 – 530 nm band pass), filter set 46 for YFP (excitation: 500/20 nm band pass, dichroic mirror: 515 nm, emission: 535/30 nm band pass), filter set 31 for mCherry (excitation: 565/30 nm band pass, dichroic mirror: 585 nm, emission: 620/60 nm band pass), and a DIC analyzer cube. The microscope and camera are controlled by Zeiss Axiovision 4.x imaging software. The spinning disc platform consists of a Nikon Ti-U inverted microscope with a motorized XY stage and a Piezo Z stage, 60x and 100x PLAN APO 1.4 NA oil objectives, a Yokogawa CSU-X1 spinning disk unit equipped with a dual dichroic mirror set for laser wavelengths 488 nm and 561 nm, 488 nm and 561 nm solid state 50 mW lasers controlled by an Andor revolution 500 series AOTF Laser modulator and combiner, Semrock 512/23 + 630/91 dual band pass emission filter, Semrock 525/30 single band pass emission filter, Semrock 617/73 single band pass filter, Semrock 4800 long pass filter (500 – 1200 pass), and an Andor iXON DU885 monochrome EMCCD+ camera. All components are controlled by MetaMorph Microscopy Automation & Image Analysis Software. Images were processed further using Adobe Photoshop or ImageJ. For live imaging of L3 and young adult hermaphrodites, animals were sedated with 10 mM Levamisole M9 and mounted on 5% agarose. To perform imaging on embryos, gravid adults were dissected in 10 mM Levamisole in M9 and mounted on 5% agarose. Phenotype clustering RNAi phenotypes were manually scored on a scale of 0 (no abnormal phenotype) to 5 (severely abnormal phenotype). Clustering was done using the Cluster 3.0 software (http://bonsai.hgc.jp/~mdehoon/software/cluster/software.htm), using Spearman rank correlation distance measure and the complete linkage hierarchical clustering method. Clustering results were visualized using Java Treeview (http:// sourceforge.net/projects/jtreeview/). 50

A combined binary interaction and phenotypic map of C. elegans cell polarity proteins

Co-localization in HeLa cells Constructs were transfected in HeLa cells grown in DMEM/Ham’s F10 (50/50%) medium containing 10% FCS and 1% penicillin/streptomycin at 37°C and 5% CO2. Two days before transfection, nearly confluent cells were plated at 1:10 in 12-wells plates on 24 mm glass coverslips. Cells were transfected with 1 mg of Avi-mCherrybait and 1 mg of pEGFP-C1-prey with polyethyleminine (PEI). After overnight incubation, cells were washed with PBS, fixed for 10 min with 4% paraformaldehyde in PBS at room temperature, washed with PBS and subsequently mounted on slides in Vectashield mounting medium with DAPI.

Results Identification of the C. elegans polarity interaction network To generate a map of interactions underlying polarity establishment in C. elegans, we selected 69 proteins known to control cell polarity in C. elegans, or that are homologous to known polarity regulators in other organisms (Fig. 1a and Table 1 (page 54)). The selected proteins include members of the PAR, Crumbs and Scribble groups of polarity regulators, cell junction components, FERM domain containing cytoskeletal linkers, Rho GTPases, signalling components, and proteins that translate cortical polarity cues to polarized cellular processes such as spindle orientation. We also included several planar cell polarity proteins, for which little functional data is available in C. elegans. Finally, several components of the Wnt/β-catenin asymmetry pathway, which plays a key role in asymmetric cell divisions in C. elegans17, were also included. For each polarity protein coding gene, a yeast two-hybrid (Y2H) bait construct was generated by cloning the full-length open reading frame (ORF) into a Gal4::DB vector. We previously showed that a greater number of interactions can be detected when multiple fragments of a protein are tested in the Y2H system16. Therefore, up to 12 additional fragment bait constructs were cloned for each protein to maximize the completeness of the polarity interaction map. Fragments were manually designed based on the size of the gene and the predicted protein domain composition (Fig. 1b and Supplementary Fig. 1). We were able to successfully generate 65 full-length bait clones, and 338 fragment bait clones, which cover all of the 69 polarity regulators (Supplementary Table 1). Each of the bait constructs was transformed into yeast, and baits able to auto-activate the HIS3 reporter were discarded (Supplementary Table 1).

Table 1. Bait proteins used to generate the polarity interaction network Apical polarity proteins PAR-3 Par3 PAR-6 Par6 PKC-3 aPKC CRB-1 Crumbs1 EAT-20 Crumbs2 CRB-3 Crumbs3 MAGU-2 Stardust/PALS1 Basolateral polarity proteins PAR-1 Par1 PAR-2 DLG-1 Discs large LET-413 Scribbled LGL-1 LGL Cytoplasmic polarity proteins PAR-4 Par4 PAR-5 STK11/LKB1 Cell junction components SAX-7 L1CAM HMP-1 α-catenin HMP-2 β-catenin HMR-1 E-cadherin JAC-1 p120 catenin MAGI-1 MAGI1 AJM-1 CLC-1 Claudin CLC-2 Claudin VAB-9 Claudin

Cytoskeletal linkers ERM-1 ERM family FRM-2 Yurt NFM-1 Merlin/NF2

LIN-2,7,10 complex LIN-2 Lin2 LIN-7 Lin7 LIN-10 Lin10 GTPases / signaling CDC-42 Cdc42 CED-10 Rac family RAL-1 Ral family RAP-1 Rap family RHEB-1 Rheb1 RHO-1 Rho family SRC-1 Src family MES-1 Nek8 PAC-1 RhoGAP PLK-1 Polo kinase PTP-3 LAR/PTPRD Spindle orientatin AGS-3 AGS3 INSC-1 Inscuteable RGS-7 RGS3

Wnt signaling DSH-1 Dishevelled DSH-2 Dishevelled MIG-5 Dishevelled CFZ-2 Frizzled LIN-17 Frizzled MIG-1 Frizzled MOM-5 Frizzled LIN-18 Ryk BAR-1 β-catenin WRM-1 β-catenin APR-1 APC GSK-3 GSK3β KIN-19 Casein Kinase I LIT-1 Nemo-like kinase MIG-14 Wntless MOM-1 Porcupine MOM-4 MAPKKK7 POP-1 TCF/LEF PRY-1 Axin SYS-1 Planar cell polarity proteins CDH-3 Fat CDH-1 Dachsous FMI-1 Flamingo PRKL-1 Prickle VANG-1 Van Gogh

C. elegans names are on the left. Common names from other species or protein family names are on the right. Figure 1. Identification and validation of a C. elegans polarity interaction network (CePIN) (a) Schematic representation of the different classes of protein selected as baits. Numbers indicate number of proteins in each class. See Table 1 for details. (b) Example of the design of bait fragments, which takes into account the predicted domain structure of each protein. Protein domains are indicated. Grey lines represent the fragments cloned. Red and blue arrows represent forward and reverse amplification primers respectively. (c) Network graph of the identified protein interactions. Green nodes are bait proteins, blue notes all other proteins. Orange edges: interactions identified using a full-length bait construct. Grey edges: interactions identified only using fragment bait constructs. (d) Enrichment of similar GO terms in interacting protein pairs compared to all possible pairs in the search space. Interactions from AD-cDNA and AD-Fragment libraries were plotted separately as the search space was highly different (baits × all C. elegans genes vs. baits × 749 early embryogenesis genes). (e) Cumulative distribution of Pearson correlation coefficients (PCCs) for mRNAs corresponding to protein pairs in the interaction dataset (red line), all possible protein pairs in the search space (blue line) and all C. elegans protein pairs (dotted grey line). Only AD-cDNA derived pairs are shown, as proteins pairs in the AD-Fragment library have such similar expression 

A combined binary interaction and phenotypic map of C. elegans cell polarity proteins Figure 1 a

7 Apical polarity regulators

5 PCP

DSH-1c

DAX

PDZ

DEP

702

11 GTPase Signaling

10 Cell Junction & Cell adhesion

20 Wnt signaling

3 Cytoskeletal linkers

3 Spindle positioning

2 Cytoplasmic PAR proteins

3 PDZ Scaffold

5 Basolateral polarity regulators

0.15

EF=4.54 p=1.04x10-8

Interactions Search space EF=1.99 p=0.0261

0.10

0.05

0.00 AD -F ra gm en t

AD -c D N A

Fraction of interactions at GO similarity 0.9 - 1.0

0.05

ra gm en t

cD N A

0.00

AP: streptavidin WB: αGFP

bait: prey: kDa 170 130 100 70 55 40 35 25

PAR-2 TYR-1 (fl.)

1 2 3

PAR-2 TYR-1 (frag.)

1 2 3

ERM-1 NRFL-1 (frag.)

1 2 3

DLG-1 TAC-1 (fl.)

1 2 3

DLG-1 TAC-1 (frag.)

1 2 3

0.8

Interactions Search space All C. elegans gene pairs

0.6

Fraction of gene pairs (cumulative)

EF=6.91 p=2.82x10-10

All baits x AD-cDNA

1.0

0.0

Fraction of interactions present in Wormnet

0.10

Interactions Search space

0.4

EF=3.64 p=5.38x10-6

0.15

0.2

−1.0

−0.5

0.0

0.5

1.0

PCC

Lanes: 1: EGFP + Avi-mCherry-bait 2: EGFP-prey + Avi-mCherry 3: EGFP-prey + Avi-mCherry-bait

profiles that further enrichment is not possible (Supplementary Fig. 3d). (f) Enrichment of the presence in WormNet for protein interactions compared to all possible pairs in the search space tested. AD-cDNA and AD-Fragment derived interactions were considered separately because of their highly different search spaces. (g) Experimental validation by co-affinity purification. Western blots detecting GFP-tagged proteins that co-purify with biotinylated mCherry-tagged proteins. Protein pairs tested are indicated above the blots, and the first protein listed is the Avi-mCherry tagged bait. Also indicated is whether the prey protein tested was full-length (f.l.) or corresponded to a shorter fragment (frag.). Lanes 1 and 2 are negative controls, lane 3 is the actual purification (1: Avi-mCherry-bait vs. EGFP alone, 2: Avi-mCherry alone vs. EGFP-tagged prey, 3: Avi-mCherry-bait vs. EGFP-tagged prey). The ~35 kDa band in lanes 2 is due to cross-reactivity of the anti-GFP antibody with the highly abundant Avi-mCherry polypeptide. Asterisks indicate bands of expected molecular mass. Expression of tagged protein in input lysates is shown in Supplementary Fig. 4.

Each of the bait constructs was used to screen two distinct Gal4::AD library types: an early embryogenesis fragment library that contains full-length and fragment clones of 749 genes that are essential for early embryonic development (ADFragment library)16, and a mixed-stage C. elegans AD-cDNA library that in theory could include all C. elegans coding genes, and was designed to contain mostly fulllength clones. We eliminated de novo auto-activators that arose during the screening process18,19, and only considered interactions where the AD-ORF fusion was in frame. To further increase the accuracy of the network, we only included AD-Fragment library-derived interactions identified in two or more yeast colonies. The AD-cDNA library is difficult to screen to completion, and many valid interactions may only be identified in a single yeast colony. Hence, we experimentally retested all interactions identified a single time from this library by transformation of the isolated AD-cDNA clone into fresh bait containing yeast. Interactions that retested positively were included in the network (see supplementary Fig. 2 for details). The final C. elegans polarity interaction network (CePIN) contains 439 interactions between 296 proteins (Fig. 1c and Supplementary Table 2). Most interactions (359/439) were detected only using non-full-length bait constructs (grey edges in Fig. 1c), confirming that the use of protein fragments increases the detectability of interactions in the Y2H system16. A searchable web interface of all interactions and fragments identified is available at http://www.projects.science.uu.nl/interactome/. Quality assessment of the polarity interaction network To assess the quality of the CePIN, we used a combination of computational and experimental approaches. We identified 53 interactions that have previously been reported (Supplementary Table 3). Moreover, interacting protein pairs were significantly enriched for similar GO terms, as well as presence in WormNet, which predicts functional linkages between C. elegans genes by integrating multiple data sources from different organisms20 (Fig. 1d,e and Supplementary Fig. 3). Finally, interactions identified from AD-cDNA library screens were highly enriched for similar mRNA expression profiles (Fig. 1f). As observed previously16, early embryogenesis genes already have such similar expression profiles that no further enrichment can be observed for interactions derived from the AD-Fragment library (Supplementary Fig. 3d). Together, these observations confirm the quality of the interactions we identified and demonstrate the ability of our Y2H screens to identify biologically relevant interactions.

A combined binary interaction and phenotypic map of C. elegans cell polarity proteins

Experimental validation of the polarity interaction network In addition to the computational comparisons above, we experimentally validated our dataset using an orthogonal binary protein interaction assay. A random sample of 40 interactions was selected to retest by co-affinity purification from Human Embryonic Kidney (HEK) 293 cells. To investigate if the use of fragments increases the chances of detecting an interaction, each interaction was tested using a fulllength prey protein construct and a shorter prey protein fragment containing the minimal region of interaction (see below). As bait, the protein fragment which most frequently detected the interaction in the Y2H system was used. Bait proteins were expressed as Avi-mCherry fusions together with the bacterial biotin ligase BirA, which recognizes and biotinylates the Avi sequence. Prey proteins were expressed as EGFP fusions. Bait proteins were purified from cell lysates using streptavidin-coated beads, and co-purification of the prey was assessed by Western blot. To control for non-specific binding we co-transfected Avi-mCherry alone with EGFP-tagged prey protein, and Avi-mCherry-tagged bait protein with the EGFP tag alone. Of the 40 pairs selected, 33 were successfully cloned and showed expression of the bait protein and at least one of the two prey protein constructs in HEK293 cells. We were able to successfully reproduce the interaction between 16 of the 33 pairs tested (48%) (Fig. 1g, Supplementary Fig. 4, and Supplementary Table 4). This is comparable to published validation rates of binary interactions in orthogonal assays, which generally are able to recover 20â&#x20AC;&#x201C;40% of interactions21,22, and is very similar to our previous validation of a series of human interactions identified by Y2H using the same co-affinity purification strategy, in which we validated 40% of novel interactions23. These results further demonstrate the quality of the CePIN. For 12 of the validated interactions, both the full-length and fragment prey were expressed. Of these, 5 were detected using both full-length and fragment prey constructs, 1 was only detected using the full-length construct, and 6 were detected only using the fragment prey protein. Thus, similar to our observations for the Y2H system, the use of fragments in addition to full-length clones provides an advantage in co-affinity purifications. Identification of minimal regions of interaction The delineation of protein regions that mediate interactions is a key goal of using protein fragments in Y2H screens. The use of multiple bait protein fragments in addition to the fragment containing prey libraries enabled us to identify minimal regions of interaction (MRIs) for both proteins involved in an interaction. For each interaction, we defined the MRI as the smallest region shared by all interacting protein fragments (Supplementary Table 5 and Supplementary Fig. 5). The AD-

Fragment library and the bait fragments were specifically designed for the unbiased identification of MRIs. By contrast, the AD-cDNA library contains only 5’ truncations. Indeed, MRIs defined from AD-Fragment library-derived interactions or from bait fragments were the most well defined, covering on average 49% and 53% of the full-length protein respectively, compared to 78% for AD-cDNA derived MRIs (Supplementary Fig. 6). The average length of all MRIs identified corresponded to 408 amino acids or 60% of the respective full-length protein (Fig. 2a,b). Figure 2 (Boxem 2015) b

30 Percentage of MRIs

20 15 10 5

20 10 0

90 99

70 80

0 10

100

10 0 0 20 0 30 0 40 0 50 0 60 0 70 0 80 0 90 10 0 >1 0 0 00 0

Percentage of MRIs

MRI size (% of full length)

MRI size (residues)

c PKC-3 binding region

PAR-6a

PB1

PDZ

309 LIN-10 binding region

LIN-2a

CKII

L27

PDZ

SH3

GuKc

961 LIN-2 binding region

LIN-10a

PTB 1

PDZ

PDZ 982

MAGI-1 binding region

β-catenin repeats

HMP-2b 1

704

MRI defined by Y2H Interaction domain from literature

Figure 2. Identification and validation of minimal regions of interaction (a) Distribution of the size of the identified MRIs as a percentage of the full-length protein. Interactions identified only as full-length are indicated separately (orange bar) (b) Distribution of the length of the identified MRIs in absolute residues. See Supplementary Figure 6a–f for separate size distributions of MRIs identified from bait proteins, from AD-Fragment clones, and from AD-cDNA clones. (c) Examples of MRIs for interactions where the binding site had already been described in the literature. See Supplementary Figure 6g for the complete set of MRIs for interactions with known interactions sites.

A combined binary interaction and phenotypic map of C. elegans cell polarity proteins

To evaluate the accuracy of the identified MRIs, we compared them with interaction domains described in the literature (Fig. 2c, Supplementary Fig. 6g, and Supplementary Table 5). For 29 MRIs, corresponding interaction domains were previously identified, either for the actual C. elegans protein or a homologous protein. Of these 29 MRIs, 27 are consistent with the interaction domain described (93%). In the two cases where the MRI differed from the interaction site described, the published interaction involved a mammalian homolog of the C. elegans protein. These cases may reflect a difference between the mammalian and C. elegans proteins. Further evidence of the accuracy of the MRIs comes from the co-affinity purification experiments (Supplementary Fig. 4). For 19 of the interactions retested by co-affinity purification, the smaller prey protein fragment tested corresponded to the MRI. Of these, 10 retested positively. Thus, the MRIs identified by Y2H were able to mediate the interaction in an orthogonal binary interaction assay as well. In some cases, the MRI borders were a near exact match to the known interaction site. Examples include the PAR-6 MRI that mediates the interaction with PKC-3, and the LIN-2 MRI that binds to LIN-10 (Fig. 2c). In other cases, the MRI we identified was larger than the published binding site. One explanation is that shorter clones were not identified or are not present in the library. For example, the AD-cDNA library was generated by poly-T priming, and can only define the N-terminal boundary of an MRI (e.g., the LIN-10 MRI that binds LIN-2, Fig. 2c). Alternatively, the interaction may be mediated by a short, linear motif that needs to be presented as part of a larger, folded polypeptide. A possible example of this latter case is the interaction between MAGI-1 and HMP-2, which is mediated by the C-terminal 4 residues of HMP-2 [ref. 24]. The identified MRI covered a much larger region, possibly because shorter constructs would fail to properly fold and express in yeast (Fig. 2c). Especially in these latter cases, the MRIs identified by Y2H can provide a starting point for follow-up experiments requiring expression of a fragment containing the interaction site. Phenotypic profiling of the polarity interaction network While the presence of a physical interaction in yeast is suggestive of a functional relationship, further evidence for a functional association can be obtained by integrating protein interaction data with phenotypic data25â&#x20AC;&#x201C;28. We therefore performed systematic phenotypic profiling of the polarity interaction network by RNAmediated interference (RNAi). Polarity regulators may act only in certain tissues or play highly distinct roles in polarity establishment. Therefore, we examined the effects of RNAi in nine different strains expressing fluorescently-tagged proteins involved in several polarity related processes (Table 2 (page 60) and Fig. 3a-i).

Table 2. Marker lines used for phenotypic profiling of the CePIN Tissue / process Early embryonic polarity

Seam cell epithelium1 Intestinal epithelium Excretory canal (tubular epithelium)2 Epithelial junctions Yolk protein trafficking3 Yolk protein receptor* PVD axon-dendrite specification PVD neuron morphology*

Marker(s) Localization mCherry::PAR-6 Anterior at 1 cell stage, contact-free surfaces from 4-cell stage GFP::H2B Histones GFP::PH Cell membrane PEPT-1::dsRed Apical membrane GFP::RAB-11 Apically enriched recycling endosomes VHA-5::GFP Canaliculi vesicles DLG-1::GFP VIT-2::GFP RME-2::GFP mCherry::RAB-3 SAD-1::GFP Myristoyl::GFP

Apical junctional complex Yolk protein, endocytosed by oocytes Oocyte membrane Axonal puncta Axonal puncta Cell membranes

*Genes that showed an RNAi defect in VIT-2::GFP trafficking or PVD axon-dendrite specification were further examined in strains expressing the VIT-2 receptor or highlighting the extensive PVD dendrite arborisation pattern, respectively. 1 The seam cells exhibit both apical-basal polarity and anterior-posterior tissue polarity, and undergo a series of asymmetric stem cell-like divisions during larval development. 2 The excretory canal is a tubular epithelium formed by a single cell, highly distinct from the multicellular seam and intestinal epithelia. 3 Uptake of yolk proteins by oocytes is an endocytic process that requires the activity of the par-3, par-6, pkc-3 and cdc-42 genes29.

To examine defects in early embryonic polarity, we used a strain expressing mCherry::PAR-6, which localizes anteriorly in the one cell embryo and to the contact free surfaces in young embryos. In the larval stages, we examined three different epithelial cell types: seam cells, intestinal cells, and the excretory canal. The seam cells exhibit both apical-basal polarity and anterior-posterior tissue polarity, and undergo a series of asymmetric stem cell-like divisions during larval development. The seam cell epithelium was examined in a line expressing a GFP::PH domain fusion that outlines the cell membrane, and a GFP::H2B fusion that marks the DNA. Intestinal polarity was followed using a line expressing PEPT-1::DsRed, which marks the apical domain, and GFP::RAB-11, which marks apically enriched recycling endosomes. In both these tissues we also examined the integrity of cell junctions using a DLG-1::GFP expressing line. The excretory canal is a tubular epithelium formed by a single cell, and therefore highly distinct from the multicellular seam and intestinal epithelia. We examined defects in excretory canal development using a strain expressing VHA-5::GFP. In addition to these epithelial tissues, we examined two processes that depend on cell polarity. Uptake of yolk proteins by oocytes is an endocytic process that requires the activity of the par-3, par-6, pkc-3 and cdc-42 58

A combined binary interaction and phenotypic map of C. elegans cell polarity proteins

te s Se tin a al C m e epi el p th l i e Yo jun the lium lk cti liu Si tra ons m ng ffi c PV le c kin D el g Ex ne l em cr ur b et on ry o or y ca na l

genes29, and we examined this process in a strain expressing the yolk-protein fusion VIT-2::GFP. All genes that showed a phenotype with this strain were re-screened in a strain expressing the fluorescently tagged VIT-2 receptor RME-2::GFP, to determine if the observed defects were due to mislocalization of the receptor. Finally, neurons represent one of the most highly polarized cell types in the body. We examined the correct polarization of the axon in PVD neurons, in a strain expressing the fluorescently tagged presynaptic molecules mCherry::RAB-3 and SAD-1::GFP. All genes that showed a defect in localization of these markers following RNAi were re-screened in a strain expressing myristoylated GFP in the PVD neuron, to examine changes to the extensive arborisation pattern of the PVD dendritic tree. Together, Figure 3 these nine strains allowed us to examine different aspects of polarity establishment in six different tissues.

b seam cell epithelium

c cell junctions (intestine)

PEPT-1::DsRed GFP::RAB-11

GFP::PH GFP::H2B

DLG-1::GFP

e yolk receptor in oocytes

f early embryonic polarity

VIT-2::GFP

RME-2::GFP

mCherry::PAR-6

g PVD axon vs. dendrite

mCherry::RAB-3 SAD-1::GFP

myristoyl::GFP

intestinal epithelium

yolk protein uptake

PVD morphology

cdh-3 frm-2 prkl-1 wrm-1 hmp-1 mom-1 bar-1 lin-10 lin-2 lin-7 hmr-1 hmp-2 magi-1 lgl-1 src-1 apr-1 gsk-3 pop-1 vang-1 mes-1 erm-1 ral-1 sys-1 lit-1 ced-10 cfz-2 rheb-1 ptp-3 kin-19 ajm-1 dlg-1 ags-3 plk-1 rho-1 mig-14 pac-1 par-1 par-6 cdc-42 par-4 par-2 par-3 par-5 pkc-3

excretory canal

VHA-5::GFP 0

severity

Figure 3. Phenotypic analysis of the bait proteins (a–i) Marker lines used for phenotypic profiling. (a) Intestinal expression of PEPT-1::dsRed (apical domain) and GFP::RAB-11 (apical recycling endosomes). (b) Seam epithelium with cell outline highlighted by GFP fused to a pleckstrin homology domain, and nuclei shown by a Histone 2B GFP fusion. (c) Expression of the junctional protein AJM-1::GFP in the intestine. (d) Uptake of the yolk protein VIT-2::GFP in oocytes. (e) Cortical localization of the RME-2::GFP, the receptor for VIT-2, in oocytes. (f) Asymmetric distribution of mCherry::PAR-6 at the two-cell stage. (g) Localization of the presynaptic molecules mCherry::RAB-3 and SAD-1::GFP in the PVD axon (arrow). SAD-1::GFP is also observed at low levels in the cell body and dendrites near the cell body. (h) Expression of myristoylated GFP under control of a PVD-specific promoter. (i) Excretory canal outline by expression of VHA-5::GFP. Shown is the head region where the excretory canal cell body is located. All scale bars represent 10 µm. (j) Clustering of the bait proteins based on observed phenotypes. Green colour indicates presence of a defect, and colour intensity indicates severity of the defect on a scale of 1–5.

We first screened each of the 69 bait proteins for defects in the localization pattern of the markers described above. For each RNAi experiment, a total of 40 possible defects across the nine marker strains were scored (Supplementary Table 6). Each defect was scored on a scale of 0 to 5 based on severity and penetrance. We identified 44 bait proteins for which down regulation by RNAi resulted in a detectable defect in at least one of the marker strains (Fig. 3j and Supplementary Table 6). Certain genes, e.g. par6, kin-19, lit-1 and dlg-1, showed defects in most tissues examined, while others, such as lgl-1 or rheb-1 affected only a single tissue. This is in line with our expectation that many genes will play tissue specific roles in polarity establishment, although it may in part also reflect incomplete inactivation by RNAi. Next, the interacting protein partners of these 44 bait proteins were screened specifically in those strains in which a defect was detected. We identified a total of 100 protein pairs for which RNAimediated depletion of both corresponding genes resulted in an abnormal phenotype in the same polarity related process (Fig. 4 and Supplementary Table 6). To evaluate the quality of the phenotypic profiling, we used hierarchical clustering to group the bait genes by phenotypic similarity (Fig. 3j). For each gene, a binary phenotypic profile was generated consisting of the highest severity score for each of the 6 tissues analysed, adding defects in localization of junctional DLG-1 as a 7th data point. The hierarchical clustering grouped together genes that were already known to act together or in a similar process, including the par genes together with pkc-3 and cdc-42, the lin-2,7,10 genes, whose protein products form a complex, and the genes encoding the interacting proteins AJM-1 and DLG-1. The accurate grouping of genes with a known functional relationship validates the quality of our RNAi-based phenotypic profiling. For 100 of the 439 interacting protein pairs, RNAi of the corresponding genes resulted in a defect in the same polarity-related process (Fig. 4a, Supplementary Fig. 7b, and Supplementary Table 6). A functional interaction between these proteins is therefore supported by both a physical interaction and an overlap in phenotype. In many cases, the interactions involved poorly characterized or uncharacterized proteins. For example, as previously reported29, inactivation of the Cadherin HMR-1 resulted in accumulation of the yolk protein VIT-2 in the body cavity (Fig. 4a-II). For one of the HMR-1 interactors, W09D10.1, RNAi of the corresponding gene resulted in a very similar phenotype (Fig. 4a-III). W09D10.1 is an ArfGAP domain-containing protein, homologous to the Arf6-specific GAP SMAP1, which has been proposed to be a key player in the clathrin dependent constitutive endocytosis of E-cadherin in Mardin-Darby canine kidney (MDCK) epithelial cells30. In addition to HMR-1, we found W09D10.1 to bind to JAC-1 p120 catenin and BAR-1 Î˛-catenin, suggesting a broader role for W09D10.1 at cell-cell junctions. Not all associations were easy

A combined binary interaction and phenotypic map of C. elegans cell polarity proteins

to interpret. For example, we identified an interaction between the Ezrin/Radixin/ Moesin homolog ERM-1 and the uncharacterized protein C30B5.4. RNAi for both genes resulted in defects in the seam epithelium, though the exact nature of the phenotype differed (Fig 4a-V, VI). C30B5.4 contains an RNA-binding motif, and is homologous to human RBMX2, which is annotated as a spliceosome component. Thus, while the physical interaction and overlap in phenotype suggest a functional association, it is not immediately obvious what the functional consequences of this association might be. As a final example, we identified an interaction between DLG-1 Discs Large and LST-1, an uncharacterized protein with no recognizable domains. RNAi for both dlg-1 and lst-1 resulted in a highly irregular morphology of the intestinal lumen (Fig. 4a VIII,IX), and a reduction in the subapical accumulation of RAB-11 vesicles (Fig. 4a XI,XII). Consistent with an intestinal defect, LST-1 was reported to be expressed in the intestine31. The PVD neuron and excretory canal are not widely used as models of cell polarity. The systematic phenotypic profiling of the 69 bait proteins in these tissues therefore led to the identification of several novel phenotypes. For example, inactivation of lgl-1, the C. elegans homolog of the Drosophila tumor suppressor lethal giant larvae, resulted in extensive spreading of RAB-3 vesicles into the dendrites of the PVD neuron (Fig. 4b-II). C. elegans lgl-1 was shown to act redundantly with par-2 in early embryonic polarity establishment32,33. However, we did not detect defects in the PVD neuron upon par-2 RNAi. These results are the first demonstration of a par2-independent function for C. elegans lgl-1. A dramatic effect on the PVD neuron was also observed upon inactivation of the Tcf transcription factor pop-1, which resulted in the formation of numerous additional PVD cell bodies and spreading of RAB3 vesicles into dendrites (Fig. 4b-IV). This may reflect additional cell divisions or cell-fate alterations in the V5 lineage that produces the PVD neuron, rather than a polarity defect, as pop-1 is known to be required for numerous cell fate decisions. Inactivation of sys-1, which encodes a Î˛-catenin protein that binds POP-1, also resulted in spreading of RAB-3 and duplication of the cell body (Fig. 4b-V).

We identified 9 genes whose inactivation resulted in defects in the appearance of the excretory canal, an H-shaped seamless tube formed by a single cell, which extends along the length of the body. Three of these, erm-1, par-6, and ral-1 had previously described roles in excretory canal formation34–36. We also identified defects in excretory canal appearance for par-2 and par-4 RNAi, which indicates a broader involvement of cell polarity regulators in excretory canal formation (Fig. 4b-VII and VIII). Finally, mes-1 and vang-1 showed a defect exclusively in the excretory canal (Fig. 4b-X and XI). MES-1 is a receptor tyrosine kinase-like protein required for accurate positioning of the mitotic spindle in the P2 and EMS cells of the early embryo37,38. Postembryonic roles for MES-1 have not been described. VANG-1 is the C. elegans homolog of the planar cell polarity regulator Strabismus/Van Gogh39–41. The identification of roles in excretory canal formation for MES-1 and VANG-1 may provide new insights into the formation of this organ.

Figure 4. Examples of RNAi phenotypes. In all panels, the gene inactivated and the marker protein  shown are indicated. EV: empty vector. Scale bars are 10 µm. (a) Examples of interacting protein pairs where RNAi for both corresponding genes resulted in a defect in the same marker strain. In panels I-III, arrows point to ectopic yolk accumulation. Arrow in panel V indicates failed cell division. Arrow in panel VI points to larger than normal distance between cells and elongated connection between cells. Additional examples are shown in Supplementary Fig. 7. (b) Examples of novel phenotypes observed in the PVD neuron and the Excretory canal during phenotypic analysis of the bait proteins. In PVD neuron panels, an asterisk indicates the location of the cell body, and arrowheads indicate spreading of RAB-3 into dendrites.

A combined binary interaction and phenotypic map of C. elegans cell polarity proteins

Figure 4

Yolk trafficking

control EV RNAi

bait I hmr-1 RNAi

GFP::VIT-2

Seam cells

EV RNAi

Intestine

VII dlg-1 RNAi

PVD neuron

mCherry::RAB-3 EV RNAi

SAD-1::GFP

Excretory canal

EV RNAi

GFP::VHA-5 EV RNAi

GFP::VHA-5

H2B::GFP GFP::PH

VIII lst-1 RNAi

PEPT-1::dsRed

X dlg-1 RNAi

XI lst-1 RNAi

GFP::RAB-11

control RNAi EV RNAi

V C30B5.4 RNAi

H2B::GFP GFP::PH

GFP::RAB-11

III

GFP::VIT-2

IV erm-1 RNAi

PEPT-1::dsRed EV RNAi

II W09D10.1 RNAi

GFP::VIT-2

H2B::GFP PH::GFP EV RNAi

prey

XII

GFP::RAB-11

RNAi of bait genes I

lgl-1 RNAi

mCherry::RAB-3

III pop-1 RNAi

SAD-1::GFP

VI par-2 RNAi

IV sys-1 RNAi

SAD-1::GFP

VII par-4 RNAi

GFP::VHA-5

IX mes-1 RNAi

X vang-1 RNAi

GFP::VHA-5

VIII

Direct interaction between PAR-6 and the RhoGap PAC-1 is required for radial polarization of PAR-6 As a demonstration of the validity of our approach to identify functionally relevant interactions, we focused on the interaction between PAR-6 and PAC-1. In addition to the physical interaction, the phenotypic profiles of par-6(RNAi) and pac-1(RNAi) clustered together (Fig. 3j). PAC-1/ARHGAP21 is a RhoGAP protein that is required for radial polarization of the C. elegans embryo42. In wild-type embryos, PAR-6 localizes to the outer, contact-free cell surface beginning at the 4-cell stage. Cortical recruitment of PAR-6 is dependent on the Rho-family member CDC-42, which localizes along the entire cortex42,43. PAC-1 localizes to cell contact sites, and is thought to trigger radial polarization by locally inactivating CDC-42, limiting the recruitment of PAR-6 by CDC-42 to the outer cell surfaces42,43. Although PAR-6 and PAC-1 do not appear to co-localize, our observations suggest that a physical interaction between PAR-6 and PAC-1 contributes to their in vivo functions. Additional fragments of PAR-6 and PAC-1 were tested in the Y2H system to narrow down the interaction sites. The PDZ domain of PAR-6 (amino acids 155 – 248 in PAR-6a) was sufficient to mediate the binding to PAC-1 (Fig. 5a,c). Although PDZ domains frequently interact with C-terminal motifs, a PAC-1 fragment lacking the final 7 residues was still able to interact with PAR-6, indicating that the PAR-6 PDZ domain interacts with an internal motif of PAC-1 (Fig. 5b,c). We found that a 100 amino acid sequence just downstream of the PAC-1 RhoGAP domain (amino acids 1221 – 1320 in PAC-1) was able to mediate binding to PAR-6 (Fig. 5b,c). Next, we tested the interaction between PAR-6 and PAC-1 by co-affinity purification from mammalian HEK293 cells. A C-terminal fragment of PAC-1b (aa 1221 – 1604), as well as the minimal PAR-6 binding domain (aa 1221 – 1320), copurified with a PAR-6a fragment containing the PDZ domain (aa 133 – 265) (Fig. 5d). Moreover, when co-expressed in HeLa cells, PAC-1 and PAR-6 co-localize in a punctate pattern that was clearly distinct from the localization pattern of PAC-1 or PAR-6 when expressed individually (Supplementary Fig. 8). These observations strongly support the presence of a physical interaction between PAC-1 and PAR-6 and confirm the minimal region of interaction we identified by Y2H. To explore the functional relevance of the interaction between PAC-1 and PAR-6, we determined whether PAC-1 lacking the PAR-6 binding domain can functionally substitute for the wild-type protein. Previously, we demonstrated that expression of a GFP-PAC-1 fusion protein can rescue the localization of PAR-6 in a pac-1 mutant embryo42. We deleted the PAR-6 binding domain (PBD) (aa 1221 – 1320) from the GFP-PAC-1 encoding plasmid and expressed GFP-PAC-1(ΔPBD) in pac-1(xn6) mutant embryos. The GFP-PAC-1(ΔPBD) protein was expressed and localized to

A combined binary interaction and phenotypic map of C. elegans cell polarity proteins

the inner cell surfaces, similar to wild-type PAC-1 (Fig. 5h). However, GFP-PAC1(ΔPBD) failed to restore the radial localization pattern of PAR-6 (Fig. 5g). Thus, the direct interaction of PAC-1 with PAR-6 represents a novel functional relationship, which Figure 5is necessary for PAC-1 to establish radial polarization. a

PAR-6a 1

Interaction

125 133 155

PB1

248 265

309

PAC-1b 1

1604

1221

PDZ

RhoGap

+ + + + +

1221

1265

1320

1416

1520

1604

Interaction

+ + + + +/+

l.) (f. -1

-1

PA C

1597

bait: PAC-1 (f.l.) prey: PAR-6 (133-265)

PA C

DB AD PAR-6 (f.l.)

PA (12 21 C -1 -1 60 (1 4) 2 PA 21 -1 C -1 32 (f. 0) l.) PA C -1 PA (12 21 C -1 -1 60 (1 4) 22 113 20 )

co-AP αGFP

kDa 70 55 40

PAC-1 (1221-1604) PAR-6 (133-265)

1 2 3

PAC-1 (1221-1320) PAR-6 (133-265)

1 2 3

35 25

PAR-6 (133-265) input αGFP

PAR-6 (155-248) -Leu -Trp +His

-Leu -Trp -His

wild type

PAR-6

n=58/58

PAR-6

input biotin

pac-1

n=62/62

40 35 25

170

170 130 100 70

pac-1; GFP::PAC-1∆PBD

PAR-6

n=61/61

70 55 40

pac-1; GFP::PAC-1∆PBD

PAC-1∆PBD

Figure 5. The PAR-6 binding domain of PAC-1 is required for radial polarization (a,b) Delineation of the region of PAR-6 responsible for binding to PAC-1 (a) and the region of PAC-1 responsible for binding to PAR-6 (b). Grey lines below the protein drawings indicate fragments tested by Y2H for interaction. Presence or absence of interaction is indicated to the left of each fragment. Yellow regions correspond to the minimal regions of interaction. (c) Y2H analysis of the PAC-1/PAR-6 interaction. The fragments of PAC-1 and PAR-6 used are indicated. Left: permissive plate containing histidine, right: selective plate lacking histidine. The full-length clones of both PAC-1 and PAR-6 were unable to interact in yeast. (d) Validation of the PAC-1/PAR-6 interaction by co-purification. Western blots show GFP::PAR-6 copurifying with biotinylated mCherry::PAC-1. Protein fragments tested are indicated above the blots. See legend of Fig. 1g for further details. (e) Localization of PAR-6 to the contact-free cell surfaces in a pac-1 wild-type background. (f) (Mis)localization of PAR-6 to inner cell surfaces in pac-1(xn6) mutant embryos. (g) Expression of PAC-1 lacking the PAR-6 binding domain (PAC-1 ΔPBD) is unable to restore normal localization of PAR-6. (h) PAC-1 ΔPBD is expressed and able to localize specifically to the inner cell surfaces. Scale bar = 10 µm.

Discussion In this study we combined physical interaction mapping with phenotypic profiling to study the network of physical interactions that underlies polarity establishment and maintenance in the nematode Caenorhabditis elegans. We identified 439 interactions between 296 proteins, and we validated the data using a combination of computational and experimental approaches. The experimental identification of the protein regions that mediated these interactions provides additional insight into the proteins in the network. In addition, fragments capable of mediating an interaction in yeast are likely to be functional in other organisms as well. Therefore, the fragment data can be of use to others wishing to express functional protein domains. Most proteins in the CePIN have not previously been linked to cell polarity. Identifying components involved in cell polarity is complicated by the variation in the mechanisms of polarity establishment in different tissues or developmental stages. For example, not all Drosophila epithelia in which Crumbs is expressed require Crumbs to establish epithelial polarity8. In the C. elegans embryo, PAR-3 is required for the assembly of intestinal cell junctions, but in epidermal epithelia, apical junctions still form in the absence of PAR-3[ref. 44]. The CePIN provides many putative interactions involved in cell polarity, which can be pursued using detailed in vivo analyses that can identify more subtle or tissue-specific roles in cell polarity for the proteins involved. While a physical interaction in yeast indicates a potential functional relationship, not all interactions identified in yeast are biologically relevant (biological false positives). Greater predictive power can be obtained through the integration of physical interaction networks with additional data sources. Here, we combined interaction mapping with phenotypic analysis by RNAi to identify potential functional relationships. 100 protein pairs physically interacted and, upon RNAimediated depletion, showed a defect in the same polarity-related process. The extent of overlap in the phenotypes differed. For some interactions, the observed phenotypes closely resembled each other (e.g., HMR-1/W09D10.1 and DLG-1/LST1 in Fig. 4a). In most cases however, the exact nature of the defect observed differed between the two interacting proteins (e.g., ERM-1/C30B5.4 in Fig. 4a). Nevertheless, since inactivation of both genes resulted in a defect in a polarity-related process, and the proteins physically interacted in yeast, these 100 interactions are promising candidates for further study. The marker strains we used for the phenotypic profiling highlight only a subset of polarity-related processes. Thus, examining the interacting protein pairs using other marker lines would likely result in the identification of additional pairs with

A combined binary interaction and phenotypic map of C. elegans cell polarity proteins

overlapping phenotypes. Moreover, many of the mechanisms that drive polarity establishment are conserved between C. elegans, flies, and human, and interactions between conserved proteins can be examined in other model systems as well. Thus, the CePIN will be a valuable resource for future studies aiming to gain novel insights into the mechanisms that control cell polarity. Acknowledgements We thank Marino Zerial for providing the MZE1 strain and Martin Harterink for providing the KyIs445 and Hrt3Is3 strains. We also thank Fried Zwartkruis, Martijn Gloerich, and Adri Thomas for critically reading of the manuscript. Some nematode strains used in this work were provided by the Caenorhabditis Genetics Center, which is funded by the NIH National Center for Research Resources (NCRR). T.K., M.v.d.V. and M.B. were supported by the Innovational Research Incentives Scheme Vidi grant 864.09.008, financed by the Netherlands Organization for Scientific Research (NWO/ALW). D.K. was supported by a National Science Foundation Graduate Research Fellowship (12-A0-00-000165-01) and J.N. was supported by grants from the NIH (R01GM078341 and R01GM098492). Author Contributions M.v.d.V. contributed to the cloning and Y2H screens. D.K. performed the PAC-1 rescue experiments. J.J.R. and S.N. contributed to the RNAi screens. M.B. performed the computational analyses. All other experiments were performed by T.K. S.v.d.H. and T.K. contributed to the design of the study. J.N. contributed to the design of the PAC-1/PAR-6 experiments. T.K. and M.B. wrote the manuscript. M.B. guided all aspects of the study. ď&#x201A;ž

References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25.

Laukoetter, M. G., Nava, P. & Nusrat, A. Role of the intestinal barrier in inflammatory bowel disease. World J. Gastroenterol. 14, 401–407 (2008). Bulgakova, N. A. & Knust, E. The Crumbs complex: from epithelial-cell polarity to retinal degeneration. J. Cell Sci. 122, 2587–2596 (2009). Lee, M. & Vasioukhin, V. Cell polarity and cancer - cell and tissue polarity as a non-canonical tumor suppressor. J. Cell Sci. 121, 1141–1150 (2008). Hanahan, D. & Weinberg, R. A. Hallmarks of cancer: the next generation. Cell 144, 646–674 (2011). Tervonen, T. A. et al. Faulty epithelial polarity genes and cancer. Adv. Cancer Res. 111, 97–161 (2011). Martin-Belmonte, F. & Perez-Moreno, M. Epithelial cell polarity, stem cells and cancer. Nat. Rev. Cancer 12, 23–38 (2012). Humbert, P. O. et al. Control of tumourigenesis by the Scribble/Dlg/Lgl polarity module. Oncogene 27, 6888–6907 (2008). Tepass, U. The apical polarity protein network in Drosophila epithelial cells: regulation of polarity, junctions, morphogenesis, cell growth, and survival. Annu. Rev. Cell Dev. Biol. 28, 655–685 (2012). St Johnston, D. & Ahringer, J. Cell polarity in eggs and epithelia: parallels and diversity. Cell 141, 757–774 (2010). Jacob, L., Opper, M., Metzroth, B., Phannavong, B. & Mechler, B. M. Structure of the l(2)gl gene of Drosophila and delimitation of its tumor suppressor domain. Cell 50, 215–225 (1987). Woods, D. F. & Bryant, P. J. The discs-large tumor suppressor gene of Drosophila encodes a guanylate kinase homolog localized at septate junctions. Cell 66, 451–464 (1991). Bilder, D. & Perrimon, N. Localization of apical epithelial determinants by the basolateral PDZ protein Scribble. Nature 403, 676–680 (2000). Laprise, P. et al. Yurt, Coracle, Neurexin IV and the Na(+),K(+)-ATPase form a novel group of epithelial polarity proteins. Nature 459, 1141–1145 (2009). Goehring, N. W. & Grill, S. W. Cell polarity: mechanochemical patterning. Trends Cell Biol. 23, 72–80 (2013). Rodriguez-Boulan, E. & Macara, I. G. Organization and execution of the epithelial polarity programme. Nat. Rev. Mol. Cell Biol. 15, 225–242 (2014). Boxem, M. et al. A protein domain-based interactome network for C. elegans early embryogenesis. Cell 134, 534–545 (2008). Sawa, H. Control of cell polarity and asymmetric division in C. elegans. Curr. Top. Dev. Biol. 101, 55–76 (2012). Vidalain, P.-O., Boxem, M., Ge, H., Li, S. & Vidal, M. Increasing specificity in high-throughput yeast two-hybrid experiments. Methods 32, 363–370 (2004). Walhout, A. J. & Vidal, M. A genetic strategy to eliminate self-activator baits prior to highthroughput yeast two-hybrid screens. Genome Res. 9, 1128–1134 (1999). Cho, A. et al. WormNet v3: a network-assisted hypothesis-generating server for Caenorhabditis elegans. Nucleic Acids Res. (2014). doi:10.1093/nar/gku367 Braun, P. et al. An experimentally derived confidence score for binary protein-protein interactions. Nat. Methods 6, 91–97 (2009). Rolland, T. et al. A proteome-scale map of the human interactome network. Cell 159, 1212–1226 (2014). Waaijers, S., Koorman, T., Kerver, J. & Boxem, M. Identification of human protein interaction domains using an ORFeome-based yeast two-hybrid fragment library. J. Proteome Res. 12, 3181– 3192 (2013). Stetak, A., Hörndli, F., Maricq, A. V., van den Heuvel, S. & Hajnal, A. Neuron-specific regulation of associative learning and memory by MAGI-1 in C. elegans. PloS One 4, e6019 (2009). Gunsalus, K. C. et al. Predictive models of molecular machines involved in Caenorhabditis elegans early embryogenesis. Nature 436, 861–865 (2005).

A combined binary interaction and phenotypic map of C. elegans cell polarity proteins

26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45.

Boulton, S. J. et al. Combined functional genomic maps of the C. elegans DNA damage response. Science 295, 127–131 (2002). Tu, Z. et al. Integrating siRNA and protein-protein interaction data to identify an expanded insulin signaling network. Genome Res. 19, 1057–1067 (2009). Wang, L., Tu, Z. & Sun, F. A network-based integrative approach to prioritize reliable hits from multiple genome-wide RNAi screens in Drosophila. BMC Genomics 10, 220 (2009). Balklava, Z., Pant, S., Fares, H. & Grant, B. D. Genome-wide analysis identifies a general requirement for polarity proteins in endocytic traffic. Nat. Cell Biol. 9, 1066–1073 (2007). Kon, S., Tanabe, K., Watanabe, T., Sabe, H. & Satake, M. Clathrin dependent endocytosis of E-cadherin is regulated by the Arf6GAP isoform SMAP1. Exp. Cell Res. 314, 1415–1428 (2008). Cuppen, E., van der Linden, A. M., Jansen, G. & Plasterk, R. H. A. Proteins interacting with Caenorhabditis elegans Galpha subunits. Comp. Funct. Genomics 4, 479–491 (2003). Hoege, C. et al. LGL can partition the cortex of one-cell Caenorhabditis elegans embryos into two domains. Curr. Biol. 20, 1296–1303 (2010). Beatty, A., Morton, D. & Kemphues, K. The C. elegans homolog of Drosophila Lethal giant larvae functions redundantly with PAR-2 to maintain polarity in the early embryo. Development 137, 3995–4004 (2010). Khan, L. A. et al. Intracellular lumen extension requires ERM-1-dependent apical membrane expansion and AQP-8-mediated flux. Nat. Cell Biol. 15, 143–156 (2013). Göbel, V., Barrett, P. L., Hall, D. H. & Fleming, J. T. Lumen morphogenesis in C. elegans requires the membrane-cytoskeleton linker erm-1. Dev. Cell 6, 865–873 (2004). Armenti, S. T., Chan, E. & Nance, J. Polarized exocyst-mediated vesicle fusion directs intracellular lumenogenesis within the C. elegans excretory cell. Dev. Biol. 394, 110–121 (2014). Berkowitz, L. A. & Strome, S. MES-1, a protein required for unequal divisions of the germline in early C. elegans embryos, resembles receptor tyrosine kinases and is localized to the boundary between the germline and gut cells. Development 127, 4419–4431 (2000). Bei, Y. et al. SRC-1 and Wnt signaling act together to specify endoderm and to control cleavage orientation in early C. elegans embryos. Dev. Cell 3, 113–125 (2002). Hoffmann, M., Segbert, C., Helbig, G. & Bossinger, O. Intestinal tube formation in Caenorhabditis elegans requires vang-1 and egl-15 signaling. Dev. Biol. 339, 268–279 (2010). Honnen, S. J. et al. C. elegans VANG-1 modulates life span via insulin/IGF-1-like signaling. PloS One 7, e32183 (2012). Sanchez-Alvarez, L. et al. VANG-1 and PRKL-1 cooperate to negatively regulate neurite formation in Caenorhabditis elegans. PLoS Genet. 7, e1002257 (2011). Anderson, D. C., Gill, J. S., Cinalli, R. M. & Nance, J. Polarization of the C. elegans embryo by RhoGAP-mediated exclusion of PAR-6 from cell contacts. Science 320, 1771–1774 (2008). Chan, E. & Nance, J. Mechanisms of CDC-42 activation during contact-induced cell polarization. J. Cell Sci. 126, 1692–1702 (2013). Achilleos, A., Wehman, A. M. & Nance, J. PAR-3 mediates the initial clustering and apical localization of junction and polarity proteins during C. elegans intestinal epithelial cell polarization. Development 137, 1833–1842 (2010). Wu, X., Pang, E., Lin, K. & Pei, Z.-M. Improving the measurement of semantic similarity between gene ontology terms and gene products: insights from an edge- and IC-based hybrid method. PloS One 8, e66745 (2013).

Supplementary information This chapter contains large datasets of both Illustrator and Excel files, tables and sheets. Therefore not all data will be shown in this thesis. Full tables are available on request or when published. Contents of the supplementary information: - Supplementary References (Materials and methods) - Supplementary figure 1: Bait protein drawings - Supplementary figure 2: Screen pipeline - Supplementary figure 3: Full GO and PCC graphs - Supplementary figure 4: Overview of the pulldowns - Supplementary figure 5: All MRIs - Supplementary figure 6: Analysis of MRIs - Supplementary figure 7: RNAi examples of baits and preys - Supplementary figure 8: Colocalization PAC-1 with PAR-6 in HeLa cells - - - - - -

Supplementary table 1: Bait information Supplementary table 2: Polarity Interaction Network Supplementary table 3: Literature comparison Supplementary table 4: Co-affinity purification Supplementary table 5: MRI data Supplementary table 6: RNAi profiling

A combined binary interaction and phenotypic map of C. elegans cell polarity proteins

Supplemental References (Materials and methods) 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19.

Chevray, P. M. & Nathans, D. Protein interaction cloning in yeast: identification of mammalian proteins that react with the leucine zipper of Jun. Proc. Natl. Acad. Sci. U. S. A. 89, 5789–5793 (1992). Vidal, M., Brachmann, R. K., Fattaey, A., Harlow, E. & Boeke, J. D. Reverse two-hybrid and onehybrid systems to detect dissociation of protein-protein and DNA-protein interactions. Proc. Natl. Acad. Sci. U. S. A. 93, 10315–10320 (1996). Yu, H. et al. High-quality binary protein interaction map of the yeast interactome network. Science 322, 104–110 (2008). Schiestl, R. H. & Gietz, R. D. High efficiency transformation of intact yeast cells using single stranded nucleic acids as a carrier. Curr. Genet. 16, 339–346 (1989). Boxem, M. et al. A protein domain-based interactome network for C. elegans early embryogenesis. Cell 134, 534–545 (2008). Fromont-Racine, M., Rain, J.-C. & Legrain, P. Building protein-protein networks by two-hybrid mating strategy. Methods Enzymol. 350, 513–524 (2002). Vidalain, P.-O., Boxem, M., Ge, H., Li, S. & Vidal, M. Increasing specificity in high-throughput yeast two-hybrid experiments. Methods 32, 363–370 (2004). Ewing, B. & Green, P. Base-calling of automated sequencer traces using phred. II. Error probabilities. Genome Res. 8, 186–194 (1998). Ewing, B., Hillier, L., Wendl, M. C. & Green, P. Base-calling of automated sequencer traces using phred. I. Accuracy assessment. Genome Res. 8, 175–185 (1998). Wu, X., Pang, E., Lin, K. & Pei, Z.-M. Improving the measurement of semantic similarity between gene ontology terms and gene products: insights from an edge- and IC-based hybrid method. PloS One 8, e66745 (2013). Hibbs, M. A. et al. Exploring the functional landscape of gene expression: directed search of large microarray compendia. Bioinforma. Oxf. Engl. 23, 2692–2699 (2007). R Core Team. R: A Language and Environment for Statistical Computing. (R Foundation for Statistical Computing, Vienna, Austria, 2014). at <http://www.R-project.org/> Cho, A. et al. WormNet v3: a network-assisted hypothesis-generating server for Caenorhabditis elegans. Nucleic Acids Res. (2014). doi:10.1093/nar/gku367 Lansbergen, G. et al. CLASPs attach microtubule plus ends to the cell cortex through a complex with LL5beta. Dev. Cell 11, 21–32 (2006). Van Spronsen, M. et al. TRAK/Milton motor-adaptor proteins steer mitochondrial trafficking to axons and dendrites. Neuron 77, 485–502 (2013). Waaijers, S., Koorman, T., Kerver, J. & Boxem, M. Identification of human protein interaction domains using an ORFeome-based yeast two-hybrid fragment library. J. Proteome Res. 12, 3181– 3192 (2013). Brenner, S. The genetics of Caenorhabditis elegans. Genetics 77, 71–94 (1974). Lamesch, P. et al. C. elegans ORFeome version 3.1: increasing the coverage of ORFeome resources with improved gene predictions. Genome Res. 14, 2064–2069 (2004). Fraser, A. G. et al. Functional genomic analysis of C. elegans chromosome I by systematic RNA interference. Nature 408, 325–330 (2000).

Figure S1 Supplementary figure 1 is not shown completely. Is available on request or when published. C35B8.4

LGL-1

TM 1

CLC-1

LLGL 1

100

941

TM TM TM

182

LIN-2a

S_TKc

L27

PDZ

SH3

GuKc

RAP-1

961

RAS 1

188

CDC-42

RHO 1

191

MES-1

STYKc

CED-10b

966

RHO 1

191

DLG-1a RHO-1

RHO 1

L27

PDZ

SH3

GuKc

967

192

RHEB-1

RAS 1

207

LIN-10a

PTB

PDZ

LIN-7a

L27

982

PDZ

209

MAGI-1a VAB-9

PDZ

GuKc

PDZ

PMP22_Claudin

1054

211

PAR-5

14_3_3 1

248

SAX-7a

IGc2

FN3

CLC-2

FN3

TM 1144

Clc-like 1

252

RAL-1b

RAS 1

254

AJM-1c

CC 1

PAR-6a

PB1 1

CC 1148

PDZ 309

Figure S1. Graphical representation of the design of the bait protein fragments. All bait proteins and their predicted protein domains are indicated. Grey lines are the fragments designed. Fragment design takes into account the size and domain composition of each protein. Red and blue arrows represent forward and reverse amplification primers respectively.

A combined binary interaction and phenotypic map of C. elegans cell polarity proteins

Figure S2 full-length bait screens

fragment bait screens

Non-self-activating bait constructs

309 (68 proteins)

Screen (Leu-, Trp-, His-)

3016

8712

primary colonies

Eliminate autoactivators Sequence max. 96 / polarity protein

1125 traces

3709 traces Remove out-of-frame Align to WS235 genome

759 alignments

2463 alignments Collapse to distinct proteins

115 pairs to verify

622 pairs to verify

Pair identified â&#x2030;Ľ 2x no

yes

Source of AD clone EE fragment

cDNA

STOP

Retest fail

Confirm bait identity pass

pass

fail STOP

STOP

439 interactions Figure S2. Overview of the yeast two-hybrid screening and computational validation pipeline used to establish the C. elegans polarity interaction network. See Online Methods for detailed descriptions of the screening procedure, computational analysis, and retesting procedure.

Figure S3 EE=0.62 p=2.81x10-7

AD-cDNA screens

Fraction of interactions

0.6

AD-Fragment screens

0.5

Interactions Search space

Fraction of interactions

0.4

EE=4.54 p=1.04x10-8

0.2

0.0 0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

0.3 0.2

EE=1.99 p=0.0261

0.1 0.0

1.0

Interactions Search space

0.4

0.0

0.1

GO semantic similarity

0.5

0.6

0.7

0.8

0.9

1.0

All baits x AD-Fragment Interactions Search space All C. elegans gene pairs

0.8 Fraction of gene pairs

0.8 0.6 0.4

0.2 0.0

0.0 −1.0

−0.5

0.0

0.5

1.0

−1.0

−0.5

PCC

0.5

1.0

Fragment baits x AD-cDNA Interactions Search space All C. elegans gene pairs

0.0

0.2

0.4

0.6

Fraction of gene pairs

0.8

Interactions Search space All C. elegans gene pairs

0.6

1.0

Full-length baits x AD-cDNA

0.8

0.0 PCC

0.4

Fraction of gene pairs

0.4

0.6

1.0

All baits x AD-cDNA Interactions Search space All C. elegans gene pairs

0.2

Fraction of gene pairs

0.3

GO semantic similarity

0.4

0.2

−1.0

−0.5

0.0 PCC

0.5

1.0

−1.0

−0.5

0.0

0.5

1.0

PCC

Figure S3. Enrichment of shared Gene Ontology (GO) terms and similar expression profiles in interacting protein pairs. Enrichments are always calculated by comparison of the interacting protein pairs with all possible protein pairs in the search space. For AD-cDNA derived interactions, this corresponds to the tested baits × all C. elegans proteins, and for AD-Fragment derived interactions to the tested baits × the 749 proteins in the AD-Fragment library. (a) GO similarity scores for interacting protein pairs identified using the AD-cDNA library. Semantic similarity scores for the BP component of GO were calculated using the HRSS software package, which scores protein pairs on a scale of 0 (dissimilar GO terms) to 1 (similar GO terms)45. Compared to the search space, protein pairs in the interaction network are depleted for pairs with a low semantic similarity score, and enriched for pairs with a high similarity score. (b) Same as (b), but for interactions identified using the AD-Fragment scores. Due to the already high semantic similarity scores for pairs in the search space (749 genes involved in early embryonic development), enrichment scores are less high than for AD-cDNA derived interactions. (c–f) Cumulative distributions of Pearson correlation coefficients (PCCs) for mRNAs corresponding to protein pairs in the indicated interaction datasets (red lines), all possible protein pairs in the search space (blue lines) and all C. elegans protein pairs (dotted grey lines). PCC values were calculated using the compendium of expression microarray data collected in Wormbase release WS236.

A combined binary interaction and phenotypic map of C. elegans cell polarity proteins

Figure S4

AP: streptavidin WB: αGFP

lysate lysate WB: streptavidin-HRP WB: αGFP

AP: streptavidin WB: αGFP

bait: NFM-1 prey: MEL-26 (frag.) kDa 170 130 100 70

1 2 3

LIT-1 CPL-1 (frag.)

ERM-1 NRFL-1 (frag.)

PAR-2 TYR-1 (fl.)

1 2 3

PAR-2 TYR-1 (frag.)

1 2 3

TAG-117 KCA-1 (fl.)

1 2 3

55 40

35 25 kDa 170 130 100 70 55 40 35 25 kDa 170 130 100 70 55 40 35 25

bait: LIN-2 prey: GMEB-3 (fl.)

LIN-2 GMEB-3 (frag.)

LIT-1 BRP-1 (frag.)

LIT-1 DLAT-1 (fl.)

LIT-1 DLAT-1 (frag.)

1 2 3

kDa

AGS-3 ZEN-4 (frag.)

1 2 3

100 70 55 40 35 25

lysate WB: streptavidin-HRP

lysate WB: αGFP

kDa 130 100 70 55 40 35 25 kDa 170 130 100 70 55 40 35

Lanes: 1: EGFP 2: EGFP-prey 3: EGFP-prey

+ Avi-mCherry-bait + Avi-mCherry + Avi-mCherry-bait

kDa

TAG-117 Y37E3.11 (frag.)

1 2 3

DSH-1 RGS-7 (fl.)

1 2 3

170 130 100 70

DLG-1 TAC-1 (frag.)

1 2 3

AGS-3 GOA-1 (fl.)

1 2 3

AGS-3 GOA-1 (frag.)

1 2 3

35 25 kDa 170 130 100 70 55 40 35 25 kDa 170 130 100 70 55 40 35 25 15

VAB-9 ZYG-12 (frag.)

1 2 3

PAC-1 IFA-4 (frag.)

1 2 3

SAX-7 PCN-1 (frag.)

1 2 3

120 85 50 35

bait: PAC-1 (f.l.) prey: PAR-6 (133-265)

lysate WB: αGFP

kDa 120 85 50 35

lysate WB: streptavidin-HRP

lysate lysate WB: streptavidin-HRP WB: αGFP

DSH-1 USO-1 (frag.)

55 40

AP: streptavidin WB: αGFP

AP: streptavidin WB: αGFP lysate WB: αGFP lysate WB: streptavidin-HRP AP: streptavidin WB: αGFP

bait: prey:

bait: DLG-1 prey: TAC-1 (fl.) kDa

kDa 120 85 50 35 15

Lanes: 1: EGFP 2: EGFP-prey 3: EGFP-prey

kDa 170 130 100 70 55 40

1 2 3

PAC-1 (1221-1320) PAC-1 (1221-1604) PAR-6 (133-265) PAR-6 (133-265)

1 2 3

35 25 15 kDa 170 130 100 70 55 40 35 25 15 kDa

170 130 100

+ Avi-mCherry-bait + Avi-mCherry + Avi-mCherry-bait

Figure S4. Western blots for all protein pairs testing positive by co-affinity purification. For every pair, three blots are shown. Top: detection of GFP-tagged proteins that co-purify with biotinylated mCherrytagged proteins using an anti-GFP antibody. Middle: detection of GFP-tagged proteins in total lysates using an anti-GFP antibody. Bottom: detection of biotinylated Avi-mCherry tagged proteins in total lysates using streptavidin coupled to horse radish peroxidase. Protein pairs tested are indicated above the blots. The first protein listed is the Avi-mCherry tagged bait. Also indicated is whether the prey protein tested was full-length (f.l.) or corresponds to a shorter fragment (frag.). In all blots, lanes 1 and 2 are negative controls, and lane 3 is the actual affinity purification (1: Avi-mCherry-bait vs. EGFP alone, 2: Avi-mCherry alone vs. EGFP-tagged prey, 3: Avi-mCherry-bait vs. EGFP-prey). The ~35 kDa band in lanes 2 of the upper blots is due to cross-reactivity of the anti-GFP antibody with the highly abundant Avi-mCherry polypeptide. Asterisks indicate bands of expected molecular mass.

A combined binary interaction and phenotypic map of C. elegans cell polarity proteins

Figure S5 Supplementary figure 5 is not shown completely. Is available on request or when published. RHO-1

PRY-1

SMO-1

RAP-1

UBQ 1

RAS 1

188

EAT-20

RLA-1

ZC449.7

Ribosomal_60s 1

111

189

NPP-8, T02C5.1

RAB-11.1, RAN-1

LEV-11, ZYG-12

CDC-42

F59C12.3, IFA-1, LET-607, MUA-6

CLC-1

RHO 1

TM TM TM

191 RAP-1

182

RHO-1

APR-1

F17E9.5

TM TM

RHO 1

192 DSH-2

184

RACK-1

FRM-2

EBP-3

WD40 WD40WD40WD40WD40WD40 WD40 1

EB1 1

325 HMP-1

187

MDT-8a

MAGI-1

CSC-1

LIT-1

Med8 1

CC 1

325 LIT-1

249

Y119C1A.1

C15H9.4, T24F1.2

RAL-1b

CCZnF_RBZ 1

RAS 1

325 MAGI-1

254

PAR-6 PAR-3

AGS-3

SYS-1

DLG-1, PAC-1 PRY-1

DNC-2

MIG-5

TAC-1

TACC 1

Dynamitin

331

260

CRB-1

APX-1

DSL EGFEGF EGF EGF

515 CLC-1 NFM-1

F59C12.3

CC 1

CC 519

Figure S5. Graphical representation of every protein in the interaction network and the MRIs identified for that protein. Grey boxes are full-length proteins. Predicted domains are shown as boxes of various colors and shapes. Identified MRIs are shown as yellow lines above the protein graphic. Protein names above each MRI indicate which proteins bind to that particular region.

Figure S6 Supplementary figure 6g is shown on the next page.

10 0 MRI size (residues)

80 90 99

100

MRI size (% of full length)

Prey MRIs, AD-Fragment library Percentage of MRIs

Percentage of MRIs

f 50 40 30 20 10 0 10 0 0 20 0 30 0 40 0 50 0 60 0 70 0 80 0 90 10 0 >1 0 0 00 0

Bait MRIs

MRI size (% of full length)

e 40

0 10 0 20 0 30 0 40 0 50 0 60 0 70 0 80 0 90 10 0 >1 0 0 00 0

Percentage of MRIs

100

30 40

MRI size (% of full length)

Percentage of MRIs

Prey MRIs, AD-cDNA library

0 10 20

100

5 0 10 20

80 90 99

50 60

MRI size (residues)

Prey MRIs, AD-cDNA library

30 20 10 0 10 0 0 20 0 30 0 40 0 50 0 60 0 70 0 80 0 90 10 0 >1 0 0 00 0

80 90 99

40 50

Percentage of MRIs

c Prey MRIs, AD-Fragment library

50 60

b Bait MRIs

0 10 20

Percentage of MRIs

MRI size (residues)

Figure S6. Analysis of MRIs (aâ&#x20AC;&#x201C;c) Distributions of MRI lengths expressed as the percentage of the corresponding full-length protein. (dâ&#x20AC;&#x201C;f) Distributions of absolute MRI lengths in amino acid residues. (a,d) MRIs identified for the bait proteins. (b,e) MRIs identified for the prey proteins from AD-Fragment library derived clones. (c,f) MRIs identified for the prey protein from AD-cDNA library derived clones. (g) Graphical representations of all MRIs where interaction sites had previously been identified in the literature. Grey boxes are full-length proteins. Predicted domains are shown as boxes of various colors and shapes. Identified MRIs are shown as yellow lines above the protein graphic. Interaction sites from the literature are shown as blue lines above the protein graphic. In cases where the literature describes an interaction domain for a non-C. elegans, the corresponding site in the C. elegans protein is shown.

HMP-1b

BAR-1

NFM-1b

HMP-2b

PRKL-1a

F59C12.3

POP-1

GSK-3

PAR-6a

LIM

FERM_C

Vinculin

NFM-1a binding region

ARM

POP-1 binding region

F59C12.3 binding region

523

519

709

704

AGS-3a

ERM-1b

VANG-1a

Vinculin

HMP-1b binding region

Vinculin

MAGI-1a binding region

VANG-1a binding region

438

Armadillo/β-Catenin repeats

LIM

HMP-1b binding region

Vinculin

362

NFM-1b binding region

LIM

HMP-2b binding region

ARM ARM

B41

PET

HMG

S_TKc

309

MRI defined by Y2H Interaction domain from literature Predicted disordered region

PDZ

PRY-1 binding region

BAR-1 binding region

PB1

PKC-3 binding region

FERM_C

930

B41

APR-1b

MAGI-1a

LIN-10a

DLG-1a

LIN-2a

TPR motifs

FERM_C

ERM-1b binding region

B41

NRFL-1b binding region

B41

TPR motifs

811

Strabismus

L27

ARM

S_TKc

LIN-10a binding region

AJM-1c binding region

532

GoLoco motifs

GOA-1 binding region

ERM

ERM-1b binding region

PRKL-1a binding region

GuKc

PDZ

579

564

ARM ARM

L27

PDZ

DLG-1a binding region

PDZ

PB1

RGS

B41

PTB

PDZ

LIN-2a binding region

SH3

PDZ

PDZ 982

967

961

PDZ

DLG-1a binding region

PDZ 1054

NFM-1b binding region

HMP-2b binding region

PDZ

GuKc

586

DIX

608

597

S_TK_X

ERM-1b binding region

S_TKc

Armadillo/β-Catenin repeats

FERM_C

PAR-6a binding region

GSK-3 binding region

HMP-1b binding region

APR-1b binding region

L27

HMP-2b

NFM-1a

NRFL-1b

PKC-3

PRY-1

1188

654

704

A combined binary interaction and phenotypic map of C. elegans cell polarity proteins

Figure S7 Additional examples of RNAi phentoypes are shown on the next page. control RNAi

Intestine

RNAi of bait genes

EV RNAi

ajm-1 RNAi

kin-19 RNAi

PEPT-1::dsRed EV RNAi

PEPT-1::dsRed ajm-1 RNAi

PEPT-1::dsRed

GFP::RAB-11

kin-19 RNAi

Figure S7. Additional examples of RNAi phenotypes. (a) Examples of novel phenotypes observed in the Intestine during phenotypic analysis of the bait proteins. (b) Examples of interacting protein pairs where RNAi for both genes resulted in a defect in the same marker strain. In all panels, the gene inactivated and the marker shown are indicated. EV: empty vector control RNAi. In the PVD neuron panels, an arrow indicates the spreading of RAB-3 positive vesicles into dendrites (mCherry::RAB-3 panels) and defects that arise during the development of the PVD neuron (Myristoyl::GFP panels). Scale bars represent 10 Îźm.

A combined binary interaction and phenotypic map of C. elegans cell polarity proteins

control

bait

prey

EV RNAi

lit-1 bait RNAi

sec24.1 prey RNAi

GFP::VIT-2

EV RNAi

cdc-42 bait RNAi

rab-11.1 prey RNAi

RME-2::GFP

EV RNAi

erm-1 bait RNAi

GFP::VHA-5

EV RNAi

par-3 bait RNAi

C18E9.10 prey RNAi

long exp. RME-2::GFP skr-9 prey RNAi

PVD neuron

Embryo

Excretory canal

Yolk trafficking

mCherry::PAR-6

EV RNAi

sys-1 bait RNAi

hipr-1 prey RNAi

SAD-1::GFP

EV RNAi

sys-1 bait RNAi

hipr-1 prey RNAi

mCherry::RAB-3

EV RNAi

sys-1 bait RNAi

hipr-1 prey RNAi

Myristoyl::GFP

Figure S8 mCherry::PAC-1_1221-1604 eGFP::EV

Merge

mCherry::EV

eGFP::PAR-6_133-265

Merge

mCherry::PAC-1_1221-1604 eGFP::PAR-6_133-265

Merge

Figure S8. Subcellular localization of mCherry::PAC-1_1221-1604 and EGFP::PAR-6_133-265 in HeLa cells. Three panels of three pictures are shown. In the top panel we co-expressed mCherry::PAC-1_1221-1604 together with GFP::EV. In the middle panel we co-expressed mCherry::EV with EGFP::PAR-6_133-265. In the bottom panel we co-expressed mCherry::PAC-1_1221-1604 with EGFP::PAR-6_133-265. The panels show the red mCherry channel, the green GFP channel and the merge of both. Scale bar represents 10 Îźm.

A combined binary interaction and phenotypic map of C. elegans cell polarity proteins

Supplementary table 1

Supplementary table 1 Sheet 1: summary of bait construct cloning and screening. Legend Bait construct that were cloned but not screened were self-activating at 2 mM 3AT Bait Protein Full length bait constructs Fragment bait constructs Cosmid ID Name designed cloned screened No. No. cloned No. designed screened K04G2.8 APR-1 yes yes at 2 mM 3AT 9 9 8 F32A6.4 AGS-3 yes yes yes 6 6 6 C25A11.4 AJM-1 yes yes yes 9 7 7 C54D1.6 BAR-1 yes yes yes 12 10 9 C35B8.4 CRB-3 yes yes yes 1 1 1

T23D8.9 T22C8.8 B0410.2 B0336.1

SYS-1 VAB-9 VANG-1 WRM-1

yes yes yes yes

no yes yes yes

9 2 5 9

8 2 4 9

6 2 4 9

Supplementary table 1 Sheet 2: Details of the bait constructs generated Legend Start and end are relative to the full-length protein as present in Wormbase release WS235. All coordinates are in amino acids. Cosmid ID Name Bait start Bait end bait_type Cloned Scr. F32A6.4a AGS-3a 1 579 full length yes yes F32A6.4a AGS-3a 1 287 fragment yes yes F32A6.4a AGS-3a 1 436 fragment yes yes F32A6.4a AGS-3a 140 287 fragment yes yes F32A6.4a AGS-3a 140 436 fragment yes yes

B0336.1 B0336.1 B0336.1 B0336.1

WRM-1 WRM-1 WRM-1 WRM-1

256 400 400 532

796 664 796 796

fragment fragment fragment fragment

yes yes yes yes

Supplementary table 2

Supplementary Table 2 Sheet 1: The complete C. elegans polarity interactome network Legend Interactions found in both orientations (DB-X/AD-Y and DB-Y/AD-X) are collapsed to unique proteins

Protein A Cosmid ID Name K04G2.8 APR-1 K04G2.8 APR-1 K04G2.8 APR-1 K04G2.8 APR-1

B0336.1 B0336.1 B0336.1

Protein B Cosmid ID Name B0336.6 ABI-1 ZK993.1 CEH-45 B0547.1 CSN-5 C25F6.2 DLG-1

WRM-1 T21B10.3 WRM-1 Y51F10.2 WRM-1 ZK1098.4

T21B10.3 Y51F10.2 ZK1098.4

No. of hits AF-Fragment AD-cDNA Total Orientations 0 1 1 1 2 0 2 1 0 12 12 1 0 1 1 1

0 0 0

1 5 4

1 1 1

Supplementary Table 2 Sheet 2: Interactions identified using full-length bait constructs Legend Interactions found in both orientations (DB-X/AD-Y and DB-Y/AD-X) are collapsed to unique proteins

Protein A Cosmid ID Name K04G2.8 APR-1 K04G2.8 APR-1 K04G2.8 APR-1 K04G2.8 APR-1

B0336.1 B0336.1 B0336.1

Protein B Cosmid ID Name B0336.6 ABI-1 F54D5.5 F54D5.5 F38A6.3 HIF-1 T05C12.6 MIG-5

WRM-1 T21B10.3 WRM-1 Y51F10.2 WRM-1 ZK1098.4

T21B10.3 Y51F10.2 ZK1098.4

No. of hits AF-Fragment AD-cDNA Total Orientations 0 1 1 1 0 1 1 1 0 2 2 1 0 2 2 1

0 0 0

1 5 4

1 1 1

A combined binary interaction and phenotypic map of C. elegans cell polarity proteins

Supplementary Table 2 Sheet 3: Interactions identified using fragment bait constructs Legend Interactions found in both orientations (DB-X/AD-Y and DB-Y/AD-X) are collapsed to unique proteins

Protein A Cosmid ID Name K04G2.8 APR-1 K04G2.8 APR-1 K04G2.8 APR-1 K04G2.8 APR-1 K04G2.8 APR-1

T22C8.8 T22C8.8 B0410.2

Protein B Cosmid ID Name ZK993.1 CEH-45 B0547.1 CSN-5 C25F6.2 DLG-1 C33G3.1 DYC-1 Y59A8B.7 EBP-1

VAB-9 F33D11.11 VAB-9 ZK546.1 VANG-1 ZK993.1

No. of hits AF-Fragment AD-cDNA Total Orientations 2 0 2 1 0 12 12 1 0 1 1 1 0 1 1 1 0 3 3 1

VPR-1 ZYG-12 CEH-45

0 4 4

1 1 0

1 5 4

1 1 1

Supplementary table 3

Only the summary sheet will be shown.

Supplementary Table 3 Sheet 4: Summary of literature evidences for interactions. Protein A Protein B Evidence source Cosmid ID Name Cosmid ID Name Intact elegans Intact Ortholog Pubmed C25F6.2 DLG-1 T22A3.3 LST-1 yes C25F6.2 DLG-1 Y54E2A.3 TAC-1 yes C27A2.6 DSH-2 T17H7.4 PAT-12 yes C01G8.5 ERM-1 C18B2.5 C18B2.5 yes F17E5.1 LIN-2 C44F1.2 GMEB-3 yes -

F32A6.4 C01G8.5 F17E5.1 F09E5.1

AGS-3 ERM-1 LIN-2 PKC-3

C26C6.2 C01F6.6 C09H6.2 T26E3.3

GOA-1 NRFL-1 LIN-10 PAR-6

yes yes yes yes

SYS-1a VAB-9 DSH-1c PAC-1b AGS-3a

DLG-1a LIN-2a LIT-1a MAGU-2 PAR-2b

T23D8.9a T22C8.8 C34F11.9c C04D8.1b F32A6.4a

1 1 422 1 140

271 100 702 417 436

811 211 702 1604 579

ZK370.3b ZK546.1c K09B11.9a K05B2.3 Y51H7C.6a

HIPR-1b ZYG-12c USO-1a IFA-4 COGC-4a

589 336 651 272 470

921 761 941 575 664

921 761 941 575 801

(prey not expressed) (prey not expressed) (prey not expressed) (prey not expressed) No

No Yes Yes Yes No

C25F6.2a F17E5.1a W06F12.1a C01B7.4 F58B6.3b

1 1 181 135 1

967 961 634 416 289

967 961 634 668 582

Y54E2A.3 C44F1.2 F23B12.5 C10H11.10 C02C2.1

TAC-1 GMEB-3 DLAT-1 KCA-1 TYR-1

48 3 154 35 348

260 376 507 407 601

Yes Yes Yes Yes Yes

Yes Yes Yes No Yes

Supplementary Table 4 Protein fragments used in co-affinity purifications, and co-affinity purification results. Bait Protein Prey Protein Results Bait Bait Bait Bait Full-length Prey splice Prey Prey Prey Full-length Full-length prey Fragment Cosmid Common Start End protein common Start End protein prey

Supplementary table 4

A combined binary interaction and phenotypic map of C. elegans cell polarity proteins

Supplementary table 5

From this table sheet 1, 3 and 5 will be shown. Sheet 2 is too large to add to the thesis.

Supplementary Table 5 Sheet 1: Minimal Regions of Interaction (MRIs) identified Legend For all interactions the MRI of both bait and prey protein is indicated, as is the full-length size of the protein.

All coordinates are in amino acids. Bait Protein Bait MRI data Prey Protein Prey MRI data Cosmid ID Name Start End Bait full Cosmid ID Name Start End Prey full length length

B0336.1 B0336.1 B0336.1 B0336.1 B0336.1

WRM-1 WRM-1 WRM-1 WRM-1 WRM-1

ZK381.5a ZK381.5a ZK381.5a

PRKL-1a PRKL-1a PRKL-1a

1 1 1 1 1

796 796 796 796 796

C06C3.1c C07G1.5 C17E7.4 C36B1.8b C39D10.7

MEL-11c HGRS-1 C17E7.4 GLS-1b C39D10.7

808 305 1 641 783

1124 729 364 1052 1171

270 132 270

523 523 523

523 Y53C12A.1 WEE-1.3 523 ZK381.5a PRKL-1a 523 ZK546.1b ZYG-12b

333 2 537

537 523 778

677 523 777

Supplementary Table 5 Sheet 3: Details of the fragments from which the Bait MRIs were derived. Legend Table of all unique bait fragment / prey protein combinations. MRIs are defined as largest start coordinate to smallest stop coordinate. All coordinates are in amino acids. Bait Protein Bait MRI data Prey Protein Source Cosmid ID Name Start End Full protein Cosmid ID Name Library Hits length

F32A6.4a F32A6.4a F32A6.4a F32A6.4a F32A6.4a

AGS-3a AGS-3a AGS-3a AGS-3a AGS-3a

B0336.1 B0336.1 B0336.1

WRM-1 WRM-1 WRM-1

140 140 288 140 140

1 1 1

436 436 579 436 436

579 579 579 579 579

Y51H7C.6 F23B12.5 C26C6.2 Y54E2A.3 M03D4.1

796 796 796

796 T21B10.3 796 Y51F10.2 796 ZK1098.4

COGC-4 DLAT-1 GOA-1 TAC-1 ZEN-4

AD-Fragment AD-Fragment AD-Fragment AD-Fragment AD-Fragment

6 2 7 18 2

T21B10.3 AD-cDNA Y51F10.2 AD-cDNA ZK1098.4 AD-cDNA

1 5 4 87

Supplementary Table 5 Sheet 4: Details of the fragments from which the Prey MRIs were derived. Legend Table of all unique bait protein / prey fragment combinations. MRIs are defined as largest start coordinate to smallest stop coordinate. All coordinates are in amino acids. Bait Protein Prey Protein Prey MRI data Source Cosmid ID Name Cosmid ID Name Start End Full protein Library Hits length

F32A6.4 F32A6.4 F32A6.4 F32A6.4 F32A6.4

AGS-3 AGS-3 AGS-3 AGS-3 AGS-3

B0336.1 B0336.1 B0336.1

WRM-1 T21B10.3 WRM-1 Y51F10.2 WRM-1 ZK1098.4

Y51H7C.6a F23B12.5 C26C6.2 Y54E2A.3 Y54E2A.3

COGC-4 DLAT-1 GOA-1 TAC-1 TAC-1

470 1 3 1 1

664 507 354 100 199

T21B10.3 Y51F10.2 ZK1098.4

592 1 1

1170 305 305

801 507 354 260 260

AD-Fragment AD-Fragment AD-Fragment AD-Fragment AD-Fragment

1170 AD-cDNA 305 AD-cDNA 305 AD-cDNA

6 2 7 4 3

1 5 4

A combined binary interaction and phenotypic map of C. elegans cell polarity proteins

Supplementary table 6

Supplementary Table 6 Sheet 1: phenotypes examined Strain name Tissue Markers Phenotypes MZE1 Intestine PEPT-1::DsRed Irregular morphology of intestine GFP::RAB-11 Diffuse RAB-11 vesicles Reduced number of RAB-11 vesicles Abnormal RAB-11 vesicle localization PEPT-1 fuzzy PEPT-1 levels reduced SV1009 Seam cells GFP::PH Altered number of cells GFP::H2B Disorganized epithelium Abnormal morphology Altered division plane BOX56 Epithelial junctions DLG-1::GFP GFP levels reduced DH1033 Yolk trafficking VIT-2::GFP VIT-2 accumulation in intestine Reduced VIT-2 levels in oocyte Diffuse VIT-2 accumulation in body cavity Droplet VIT-2 accumulation in body cavity Dense VIT-2 clusters in embryo RT408 Yolk trafficking RME-2::GFP Cytoplasmic cluster of RME-2 Cortical RME-2 levels reduced JH2647 Embryo mCherry::PAR-6 PAR-6 levels reduced Embryonic development abnormal PAR-6 distribution symmetric First division symmetric Multiple nuclei PAR-6 visible in cytoplasmic punctae CX9797 PVD neuron mCherry::RAB-3 Spreading of RAB-3 vesicles in 1st through 3rd degree dendrites

SAD-1::GFP

STR62 PVD neuron Myristoyl::GFP ML846 Excretory canal VHA-5::GFP

Spreading of RAB-3 vesicles to 4th degree dendrites

Extra PVD cell bodies SAD-1 spreading into dendrites RAB-3 vesicle clusters in cell body or dendrite Ectopic dendrite branching Fused dendrite branches Under-developed 4th degree dendrites Overlapping dendrites Retrograde growth of dendrites Extra PVD neurons No or little canal development Shorter canals Ruptured or interrupted canals Cyst-like structures in canals Irregular or twisted canals

From this table sheet 2 is too large to show in this thesis.

Supplementary Table 6 Sheet 3: strongest phenotype for each tissue examined Legend 0: no phenotype observed. -: not tested. Gene Tissue examined Cosmid ID Name Intestine Seam Epithelial Yolk Embryo PVD Excretory cells junctions trafficking canal B0041.2 B0303.14 B0336.1 B0336.6 B0336.7

ain-2 B0303.14 wrm-1 abi-1 B0336.7

0 -

3 0 1 0 0

0 -

0 0 0 -

0 0

0 0 1 0

0 0 -

ZK593.5 ZK858.4 ZK993.1

dnc-1 mel-26 ceh-45

3 0 0

2 0

1 -

3 0

0 0

Supplementary Table 6

Sheet 4: Interacting proteins where RNAi of corresponding genes results in defects in overlapping tissues. Legend

‘yes’ indicates that RNAi for both genes results in a defect in the indicated tissue / process Gene A

Cosmid ID

Gene B

Name Cosmid ID

W06F12.1 lit-1

F12F6.6

C25A11.4 ajm-1 C25F6.2

Name

Intestine Seam Epithelial Yolk Embryo PVD Excretory cells junctions trafficking canal

sec-24.1

yes

dlg-1

yes

F09E5.1

pkc-3 T26E3.3

par-6

yes

C04D8.1

pac-1 ZK593.5

dnc-1

yes

C04D8.1

pac-1 T26E3.3

par-6

Tissue

yes

C01G8.5

erm-1 C52D10.7 skr-9

yes

C01G8.5

erm-1 B0464.5

yes

C01G8.5

erm-1 Y39B6A.1 Y39B6A.1 spk-1

yes

A combined binary interaction and phenotypic map of C. elegans cell polarity proteins

Supplementary Table 6

Sheet 4: Interacting proteins where RNAi of corresponding genes results in defects in overlapping tissues. Legend

‘yes’ indicates that RNAi for both genes results in a defect in the indicated tissue / process Gene A

Gene B

Cosmid ID

Name Cosmid ID Name

W06F12.1

lit-1

C25A11.4

F09E5.1

C04D8.1

F12F6.6

ajm-1 C25F6.2

pkc-3 T26E3.3

pac-1 T26E3.3

sec-24.1

dlg-1

par-6

Tissue

yes

dnc-1

yes

C25F6.2

dlg-1 T20G5.1

chc-1

yes

C04D8.1

W06F12.1

F23B12.5

pac-1 T20G5.1

pac-1 W10G6.3

lit-1

Y47D3A.6

dlat-1

chc-1

mua-6

tra-1

yes

pac-1 ZK593.5 lit-1

Intestine Seam Ep. Yolk Embryo PVD Excr. cells junctions trafficking canal

C04D8.1

W06F12.1

yes

F54E7.3

par-3 Y44E3A.6

Y44E3A.6

yes

F54E7.3

par-3 T22A3.3

lst-1

yes

C04D8.1 F58B6.3

C04D8.1 C04D8.1 C25F6.2

W06F12.1 C25A11.4

F27E11.3

F54E7.3

C25A11.4

F54E7.3

C25F6.2

C04D8.1

pac-1 Y54E2A.3 par-2 F56C9.1

pac-1 F33G12.5 pac-1 K09B11.9 dlg-1 T22A3.3 lit-1

T22A3.3

cfz-2

T07A9.9

ajm-1 ZC247.1

par-3 F54D5.15

ajm-1 T20G5.1

par-3 K09B11.9

dlg-1 Y54E2A.3

tac-1

gsp-2

F33G12.5 uso-1 lst-1 lst-1

ZC247.1

T07A9.9

F54D5.15

chc-1

uso-1

tac-1

pac-1 W03D8.2

W03D8.2

pac-1 F29G9.2

F29G9.2

pac-1 Y65B4BR.4 wwp-1

pac-1 F38B2.1

pac-1 F07A5.7

ifa-1

unc-15

yes yes yes yes yes yes yes

yes

yes yes

yes

F58B6.3

par-2

Y37E11AR.2 siah-1

yes

W06F12.1

lit-1

D1037.1

yes

W06F12.1 W06F12.1 R07G3.1

W10C8.2

W10C8.2 W02B9.1

lit-1 lit-1

Y119C1A.1

Y119C1A.1 D1037.1

Y57A10A.24 Y57A10A.24

cdc-42

K01G5.4

ran-1

pop-1

R119.4

pqn-59

pop-1

hmr-1

K04H4.1 B0041.2

emb-9 ain-2

Y92H12A.1 src-1

T05G5.9

Y92H12A.1 src-1

ZK381.5

prkl-1

C01G8.5

C30B5.4

R13H4.4

hmp-1 K05C4.6

C01G8.5

erm-1

W06F12.1

lit-1

W10C8.2 C25F6.2

C04D8.1

erm-1

pac-1

W04D2.1

atn-1

pop-1

Y105E8B.1

lev-11

dlg-1

T23D8.9

sys-1

W06F12.1 F17E5.1

spd-5

F59A2.1

gsk-3

T23D8.9

F56A3.4

C18B2.5

pop-1

Y18D10A.5 W10C8.2

C18B2.5

hmp-2

sys-1 lit-1

Y105E8B.1 F44B9.7

Y105E8B.1 Y105E8B.1

atn-1

lev-11 mdt-30 lev-11 lev-11

F55F8.4

cir-1

Y54G11A.10 lin-7

R12B2.4

him-10

C54D1.6

bar-1

W09D10.1

F17E5.1

lin-2

F17E5.1

lin-2

W04D2.1

npp-9

lin-2

Y54G11A.10 lin-7 C09H6.2

lin-10

C12C8.3

F11C3.3

F29G6.3

C07G1.5

K05C4.6

hmp-2 Y110A7A.1

W02B9.1

hmr-1

W02B9.1 R07G3.1 F17E5.1

unc-54

hpo-34 hgrs-1 hcp-6

hmr-1

W09D10.1

cdc-42

F53G12.1

rab-11.1

lin-2

Y54G11A.10 lin-7 C54D1.6

bar-1

F17E5.1

lin-2

F17E5.1

lin-41

lin-2

T23G7.1 F02A9.6

T05G5.9

W04D2.1 W04D2.1

dpl-1 glp-1

T05G5.9 atn-1

yes

yes yes yes yes yes yes yes

yes

yes yes yes yes yes yes yes yes yes yes yes yes yes yes yes yes yes yes yes yes

yes yes yes yes yes yes

yes

C54D1.6

pop-1

yes

W10C8.2

Y105E8B.1

yes

hmp-2 lev-11

par-1

yes

Y105E8B.1

H39E23.1

yes

Y54G11A.10 lin-7

K05C4.6

lin-10

lin-2

yes

atn-1

C09H6.2

F17E5.1

yes

lev-11

bar-1

yes yes yes yes

A combined binary interaction and phenotypic map of C. elegans cell polarity proteins

F17E5.1

K05C4.6

T23D8.9

W06F12.1

F32A6.4

F54E7.3

F32A6.4

F54E7.3

F58B6.3

T23D8.9

K01A6.2

K04G2.8

K01A6.2

lin-2

sys-1

lit-1

ags-3

par-3

ags-3

par-3

par-2

W06F12.1

C01G8.5

Y54E2A.3

C14B9.4

ZK858.4

B0336.6

magi-1 W04D2.1

magi-1 F35B12.5

pop-1

F58B6.3

C18E9.10

T17H7.4

apr-1

W10C8.2

T23D8.9

M03D4.1

T02G5.9

pop-1

W06F12.1

F54C8.3

C26C6.2

sys-1

W10C8.2

K04G2.8

ZK381.4

C32E8.10

magi-1 F39C12.2

Y92H12A.1 src-1

K04G2.8

B0464.5

hmp-2 B0464.5

apr-1

sys-1

lit-1

par-2

erm-1

K09B11.9

ZK1248.3

spk-1

pgl-1

unc-11

emb-30

goa-1

zen-4

C18E9.10

pat-12

tac-1

plk-1

mel-26

kars-1

add-1

abi-1

atn-1

sas-5

uso-1

ehs-1

ZC317.7

C25F6.2

dlg-1

Y105E8B.1 lev-11

ZK370.3

T02E1.3

ZC317.7

DY3.2

C52D10.7

Y39B6A.1

B0464.5

hipr-1

gla-3

ZC317.7

lmn-1

skr-9

Y39B6A.1

spk-1

yes

Chapter 3 Identification of human protein interaction domains using an ORFeome-based yeast two-hybrid fragment library

Selma Waaijers1,2, Thijs Koorman1,2, Jana Kerver2, and Mike Boxem2 1. These authors contributed equally to this work 2. Utrecht University, Department of Biology, Padualaan 8, 3584CH Utrecht, The Netherlands

Adapted from: Journal of Proteome Research (2013) Jul 5;12(7): 3181-92

Abstract

hysical interactions between proteins are essential for biological processes. Hence, there have been major efforts to elucidate the complete networks of protein-protein interactions, or â&#x20AC;&#x2DC;interactomesâ&#x20AC;&#x2122;, of various organisms. Detailed descriptions of protein interaction networks should include information on the discrete domains that mediate these interactions, yet most large-scale efforts model interactions between whole proteins only. We previously developed a yeast twohybrid based strategy to systematically map interaction domains, and generated a domain-based interactome network for 750 proteins involved in C. elegans early embryonic development. Here, we expand the concept of Y2H based interaction domain mapping to the genome-wide level. We generated a human fragment library by randomly fragmenting the full-length open reading frames (ORFs) present in the human ORFeome collection. Screens using several proteins required for cell division or polarity establishment as baits demonstrate the ability to accurately identify interaction domains for human proteins using this approach, while the experimental quality of the Y2H data was independently verified in co-affinity purification assays. The library generation strategy can easily be adapted to generate libraries from fulllength ORF collections of other organisms.

Protein identification using an ORFeome-based Y2H fragment library

Introduction Eukaryotic proteins are modular in nature: most are predicted to contain multiple domains that can fold independently and retain their specific biological activity when expressed in isolation1,2. Domains are thought to represent functional units that facilitate the evolution of proteins with new or modified functions3. An important role of protein domains is the mediation of molecular interactions4,5. Interaction domains can interact with other interaction domains, with short linear peptide motifs, and with other macromolecules such as DNA or lipids. The specific combination of interaction domains and motifs determines the network of interactions a protein engages in, partly dictating its function. Large-scale efforts using affinity purification/mass spectrometry or the yeast two-hybrid (Y2H) system have made considerable progress describing the protein interactomes of several model organisms, as well as humans6â&#x20AC;&#x201C;13. However, large-scale interaction mapping approaches largely disregard the modular nature of proteins, and current domain-domain interaction database entries are either inferred from protein structures in the Protein Data Bank (PDB), or computationally predicted14â&#x20AC;&#x201C;16. Capturing information on interaction domains is essential for an accurate description of interactome networks, and facilitates generating biological hypotheses from such networks. In addition to generating more accurate descriptions of protein interactions, domain boundaries that are experimentally defined can be a valuable resource for approaches that rely on expressing large amounts of purified protein, such as structure determination by crystallography and high-throughput screens for drug discovery5,17,18. The yeast two-hybrid system is one of the most powerful and widely used methods available to date to identify protein interactions: it is cost-effective, easy to perform, scalable to whole-genome levels, and can be applied to map interactions of any species, including cross-species interactions such as virus-host interactions19,20. Since its inception, major improvements have been made to the Y2H system to increase the data quality and effectiveness of the approach21â&#x20AC;&#x201C;26. We previously demonstrated that the Y2H system can be used to systematically identify protein interaction domains for a set of 750 proteins involved in C. elegans early embryogenesis27. Here, we expand our approach and demonstrate the use of a genome-wide fragment library to accurately identify interaction domains of human proteins. We generated a human Y2H fragment library by mechanically fragmenting open reading frame (ORF) clones from the human ORFeome 5.128, a collection of 15,483 human ORF clones, and screened the library with a series of bait proteins with well described interactions. For each of 7 interactions where an interaction domain had previously

been defined, the minimal binding region identified by our approach matched the published interaction domain. Moreover, independent co-affinity purification assays validated 50% of the interaction domains we identified. Our experiments demonstrate that interacting regions of proteins can be systematically identified by Y2H from genome-wide ORFeome-derived random fragment libraries. The identification of interaction domains adds additional detail to interactome networks, while the experimentally defined domain boundaries that result in functional proteins in yeast are a valuable starting point for structurefunction analyses, and other approaches requiring expression of protein domains. The pipeline we used to generate the library can be used to generate comparable random fragment libraries for other systems for which ORF collections are available.

Materials and methods Y2H Gal4-AD vector Vector pMB26 is a modified version of pPC86[ref. 50] that contains a flexible linker (GGSSGA) between the Gal4-AD and the cloned ORF, and facilitates blunt-end cloning using SmaI. Gal4-AD is fused to the N-terminus of the ORF. pMB26 was generated by inserting an oligonucleotide linker into pPC86 digested with SalI and NotI. The linker was created by annealing oligos pMB26_F1: 5’- TCGAGTGGCGCGCCCGGGTAGC and pMB26_R1: 5’- GGCCGCTACCCGGGCGCGCCAC. pMB26 also contains the CYH2 gene from Clontech vector pAS2-1 (http://www.clontech.com). An EcoRV fragment containing CYH2 was cut from pAS2-1, and ligated into pPC86 digested with Acc65I and treated with Klenow + dNTPs to generate blunt ends. Amplification of ORFeome clones Each of the 15,483 ORF clones contained in the ORFeome 5.1 collection was PCR amplified using universal primers that anneal to the pDonr223 Gateway vector backbone (pDonr223_F: 5’-CCCAGTCACGACGTTGTAAAACG and pDonr223_R: 5’-GTAACATCAGAGATTTTGAGACAC). All PCR reactions were done in 384well plates in a 15 µl volume, using Novagen KOD Hot Start DNA Polymerase. The reactions were assembled and run according to the manufacturer’s instructions, using an annealing temperature of 58°C and an extension temperature of 68°C. The PCR products were collected in 5 pools based on size (<500 bp: 2610 ORFs, 500 – 1000 bp: 4811 ORFs, 1000 – 1500 bp: 3784 ORFs, 1500 – 2000 bp: 2116 ORFs, >2000 bp: 2162 ORFs), and purified using a Gel and PCR Clean-up kit (Machery-Nagel).

Protein identification using an ORFeome-based Y2H fragment library

Removal of vector sequences The PCR products contain ~75 bp of vector sequence preceding and following the ORF. These were removed by an Exonuclease III (Fermentas) treatment titrated to remove approximately 75 bp from each end. Each of the 5 PCR pools was incubated with 15 units Exonuclease III / pmol PCR product at 37°C for 8 minutes, and the reaction was stopped by heat inactivation at 75°C for 15 minutes. A 1 hr. treatment with mung bean nuclease (1 Unit / µg DNA) (NEB) was used to remove the single strand overhangs left by the Exonuclease III treatment. The Exonuclease treated PCR pools were not further purified before proceeding with the fragmentation. Fragmentation of PCR pools To fragment PCR pools to the desired size range, 10 µg of each Exonuclease III treated pool of PCR products was sheared using a Covaris S2 Focused-ultrasonicator in a 130 µl microTUBE. Settings for 700 bp fragments: Intensity = 3, Duty Cycle = 5%, Cycles per Burst = 200, Treatment Time = 60 s. Settings for 500 bp fragments: Intensity = 3, Duty Cycle = 5%, Cycles per Burst = 200, Treatment Time = 80 s. Settings for 250 bp fragments: Intensity = 4, Duty Cycle = 10%, Cycles per Burst = 200, Treatment Time = 130 s. Each pool of fragments was then subjected to electrophoresis on an agarose gel, and DNA in the desired size range was isolated and purified using a Gel and PCR Clean-up kit (Machery-Nagel). Ligation into pMB26 Fragmented DNA from each PCR pool was ligated into pMB26. To add a full-length component to the library, we also ligated into pMB26 DNA that had only been exonuclease treated, but not fragmented. The ends of the DNA to be ligated were converted to 5’-phosphorylated blunt ends using the Epicentre End-It DNA EndRepair Kit. Insert and vector were mixed in a 5:1 molar ratio, starting with 2 µg of insert DNA, in a final volume of 200 µl 1X T4 DNA ligase buffer containing 10 µl T4 ligase (50U) (Fermentas). For optimal ligation of blunt ends, the ligation reaction was cycled between 10°C and 30°C for 10 seconds each, for a total of 12 hours. Transformation of ligated fragments into bacteria Ligation mixtures were transformed into E. coli strain DH5α, made competent using the method of Inoue et al.51 100 µl of ligation mixture was mixed with 2 ml of competent bacterial cells, and split into 50 µl aliquots in a 96-well plate. The plate was incubated on ice for 30 min, heat shocked at 42°C for 45 s in a thermocycler, and placed back on ice for 2 min before adding 150 µl of SOC medium to each well. All bacterial suspensions were collected and pooled in a 50 ml Erlenmeyer flask, and

allowed to recover in a shaking incubator at 37°C for 1 hr. Finally, the transformations were plated on 15 cm ø LB-Agar plates containing 50 µg/ml Ampicillin (25 plates, 320 µl cells each), and grown overnight at 37°C. For each pool, we determined the desired number of colonies to obtain. First, we calculated the number of fragments per ORF needed such that there is a 99% probability that every part of the full-length ORF is covered. The number of fragments needed is given by the expression N = In (1-P)/In(1-f) where N is the number of colonies, P is the probability, and f is the fraction of the full-length ORF covered by the fragment.52 Thus, for the first pool (ORFs < 500 bp) aiming for a fragment size of 250 bp, N = In (1-0.99)/In(1-250/500), or 6.64 fragments/ORF. To obtain the number of colonies desired following ligation, the number of fragments/ ORF is multiplied with the number of ORFs present in the pool, and multiplied by six to account for only 1/6 clones being in frame. For the full-length pools, we aimed for 10x the number of clones in the pool. When necessary, we repeated each ligation and transformation until sufficient colonies were reached. Preparation of fragment library DNA After O/N growth, bacterial colonies were washed off each plate in 5 ml LB medium containing 50 µg/ml Ampicillin, using a sterile cell scraper. Colonies from the same ligation reaction were all combined in a single Erlenmeyer flask, and LB medium containing 50 µg/ml Ampicillin was added to a final volume of 100 ml. After growing the culture for 2 hrs. at 37°C in a shaking incubator, plasmids were purified using a Maxi-prep kit (Machery-Nagel). DNA from each of the 5 fragmented PCR pools cloned into pMB26 was mixed such that all ORFs are represented at similar levels. DNA from each of the 5 full-length pools was mixed in a similar fashion. Full-length and fragment components were then mixed in a 1:3 ratio to obtain the final library. Generating the Gal4-AD yeast mating library To generate AD mating libraries for screening, yeast strain Y8800 (genotype MATa trp1-901 leu2-3,112 ura3-52 his3-200 gal4Δ gal80Δ cyh2R GAL1::HIS3@LYS2 GAL2::ADE2 GAL7::LacZ@met2) was transformed with 90µg of the complete AD fragment library, using the LiAc method53. Yeast were plated on SC-Trp plates to select for transformants. After two days of growth at 30°C, all colonies were harvested in 200 ml YEPD medium containing 20% glycerol (w/v), and frozen in 2 ml aliquots at -80°C.

100

Protein identification using an ORFeome-based Y2H fragment library

Generating the Gal4-DB bait strains For each of the ORFs to be used as a Gal4-DB bait fusion, the corresponding Gateway Entry clone plasmid was isolated from the ORFeome 5.1 collection. The correct identity of the ORF was verified by sequencing. Each ORF was then transferred into the pDest-DB Y2H vector by Gateway recombinational cloning (Life Technologies). The bait constructs were transformed into yeast strain Y8930 (genotype MATα trp1901 leu2-3,112 ura3-52 his3-200 gal4Δ gal80Δ cyh2R GAL1::HIS3@LYS2 GAL2::ADE2 GAL7::LacZ@met2) using the LiAc method53. To test for autoactivation of Y2H reporter genes in the absence of an interacting AD-Y fusion protein, each bait strain was plated on an SC plate lacking leucine and histidine. Strains able to grow are able to activate the HIS3 reporter, and were eliminated from the screening process. Y2H screening All Y2H screens were done using a mating-based approach as previously described27. Each Gal4-DB bait strain was grown overnight at 30°C in 5 ml of YEPD medium in a shaking incubator. An equal amount of Gal4-DB cells and thawed Gal4-AD fragment library was then mixed in a 15 ml conical tube. For each screen, we used an amount of yeast that would yield an O.D. 600 value of 6 if suspended in 1 ml (e.g., if the O.D. 600 of the overnight culture is 3, then 2 ml of this culture was used). Mixed yeast cells were spun down at 650 g for 3 min, resuspended in 200 µl of sterile water, and plated on a 10cm ø YEPD media plate. The plates were incubated at 30°C for 4 hrs, which results in a low (~1%) mating efficiency. This minimizes the chance that yeast cells undergo mitosis after mating, which could lead to the identification of interactions from multiple colonies derived from the same parent yeast. After the incubation, yeast cells were washed off the plate in 200 µl sterile water, and plated on a 15cm ø SC –Leu –Trp –His media plate. After 4 days of growth at 30°C, colonies growing on the SC –Leu –Trp –His plates were picked into a 96-well plate with 25 µl of water, and plated on two fresh SC – Leu –Trp –His plates. Controls of known reporter activity strength were also added to these plates. After 3 days of growth at 30°C, yeast from one of the two plates was used to identify the identity of the interacting AD fusion by PCR and sequencing. The other plate was replica plated to four different assay plates: SC –Leu –Trp –His and SC –Leu –Trp –His + 2 mM 3AT to gauge the strength of activation of the HIS3 reporter gene, SC –Leu –Trp –Ade to gauge the strength of activation of the ADE2 reporter gene, and SC –Leu –His + 1µg/ml cycloheximide to identify autoactivating yeast colonies that express the HIS3 reporter gene even in the absence of an AD fusion protein21. Yeast colonies that express at least one reporter gene, and which do not show growth on the corresponding autoactivation plates, are considered positive. 101

PCR amplification and sequencing of AD clones From each positive yeast colony a small amount of cells, roughly a sphere of 0.5 mm diameter, was resuspended in 20 µl lysis buffer (0.1 M NaPO4 buffer pH 7.4, 2.5 mg/ml Zymolase 20T, Seikagaku Corporation) in a 96-well plate, using 200 µl pipet tips. Plates with resuspended yeast were then incubated for 5 minutes at 37°C followed by 5 minutes at 94°C. Next, 80 µl of sterile water was added. From the lysed yeast cells, the ORF insert in the Gal4-AD plasmid was amplified using primers AD: 5’-CGCGTTTGGAATCACTACAGGG and TERM: 5’-GGAGACTTGACCAAACCTCTGGCG. The PCR reactions were done using Novagen KOD Hot Start DNA Polymerase, using 2 µl of lysed yeast in a 25 µl reaction. The reactions were assembled and run according to the manufacturer’s instructions, using an annealing temperature of 58°C, an extension temperature of 68°C, and an extension time of 5 min. The amplified ORF inserts were then sent for sequencing using primer AD. All sequencing was done by Macrogen, in 96-well plate format with purification of PCR product performed by Macrogen. Sequence data analysis Each sequence read was analyzed as follows: First, phred54 was used to find the high quality segment of the read, using the -trim_alt option. Next, the vector sequences were clipped from the 5’- and 3’- ends of the read, leaving only the insert sequence. Traces where we were unable to identify an exact match to the 12 bases of vector sequence directly preceding the insert were eliminated. Traces where the Gal4AD sequence was not in frame with the insert sequence were also eliminated. The remaining traces were used to determine the identity of the interacting protein, by comparing the trace sequence to the ORF predictions in the human genome reference consortium release GRCh37_61 (obtained from the ensemble ftp site: ftp://ftp. ensembl.org) by BLAST+[ref. 55]. The start and endpoints of each insert relative to the corresponding full-length ORF were determined by direct comparison of the sequence read with the predicted ORF sequence, or using phrap56 when an exact match could not be found (e.g., when the sequence trace is not fully accurate). For comparison of randomly sequenced fragments from the library with the full-length ORF, when multiple splice variants are predicted to exist, we plotted the fragment against the first splice variant in the database (alphabetically) that contains the entire fragment. Mammalian expression constructs Vectors used were pCI-NEO-BirA57, Avi-mCherry-C1[ref. 58], and pEGFP-C1 (Clontech). Avi-mCherry-C1 contains the sequence MASGLNDIFEAQKIEWHEGGG, which is a

102

Protein identification using an ORFeome-based Y2H fragment library

substrate for the biotin ligase BirA, upstream of mCherry58. All ORFs were amplified from ORFeome clones by PCR, and cloned into Avi-mCherry-C1 or pEGFP-C1 using EcoRI/SalI, EcoRI/KpnI, SalI/KpnI, or SalI/BamHI sites added to the PCR primers. Co-affinity purification Constructs were transfected into HEK293 cells grown in DMEM/Ham’s F10 (50/50%) medium containing 10% FCS and 1% penicillin/streptomycin. One day before transfection, nearly confluent cells were plated at 1:10 in 10 cm tissue culture dishes. Cells were transfected using polyethylenimine (PEI) as follows: plasmid DNA was diluted in 500 µl Ham’s F10 medium, and PEI was added in a 3:1 PEI(µg):DNA(µg) ratio. After a 20 minute incubation at room temperature during which the cell culture medium was refreshed, the PEI/DNA mixture was added to the cells in a dropwise fashion. For every protein pair to be tested, 3 sets of plasmids were transfected. 1: Avi-mCherry-bait (7 µg) + empty EGFP vector (2 µg) + BirA (5 µg), 2: empty Avi-mCherry vector (2 µg) + EGFP-prey (7 µg) + BirA (5 µg), 3: AvimCherry-bait (7 µg) + EGFP-prey (7 µg) + BirA (5 µg). Cells were harvested 24 hours after transfection, by washing and then scraping the cells in 2 ml ice-cold TBS (20 mM Tris, 150 mM NaCl, pH 8.0). Cells pelleted by centrifugation at 1000 g for 3 minutes were lysed in 500 µl lysis buffer (20 mM TrisHCl, pH 8.0, 150 mM NaCl, 1.0% Triton X-100, and protease inhibitors; Roche), by freezing the cells at -80°C for 30 minutes, and incubating them for 30 minutes on ice after thawing at room temperature. Cell debris was removed by centrifugation at 13,200 rpm for 15 minutes. A 50 µl sample of cell lysate was used to check for expression of EGFP and mCherry constructs. The remainder was mixed with 25 µl of Dynabeads M-270 streptavidin (Invitrogen) first blocked by a 45 minute incubation with 0.2% chicken egg white in lysis buffer, and incubated for 1.5 hours rotating at 4°C. Beads were separated by centrifugation at 9000 rpm for 30 seconds, and washed three times in wash buffer (20 mM Tris-HCl, pH 8.0, 150 mM NaCl, 0.5% Triton X-100, and protease inhibitors; Roche). Bound proteins were eluted with 2x SDS sample buffer. All protein samples were separated on 10% acrylamide gels, and subjected to western blotting on polyvinylidene difluoride membrane (Immobilon-P; Millipore). Blots were blocked with 5% skim milk in PBST (7 mM Na2HPO4, 3 mM NaH2PO4, 140 mM NaCl, 5 mM Kcl, 0.05% Tween-20) for 1 hour at room temperature. For detection of EGFP, blots were incubated with rabbit polyclonal anti-GFP (Abcam ab6556, 1:1000) in PBST + 5% skim milk for 1 hour at room temperature, washed with PBST three times for 10 minutes at room temperature, incubated with antirabbit IgG antibody conjugated to horseradish peroxidase (Jackson Immuno

103

Research 111035003, 1:10,000) for 45 minutes at room temperature, washed with PBST three times for 10 minutes at room temperature, and finally washed once with PBS at room temperature for 10 minutes. For detection of biotinylated Avi-mCherry, blots were washed in PBST twice for 10 minutes at room temperature, incubated with Streptavidin coupled to horseradish peroxidase (Pierce High Sensitivity Streptavidin HRP, 1:20,000) in PBST for 1.5 hours at room temperature, washed in PBST five times for 10 minutes at room temperature, and finally washed once with PBS at room temperature for 10 minutes. Blots were developed using enhanced chemiluminescent Western blotting substrate (BioRad). Domain predictions Computationally predicted protein domains, as well as regions of low complexity and coiled coils, were obtained from SMART 7 (http://smart.embl-heidelberg.de/)59 and from Pfam 26.0[ref. 60].

Results ORFeome clones are an ideal resource to generate genome-wide random fragment libraries To enable the systematic mapping of interaction domains on a genome-wide scale, we set out to generate a genome-wide human AD-fragment library. We had previously used a PCR-based approach to generate a fragment library for 750 C. elegans proteins27. The attractive feature of this method is that it yields complete control over the contents of the library: each protein can be systematically covered by a series of fragments of the different sizes, each clone will be in frame, and the library will be normalized. However, using PCR to generate a genome wide library is both costly and time consuming, since each fragment has to be amplified individually. As alternative strategies, we considered the use of random-primed cDNA libraries and the random fragmentation of protein coding sequences, both of which have been used to generate Y2H libraries29,30 and should be able to identify interaction domains. One of our objectives was to generate a library in which all genes are equally represented. Since random-primed cDNA libraries will contain widely varying levels of clones for different genes, dependent on their expression levels, we decided to generate a library by random fragmentation of full-length open reading frame (ORF) clones. This approach is made possible by the development of ORFeome projects, which aim to provide complete sets of protein-encoding open reading frames (ORFs) for organisms including viruses, bacteria, nematodes and

104

Protein identification using an ORFeome-based Y2H fragment library

humans28,31–36. Using an ORFeome library as the starting material offers several advantages. First, library generation starts with an equal amount of each ORF construct, resulting in a normalized representation of all genes in the library. Second, the library can contain both fragmented and full-length constructs, which increases the chances of identifying a given interaction. Third, library construction can be efficiently accomplished by working with large pools of PCR products generated with primers annealing to vector sequences flanking the ORF. The main drawback of an ORFeome based library is that genes not present in the ORFeome collection will also be absent from the library. The library constructed here is based on version 5.1 of the human ORFeome28, which contains 15,483 different ORFs corresponding to 12,794 distinct genes. Construction of the human AD-fragment library The strategy used to generate the human AD-fragment library is shown in Figure 1. We PCR amplified each of the 15,483 full-length ORF constructs using a set of universal primers annealing to the vector sequences flanking the Gateway attL sites. PCR amplifications were performed in 384 well plates, and 12 samples from each plate were analyzed on a gel to monitor the PCR success rate, which was >95%. The PCR products were pooled into 5 pools based on size (<500 bp, 500 – 1000 bp, 1000 – 1500 bp, 1500 – 2000 bp, >2000 bp) (Figure 2A). After a short Exonuclease III treatment to remove the Gateway attL tails flanking the ORFs, the PCR pools were mechanically fragmented using a Covaris S2 ultrasonicator (Figure 2B). We chose mechanical fragmentation over enzymatic approaches such as frequently cutting restriction enzymes, as this approach fragments DNA without any specific sequence preference. Taking into account that most self-folding protein domains are estimated to be between 100 and 200 amino acid residues long2, we aimed for a median fragment size of 700 bp for ORFs > 1000 bp in length, a median fragment size of 500 bp for ORFs 500 – 1000 bp, and a median fragment size of 250 bp for ORFs < 500 bp. Settings for the ultrasonicator were chosen to maximize the yield of fragments in the desired size ranges (Figure 2C). To further ensure cloning of fragments of the correct size, we gel-purified the desired range from the ulrasonicated PCR pools (Figure 2D). The resulting pools of fragments were ligated into the pMB26 Gal4-AD Y2H vector and transformed into E. coli. From each pool we generated sufficient transformants to have a >99% chance that each ORF is fully represented in the library (see methods). Finally, the 5 pools of cloned fragments were mixed together such that all ORFs are represented equally. The final library contains a total of 1.6x106 clones. Assuming an 85% cloning success rate (see below) and taking into account that 1/6 of the clones are in frame with the Gal4-AD sequence, this corresponds to 2.3x105 in frame fragment clones, or an average of 14 clones/ORF. 105

ORF

attL1

attL2

ORFeome 5.1 clone PCR amplify 15,483 full-length ORFs, pool by size

<500 bp 500 - 1000 - 1500 - >2000 bp 1000 1500 2000 attL1

ORF

attL2

Remove attL tails with ExoIII

ORF

Fragment using ultrasonicator ORF ORF ORF

Gel purify desired size range, Clone into pMB26 Gal4-AD vector Gal4-AD

ORF

pMB26

Clone into pMB26 Gal4-AD vector Gal4-AD

ORF ORF pMB26

Mix 3:1, transform into S. cerevisiae strain Y8800 Figure 1. Schematic representation of the pipeline used to generate the human fragment library. Fulllength ORFs were PCR amplified from the ORFeome 5.1 resource using primers annealing just outside the Gateway attL sequences. After removal of vector sequences by Exonuclease and mechanical fragmentation by ultrasonication, fragments were cloned into a Y2H Gal4-AD vector. A subset of PCR products was cloned without fragmentation to add a full-length component to the library.

106

Protein identification using an ORFeome-based Y2H fragment library

A. Pooled PCR products

10000 bp 8000 6000 5000 4000 3500 3000 2500 2000 1500

B. Exonuclease treatment

1000 750 500 250

C. Ultrasonicated PCR pools

D. Gel-purified pools

Figure 2. Agarose gel analysis of DNA samples at several steps in the pipeline shown in Figure 1. (A) PCR products collected in 5 pools based on size. (B) Example of the elimination of attL tails in a brief exonuclease III treatment. Left: pool 4 prior to treatment, right: pool 4 after exonuclease treatment. (C) DNA pools after ultrasonication on a Covaris S2. (D) DNA pools after gel-purification of desired size range. This represents the final step before ligation into vector pMB26. In all gels, lanes are as follows: 1: pool 1 (ORFs < 500 bp), 2: pool 2 (ORFs 500 – 1000 bp), 3: pool 3 (ORFs 1000 – 1500 bp), 4: pool 4 (ORFs 1500 – 2000 bp), 5: pool 5 (ORFs >2000 bp).

In addition to the generation of fragments, we cloned each PCR pool without fragmentation, to add a set of full-length clones to the library. We generated a total of 69,000 in frame full length clones, again taking into account the average cloning success rate and 1/6 chance of an in frame clone due to the exonuclease treatment. The final human fragment library was made by mixing the fragment and full-length clones in a 3:1 ratio. The library was transformed into S. cerevisiae strain Y8800 to generate a yeast mating library, with which all screens were performed. The final yeast mating library contains >3 x 106 clones. To verify the cloning success rate and examine the distribution of fragments relative to the full-length ORFs, we PCR amplified and sequenced the insert from 96 randomly selected yeast clones obtained by plating a 1:100.000 dilution of the final mating library on selective plates (SC –Trp). Of these 96 colonies, 15 could not

107

Coverage of ORF by AD insert

be analyzed due to poor sequence quality. Of the remaining 81 clones, 67 (85%) contained an insert matching a human ORF, of which 14 (21%) were in frame. The average insert length was 400 bp, which is less than the expected median length of 550 bp for the entire library. This indicates either a bias in our random picking of 96 clones, or preferential cloning of a subset of smaller fragments, despite size selection by agarose gel. We next examined the distribution of each fragment relative to the full-length ORF, by plotting the region of the full-length ORF covered by the cloned fragment (Figure 3). The fragments covered from 3% up to 100% of the full-length ORF, with a median of 31%, and the start and end points of the fragments were distributed along the entire length of the ORFs. Together, these results indicate no particular bias in fragmenting of full-length ORFs and subsequent cloning, with the exception of a potential preference for smaller fragments.

0% 10%

20%

30%

40%

50%

60%

70%

80%

90% 100%

Normalized full-length ORF size Figure 3. Distribution of a random selection of fragments from the library relative to the full-length ORF size. AD vector inserts were PCR amplified from randomly selected yeast colonies obtained by plating the final Y2H mating library, and analyzed by sequencing. Each blue line represents the area of the corresponding full-length ORF that is covered by the fragment identified from the library.

108

Protein identification using an ORFeome-based Y2H fragment library

Identification of known and novel interactions To test the human fragment library, we screened the library with 44 bait proteins involved in cell-cycle regulation and cell polarity control (Table 1). For each protein, the corresponding full-length ORF was transferred from the ORFeome entry clone into the pDEST-Gal4-DB vector by Gateway recombinational cloning, and the identity was confirmed by sequencing. Gal4-DB bait clones were then transformed into yeast, and assayed for autoactivation of reporter genes in the absence of an interacting AD-ORF clone. Eight bait strains showed autoactivation of the Y2H reporter genes and were not used further. The remaining 36 bait strains were screened against the human fragment library using a mating-based procedure. From yeast colonies growing on selective plates, we first eliminated the most common source of false-positives: de novo autoactivators that arise in the screening process. These are yeast cells that are able to activate reporter genes irrespective of the presence of an interaction. These can efficiently be eliminated by removing all colonies that still activate reporter genes after the AD-ORF plasmid is lost through a counterselection step based on the sensitivity to cycloheximide conferred by the CYH2 gene present on the AD plasmid.21 Next, the AD plasmid inserts were PCR amplified and sequenced to determine the identity and end points of the interacting ADfragments. Sequence traces that were in the wrong orientation or out of frame with the AD coding segment were eliminated. Because it is unlikely that a prey protein is identified twice by chance, only interactions identified in two or more independent yeast colonies were considered valid Y2H interactions. Finally, to ensure that each interaction is reproducible, we retested all interactions in yeast. For every interacting protein pair we isolated a representative Gal4-AD prey plasmid from yeast, and verified the identity by sequencing. Next, the isolated Gal4-AD prey plasmid was transformed into fresh yeast, and mated with the corresponding Gal4-DB bait yeast strain to confirm the interaction.

109

Table 1. Bait proteins screened Protein Interactions ASPM 1 CCND3 3 CDK4 2 CDKN2B CTNNA2 4 CTNNA3 SA CTNNB1 SA DLG2 DLG5 DLGAP5 DVL2 1 DVL3 GNAI1 GNAO1 4 GPSM1 SA: Self-activating bait

Protein GPSM2 LLGL1 LNX1 MARK2 MPP4 MPP5 NUMA1 NUMB NUMBL PARD3 PARD6B PLK1 PLK2 PLK4 PRKAA2

Table 2. Protein interactions identified DB-X CTNNA3 CDK4 CDKN2B PROX1 DVL3 CTNNA3 CCND3 CDK4 NUMA1 CTNNA3 GPSM1 PRKCI CDK4 GPSM1 LNX1 MPP5 CDKN2B CTNNA3 GPSM1 GPSM1 MPP5 PRKCI

110

AD-Y CTNNB1 CDKN2D PYCRL PROX1 HOMEZ EHMT2 CDKN1B CCND1 CCDC57 JUP PPP6R3 PARD6B KLHL32 TRIM23 PRPH LIN7A RNF20 SPRY2 CBS RALBP1 ZNF451 CRX

Hits 77 34 30 26 17 12 11 8 6 5 5 5 3 3 3 3 2 2 2 2 2 2

Interactions SA 1 2 1 SA SA -

Protein PRKCD PRKCI PRKCZ PROX1 PTEN RHOA RIC8A RIC8B STAU1 STAU2 STK11 VHL YWHAH YWHAZ

Additional evidence HI-2012 IntAct no no yes yes no no no no no no no no yes no yes yes no yes no no no no yes yes no no no no no no no yes no no no no no no no no no no no no

Interactions 2 1 SA SA SA

Literature yes yes no no no no yes yes no yes no yes no no no yes no no no no no no

Protein identification using an ORFeome-based Y2H fragment library

Together, 11 bait proteins (31%) identified 22 interactions, corresponding to 2 interactors per bait (Table 2). These numbers are comparable to our previous results screening a C. elegans fragment library, where 37% of full-length bait proteins identified an average of 2.2 interacting proteins each27. To assess the relevance of the interactions we identified, we compared our results with the human interactome HI-2012 (http://interactome.dfci.harvard.edu/H_sapiens/), with interactions in IntAct37, and with the literature (Table 2). Of the 22 interactions, 5 were present in HI-2012, 4 were present in IntAct, and 7 were present in the literature. Overall, 8 interactions were supported by these additional sources of evidence, demonstrating that our approach identifies valid interactions. Identification of minimal interacting regions For each interaction, the minimal interacting region for an interacting prey protein is defined as the smallest region present in all AD-fragments interacting with a particular bait (Figure 4 and Figure S1). Of the 22 interactions we identified, 7 have already been described in small-scale literature experiments (Table S1). For each of these 7 interactions, detailed information on the sites that mediate the interaction is known, allowing us to assess the accuracy of the interaction domains we identify. Except for Par6, for which we only found nearly full-length fragments in our screens, we were able to delineate a minimal interacting region for each of the 7 prey proteins involved in the literature-supported interactions (Figure 4 and Table S2). Importantly, the minimal interacting regions we identified closely match the published interaction sites (compare blue literature interaction sites in Figure 4 with yellow interacting regions we identified). For example, for the homologous proteins α-Catenin and Junction Plakoglobin, we found that a small region upstream of a series of Armadillo repeats mediates their interaction with α-Catenin. These regions overlap the stretch of 29 amino acids previously reported to mediate these interactions38. Similarly, we found that CDK4 binds to the N-terminal lobe of Cyclin D1 and to the first 4 Ankyrin repeats of p19Ink4d, that Cyclin D3 binds to the CDI domain of p27Kip1, and that the human Pals1 homolog MPP5 binds to the L27 domain of LIN7. In all cases this is in accordance with the published interaction sites39–43. The smallest interacting region, expressed as a fraction of the full-length protein, was the α-Catenin region that binds to β-Catenin (60 amino acids, or 7.7% of full-length α-Catenin), while the largest interacting region we found was the region of p19Ink4d that mediates binding to CDK4 (142 amino acids, or 85% of full-length p19Ink4d). Thus, both small interaction domains and larger interacting stretches can be accurately identified by our approach. We also examined novel interacting regions for overlap with known protein domains. As we observed in our previous screens with C. elegans proteins27,

111

some interacting regions overlap well with predicted domains, while others overlap only partially or match a region of the protein for which no domain predictions are known. Five of the interacting regions we identified correlate well with specific protein domain predictions. CCND3 binding region 7 129 3 2 5

17 13

Cyclin_N

2 1

L27

295 1

CTNNA3 binding region 82 141 6 1 35 1 1 5 1 2 1

PDZ LIN7A

233

PRKCI binding region 28

166

MPP5 binding region 22 100 1 2

Cyclin_C

CCND1

ANK

CDKN2D

CDK4 binding region 44 169 8 1

ANK

198

CDKN1B

ANK

CDI 1

CDK4 binding region 1 142

373

PB1 1

PDZ PARD6B

372

ARM ARM ARM ARM ARM ARM ARM ARM ARM ARM ARM ARM 1

781

CTNNB1

CTNNA3 binding region 53 138 1

ARM ARM ARM ARM ARM ARM ARM ARM ARM ARM ARM ARM 1

JUP

745

Figure 4. Comparison of minimal interacting regions identified by our Y2H screens (yellow area) with interaction site information present in the literature (blue boxes). Grey lines above the prey protein cartoon indicate distinct prey fragments that we identified, and the numbers before the lines show how often a particular fragment was identified independently in different yeast colonies. Labeled red and blue boxes are SMART domain predictions.

112

Protein identification using an ORFeome-based Y2H fragment library

The interaction of GPSM1 with TRIM23 is mediated by two B-Box zinc-finger domains, which can be involved in protein interactions (for example in the binding of alpha 4 to the B-Box domain protein MID1[ref. 44,45]). GPSM1 also interacts with the C-terminal domain of CBS, which is involved in interacting with the known CBS inhibitor LanCL1.46 The CDK4 binding site in KLHL32 overlaps with a predicted BTB domain, which is a known protein interaction domain that can mediate heterodimeric interactions47. MPP5 binds to two predicted zinc-fingers of the multiple zinc-finger protein ZNF451, and zinc-fingers are known to be able to mediate protein interactions48. Finally, Îą-catenin binds to a region of the methyltransferase EHMT2 encompassing a SET domain with flanking pre- and post-SET domains. SET domains catalyze the methylation of substrate lysines, but together with adjacent pre- and post-SET domains also mediate binding to substrates or other protein binding partners49. In summary, the domain predictions overlapping our experimentally defined interacting regions are consistent with a role in mediating a protein-protein interaction. Independent validation of interactions and interaction domains by co-affinity purification To further demonstrate the validity of the interactions and interaction domains we identified, we tested each interaction in a co-affinity purification assay. For each interaction, we tested the full-length bait ORF against both the full-length prey ORF and the shortest prey fragment we identified in the Y2H screens. Bait and prey ORFs were transfected into HEK293 cells as Avi-tagged mCherry (Avi-mCherry) and EGFP fusions, respectively, together with the bacterial biotin ligase BirA, which recognizes and biotinylates the Avi tag. Binding of the tagged bait and prey proteins was assessed by affinity purification of biotinylated Avi-mCherry-bait protein using streptavidin coated beads, followed by detection of co-purified EGFP-prey protein on Western blot. As negative controls, we tested for interaction between Avi-mCherry tagged bait protein and the EGFP tag, as well as between EGFP tagged prey protein and the Avi-mCherry tag. An interaction was considered positive only if no interaction was detected in the negative controls. In total, we were able to reproduce 12 of the 22 protein interactions (55%) in this orthologous assay (Figure 5, Table 3, Figure S2). Of these, 6 correspond to interactions previously known from the literature, and 6 are novel interactions. A reproducibility rate of 55% (or 40% if considering only the novel interactions) compares favorably with previous results testing interactions obtained using one method in an orthologous assay. For example, in a study testing a panel of known literature-derived interactions in 5 different protein interaction assays, the maximum reproducibility rate was 36%[ref. 22]. Similarly, we previously

113

tested a series of interactions identified by Y2H in the mammalian MAPPIT assay, and obtained a maximum reproducibility rate of 40%[ref. 27]. Importantly, of the 12 interactions confirmed by co-affinity purification, 11 were also identified using the fragment prey ORF. This result demonstrates that the minimal interacting regions we identify by Y2H represent valid interaction domains.

AP: streptavidin WB: αGFP

kDa 170 130 100 70

CTNNA3 CTNNB1 (frag.) 1 2 3

CDK4 CDKN2D (f.l.) 1 2 3

CDK4 CDKN2D (frag.) 1 2 3

CDKN2B PYCRL (f.l.) 1 2 3

CDKN2B PYCRL (frag.) 1 2 3

PROX1 PROX1 (f.l.) 1 2 3

PROX1 PROX1 (frag.) 1 2 3

CCND3 CDKN1B (frag.) 1 2 3

CTNNA3 JUP (f.l.) 1 2 3

CTNNA3 JUP (frag.) 1 2 3

GPSM1 PPP6R3 (frag.) 1 2 3

PRKCI PARD6B (f.l.) 1 2 3

PRKCI PARD6B (frag.) 1 2 3

LNX1 PRPH (f.l.) 1 2 3

LNX1 PRPH (frag.) 1 2 3

MPP5 LIN7A (f.l.) 1 2 3

GPSM1 RALBP1 (frag.) 1 2 3

MPP5 ZNF451 (f.l.) 1 2 3

MPP5 ZNF451 (frag.) 1 2 3

55 40 35 25 15

kDa

AP: streptavidin WB: αGFP

Lanes: 1: EGFP + Avi-mCherry-bait 2: EGFP-prey + Avi-mCherry 3: EGFP-prey + Avi-mCherry-bait

CTNNA3 CTNNB1 (f.l.) 1 2 3

170 130 100 70 55 40 35 25

AP: streptavidin WB: αGFP

kDa 170 130 100 70 55 40 35 25

Figure 5. Interactions testing positive by co-affinity purification. Bait proteins are tagged with AvimCherry and prey proteins with EGFP. Depicted are αGFP Western blots on proteins purified using streptavidin coated beads. The tested proteins are indicated above each blot, the first protein listed is the bait, the second protein is the prey. Also indicated is whether the prey protein tested was full-length (f.l.) or corresponds to the shortest fragment identified by Y2H (frag.). In each blot, lanes 1 and 2 contain the negative control purifications, while lane 3 contains the actual co-affinity purification (1: Avi-mCherrybait tested against the EGFP tag, 2: the EGFP-prey tested against the Avi-mCherry tag, 3: the interacting Avi-mCherry-bait and EGFP-prey). Asterisks indicate bands of the expected molecular mass. In some lanes, degradation products are visible as bands of lower size. Expression of tagged protein in input lysates is shown in Figure S2.

114

Protein identification using an ORFeome-based Y2H fragment library

Discussion In this work, we have used physical fragmentation of PCR products generated from an ORFeome resource to generate a human random fragment Y2H library. Using a series of bait proteins involved in cell-cycle regulation of cell polarity establishment, we identified 22 protein-protein interactions and corresponding minimal interacting regions. A large fraction (32%) of the interactions we identified are supported by the literature. We did notice that, compared to the literature supported interactions, many novel interactions were identified in relatively few independent yeast colonies. One interpretation of this is that a more stringent cutoff could be used to increase the fraction of biologically relevant interactions. For example, for this particular dataset, keeping only those interactions that were found in 5 or more independent yeast colonies would raise the number of literature supported interactions from 32% to 50%. It is also possible, however, that interactions found in fewer independent yeast colonies represent interactions that are more difficult to detect, for example because the interaction is transient or weak. Literature derived interactions may have an inherent bias for more easily detected interactions. Removing the 10 interactions found fewer than 5 times would have eliminated one interaction supported by the literature, and three interactions validated by co-affinity purification. For this reason, and because there is no experimental rationale for choosing a particular cutoff, we present all interactions, with the caveat that the interactions identified more frequently may be more easily reproduced in other assays. In addition to the literature supported known interactions, 40% of the novel interactions are supported by independent co-affinity purification assays. The validity of the minimal interacting regions we identify is supported by the finding that we recapitulated the known interaction site for each of the literature described interactions. Moreover, 11 of the 12 interactions validated by co-affinity purification could be reproduced using the shortest fragment identified by Y2H. The single co-affinity purification where only the full-length ORF interacted (MPP5/LIN7A) represents an interaction where our interaction domain is validated by the literature, indicating that the lack of interaction in tissue culture is a technical false-negative, and is not due to an incorrectly identified interaction domain. Together, these results validate that we have established a reliable procedure for the identification of minimal interacting regions on a genome-wide scale.

115

Table 3. Validation of interactions by co-affinity purification. bait*

Prey name

Full-length size (a.a.)* CTNNA3 (130 kDa) CTNNB1 781 (114 kDa) CDKN2D 166 (46 kDa) CDK4 (64 kDa) 286 (57 kDa) CDKN2B (45 kDa) PYCRL 737 (111 kDa) PROX1 PROX1 (113 kDa) HOMEZ 550 (87 kDa) DVL3 (108 kDa) 1210 (140 kDa) CTNNA3 (130 kDa) EHMT2 CDKN1B 198 (50 kDa) CCND3 (63 kDa) 295 (62 kDa) CCND1 CDK4 (64 kDa) NUMA1 (139 kDa) CCDC57 916 (114 kDa) 745 (110 kDa) CTNNA3 (130 kDa) JUP 873 (126 kDa) PPP6R3 GPSM1 (78 kDa) PARD6B 372 (69 kDa) PRKCI (98 kDa) KLHL32 620 (82 kDa) CDK4 (64 kDa) TRIM23 574 (92 kDa) GPSM1 (78 kDa) 470 (82 kDa) PRPH LNX1 (100 kDa) 233 (54 kDa) LIN7A MPP5 (108 kDa) 975 (142 kDa) CDKN2B (45 kDa) RNF20 315 (67 kDa) CTNNA3 (130 kDa) SPRY2 551 (89 kDa) CBS GPSM1 (78 kDa) RALBP1 655 (104 kDa) GPSM1 (78 kDa) ZNF451 1061 (135 kDa) MPP5 (108 kDa) 299 (61 kDa) CRX PRKCI (98 kDa)

Fragment region (a.a.)* 64−165 (40 kDa) 1−142 (43 kDa) 8−273 (56 kDa) 184−381 (51 kDa) 52−550 (84 kDa) 907−1210 (63 kDa) 7−173 (47 kDa) 44−169 (43 kDa) 1−752 (0 kDa) 53−173 (41 kDa) 522−628 (40 kDa) 10−372 (68 kDa) 11−206 (50 kDa) 146−262 (41 kDa) 40−470 (78 kDa) 22−100 (38 kDa) 232−466 (56 kDa) 39−309 (58 kDa) 371−551 (49 kDa) 403−634 (56 kDa) 306−395 (39 kDa) 13−272 (56 kDa)

Interaction FullFragment length yes yes yes yes yes yes yes yes no no no no yes ND no no NA no yes yes yes ND yes yes no no no no yes yes no yes no no no no no no yes no yes yes no no

ND: Experiment could not be performed for technical reasons. NA: Not applicable, no fragment smaller than full-length was identified by Y2H. * molecular mass in kDa includes the EGFP or Avi-mCherry tag. Sizes and region coordinates are in amino acids.

One of the features that make the Y2H system such an attractive approach to identify protein interactions is that Y2H screens can be performed relatively quickly and inexpensively. The method outlined here offers the advantage of identifying minimal interacting regions, without increasing the workload compared to traditional cDNA library screens. Initial library generation is the most time-consuming step. However, except for access to an ultrasonicator, the procedure uses basic molecular biology techniques and enzymes and can easily be adopted by others to generate random fragment libraries for any organism for which a source of cloned ORFs is available. The cloning of bait proteins and the screening procedure itself do not take any additional time compared with other library screens. Thus, our approach offers significant advantages without increasing the screening workload, apart from the initial investment in library generation.

116

Protein identification using an ORFeome-based Y2H fragment library

Acknowledgements We thank Ewart de Bruin and Edwin Cuppen for use of and help with the Covaris ultrasonicator, Marina Mikhaylova and Anna Akhmanova for reagents and protocols for co-affinity purification, and David Hill and Marc Vidal for sharing the human ORFeome collection. We also thank Sander van den Heuvel for helpful discussions. The Human Interactome HI-2012 dataset was generated by the Center for Cancer Systems Biology (CCSB) at the Dana-Farber Cancer Institute, with funding from The National Human Genome Research Institute (NHGRI) of NIH; The Ellison Foundation, Boston, MA; and The Dana-Farber Cancer Institute Strategic Initiative. This work was funded by a Horizon grant from the Netherlands Organization for Scientific Research (NWO)/Netherlands Genomics Initiative (NGI) to M.B. ď&#x201A;ž

117

References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24.

118

Ponting, C. P. Russell, R. R. The natural history of protein domains. Annu. Rev. Biophys. Biomol. Struct. 2002, 31, 45–71. Bornberg-Bauer, E. Beaussart, F. Kummerfeld, S. K. Teichmann, S. A. Weiner, J., 3rd The evolution of domain arrangements in proteins and interaction networks. Cell. Mol. Life Sci. Cmls 2005, 62, 435–445. Jin, J. Xie, X. Chen, C. Park, J. G. Stark, C. James, D. A. Olhovsky, M. Linding, R. Mao, Y. Pawson, T. Eukaryotic protein domains as functional units of cellular evolution. Sci. Signal. 2009, 2, ra76. Pawson, T. Nash, P. Assembly of cell regulatory systems through protein interaction domains. Science 2003, 300, 445–452. Jadwin, J. A. Ogiue-Ikeda, M. Machida, K. The application of modular protein domains in proteomics. Febs Lett. 2012. Ho, Y. et al. Tyers, M. Systematic identification of protein complexes in Saccharomyces cerevisiae by mass spectrometry. Nature 2002, 415, 180–183. Gavin, A.-C. et al. Functional organization of the yeast proteome by systematic analysis of protein complexes. Nature 2002, 415, 141–147. Giot, L. et al. A protein interaction map of Drosophila melanogaster. Science 2003, 302, 1727–1736. Li, S. et al. Map of the interactome network of the metazoan C. elegans. Science 2004, 303, 540–543. Rual, J.-F. et al. Towards a proteome-scale map of the human protein-protein interaction network. Nature 2005, 437, 1173–1178. Stelzl, U et al. A human protein-protein interaction network: a resource for annotating the proteome. Cell 2005, 122, 957–968. Yu, H. et al. High-quality binary protein interaction map of the yeast interactome network. Science 2008, 322, 104–110. Arabidopsis Interactome Mapping Consortium Evidence for network evolution in an Arabidopsis interactome map. Science 2011, 333, 601–607. Luo, Q. Pagel, P. Vilne, B. Frishman, D. DIMA 3.0: Domain Interaction Map. Nucleic Acids Res. 2011, 39, D724–729. Stein, A. Céol, A. Aloy, P. 3did: identification and classification of domain-based interactions of known three-dimensional structure. Nucleic Acids Res. 2011, 39, D718–723. Yellaboina, S. Tasneem, A. Zaykin, D. V. Raghavachari, B. Jothi, R. DOMINE: a comprehensive collection of known and predicted domain-domain interactions. Nucleic Acids Res. 2011, 39, D730– 735. Reich, S. et al. Combinatorial Domain Hunting: An effective approach for the identification of soluble protein domains adaptable to high-throughput applications. Protein Sci. Publ. Protein Soc. 2006, 15, 2356–2365. Littler, E. Combinatorial Domain Hunting: solving problems in protein expression. Drug Discov. Today 2010, 15, 461–467. Neveu, G. et al. Jacob, Y. Comparative analysis of virus-host interactomes with a mammalian highthroughput protein complementation assay based on Gaussia princeps luciferase. Methods San Diego Calif 2012. Rozenblatt-Rosen, O. et al. Interpreting cancer genomes using systematic host network perturbations by tumour virus proteins. Nature 2012, 487, 491–495. Vidalain, P.-O. Boxem, M. Ge, H. Li, S. Vidal, M. Increasing specificity in high-throughput yeast two-hybrid experiments. Methods San Diego Calif 2004, 32, 363–370. Braun, P. et al. An experimentally derived confidence score for binary protein-protein interactions. Nat. Methods 2009, 6, 91–97. Venkatesan, K. et al. An empirical framework for binary interactome mapping. Nat. Methods 2009, 6, 83–90. Chen, Y.-C. Rajagopala, S. V. Stellberger, T. Uetz, P. Exhaustive benchmarking of the yeast twohybrid system. Nat. Methods 2010, 7, 667–668.

Protein identification using an ORFeome-based Y2H fragment library

25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48.

Stellberger, T. Häuser, R. Baiker, A. Pothineni, V. Haas, J. Uetz, P. Improving the yeast two-hybrid system with permutated fusions proteins: the Varicella Zoster Virus interactome. Proteome Sci. 2010, 8, 8. Worseck, J. M. Grossmann, A. Weimann, M. Hegele, A. Stelzl, U. A Stringent Yeast Two-Hybrid Matrix Screening Approach for Protein–Protein Interaction Discovery. In Two Hybrid Technologies Suter, B. Wanker, E. E., Eds. Humana Press: Totowa, NJ, 2012 Vol. 812, pp. 63–87. Boxem, M. et al. A protein domain-based interactome network for C. elegans early embryogenesis. Cell 2008, 134, 534–545. Lamesch, P. et al. Vandenhaute, J. Dunham, I. Hill, D. E. Vidal, M. hORFeome v3.1: a resource of human open reading frames representing over 10,000 human genes. Genomics 2007, 89, 307–315. DeGrado-Warren, J. Dufford, M. Chen, J. Bartel, P. L. Shattuck, D. Frech, G. C. Construction and characterization of a normalized yeast two-hybrid library derived from a human protein-coding clone collection. BioTechniques 2008, 44, 265–273. Rain, J.-C. Cribier, A. Gérard, A. Emiliani, S. Benarous, R. Yeast two-hybrid detection of integrasehost factor interactions. Methods San Diego Calif 2009, 47, 291–297. Dricot, A. et al. Hill, D. E. Letesson, J.-J. De Bolle, X. Generation of the Brucella melitensis ORFeome version 1.1. Genome Res. 2004, 14, 2201–2206. Lamesch, P. et al. C. elegans ORFeome version 3.1: increasing the coverage of ORFeome resources with improved gene predictions. Genome Res. 2004, 14, 2064–2069. Brandner, C. J. Maier, R. H. Henderson, D. S. Hintner, H. Bauer, J. W. Onder, K. The ORFeome of Staphylococcus aureus v 1.1. Bmc Genomics 2008, 9, 321. Pellet, J. et al. ViralORFeome: an integrated database to generate a versatile collection of viral ORFs. Nucleic Acids Res. 2010, 38, D371–378. Rajagopala, S. V.et al. The Escherichia coli K-12 ORFeome: a resource for comparative molecular microbiology. Bmc Genomics 2010, 11, 470. Yang, X. et al. A public genome-scale lentiviral expression library of human ORFs. Nat. Methods 2011, 8, 659–661. Kerrien, S. et al. The IntAct molecular interaction database in 2012. Nucleic Acids Res. 2012, 40, D841–846. Aberle, H. Schwartz, H. Hoschuetzky, H. Kemler, R. Single amino acid substitutions in proteins of the armadillo gene family abolish their binding to alpha-catenin. J. Biol. Chem. 1996, 271, 1520– 1526. Hirano, Y. Yoshinaga, S. Takeya, R. Suzuki, N. N. Horiuchi, M. Kohjima, M. Sumimoto, H. Inagaki, F. Structure of a cell polarity regulator, a complex between atypical PKC and Par6 PB1 domains. J. Biol. Chem. 2005, 280, 9653–9661. Jeffrey, P. D. Russo, A. A. Polyak, K. Gibbs, E. Hurwitz, J. Massagué, J. Pavletich, N. P. Mechanism of CDK activation revealed by the structure of a cyclinA-CDK2 complex. Nature 1995, 376, 313–320. Russo, A. A. Jeffrey, P. D. Patten, A. K. Massagué, J. Pavletich, N. P. Crystal structure of the p27Kip1 cyclin-dependent-kinase inhibitor bound to the cyclin A-Cdk2 complex. Nature 1996, 382, 325–331. Russo, A. A. Tong, L. Lee, J. O. Jeffrey, P. D. Pavletich, N. P. Structural basis for inhibition of the cyclin-dependent kinase Cdk6 by the tumour suppressor p16INK4a. Nature 1998, 395, 237–243. Zhang, J. Yang, X. Wang, Z. Zhou, H. Xie, X. Shen, Y. Long, J. Structure of an L27 domain heterotrimer from cell polarity complex Patj/Pals1/Mals2 reveals mutually independent L27 domain assembly mode. J. Biol. Chem. 2012, 287, 11132–11140. Massiah, M. A. Simmons, B. N. Short, K. M. Cox, T. C. Solution structure of the RBCC/TRIM B-box1 domain of human MID1: B-box with a RING. J. Mol. Biol. 2006, 358, 532–545. Trockenbacher, A. et al. Schweiger, S. MID1, mutated in Opitz syndrome, encodes an ubiquitin ligase that targets phosphatase 2A for degradation. Nat. Genet. 2001, 29, 287–294. Zhong, W. et al. Lanthionine synthetase C-like protein 1 interacts with and inhibits cystathionine β-synthase: a target for neuronal antioxidant defense. J. Biol. Chem. 2012, 287, 34189–34201. Ahmad, K. F. et al. Privé, G. G. Mechanism of SMRT corepressor recruitment by the BCL6 BTB domain. Mol. Cell 2003, 12, 1551–1564. Gamsjaeger, R. Liew, C. K. Loughlin, F. E. Crossley, M. Mackay, J. P. Sticky fingers: zinc-fingers as protein-recognition motifs. Trends Biochem. Sci. 2007, 32, 63–70.

119

49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60.

120

Schapira, M. Structural Chemistry of Human SET Domain Protein Methyltransferases. Curr. Chem. Genomics 2011, 5, 85–94. Chevray, P. M. Nathans, D. Protein interaction cloning in yeast: identification of mammalian proteins that react with the leucine zipper of Jun. Proc. Natl. Acad. Sci. U. S. A. 1992, 89, 5789–5793. Inoue, H. Nojima, H. Okayama, H. High efficiency transformation of Escherichia coli with plasmids. Gene 1990, 96, 23–28. Clarke, L. Carbon, J. A colony bank containing synthetic Col El hybrid plasmids representative of the entire E. coli genome. Cell 1976, 9, 91–99. Schiestl, R. H. Gietz, R. D. High efficiency transformation of intact yeast cells using single stranded nucleic acids as a carrier. Curr. Genet. 1989, 16, 339–346. Ewing, B. Hillier, L. Wendl, M. C. Green, P. Base-calling of automated sequencer traces using phred. I. Accuracy assessment. Genome Res. 1998, 8, 175–185. Camacho, C. Coulouris, G. Avagyan, V. Ma, N. Papadopoulos, J. Bealer, K. Madden, T. L. BLAST+: architecture and applications. BMC Bioinformatics 2009, 10, 421. Rieder, M. J. Taylor, S. L. Tobe, V. O. Nickerson, D. A. Automating the identification of DNA variations using quality-based fluorescence re-sequencing: analysis of the human mitochondrial genome. Nucleic Acids Res. 1998, 26, 967–973. Lansbergen, G. et al. CLASPs attach microtubule plus ends to the cell cortex through a complex with LL5beta. Dev. Cell 2006, 11, 21–32. Van Spronsen, M. et al. TRAK/Milton motor-adaptor proteins steer mitochondrial trafficking to axons and dendrites. Neuron 2013, 77, 485–502. Letunic, I. Doerks, T. Bork, P. SMART 7: recent updates to the protein domain annotation resource. Nucleic Acids Res. 2012, 40, D302–305. Punta, M. et al. The Pfam protein families database. Nucleic Acids Res. 2012, 40, D290–301.

Protein identification using an ORFeome-based Y2H fragment library

Supplementary information Figure S1 A and B CDKN2B binding

CDK4 binding

CTNNA3 binding

AnkAnk Ank Ank Ank 1

CDKN2D

Sprouty

CCND3 binding

Cyclin C 295

CREB3

PRKCI binding HOX

PDZ

LIN7A

233

PB1 299

LNX1 binding Fil. head

470

BBOX BBOX

BBC

574

CDK4 binding HOX

BTB

HOMEZ

550

BACK

Kelch Kelch Kelch Kelch Kelch

KLHL32

GPSM1 binding PALP 1

Full-length prey protein

CBS

372

ARF

TRIM23

DVL3 binding HOX

PDZ

PARD6B

GPSM1 binding RING

Filament

PRPH

371

PRKCI binding

TF Otx

CRX

315

BRLZ

CCND1

198

MPP5 binding L27

SPRY2

PARD3 binding

Cyclin N

CDKN1B

286

CDK4 binding

CDI 1

PYCRL

166

620

GPSM1 binding

CBS

RhoGAP 551

Minimal Region of Interaction

RALBP1

Pfam-A / SMART

Low Complexity

655

Coiled-Coil

121

PROX1 binding Prox1

Prox1

PROX1

737

CTNNA3 binding Arm Arm Arm Arm Arm Arm Arm Arm Arm

Arm

Arm Arm

JUP

745

CTNNA3 binding Arm Arm Arm Arm Arm Arm Arm Arm Arm

Arm

Arm Arm

CTNNB1

781

GPSM1 binding SAPS

SAPS

PPP6R3

873

NUMA1 binding

CCDC57

916

CDKN2B binding RING

RNF20

975

MPP5 binding ZF ZF 1

ZF ZF

ZF ZF ZF

ZF ZF

ZNF451

1061

CTNNA3 binding Ank Ank Ank Ank Ank Ank 1

Full-length prey protein

PreSET

SET

EHMT2

Minimal Region of Interaction

Pfam-A / SMART

1210

Low Complexity

Coiled-Coil

Figure S1 A and B. Graphical representation of minimal region of interaction for all interacting protein pairs. Green bars represent full-length proteins. Yellow bars represent regions of the full-length protein required for interaction with the indicated binding partner. Pfam-A and SMART domain signatures are drawn as red boxes, coiled-coil predictions as green boxes, and low complexity regions as blue boxes. Where multiple predictions overlapped, display preference for these domains was in the following order: Pfam-A/SMART, coiled-coil, low complexity.

122

Protein identification using an ORFeome-based Y2H fragment library

lysate WB: αGFP

AP: streptavidin WB: αGFP

Figure S2 A, B and C

kDa 170 130 100 70

CTNNA3 CTNNB1 (frag.) 1 2 3

CDK4 CDKN2D (f.l.) 1 2 3

CDK4 CDKN2D (frag.) 1 2 3

CDKN2B PYCRL (f.l.) 1 2 3

CDKN2B PYCRL (frag.) 1 2 3

PROX1 PROX1 (f.l.) 1 2 3

55 40 35 25 15

kDa 170 130 100 70 55 40 35 25 15

lysate WB: streptavidin-HRP

CTNNA3 CTNNB1 (f.l.) 1 2 3

†

kDa 170 130 100 70 55 40 35 25 15

Lanes: 1: EGFP + Avi-mCherry-bait 2: EGFP-prey + Avi-mCherry 3: EGFP-prey + Avi-mCherry-bait

123

AP: streptavidin WB: αGFP lysate WB: αGFP

kDa 170 130 100 70

PROX1 PROX1 (frag.) 1 2 3

CCND3 CDKN1B (frag.) 1 2 3

CTNNA3 JUP (f.l.) 1 2 3

CTNNA3 JUP (frag.) 1 2 3

GPSM1 PPP6R3 (frag.) 1 2 3

PRKCI PARD6B (f.l.) 1 2 3

PRKCI PARD6B (frag.) 1 2 3

55 40 35 25 15 kDa 170 130 100 70 55 40 35

lysate WB: streptavidin-HRP

25 15 kDa 170 130 100 70 55 40 35 25 15

Lanes: 1: EGFP + Avi-mCherry-bait 2: EGFP-prey + Avi-mCherry 3: EGFP-prey + Avi-mCherry-bait

Figure S2 A, B and C. Interactions testing positive by co-affinity purification. Bait proteins are tagged with Avi-mCherry and prey proteins with EGFP. For each protein pair, 3 blots are shown. Top: αGFP Western blots on proteins purified using streptavidin coated beads, middle: αGFP Western blots on input lysates (to determine EGFP expression levels), bottom: streptavidin-HRP Western blots on input lysates (to determine expression and biotinylation of Avi-mCherry). The tested proteins are indicated above each series of blots, the first protein listed is the bait, the second protein is the prey. Also indicated is whether the prey protein tested was full-length (f.l.) or corresponds to the shortest fragment identified by Y2H (frag.). In each blot, negative control mixtures are in lanes 1 and 2, while lane 3 contains the bait and prey proteins that interacted in the actual co-affinity purification (1: Avi-mCherry-bait tested against the EGFP tag, 2: the EGFP-prey tested against the Avi-mCherry tag, 3: the interacting Avi-mCherry-bait and 

124

AP: streptavidin WB: αGFP

Protein identification using an ORFeome-based Y2H fragment library

kDa 130 100

LNX1 PRPH (f.l.) 1 2 3

LNX1 PRPH (frag.) 1 2 3

MPP5 LIN7A (f.l.) 1 2 3

GPSM1 RALBP1 (frag.) 1 2 3

MPP5 ZNF451 (f.l.) 1 2 3

MPP5 ZNF451 (frag.) 1 2 3

70 55 40 35

kDa 130

lysate WB: αGFP

100 70 55 40

lysate WB: streptavidin-HRP

†

kDa 170 130 100 70 55 40 35 25 15

Lanes: 1: EGFP + Avi-mCherry-bait 2: EGFP-prey + Avi-mCherry 3: EGFP-prey + Avi-mCherry-bait EGFP-prey). Asterisks indicate EGFP-prey and Avi-mCherry-bait bands of the expected molecular mass. Empty Avi-mCherry (lane 2 in bottom blots) and EGFP (lane 1 in middle blots) both have a mass of 29 kDa, and are not indicated. In some lanes, degradation products are visible as bands of lower size. In the streptavidine-HRP Westerns (bottom row of blots), endogenously biotinylated proteins result in the detection of background bands in addition to the biotinylated Avi-mCherry fusion proteins. In two experiments, expression of the EGFP-prey fusion proteins was low, and 1/5 of empty-EGFP input lysate was loaded to prevent overexposure in the overnight exposure used to detect EGFP-prey fusion levels (indicated with a † in the blot). Finally, in the PROX-1 – PROX-1(frag) transfections, levels of biotinylated Avi-mCherry::PROX-1 were below the limits of detection in input lysate, but detectable by immunofluorescence microscopy and sufficient for affinity purification of EGFP::PROX-1.

125

Supplementary Tables Table S1: Interactions identified in our screens for which interaction domain information is present in the literature. Bait name

CTNNA3

Bait Prey name Prey Link to publication detailing description description interaction domains Alpha Catenin CTNNB1 Beta Catenin http://www.ncbi.nlm.nih.gov/ pubmed/8576147

CDK4

Cdk4

CDKN2D

p19Ink4d

CCND3

Cyclin D3

CDKN1B

p27Kip1

CDK4

Cdk4

CCND1

Cyclin D1

CTNNA3

Alpha Catenin JUP

PRKCI

Protein kinase PARD6B C, iota

MPP5

126

Pals1

LIN7A

Junction Plakoglobin

Par6

LIN7

http://www.ncbi.nlm.nih.gov/ pubmed/9751050

http://www.ncbi.nlm.nih.gov/ pubmed/8684460

http://www.ncbi.nlm.nih.gov/ pubmed/7630397

http://www.ncbi.nlm.nih.gov/ pubmed/8576147

http://www.ncbi.nlm.nih.gov/ pubmed/15590654

http://www.ncbi.nlm.nih.gov/ pubmed/22337881

Protein identification using an ORFeome-based Y2H fragment library

Table S2: Minimal interacting regions identified. All coordinates in basepairs. Corresponding images are Figure S1 A and B. Bait CDK4 CCND3 MPP5 CDKN2B CDK4 PRKCI CTNNA3 PARD3 PRKCI LNX1 DVL3 GPSM1 GPSM1 CDK4 GPSM1 PROX1 CTNNA3 CTNNA3 GPSM1 NUMA1 CDKN2B MPP5 CTNNA3

Prey CDKN2D CDKN1B LIN7A PYCRL CCND1 CRX SPRY2 CREB3 PARD6B PRPH HOMEZ CBS TRIM23 KLHL32 RALBP1 PROX1 JUP CTNNB1 PPP6R3 CCDC57 RNF20 ZNF451 EHMT2

Prey splice CDKN2D-201 CDKN1B-001 LIN7A-201 PYCRL-001 CCND1-201 CRX-201 SPRY2-001 CREB3-001 PARD6B-001 PRPH-001 HOMEZ-201 CBS-001 TRIM23-001 KLHL32-001 RALBP1-001 PROX1-001 JUP-001 CTNNB1-001 PPP6R3-001 CCDC57-001 RNF20-001 ZNF451-001 EHMT2-001

MRI start 1 7 22 20 44 13 39 74 10 40 58 371 146 11 403 189 53 79 522 4 232 306 914

MRI end 142 129 100 285 169 272 309 271 373 471 551 551 262 206 634 266 138 157 628 917 466 395 1211

Prey ORF length 167 199 234 287 296 300 316 372 373 471 551 552 575 621 656 738 746 782 874 917 976 1062 1211

127

Chapter 4

Controlled sumoylation of the Mevalonate pathway enzyme HMGS-1 regulates metabolism during aging

Amir Sapira, Assaf Tsurb, Thijs Koormanc, Kaitlin Chinga, Prashant Mishraa, Annabelle Bardenheiera, Lisa Podolskyb, Ulrike Bening-Abu-Shachb, Mike Boxemc, Tsui-Fen Choud, Limor Brodayb,1, Paul W. Sternberga,1 : Howard Hughes Medical Institute and Division of Biology and Biological Engineering,

California Institute of Technology, Pasadena, CA 91125, USA.

: Department of Cell and Developmental Biology, Sackler School of Medicine, Tel Aviv

University, Tel Aviv 69978, Israel

: Developmental Biology, Utrecht University, 3584 CH Utrecht, The Netherlands

: Division of Medical Genetics, Department of Pediatrics, Harbor-UCLA Medical Center

and Los Angeles Biomedical Research Institute,

Torrance, CA 90502, USA. : Corresponding authors.

Adapted from: Proceedings of the National Acadamy of Sciences of The United States of America (PNAS). 2014 Sep 16; 111 (37): E3880-9

Abstract

any metabolic pathways are critically regulated during development and aging but little is known about the molecular mechanisms underlying this regulation. One key metabolic cascade in eukaryotes is the mevalonate pathway. It catalyzes the synthesis of sterol and non-sterol isoprenoids, such as cholesterol and ubiquinone, as well as other metabolites. In humans, an age-dependent decrease in ubiquinone levels and changes in cholesterol homeostasis suggest that mevalonate pathway activity changes with age. However, our knowledge of the mechanistic basis of these changes remains rudimentary. We have identified a regulatory circuit controlling the sumoylation state of HMGS1, the C. elegans ortholog of human HMGCS1 protein, which mediates the first committed step of the mevalonate pathway. In vivo, HMGS-1 undergoes an agedependent sumoylation that is balanced by the activity of ULP-4 SUMO protease. ULP-4 exhibits an age-regulated expression pattern and a dynamic cytoplasm-tomitochondria translocation. Thus, spatiotemporal ULP-4 activity controls HMGS1 sumoylation state in a mechanism that orchestrate mevalonate pathway activity with the organismâ&#x20AC;&#x2122;s age. To expand the HMGS-1 regulatory network, we combined proteomic analyses with knockout studies and found that HMGS-1 level is also governed by the ubiquitin-proteasome pathway. We propose that these conserved molecular circuits have evolved to govern the level of mevalonate pathway flux during aging, a flux whose dysregulation is associated with numerous agedependent cardio-vascular and cancer pathologies.

130

Sumoylation of HMGS-1 regulates metabolism during aging

Introduction Many metabolic pathways are critically regulated during development and aging, but little is known about the molecular mechanisms underlying this regulation. The mevalonate pathway is a key metabolic cascade that converts acetyl-CoA and acetoacetyl-CoA to farnesyl diphosphate, a precursor of sterol isoprenoids such as cholesterol, steroid hormones, and bile acids. In addition, farnesyl diphosphate feeds into cascades that synthesize non-sterol isoprenoids, such as heme-A and ubiquinone, required for electron transfer during respiration1. Moreover, the mevalonate pathway catalyzes the synthesis of essential intermediates for tRNA modification, protein glycosylation, and protein prenylation. Protein prenylation is a requisite step in the activation of proteins involved in many intracellular signaling pathways that control cell growth and differentiation. For example, prenylation of small G proteins from the Ras, Rho, and Rac super-families dictates the membrane localization of these proteins that is essential for their activation2. Although the main trunk of the pathway is conserved in eukaryotes, some of the downstream branches vary between organisms. For example in fungi the main structural sterol produced by the pathway is ergosterol instead of cholesterol in vertebrates whereas in some invertebrates, including C. elegans, the sterol synthesis branch is absent3. Many cellular processes and physiological states rely on variable levels of mevalonate pathway metabolites, suggesting that the pathway activity is highly regulated during the organismâ&#x20AC;&#x2122;s life cycle. Elucidating the molecular details of this regulation is the first step in understanding how pathway dysregulation leads to diseases. For example, as part of the tumorigenic process of breast cancer cells, the expression of many mevalonate pathway enzymes is upregulated by a mutant p53 protein4. This over-activation of the mevalonate pathway is both necessary and sufficient to induce the mutant p53 phenotype in cancerous breast tissue architecture. Studies of mevalonate pathway regulation have primarily focused on regulators of HMG-CoA reductase (HMGCR), which converts HMG-CoA into mevalonate5. HMGCR catalyzes the primary rate-limiting step of the mevalonate pathway and is therefore highly regulated by transcriptional and post-transcriptional control6. The regulation of other enzymes in the pathway is largely unexplored, but growing evidence suggests that enzymes beyond HMGCR can serve as flux controlling points7. One potential regulatory node is the first committed enzyme of the pathway, HMGCoA synthase (HMGCS1). This cytoplasmic enzyme mediates the condensation of acetyl-CoA and acetoacetyl-CoA to HMG-CoA. hmgcs1 transcription is highly controlled by cholesterol levels in humans8. In addition, whole proteome studies of post-translational modifications have identified specific HMGCS1 residues that

131

undergo phosphorylation9, acetylation10, and ubiquitination11-13. These findings suggest that HMGCS1 undergoes complex post-translational regulation, but the biological significance of these modifications remains unclear. Among the growing family of ubiquitin like modifiers (UBLs), ubiquitin and SUMO (Small Ubiquitin-like MOdifier) are the most studied proteins. These two proteins share analogous enzymatic cascades that mediate, as a final step, covalent conjugation of the modifier to lysine residues of the substrate. In contrast to the canonical role of ubiquitin as mediator of protein degradation, conjugation of SUMO (sumoylation) can alter protein activity, localization, and stability without directly causing degradation. SUMO has emerged as a critical regulator in a variety of processes including cell cycle regulation, transcription, nuclear architecture control, chromosome stability regulation, and subcellular transport14,15. So far, the role of sumoylation in metabolic control has been attributed primarily to its transcriptional regulation activity as in the regulation of the metabolic transcription factor HIF-1a. However, SUMO is also conjugated to the mitochondria fission protein DRP-1[ref. 16], and some metabolic enzymes have been identified among the pool of potential SUMO targets17. These findings suggest that SUMO might play a direct role in metabolic regulation. SUMO modification is a highly dynamic and reversible process in part due to the activity of SUMO proteases that cleave SUMO from the substrate18. This highly conserved family of cysteine proteases includes Ulp1 and Ulp2 of S. cerevisiae, ULP1-2, 4-5 of C. elegans, and SENP1-3, 5-7 of humans18,19. SUMO proteases share the same domain organization of a conserved catalytic domain close to the C-terminus end and a variable N-terminus domain that dictates the protease sub-cellular localization and substrate recognition. Studies in yeast and mammals have revealed that most SENPs localize to the nucleus in recognizable sub-nuclear compartments18. In recent years, however, emerging data have demonstrated that some SENPs/ULPs also localize to the cytoplasm and intracellular organelle surfaces. For example, Ulp1 localizes primarily in the nuclear pore complex but translocates to the cytoplasm in a cell cycle-specific manner20. Even more strikingly, SENP5 localization cycles between the nucleus and the outer surface of the mitochondria. At the mitochondria, it regulates organelle morphology by controlling the sumoylation state of the mitochondria fission protein DRP-1[ref 21]. Beyond these few cases, the role of SENPs/ ULPs in the regulation of cytoplasmic and mitochondrial processes has yet to be explored. Here we describe a novel regulatory circuit that orchestrates HMGS-1 levels with age. We found that HMGS-1 undergoes an age-dependent sumoylation that is temporally balanced by the activity of ULP-4 SUMO protease. In addition to a role for sumoylation, we discovered that the ubiquitin-proteasome system also controls

132

Sumoylation of HMGS-1 regulates metabolism during aging

HMGS-1 levels. Evolutionary conservation of both mechanisms suggests that this regulatory network is fundamental to mevalonate pathway activity in eukaryotes. Our findings place the first committed enzyme of the pathway as a novel target for complex post-translational regulation.

Materials and methods Yeast 2-hybrid screen Yeast two-hybrid screens were performed by mating as previously described54. Four distinct regions of ulp-4 were used as bait: FL-ULP-4 (amino acids 1 – 382), N-ULP-4 (amino acids 1 – 110), Ex2-3ULP-4 (amino acids 58 – 145), and C-ULP-4 (amino acids 111 – 382). To avoid ULP-4-related toxicity, the predicted catalytic cysteine of the active site was mutated to serine (C743S) in constructs that bore the ULP-4 catalytic domain. To generate the Gal4-DB::ulp-4 bait clones, the corresponding regions of the ulp-4 coding sequence were PCR amplified from cDNA synthesized from mRNA of mixed-stage wild-type worms. A forward primer containing an AscI restriction site immediately upstream of the coding region was used with a reverse primer containing a stop codon followed by a NotI restriction site immediately downstream of the ulp-4 coding region. These sites were used to clone each of the four fragments into the Gal4 DB vector pMB27, a modified version of pPC97[ref. 55] which encodes a flexible linker (GGGG) upstream of the cloned ORF, and contains a polylinker that facilitates cloning using AscI and NotI restriction enzymes. Each of the Gal4DB::ulp-4 bait constructs were transformed into yeast strain Y8800 and mated to a Gal4-AD-cDNA library (a generous gift from X. Xin and C. Boone, U. Toronto) and a Gal4-AD-Fragment library containing fragments of 749 genes required for C. elegans early embryonic development54,55,56. Both libraries were in yeast strain Y8930[ref. 56]. Candidate interactors were selected on Sc –Leu –Trp –His plates, and assayed on three additional plates: Sc –Leu –Trp –His + 2 mM 3AT to gauge the strength of activation of the HIS3 reporter gene, Sc –Leu –Trp –Ade to gauge the strength of activation of the ADE2 reporter gene, and Sc –Leu –His + 1µg/ml cycloheximide to identify yeast clones that are able to activate the HIS3 reporter gene in the absence of an AD fusion protein as described57. Clones that activate the HIS3 reporter in the absence of a Gal4-AD interaction partner were eliminated. The identity of the remaining interactors was determined by PCR amplification from yeast cells and sequencing. As a final confirmation, we isolated the Gal4-AD plasmids, and transformed them into fresh Y8930 cells, which were mated with Gal4-DB-ULP-4 expressing Y8800 cells and again assayed for activation of the HIS3 and ADE2 reporter genes. Eight independent HMGS-1 clones were isolated from the Gal4-AD-cDNA library. 133

Detection of in vivo sumoylation Worms were washed from starved plates and transferred to growing media (M9 + 5µg/ml cholesterol + 1mM MgSO4 + HB101 bacteria) and grown in liquid culture in 1-3 worms/microliter density at 20°C or 25°C. For synchronized cultures beyond three days of adulthood, the fer-15(b26)II; fem-1(hc17)IV background was used. In this background, worms were grown at 25°C until day 1 of adulthood and then were transferred to the desired temperature. Worms were harvested by three to four cycles of M9 washing and collection by spinning 2600 rpm for 2 minutes. When the M9 became clear, worms were divided into micro-centrifuge tubes in aliquots of 300 mg compact worms. These aliquots were flashed-frozen in liquid nitrogen and stored at -80C until analyzed. From this point forward, all steps were carried out at 4°C unless otherwise stated. Worm pellets were resuspended in two volumes of cold worm lysis buffer (50 mM Tris HCl, pH 7.4, 150 mM NaCl, 1 mM EDTA, 1% Triton X-100, 1mM PMSF, 200 μM iodoacetamide, one mini tablet of Roche protease inhibitor –EDTA/10 ml buffer, and 25 mM N-ethylmaleimide) prepared fresh and added to every tube just before sonication. Immediately after resuspension, worms were sonicated in five ten-second intervals with 50 seconds between each sonication interval. After sonication, each lysate was spun twice at 14,000 rpm for 30 minutes and finally 14,000 rpm for five minutes, transferring the supernatant to a new tube each time. After determination of protein concentration, 7 milligram of total protein from each lysate was diluted in 800 microliter of lysis buffer. For the IP, FLAG beads (Sigma A2220) were thawed on ice, and 20 microliters of compact beads were used for every 300 microliters of compact worms. Lysate was placed on a roller over night; then beads were pelleted by centrifugation at 1800 rpm and washed three times in lysis buffer. Next, protein was eluted from the beads by adding 100 microliter of 150ng/µl 3xFLAG elution peptide in FLAG washing buffer. Samples were incubated for 30 minutes before the supernatants were transferred to a new microcentrifuge tube and 900 microliters of RIPA were added to each tube. Anti-GFP antibody (Roche) was added to the tubes before incubating for one hour. Next, protein-G beads were added and tubes were gently rolled for seven hours at 4°C. As a final step, the beads were washed three times in FLAG washing buffer and sample buffer was added directly to the bead pellet. Sumoylation prediction and in vitro sumoylation assays Potential sumoylation sites were identified by integrating the predictions of three sumoylation prediction programs: SUMOplot58, SUMOsp2.0[ref 59], and SUMORF. Lysine-to-arginine mutagenesis of candidate sites was done using Agilent technologies Quick Change site Directed Mutagenesis kit. Wild-type and mutant

134

Sumoylation of HMGS-1 regulates metabolism during aging

HMGS-1 proteins were cloned into pET28B, transformed into BL-21 (DE3). Colonies were grown over one day at 37°C, and protein expression was induced by 1mM IPTG over night at 20°C. Proteins were purified using standard techniques and were subjected to buffer exchange using Amicon centrifugal filters MWCO10kDa. After protein determination, 8 ng protein were used per sumoylation reaction preformed according the manufacture’s protocol. In vitro sumoylation reactions took place at 37°C with an optimal incubation time of 30 minutes. Mass spectrometry analyses Immunoprecipitation Worms were lysed in lysis buffer (1% Triton, 50mM Tris pH7.4, 150mL NaCl, 1mM EDTA, 250uM iodoacetamide, 25mM N-ethylmaleimide, 1mM PMSF, 10uM bortezomib, 1 tablet of Roche complete protease inhibitor) by sonication using parameters similar to the in vivo sumoylation assays. Cell lysates were cleared by centrifugation for 30 minutes at 13000 rpm in an Eppendorf centrifuge (5417R), and supernatants were transferred to a new tube to repeat the centrifugation. For each IP, FLAG beads (50µL) was washed twice with 1mL lysis buffer. Clear cell lysate was incubated with FLAG beads at 4˚C for 1 hour. Beads were washed with 1mL lysis buffer three times and 100mM Tris buffer (pH 8.5) twice. Proteins were eluted with 10M Urea in 100mM Tris buffer (pH 8.5) by incubating at 37˚C for 15 minutes. In-solution tryptic digest followed by Nano-LC-MS/MS Analysis Eluted proteins (100µL) were diluted with 25µL of 100mM Tris buffer (pH 8.5), reduced with 1µL of 0.5M Tris (2-carboxyethyl) phosphine hydrochloride (Thermo Fisher Scientific) for 20 minutes at 37˚C, alkylated with 3µL of 0.5M chloroacetamide (Fisher) for 15 minutes at 37˚C, and digested with 2µL of 100ng/µL Lysyl endopeptidase (Lys-C, Wako Chemicals) for 4 hours at 37oC. Samples were diluted to a final concentration of 2M urea by adding 375µL of 100mM Tris-HCl pH 8.5 and digested with 3µL of 100ng/µL trypsin (Thermo Fisher Scientific) with 5µL 100mM CaCl2 for 18 hours at 37oC. After desalting with Pierce C18 Tip (Thermo Fisher Scientific), peptides were eluted with 200µL of 75% CAN 0.1% TFA. Solvent was removed using a SpeedVac. Dried samples were acidified by 0.2% formic acid and loaded onto an Easy Nano-LC Q-Exactive Orbitrap. Peptides were loaded onto to a trap column (PepMap 100 C18 75uM 3uM Nanovip) and separated with a EasySpray column (PepMap C18, 3 uM 100Å, 75um x 15cm (ES800) with a 2-30% gradient (A: 0.1% FA in water, B: 0.1% FA in 100% acetonitrile) at 300nl/min for 60 minutes followed by 30% B to 100% B for 5 minutes and incubated at 100% B for 10 minutes.

135

The Q-Exactive acquisition method was created in data-dependent mode with one precursor scan in the Orbitrap, followed by fragmentation of the 15 most abundant peaks. Resolution of the precursor scan was set to 70,000, AGC target 3e6, Maximum IT 100 ms, scanning from 350â&#x20AC;&#x201C;2000 m/z. MS2 resolution was 17,500, AGC target 5e4, isolation window 2.0 m/z, Fixed first mass 100.0 m/z, NCE 25. Underfill ratio 0.1%. Charge exclusion unassigned, 1. Peptide match preferred. Exclude isotope on, Dynamic exclusion 30.0s. The peak list generating software used was Proteome Discoverer Software 1.4.1.14 (Thermo Scientific Inc., San Jose, CA). The outline of the method used in the spectrum selector mode of the Proteome Discoverer software. All were set to default settings to generate MS/MS spectra. The precursor charge state (high/low), retention time, minimum peak count, total intensity threshold value was all set to default settings. The Max and Min precursor mass settings were 5000 Da and 350 Da, respectively. The Caenorhabditis elegans Fasta file (canonical and isoform) was downloaded from UniProt (http://www. uniprot.org). The Sequest HT search parameters included in this study were: Two missed cleavages allowed, fully tryptic peptides only, min. peptide length 6, max. peptide length 144, fixed modification of carbamidomethyl cysteine (+57.021 Da), and dynamic modifications of oxidized methionine (+15.995Da), N-terminal acetylation (+42.011Da). The precursor ion mass tolerance was 10 ppm and fragment mass tolerance was 0.05 Da. Input data maximum Delta Cn 0.05. The decoy database search was applied with Strict FDR of 0.01 and Relaxed FDR of 0.05, Valdiation based on q-value. Alignment of protein sequences and structures The EBI ClustalW multiple sequence alignment server was used to align all protein sequences with the default setting. Sequences were further annotated using the JalView 2.8 software60. Human HMGCS1 modifications are based on data available at PhosphoSite Plus http://www.phosphosite.org/homeAction.do. The Phyre server61 was used for the search of proteins with a HMGS-1 related structure in the PBD using default parameters. For structural alignment of HMSG-1 and HMGCS1, we used the PyMOL 1.3 software. Phylogenetic analysis of C. elegans ULPs and other Ulp/SENP family members Protein sequences of SENP/Ulp were obtained by psi-BLAST search using NCBI Entrez non-redundant database or directly from genome sequencing projects. 250 sequences from 75 organisms representing the four kingdoms of eukarya were aligned using Multiple Sequence Comparison by the Log-Expectation program62.

136

Sumoylation of HMGS-1 regulates metabolism during aging

Bootstrapping and determination of test estimate of ML tree topology were conducted with Rapid bootstrapping algorithm from RaxML version 8.0.9 under the GTR+I model running on the CIPRES portal63. Consensus tree images were generated using FigTree (http://tree.bio.ed.ac.uk/software/figtree/) which were then manipulated using the Adobe Photoshop software. C. elegans genetics Worm strains were cultured as in64. All integrated strains were out-crossed at least four times before experiments. The tm33688 ulp-4 deletion was out-crossed at least seven times. Strains used in this study were: PS6355: ulp-4 (tm3688) / mIn1 [mIs14 dpy-10(e128)] PS6366: ulp-4 (tm3688) / mIn1 [mIs14 dpy-10(e128)]; tvIs92 [ulp-4p::ulp-4::GFP] PS6709: ulp-4 (tm3688) / mIn1 [mIs14 dpy-10(e128)]; syIs261 [HMGS-1::GFP 1ng/µl+ pha-1 50ng/µl +myo-2::mRFP 5ng/µl] CF152: fer-15(b26)II; fem-1(hc17)IV) NX349: fer-15(b26)II; fem-1(hc17)IV); syIa261 [HMGS-1::GFP 1ng/µl+ pha-1 50ng/µl +myo-2::mRFP 5ng/µl]; him-5(e1490) V PS6309: syIs268 [myo-3p::TOM20::mRFP (50ug/µl) +unc-119 (50ng/ µl)] V, tvIs92[ulp4p::ulp-4::gfp; rol-6] NX92: tvIs92[ulp-4p::ulp-4::gfp; rol-6] PS6332 tvIs92 [ulp-4p::ulp-4::gfp], ljEx446 [pclh-3::mcherry (75ng/µl), pRF4 (75 ng/µl] ulp-4 molecular analysis and constructs The tm3668 deletion in ulp-4 locus was characterized by PCR of genomic DNA. For analysis of ulp-4 transcripts in a tm3688 background, total mRNA was purified from tm3668 homozygotes, heterozygotes, and wild type worms. The mRNA was subjected to RT-PCR amplification with ulp-4 specific primers. To generate ulp-4 constructs, ulp-4 DNA was amplified using DNA Expand Taq (Roche) or PFU Ulta (Agilnet technologies), cloned into relevant plasmids, and sequenced. Injection, handling, and characterization of transgenic strains was done according to standard procedures. HMGS-1/ HMGCS1 constructs and worm strains HMGS-1 cDNA was amplified from an open-bio-system RNAi library and cloned into pET28B plasmid for bacterial expression and in vitro sumoylation. The Quick Change Site Directed Mutagenesis technique (Agilent technologies) was used for human HMGCS1 mutagenesis. For HMGCS1 cloning, total mRNA from human

137

GM12878 cells (pre B cell lymphoblastoid) was amplified with hmgcs1 specific primers and was cloned to the pET28B plasmid as. To follow the expression pattern and sumoylation state of HMGS-1 worms were injected with a fosmid that harbors a tagged version of HMSG-1. This fosmid, obtained from the TransgeneOme project, had HMGS-1 followed by the 2XTY, GFP, and 3XFLAG tags. The GFP tag was used for expression analysis while GFP and FLAG tags were used for biochemical analyses. Due to toxicity, the fosmid was injected at a final concentration of 1ng/µl. RNAi RNAi knockdown was performed according to65 with minor modifications. Relevant RNAi clones were sequenced to verify clone identity. To induce RNAi in liquid cultures, worms were incubated with the RNAi expression bacteria of interest in the presence of 1mM IPTG and 25ug/ml of carbenicillin. Microscopy Light microscopy Larvae and adults were anesthetized by 10mM levamisole, mounted on 2% agar slides, and stored in a humid box until use. Images were captured using Zeiss AxioImager M1 microscope and the Axiovision software. Images for intensity comparison were taken with the same imaging parameters. Confocal and super-resolution microscopy Embryos were harvested from gravid adults and mounted on a 2% agar pad, covered with a 50µl M9 drop and a small cover slide. Larvae and adults were immobilized on slides either using levamisole anesthetics or by the dry agar slide technique described in66. Images were captured using a Ziess LSM 510 confocal microscope. For super resolution analysis, worms were immobilized on dry agar slides and visualized using a Leica TCS sp5 II super resolution microscope (STED). After acquisition, images were deconvoluted using the built-in deconvolution algorithms of the Leica LAS-AF software. The PSF was generated by using a 2D Lorentz function with the dull-width half-maximum set to 150 nm (as calculated on the image using the signal energy algorithm; regularization parameter: 0.1). For each stage/condition number of animals analyzed, n ≥15. Oil-O-Red staining and measurement of total fat Oil-O-Red (Sigma, O9755) staining method was followed67 with minor modifications.

138

Sumoylation of HMGS-1 regulates metabolism during aging

TMRE staining TMRE powder (tetramethylrhodamine, ethyl ester; Molecular probes, T669) was dissolved in DMSO to make a 1mM stock solution. This stock was further diluted 100X in M9 and used to stain worms for 30 minutes in microcentrifuge tubes in the dark. Then, worms were washed 2-3 times in M9 and mounted on dry agar slides. Because TMRE signal bleaches quickly samples were visualized using confocal microscopy. Oxygen consumption measurements Oxygen consumption rate was measured in a Seahorse Biosciences Extracellular Flux Analyzer (model XF96) utilizing a protocol adapted from68. One-day old worms of the indicated genotype were washed in M9 medium (to remove bacteria) and placed on bacteria-free NGM agar plates. 10 worms were then manually transferred to each well in an XF96 measurement plate loaded with M9. Worms were allowed to settle to the bottom of the well, and oxygen consumption was measured four times per well. At least three different wells were measured for each condition or treatment. Pharyngeal pumping and swimming measurements To measure pharyngeal pumping, plates of worms were placed on a high magnification stereomicroscope. Pharyngeal pumping was captured in one-minute movies of individual worms at 30 frames/sec. These movies were opened in Image J and played at eight frames/second. Pharyngeal pumping in 450 frames (15 seconds) was counted manually. To measure worm swimming, a single worm was placed in a drop of 40Âľl M9 on an NGM plate using an eyelash. After 10 minutes, the swimming worm was captured in a video as above. Analysis was done as above but the number of body turns was counted in 30 seconds. A standard body turn was defined as a complete change in direction of bending at mid-body.

Results ULP-4 SUMO protease exhibits cytoplasm-to-mitochondria translocation As a first step in the study of SUMO proteases role in whole-organism development, we have focused on the family of C. elegans ULPs (Fig. S1). In the process of exploring the expression patterns of these ULPs, we found that a GFP-labeled ULP-4 exhibits a developmentally-regulated expression pattern that is dependent on tissue type and worm age (Fig. 1). Notably, ULP-4::GFP rescues ulp-4 deletion mutant phenotypes (see below); furthermore, its GFP signal is reduced by ulp-4 RNAi (Fig.

139

S2). These suggest that the construct represents endogenous ULP-4 activity. ulp-4 expression initiates in body-wall muscles and hypodermal cells during embryonic development (Fig. 1A-C) and is maintained throughout the wormâ&#x20AC;&#x2122;s life (Fig. 1D-H). During the development from larval stages 1 to 3, ulp-4 is expressed in the pharynx and hypodermis (Fig. 2D). A second phase of ulp-4 expression in body wall muscles (BWM) and hermaphroditic-specific neurons (HSN) starts at L4 stage (Fig. 1F) and continues through adulthood (Fig. 1G-H). Although ULP-4 localization pattern varies between tissue types, we found a common pattern of sub-cellular localization with time. In each tissue, ULP-4 protein initially localizes to the cellâ&#x20AC;&#x2122;s nucleus and cytoplasm but accumulates later in development in intra-cellular compartments that resemble mitochondria (Fig. 1K). We confirmed the mitochondrial localization by co-localization of ULP-4::GFP with Mitotracker Red, a dye that selectively and stereotypically labels C. elegans mitochondria (data not shown). To determine the sub-organelle localization of ULP-4, we used high-resolution microscopy to image worms co-expressing ULP-4::GFP and TOM-20::mRFP, a protein that labels the outer mitochondria membrane (Fig. 1 I-J). Optical cross sections through the organelle suggest that ULP-4 accumulates in the mitochondrial matrix. Likely, this age-regulated sub-cellular localization determines ULP-4 substrate specificity by changing its site of action. ULP-4 interacts with C. elegans HMGS-1 Identification of substrates is crucial for understanding the role of SENPs/ULPs in cellular processes. Intrigued by the ULP-4 localization pattern, we screened C. elegans two-hybrid libraries to identify potential substrates. Because the N-terminus portion of many SENPs/ULPs has been identified as the substrate recognition domain22, we used different regions of this domain as bait (Fig. 2A). The screen revealed eight independent interaction events between ULP-4 and the HMGS-1 protein (Fig. 2A). C. elegans HMGS-1 is the ortholog of two metabolic enzymes of mammals: HMGCS1 and HMGCS2. HMGCS1 is the first committed enzyme of the mevalonate pathway3 whereas HMGCS2 functions in ketone body synthesis23. Reports about ketone body metabolism in invertebrates are scarce24,25, so we focused on whether HMGS-1 is required for mevalonate pathway metabolism. hmgs-1 RNAi induces strong and robust phenotypes including paralysis, reduced pharyngeal pumping, and sterility (Fig. S3B, Movies S1). In agreement with previous reports about the involvement of hmgs-1 in the mevalonate pathway26,27, these hmgs-1 RNAi phenotypes were fully rescued by mevalonate supplementation, indicating that C. elegans HMGS-1 is functionally equivalent to its ortholog HMGCS1 (Fig. S3B, Movies S2).

140

Sumoylation of HMGS-1 regulates metabolism during aging

Figure 1. ULP-4 exhibits a cytosol-to-mitochondria translocation (A-C) ULP-4 expression initiates at mid-embryogenesis in muscles. ULP-4 protein is distributed in the cytoplasm. (D) ULP-4 protein cytoplasm-to-mitochondria translocation in hypodermal cells at L1 stage. Pharyngeal cells also start to express ULP-4 that is sorted to the mitochondria. (E) At L3 body wall muscle cells express ULP-4 as a cytoplasmic protein (arrow). (F) At L4, ULP-4 expression initiates at the hermaphrodite-specific neurons (HSN) concomitant with ULP-4 translocation to the mitochondria of head and tail muscles. (G) Day one adult worms exhibit mitochondrial localization of ULP-4 in HSNs and most of the body wall muscle cells. Inset is a muscle anterior to the HSN soma where ULP-4 is still cytoplasmic. (H) In day ten adult worms, ULP-4 localizes to the mitochondria of all muscle cells. Inset shows mitochondrial ULP-4 in the region as in (G). (I-J) Co-localization of ULP-4 with the mitochondrial outer-membrane marker TOM-20::mRFP in muscles demonstrates ULP-4 matrix localization. (K) A comprehensive diagram of ULP-4 expression and localization pattern. In all experiments n≥15 worms. Scale bars: (A-C) 10µm, (D-H) 50µm, G, H inserts 10µm, I 10µm, J 2.5 µm.

141

Figure 2. HMGS-1 is sumoylated in an age-dependent manner in vivo (A) ULP-4 and HMGS-1 protein fragments interact in a yeast 2-hybrid assay. (B-D) In vitro sumoylation reactions. Lanes labeled by â&#x20AC;&#x201C;E2 are control reactions in which UBC9 is absent, â&#x20AC;&#x201C;S are substrate-absent controls. (B) Time course (in minutes) of HMGS-1 sumoylation. (C) Similar to HMGS-1, human HMGCS1 is sumoylated in vitro. (D) The sumoylation pattern of wild-type and various lysine-to-arginine HMGS-1 mutants constructed based on sumoylation prediction (Fig. S4A). 4KR is a HMGS-1 variant whose four predicted sumoylated lysines were mutated to arginines. Arrows mark the sumoylated-HMGS-1 bands, and asterisks label sumoylation-abolished HMGS-1 variants. (E) An assay for HMGS-1 sumoylation detection in vivo. IP: immunoprecipitation; IB: Immunoblot (F) HMGS-1 sumoylation increases with age and temperature until day 10 of adulthood. Total levels of HMGS-1 were determined by HMGS1::FLAG::GFP detection from the total lysates. (G) ULP-4 knockdown resulted in increased HMGS1 sumoylation whereas ULP-4 over expression (ulp-4 OEx) resulted in a mild decrease of HMGS-1 sumoylation.

142

Sumoylation of HMGS-1 regulates metabolism during aging

HMGS-1 is sumoylated in vitro at lysine 408 The interaction between ULP-4 SUMO protease and HMGS-1 suggests that HMGS1 is sumoylated. To understand the molecular details of HMGS-1 modification, we first tested whether HMGS-1 is sumoylated in vitro. We found that a recombinant, bacterial-expressed, C. elegans HMGS-1 protein is efficiently sumoylated in vitro when incubated with human SUMO-1, E1, UBC9, and ATP (Fig. 2B). To examine a possible evolutionary conservation of this mechanism, we employed the same in vitro sumoylation assay on recombinant human HMGCS1 and found that it is also sumoylated. This observation suggests that this mode of regulation is evolutionarily conserved in humans (Fig. 2C). To identify the sumoylated residue, we surveyed the HMGS-1 protein sequence for predicted sumoylation sites. Of the twenty-seven HMGS-1 lysines, sumoylation site prediction programs identified four potential sites (Fig. S4A). In agreement with this prediction, the HMGS-1 sumoylation level was reduced when we mutated all four predicted lysines to arginines (Fig. 2D). Single Lys to Arg mutations pinpointed the sumoylated residue as Lys408. Lys408 is a strong candidate to be the site of HMGS-1 sumoylation in vivo as it has the highest sumoylation prediction score and is sumoylated in vitro. Moreover, this residue is conserved in the mammalian HMGCS1 sequence, and undergoes ubiquitination12 (Figs. S4B, S5). This conservation may represent an evolutionarily conserved cross talk between sumoylation and ubiquitination. Sumoylation of Lys408 might protect the protein from ubiquitination that occurs at the same site28, or it might elicit sumoylation-dependent ubiquitination (e.g.29). in vivo, HMGS-1 is sumoylated in an age-dependent manner To test whether C. elegans HMGS-1 is sumoylated in vivo, we immuno-precipitated a GFP-3XFLAG tagged version of HMGS-1 and assessed its sumoylation state by antiSUMO antibodies (Fig. 2E). We found that HMGS-1 is sumoylated when the cultures include worms grown for at least four days as gravid adults (Fig. 2F). The size shift of the modified HMGS-1 is about 20kDa, corresponding to mono-sumoylation. The fact that HMGS-1 sumoylation was detected only in cultures with worms older than day four gravid adult suggests that HMGS-1 is being sumoylated in an age-dependent manner. To test the dynamics of HMGS-1 SUMO conjugation, we grew synchronized, sterile cultures of worms and examined the degree of HMGS1 sumoylation in five-day intervals during adulthood. Strikingly, we found that the level of HMGS-1 sumoylation increases with age from day 1 of adulthood (no detectable sumoylation) until day 10 (maximum sumoylation; Fig. 2F). To test the possible effect of environmental conditions on HMGS-1 modification, we repeated the analysis on synchronized cultures grown at 25Ë&#x161;C, a mild stress condition. We

143

found the same pattern of age-dependent HMGS-1 modification; however, the level of HMGS-1 sumoylation increased relative to worms of the same age grown at 20Ë&#x161;C. This suggests that environmental factors that change the physiological state of the organism can also change the level of HMGS-1 modification. Worms fed with smo-1 RNAi exhibited an almost complete elimination of the SUMO band, demonstrating the specificity of this assay (Fig. 2G). To study the effect of ULP-4 activity on HMGS1 modification state, we knocked down ULP-4 SUMO protease by RNAi. Strikingly, this knockdown resulted in an increased level of HMGS-1 sumoylation in vivo (Fig. 2G). This finding accords with both the interaction between HMGS-1 and ULP-4 found in the 2-hybrid screen and the proposed enzymatic activity of ULP-4 as a SUMO protease. Taken together, these results reveal a novel molecular circuit in which controlled activity of ULP-4 SUMO protease balances the level of HMGS-1 sumoylation. ulp-4 mutants worms are metabolically impaired To unveil the function of ULP-4 SUMO protease we characterized the phenotypes resulting from a deletion in the ulp-4 locus. This allele introduces a frame shift followed by a stop codon at the beginning of exon 4 (Fig. S6), presumably resulting in a severe decrease of ULP-4 function. Worms homozygous for this allele are viable but exhibit phenotypes associated with impaired metabolism. ulp-4 mutants are almost completely sterile; furthermore, their pumping rate and locomotion undergo a notable decline with age (Fig. 3A). In addition, ulp-4 mutants have significantly less fat (Fig. S6C). This effect can be partially rescued by the ULP-4::GFP construct, demonstrating that these phenotypes stem from ULP-4 loss. Remarkably, these phenotypes show a strong age dependency consistent with ulp-4 age-dependent expression as well as with the proposed role of ULP-4 as a regulator of the mevalonate pathway. Consistent with this observation, ulp-4 mutant worms are long-lived, whereas ULP-4::GFP worms, in which ULP-4 is mildly overexpressed, are shortlived (Fig. S7). Through ubiquinone and heme-A synthesis, mevalonate pathway flux is directly required for proper electron transport (Fig. S3A). To examine the possible effect of ulp-4 loss on mitochondria homeostasis, we measured the degree of mitochondria membrane potential and respiration. ulp-4 mutant mitochondria fail to maintain normal membrane potential as determined by TMRE staining (Fig. 3D). In addition, ulp-4 loss results in decreased oxygen consumption relative to wild-type worms (Fig. 3E) whereas ULP-4::GFP worms consume more oxygen than wild-type worms. Dysregulation of the mevalonate pathway likely impairs mitochondrial electron transport rate by affecting CoQ production, thereby affecting life span and the rate of respiration30. Although ULP-4 probably has many targets, it is possible that the 144

Sumoylation of HMGS-1 regulates metabolism during aging

ulp-4 mutant phenotypes described above stem, at least in part, from dysregulated HMGS-1 activity. To examine this hypothesis, we tried to rescue the ulp-4 loss of function phenotypes by adding mevalonate exogenously. Indeed, mevalonate supplementation rescues ulp-4 mutants’ reduction of pumping in a dosagedependent manner while only marginally affecting the pumping rates of wildtype worms (Fig. 3C). Taken together, these results show the role ULP-4 plays in controlling the mevalonate pathway and worm metabolism. A partial rescue of ulp-4 mutant phenotypes by mevalonate demonstrates that over-sumoylated HMGS-1 is less active, placing ULP-4 as a positive regulator of HMGS-1 activity. This uncovers the mechanistic basis by which HMGS-1 sumoylation and ULP-4-mediated desumoylation control the mevalonate pathway.

Figure 3. ULP-4 loss results in age-dependent metabolic phenotypes (A) Pharyngeal pumping rate is reduced in ulp-4(-) mutant worms relative to the pumping rate of wild type worms. (B) ulp-4(-) worms exhibit an accelerated age-dependent reduction of locomotion. (C) Nutritional supplementation of mevalonate rescues the ulp-4(-) mutant phenotype. Worms were grown at 20°C and tested at day 1 of adulthood. Bars correspond to standard errors. P-values: *≤ 0.05; **≤ 0.01; ***≤ 0.001; ****≤ 0.0001. (D) Lack of TMRE accumulation demonstrates that mitochondria homeostasis is impaired in worms lacking ulp-4. (E) ULP-4 activity is required for proper respiration. ulp-4(-) worms exhibit a significant reduction in oxygen consumption. This reduction was rescued by the ULP-4::GFP construct described in Figure 1. Scale bars: D 20µm.

145

HMGS-1 interacts with the ubiquitin-proteasome pathway To characterize the molecular circuits that control HMGS-1 more comprehensively, we searched for HMGS-1 interactors by immunoprecipitation (IP) followed by mass spectrometry analysis. We identified 663 proteins that were enriched in an HMGS1 immunoprecipitated fraction in comparison to the control (Dataset 1). The list of HMGS-1 interactors was enriched in five KEGG pathways (Fig. 4A, and Fig. S8). Three of these (valine/leucine/isoleucine degradation, citrate cycle, and propanoate metabolism) feed substrates into the mevalonate pathway (acetyl-CoA or acetoacetylCoA, explicitly) that are metabolized by HMGS-1. This raises the possibility that HMGS-1 functions as part of a high-order metabolic scaffold (Fig. 4A). A fifth enriched pathway is the ubiquitin-proteasome, which may lead to degradation of HMGS-1. Strikingly, twenty proteins that co-precipitated with HMGS-1 are in the broader ubiquitin-proteasome pathway (Fig. S8). These include an E2 ubiquitin carrier (UBC-13), an ubiquitin fusion degradation enzyme (UFD-1), a proteasomal ubiquitin adaptor (ADRM-1 like)31, and an ubiquitin carboxyl-terminal hydrolase UBH-4. To test the potential role of HMGS-1 ubiquitination, these interactors were knocked down by RNAi. Knockdown of either UBA-5 or UBC-13 interactors resulted in an increased HMGS-1::GFP signal in comparison to an empty vector control (Fig. 4B). This demonstrates a role of the ubiquitin pathway in HMGS-1 regulation. To examine whether the identified HMGS-1 regulatory network is conserved in sterolproducing organisms, we searched the protein-protein interaction database BioGRID (http://thebiogrid.org/)32. Consistent with our findings in C. elegans, the HMGS-1 ortholog ERG13 of the sterol-producing yeast S. cerevisiae is both sumoylated33 and ubiquitinated34. ERG-13 interacts with ubiquitin (UBI4), ubiquitin-specific protease (UBP3), SUMO (SMT3), a SUMO protease (ULP-1)35 (Fig. 4D), as well as with various proteasomal subunits. So far, only a small set of interactors have been identified for HMGCS1, the human ortholog of HMGS-1. Although limited in number, the pool of HMGCS1 interactors includes the polyubiquitin-C protein11,12. High throughput screens have suggested that human HMGCS1 undergoes ubiquitination at multiple sites (Fig. S4), but the outcome of this modification is not yet clear. This interaction, along with our finding that HMGCS1 is sumoylated in vitro, suggest that this regulatory circuit is also conserved in humans.

146

Sumoylation of HMGS-1 regulates metabolism during aging

Figure 4. HMGS-1 interaction network and a proposed regulatory circuit (A) HMGS-1 interactors, revealed by HMGS-1 immunoprecipitation and mass spectrometry. Shown are KEGG pathways significantly enriched (see also Fig. S8A). The four enriched metabolic pathways are involved in either acetyl-CoA or acetoacetyl-CoA metabolism suggesting that HMGS-1 is part of high-order metabolic scaffolds. Twenty proteins from the proteasome and ubiquitin pathway were also identified (Fig. S8C). (B) HMGS-1::GFP over-expression worms were subjected to RNAi against candidate genes corresponding to proteins identified as HMGS-1 interactors from the proteomic screen. Depletion of UBA-5 and UBC-13 leads to an elevated levels of HMGS-1::GFP. (C) A survey of protein-protein databases demonstrated that HMGS-1 regulatory network is conserved in sterol-producing organisms. ULB-Tag are either Ubiquitin or SUMO modifiers and UBL-proteases are either Ubiquitin or SUMO proteases. (D) A model for HMGS-1 regulatory circuit. Thsi model suggest that balance between HMGS-1 sumoylation and de-sumoylation affects HMSG-1 activity and mevalonate pathway flux. H: C. elegans HMGS-1; [Mev]: Mevalonate pathway flux; Ub: C. elegans Ubiquitin-A; Su: C. elegans SUMO Scale bars: B 200Âľm.

147

A model for the HMGS-1 regulatory circuit Collectively, the molecular, genetic, and metabolic data suggest a model in which HMGS-1 activity is governed by two modes of post-translational modification (Fig. 4C). We propose that the balance between HMGS-1 sumoylation and desumoylation is highly regulated during the life span of the organism and is used to control the level of HMGS-1 activity. This balance orchestrates HMGS-1 activity with respect to tissue type, developmental stage, and the organismâ&#x20AC;&#x2122;s age. Early in the development of C. elegans, the activity of ULP-4 in the cytoplasm results in less sumoylated and presumably more active HMGS-1. As the worm ages, mitochondrial sequestration of ULP-4 results in HMGS-1 over-sumoylation and inactivation (Fig. 4D). In addition, the ubiquitin-proteasome pathway also governs HMGS-1 levels, highlighting how critical control of HMGS-1 levels is. It is possible that a cross talk between sumoylation and ubiquitination dictates HMGS-1 levels or that these two pathways are acting as two parallel branches within the HMGS-1 regulatory network. Nevertheless, these regulatory circuits not only explain how spatial and temporal control of mevalonate pathway flux evolved; they also place HMGS-1/ HMGCS1 as a target for age-dependent post-translational regulation. In addition to the dynamic control of HMG-CoA reductase levels by ubiquitination36, HMGS1/HMGCS1 post-translational regulation presumably plays a more long-term role in mevalonate pathway control. One potential implication of our findings is a mechanistic understanding of how the mevalonate pathway flux is controlled with age. We hypothesize that in humans, perturbations of this regulatory circuit may result in an age-dependent over-activation of the mevalonate pathway that leads to cholesterol over-production and cancer progression.

Discussion Despite their vital role in numerous cellular processes related to health and disease, little is known about post-translational mechanisms that control the mevalonate pathway beyond HMGCR. We have applied genetic and proteomic screens along with developmental, metabolic, and biochemical approaches to discover an evolutionarily conserved mechanism that controls HMGS-1 levels. This molecular circuit includes age-dependent HMGS-1 sumoylation that is balanced by the spatiotemporally regulated activity of the ULP-4 SUMO protease. We also identified the ubiquitin-proteasome pathway as part of the HMGS-1 regulatory network and demonstrated its role in controlling HMGS-1 levels. By identifying HMGS-1 as the target for both ubiquitination and age-controlled sumoylation and in vivo, we

148

Sumoylation of HMGS-1 regulates metabolism during aging

set the foundation for a mechanistic understanding of HMGS-1 post-translational regulation within the context of the dynamic metabolism of the organism. Studies on sumoylation were carried out primarily in single cell systems such as yeast and mammalian cells in culture19,37. It is not clear how sumoylation is controlled in a whole organism with various tissue types and a complex life cycle. The unexpectedly complex but stereotypic expression and sub-cellular localization pattern of ULP-4 (Fig. 1) may serve as a mechanism by which SUMO proteases temporally govern developmental processes. In contrast to SENP5 that localizes at some stages of the cell cycle on the mitochondria outer surface21, our high-resolution analyses show that ULP-4 translocates to the mitochondrial matrix (Fig. 1K). In addition, ULP-4 depletion does not seem to affect C. elegans DRP-138 localization or levels (data not shown), suggesting that ULP-4 is not a functional ortholog of SENP5. Although we cannot rule out the possibility that ULP-4 de-conjugates sumoylated mitochondrial matrix proteins, our results suggest that localization to the mitochondrial matrix sequesters ULP-4, preventing its cytoplasmic function. The ULP-4 protein does not have a defined mitochondria sorting signal, but it may harbor a non-canonical signal as do more than 50% of the mitochondrially-sorted proteins39. Consistent with reports on other SENPs/ULPs20, substrate recognition by ULP-4 was dependent on its N-terminus portion (Fig. 2A). This region harbors a signature acidic stretch that also exists in the N-terminus portion of human SENP-3 and less evidently in the ULP-4 orthologs SENP6/SENP7 (Fig. S9). This acidic stretch may play a conserved role either in sub-cellular sorting or in substrate recognition. Our temporal biochemical analyses clearly show that HMGS-1 undergoes age-dependent sumoylation (Fig. 2) and that this sumoylation is regulated by the activity of an age-controlled protease (Figs. 1-2). Although sumoylation may be a powerful mechanism to regulate protein activity with age, only a few cases of regulated sumoylation have been reported. Global sumoylation analyses comparing 3- and 25-month-old mice spleens revealed that a few uncharacterized proteins undergo age-enhanced sumoylation in a SUMO-1 specific manner40. In addition, SUMO-3 conjugation is reported to play some role in age-dependent learning and memory decay. In the AÎ˛PP Tg2576 mouse model of Alzheimerâ&#x20AC;&#x2122;s disease, SUMO1 conjugation of a yet-to-be identified substrate increased at 3 and 6 months in comparison to 1.5 months of age41. To our knowledge, age-dependent sumoylation of HMGS-1 is the first reported case of well-characterized, regulated sumoylation with age. Presumably, age-dependent SUMO conjugation is a widespread mechanism controlling many metabolic and signaling networks during aging. In contrast to the still uncharacterized role sumoylation plays during aging, the involvement of the ubiquitin-proteasome pathways is well established42,43. In

149

C. elegans for example, proteasomal E3 ligase activity regulates life span44, and elevated levels of proteasomal protein RPN-6 mediate somatic life span extension45. This demonstrates how changes of the activity of the protein degradation machinery may consequently affect the life span of the organism. By its nature, mass-spectrometry analyses (Figs. 4B, S8) are not tissue specific and represent the sum of interactors in the whole organism. As with ULP-4 (Fig. 1), it is likely that a spatiotemporal activity of specific interactors dictates the level HMGS-1 ubiquitination and activity in a tissue and age specific manner. The cross talk between the ubiquitin-proteasome and SUMO pathways may serve as a central mode of post-translational modification. For example, substrate sumoylation triggers ubiquitination and protein degradation via the activity of SUMO-targeted ubiquitin E3 ligases (STUbLs; e.g.46). Likewise, HMGS-1 sumoylation at old age may lead to the binding of SUMO-dependent ubiquitin E3 ligases and protein degradation. One fundamental characteristic of metabolism is its gross dependency and effect on the organism’s life cycle. In many cases, metabolism is modulated and even rewired as the organism develops, ages, and copes with environmental and physiological stress. For example, an expression analysis of the C. elegans stressresistant dauer larvae identified a switch from oxygenic respiration to anaerobic fermentation47,48. Even within the same organism, different tissues exhibit gross variability in metabolic profile. A comparison among ten human tissues revealed that post-transcriptional regulation plays a central role in shaping tissue-specific metabolic profiles49. Due to its vital role in cellular growth, differentiation, and basic activities such as cellular respiration, the mevalonate pathway is probably highly regulated during the life cycle of the organism. In humans, an age-dependent decrease in ubiquinone levels50 and changes in cholesterol homeostasis51 suggest that the pathway’s activity changes with age. The dynamic regulation of HMGR1 by cholesterol and cholesterol-related metabolites cannot account for the more long-term control required during the organism’s life cycle36. Notably, large-scale proteomic studies reveal that almost every enzyme in the pathway is subjected to ubiquitination7. This suggests that ubiquitin-dependent degradation by the proteasome may play a role at many rate-limiting points of the pathway. Only one modified enzyme, squalene monooxygenase, has been investigated in detail so far though52,53. The fact that HMGS-1/HMGCS1 activity allows substrates to enter the mevalonate pathway places its regulation as an effective control of flux from the pathway’s starting point. Consistent with this notion, the level of HMGCS-1 transcription is highly controlled by the sterol-responding transcription factor SREBP6 and by mutant p53[ref 4]. Our findings highlight a HMGS-1 post-translational modification network as an important regulatory mechanism for controlling the

150

Sumoylation of HMGS-1 regulates metabolism during aging

mevalonate pathway in C. elegans. Conservation of this network in the sterolproducing yeast S. cerevisiae (Fig. 4C) as well as the sumoylation of human HMGCS1 in vitro (Fig. 2C) suggests that this evolutionarily conserved mechanism may play a similar role in humans. Based on these data, we hypothesize that regulated HMGCS1 sumoylation along with ubiquitin-proteasome degradation control HMGS-1 activity with age. Perturbation of this regulatory network may contribute to pathway overactivation in various diseases such as p53-driven tumorigenesis and age-dependent loss of cholesterol homeostasis. This novel and conserved molecular circuit could serve as a handle for targeting the mevalonate pathway in future therapeutics.

Acknowledgements This research was supported by the Howard Hughes Medical Institute, with which P.W.S is an Investigator, and by the Israel Science Foundation (ISF 1617/11) for L.B.; A.T. was supported by the Ori Foundation-in memory of Ori Levi, the Israeli mitochondrial disease foundation www.orifund.org. T.F.C. is a member of UCLA Jonsson Comprehensive Cancer Center. We thank WormBase and the Caenorhabditis Genetics Center for C. elegans genetic annotation and strains, Shohei Mitani for a knockout allele, Ann Wang for the ulp-4 RNAi on ulp-4::gfp experiments, Adam Kolawa and Kevin Yu for their help with worm functional assays, Robyn Branicky and William Schafer for sharing the clh-3::mCherry worms, Brian Williams for total human RNA, and David Chan for use of his Seahorse oxygen consumption analyzer. The Henry L. Guenther Foundation supported the Q-Exactive Mass Spectrometer. We thank Domenico Fasci for his help with in vitro sumoylation assays and Jennifer Watt for fat measurements.

Author contributions A.S., A.T., L.B., and P.W.S. conceived and designed the experiments. T.K. and M.B. performed the 2-hybrid analysis. K.C., A.B., L.P., U.B., A.S., A.S, and L.B. performed the genetic, developmental, and metabolic experiments. A.S. and L.B performed the sumoylation assays, P.M. preformed the oxygen consumption assays, A.S., A.T., and L.B. performed the imaging; K.C. performed life span experiments. T.F.C. performed the mass spectrometry analyses. All of the authors discussed the data and helped manuscript preparation. A.S. wrote the manuscript with intellectual input from all authors. P.W.S. and L.B. supervised the project. ď&#x201A;ž

151

References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25.

152

Brown MS & Goldstein JL (1980) Multivalent feedback regulation of HMG CoA reductase, a control mechanism coordinating isoprenoid synthesis and cell growth. J Lipid Res 21(5):505-517. Zhang FL & Casey PJ (1996) Protein prenylation: molecular mechanisms and functional consequences. Annu Rev Biochem 65:241-269. Rauthan M & Pilon M (2011) The mevalonate pathway in C. elegans. Lipids Health Dis 10:243. Freed-Pastor WA, et al.(2012) Mutant p53 disrupts mammary tissue architecture via the mevalonate pathway. Cell 148(1-2):244-258. DeBose-Boyd RA (2008) Feedback regulation of cholesterol synthesis: sterol-accelerated ubiquitination and degradation of HMG CoA reductase. Cell Res 18(6):609-621. Goldstein JL & Brown MS (1990) Regulation of the mevalonate pathway. Nature 343(6257):425-430. Sharpe LJ & Brown AJ (2013) Controlling cholesterol synthesis beyond 3-hydroxy-3-methylglutarylCoA reductase (HMGCR). J Biol Chem 288(26):18707-18715. Smith JR, Osborne TF, Brown MS, Goldstein JL, & Gil G (1988) Multiple sterol regulatory elements in promoter for hamster 3-hydroxy-3-methylglutaryl-coenzyme A synthase. J Biol Chem 263(34):18480-18487. Van Hoof D, et al. (2009) Phosphorylation dynamics during early differentiation of human embryonic stem cells. Cell Stem Cell 5(2):214-226. Choudhary C, et a. (2009) Lysine acetylation targets protein complexes and co-regulates major cellular functions. Science 325(5942):834-840. Kim W, et al. (2011) Systematic and quantitative assessment of the ubiquitin-modified proteome. Mol Cell 44(2):325-340. Wagner SA, et al. (2011) A proteome-wide, quantitative survey of in vivo ubiquitylation sites reveals widespread regulatory roles. Mol Cell Proteomics 10(10):M111 013284. Wagner SA, et al. (2012) Proteomic analyses reveal divergent ubiquitylation site patterns in murine tissues. Mol Cell Proteomics 11(12):1578-1585. Flotho A & Melchior F (2013) Sumoylation: a regulatory protein modification in health and disease. Annu Rev Biochem 82:357-385. Gareau JR & Lima CD (2010) The SUMO pathway: emerging mechanisms that shape specificity, conjugation and recognition. Nat Rev Mol Cell Biol 11(12):861-871. Braschi E, Zunino R, & McBride HM (2009) MAPL is a new mitochondrial SUMO E3 ligase that regulates mitochondrial fission. EMBO Rep 10(7):748-754. Kaminsky R, et al. (2009) SUMO regulates the assembly and function of a cytoplasmic intermediate filament protein in C. elegans. Dev Cell 17(5):724-735. Mukhopadhyay D & Dasso M (2007) Modification in reverse: the SUMO proteases. Trends Biochem Sci 32(6):286-295. Yeh ET (2009) SUMOylation and De-SUMOylation: wrestling with lifeâ&#x20AC;&#x2122;s processes. J Biol Chem 284(13):8223-8227. Li SJ & Hochstrasser M (2003) The Ulp1 SUMO isopeptidase: distinct domains required for viability, nuclear envelope localization, and substrate specificity. J Cell Biol 160(7):1069-1081. Zunino R, Braschi E, Xu L, & McBride HM (2009) Translocation of SenP5 from the nucleoli to the mitochondria modulates DRP1-dependent fission during mitosis. J Biol Chem 284(26):17783-17795. Dou H, Huang C, Singh M, Carpenter PB, & Yeh ETH (2010) Regulation of DNA Repair through DeSUMOylation and SUMOylation of Replication Protein A Complex. (Translated from English) Molecular Cell 39(3):333-345 (in English). Hegardt FG (1999) Mitochondrial 3-hydroxy-3-methylglutaryl-CoA synthase: a control enzyme in ketogenesis. Biochem J 338 ( Pt 3):569-582. Weltzien FA, Hemre GI, Evjemo JO, Olsen Y, & Fyhn HJ (2000) beta-Hydroxybutyrate in developing nauplii of brine shrimp (Artemia franciscana K.) under feeding and non-feeding conditions. Comp Biochem Physiol B Biochem Mol Biol 125(1):63-69. Beis A, Zammit VA, & Newsholme EA (1980) Activities of 3-hydroxybutyrate dehydrogenase, 3-oxoacid CoA-transferase and acetoacetyl-CoA thiolase in relation to ketone-body utilisation in muscles from vertebrates and invertebrates. Eur J Biochem 104(1):209-215.

Sumoylation of HMGS-1 regulates metabolism during aging

26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51.

Shi Z & Ruvkun G (2012) The mevalonate pathway regulates microRNA activity in Caenorhabditis elegans. Proc Natl Acad Sci U S A 109(12):4568-4573. Liu Y, Samuel BS, Breen PC, & Ruvkun G (2014) Caenorhabditis elegans pathways that surveil and defend mitochondria. Nature 508(7496):406-410. Kim MJ, Chia IV, & Costantini F (2008) SUMOylation target sites at the C terminus protect Axin from ubiquitination and confer protein stability. FASEB J 22(11):3785-3794. Hickey CM, Wilson NR, & Hochstrasser M (2012) Function and regulation of SUMO proteases. Nat Rev Mol Cell Biol 13(12):755-766. Dillin A, et al. (2002) Rates of behavior and aging specified by mitochondrial function during development. Science 298(5602):2398-2401. Husnjak K, et al. (2008) Proteasome subunit Rpn13 is a novel ubiquitin receptor. Nature 453(7194):481-488. Stark C, et al. (2006) BioGRID: a general repository for interaction datasets. Nucleic Acids Res 34(Database issue):D535-539. Hannich JT, et al. (2005) Defining the SUMO-modified proteome by multiple approaches in Saccharomyces cerevisiae. J Biol Chem 280(6):4102-4110. Peng J, et al. (2003) A proteomics approach to understanding protein ubiquitination. Nat Biotechnol 21(8):921-926. Sung MK, et al. (2013) Genome-wide bimolecular fluorescence complementation analysis of SUMO interactome in yeast. Genome Res 23(4):736-746. Sever N, Yang T, Brown MS, Goldstein JL, & DeBose-Boyd RA (2003) Accelerated degradation of HMG CoA reductase mediated by binding of insig-1 to its sterol-sensing domain. Mol Cell 11(1):2533. Broday L, et al. (2004) The small ubiquitin-like modifier (SUMO) is required for gonadal and uterine-vulval morphogenesis in Caenorhabditis elegans. Genes Dev 18(19):2380-2391. Labrousse AM, Zappaterra MD, Rube DA, & van der Bliek AM (1999) C. elegans dynamin-related protein DRP-1 controls severing of the mitochondrial outer membrane. Mol Cell 4(5):815-826. Bolender N, Sickmann A, Wagner R, Meisinger C, & Pfanner N (2008) Multiple pathways for sorting mitochondrial precursor proteins. EMBO Rep 9(1):42-49. Zhang L, Li F, Dimayuga E, Craddock J, & Keller JN (2007) Effects of aging and dietary restriction on ubiquitination, sumoylation, and the proteasome in the spleen. FEBS Lett 581(28):5543-5547. Nistico R, et al. (2014) Age-related changes of protein SUMOylation balance in the AbetaPP Tg2576 mouse model of Alzheimerâ&#x20AC;&#x2122;s disease. Front Pharmacol 5:63. Low P (2011) The role of ubiquitin-proteasome system in ageing. Gen Comp Endocrinol 172(1):3943. Kevei E & Hoppe T (2014) Ubiquitin sets the timer: impacts on aging and longevity. Nat Struct Mol Biol 21(4):290-292. Ghazi A, Henis-Korenblit S, & Kenyon C (2007) Regulation of Caenorhabditis elegans lifespan by a proteasomal E3 ligase complex. Proc Natl Acad Sci U S A 104(14):5947-5952. Vilchez D, et al. (2012) RPN-6 determines C. elegans longevity under proteotoxic stress conditions. Nature 489(7415):263-268. Tatham MH, et al. (2008) RNF4 is a poly-SUMO-specific E3 ubiquitin ligase required for arsenicinduced PML degradation. (Translated from English) Nat Cell Biol 10(5):538-546 (in English). Burnell AM, Houthoofd K, Oâ&#x20AC;&#x2122;Hanlon K, & Vanfleteren JR (2005) Alternate metabolism during the dauer stage of the nematode Caenorhabditis elegans. Exp Gerontol 40(11):850-856. Madi A, et al. (2008) Mass spectrometric proteome analysis suggests anaerobic shift in metabolism of Dauer larvae of Caenorhabditis elegans. Biochim Biophys Acta 1784(11):1763-1770. Shlomi T, Cabili MN, Herrgard MJ, Palsson BO, & Ruppin E (2008) Network-based prediction of human tissue-specific metabolism. Nat Biotechnol 26(9):1003-1010. Bentinger M, Tekle M, & Dallner G (2010) Coenzyme Q--biosynthesis and functions. Biochem Biophys Res Commun 396(1):74-79. Martin M, Dotti CG, & Ledesma MD (2010) Brain cholesterol in normal and pathological aging. Biochim Biophys Acta 1801(8):934-944.

153

52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68.

154

Foresti O, Ruggiano A, Hannibal-Bach HK, Ejsing CS, & Carvalho P (2013) Sterol homeostasis requires regulated degradation of squalene monooxygenase by the ubiquitin ligase Doa10/Teb4. Elife 2:e00953. Zelcer N, et al. (2014) The E3 ubiquitin ligase MARCH6 degrades squalene monooxygenase and affects 3-hydroxy-3-methyl-glutaryl coenzyme A reductase and the cholesterol synthesis pathway. Mol Cell Biol 34(7):1262-1270. Boxem M, et al. (2008) A protein domain-based interactome network for C. elegans early embryogenesis. Cell 134(3):534-545. Vidal M, Brachmann RK, Fattaey A, Harlow E, & Boeke JD (1996) Reverse two-hybrid and onehybrid systems to detect dissociation of protein-protein and DNA-protein interactions. Proc Natl Acad Sci U S A 93(19):10315-10320. Yu H, et al. (2008) High-quality binary protein interaction map of the yeast interactome network. Science 322(5898):104-110. Vidalain PO, Boxem M, Ge H, Li S, & Vidal M (2004) Increasing specificity in high-throughput yeast two-hybrid experiments. Methods 32(4):363-370. Xue Y, Zhou F, Fu C, Xu Y, & Yao X (2006) SUMOsp: a web server for sumoylation site prediction. Nucleic Acids Res 34(Web Server issue):W254-257. Ren J, et al. (2009) Systematic study of protein sumoylation: Development of a site-specific predictor of SUMOsp 2.0. Proteomics 9(12):3409-3412. Waterhouse AM, Procter JB, Martin DM, Clamp M, & Barton GJ (2009) Jalview Version 2--a multiple sequence alignment editor and analysis workbench. Bioinformatics 25(9):1189-1191. Kelley LA & Sternberg MJ (2009) Protein structure prediction on the Web: a case study using the Phyre server. Nat Protoc 4(3):363-371. Edgar RC (2004) MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Res 32(5):1792-1797. Stamatakis A, Hoover P, & Rougemont J (2008) A rapid bootstrap algorithm for the RAxML Web servers. Syst Biol 57(5):758-771. Brenner S (1974) The genetics of Caenorhabditis elegans. Genetics 77(1):71-94. Timmons L, Court DL, & Fire A (2001) Ingestion of bacterially expressed dsRNAs can produce specific and potent genetic interference in Caenorhabditis elegans. Gene 263(1-2):103-112. Han SM, et al. (2012) Secreted VAPB/ALS8 major sperm protein domains modulate mitochondrial localization and morphology via growth cone guidance receptors. Dev Cell 22(2):348-362. Yen K, et al. (2010) A comparative study of fat storage quantitation in nematode Caenorhabditis elegans using label and label-free methods. PLoS One 5(9). Houtkooper RH, et al. (2013) Mitonuclear protein imbalance as a conserved longevity mechanism. Nature 497(7450):451-457.

Sumoylation of HMGS-1 regulates metabolism during aging

Associated content Supplementary movies S1 and S2. This material is available free of charge via the Internet.

Supplementary information

SUPPLEMENTARY INFORMATION Figure S1

Figure S1

Figure S1 Phylogenetic tree of C. elegans ULPs and other ULPs/SENPs Figure S1 Phylogenetic of C.inelegans ULPs and other ULPs/SENPs C. elegans ULPs aretree labeled bold font. Colors represent different ULP/SENP families. C. elegans are 1-3; labeled bold font. Colors ULP/SENP families. HumanULPs SENPs 5-8inare shown as wellrepresent as ULPsdifferent of D. melanogaster (Dme) and S. cerevisiae (Sce). Scale bar Human SENPsbranch 1-3; 5-8 are shown as well as ULPs of D.number melanogaster (Dme) and cerevisiae (Sce). Scale bar indicates indicates length, expressed as the expected of substitutions perS.site. branch length, expressed as the expected number of substitutions per site.

155

Figure S2 Figure S2

Figure ulp-4::gfp signal is reduced by ulp-4 RNAi Figure S2.S2. ulp-4::gfp signal is reduced by ulp-4 RNAi tested activityofofulp-4 ulp-4RNAi RNAion onworms wormsexpressing expressingulp-4p::ulp-4::GFP. ulp-4p::ulp-4::GFP.(A) (A)The Thestereotypic stereotypicexpression expressionpattern patternof WeWe tested thetheactivity ulp-4::GFP in dayin1 day adult1 hermaphrodites fed with expression (white arrowhead) and expression of ulp-4::GFP adult hermaphrodites fedempty with vector empty RNAi. vector HSN RNAi. HSN expression (white arrowhead) and in expression head muscles arrows) is shown. (B) ulp-4 RNAi activity resulted the elimination signal from body wall in (white head muscles (white arrows) is shown. (B) ulp-4 RNAiinactivity resultedofinGFP the elimination of GFP muscle HSN neurons. the HSN wormsneurons. were treated ulp-4 RNAi morewith thanulp-4 two RNAi generations, the than signal signalcells fromand body wall muscleWhen cells and Whenwith the worms were for treated for more was also reduced in the pharyngeal cellsalso (data not shown). Scale bar,cells 50Âľm. two generations, signal was reduced in pharyngeal (data not shown). Scale bar, 50Âľm.

Figure S3 Figure S3

Figure S3. hmgs-1 knock down is rescued by mevalonate supplementation Figure S3. hmgs-1 knock down is rescued by mevalonate supplementation (A) of mevalonate mevalonatepathway pathwaymetabolism metabolism . Cholesterol production is absent in elegans C. elegans (A)AAschematic schematic representation representation of . Cholesterol production is absent in C. but but other branches ofpathway the pathway are conserved. (B) Worms with RNAi hmgs-1 RNAi show a significant in other branches of the are conserved. (B) Worms fed withfed hmgs-1 show a significant reduction inreduction the level of the level of pharyngeal pumping. Thisbe effect be fully by supplementation of 20mM mevalonate indicating pharyngeal pumping. This effect can fullycan rescued by rescued supplementation of 20mM mevalonate indicating that hmgs-1 knock down phenotypes stem from impaired pathway flux. 20mM mevalonate not affect the pumping that hmgs-1 knock down phenotypes stemmevalonate from impaired mevalonate pathway flux.does 20mM mevalonate does rate not of wild worms; precluding a non-specific rescue bya mevalonate Bars represent standard errors. affect thetype pumping rate of wild type worms; precluding non-specificsupplementation. rescue by mevalonate supplementation. Bars Number of worms, n= 10 wild type worms in each condition. represent standard errors. Number of worms, n= 10 wild type worms in each condition.

156

Sumoylation of HMGS-1 regulates metabolism during aging

FigureS4S4 Figure

Figure S4. A comparison between HMGS-1 and human HMGCS1 modifications Figure S4. A comparison between andand human HMGCS1 modifications (A) Prediction of sumoylation sites HMGS-1 in HMGS-1 HMGCS1 proteins. (A) Prediction of sumoylation sites in HMGS-1 and HMGCS1 proteins. (B) HMGCS1 undergoes ubiquitination and acetylation at multiple sites. Ubiquitination UbiquitinationofofK426 K426may mayrepresent representa (B) HMGCS1 undergoes ubiquitination and acetylation at multiple sites. acompetition competitionbetween between sumoylation ubiquitination in HMGCS1. Human HMGCS1 modifications are on based sumoylation andand ubiquitination in HMGCS1. Human HMGCS1 modifications are based data on data available in PhosphoSite Plus http://www.phosphosite.org/homeAction.do. Sequences using available in PhosphoSite Plus http://www.phosphosite.org/homeAction.do. Sequences were aligned were usingaligned ClustalW and visualizedand using the Jalview software. ClustalW visualized using the Jalview software.

157

FigureS5 S5 Figure

Figure S5. A model of HMGS-1 structure and its alignment with human HMGCS1 Figure S5. A model of HMGS-1 structure and its with human (A) HMGS-1 predicted 3D structure based on alignment PDB structural search.HMGCS1 HMGS-1 shares the same structure with (A)the HMGS-1 predictedHMGCS2 3D structure based PDB structural HMGS-1 shares the same structure with the human human proteins (42%) andonHMGCS1 (40%)search. with confidence of 100%. The sumoylated lysine 408 proteins HMGCS2 (42%) and HMGCS1 (40%) with confidence of 100%. The sumoylated lysine 408 is labeled in red. (B) is labeled in red. (B) Alignment between HMGS-1 and its functional human ortholog HMGCS1. On HMGCS1 Alignment between HMGS-1 and its functional human ortholog HMGCS1. On HMGCS1 structure, lysine 426 which is the structure, 426 which is of theHMGS-1 closest tois the sumoylated of HMGS-1 is labeled in blue. This residue was closest to the lysine sumoylated residue labeled in blue. residue This residue was found to undergo ubiquitination in human 1 found toThe undergo in human HMGCS1. The Phyre wasrelated used for the search of proteins with a HMGCS1. Phyreubiquitination server1 was used for the search of proteins with aserver HMGS-1 structure. HMGS-1 related structure.

158

Sumoylation of HMGS-1 regulates metabolism during aging

Figure S6

Figure S6. Molecular characterization of the ulp-4 locus and its requirement for normal fat homeostasis (A) RT-PCR analyses of tm3886 withinofthe ulp-4 N-terminus region is infat green (exons 1-3). This is Figure S6. Molecular characterization theulp-4 ulp-4 locus. locus and its requirement for normal homeostasis (A) RT-PCR analyses of tm3886 within the 4-5) ulp-4that locus. ulp-4 N-terminus region is inharbors green (exons 1-3).catalytic This is followed followed by the C-terminus region (exons is labeled in gray. This region the ulp-4 domain by the C-terminus (exons 4-5) that is analyses labeled inrevealed gray. This the ulp-4 catalytic (conserved cysteineregion is labeled “c”). RT-PCR thatregion in theharbors strain homozygous for thedomain tm3886(conserved deletion, cysteine is labeled “c”). RT-PCR analyses revealed that in the strain homozygous for the tm3886 deletion, a truncated a truncated transcript still exists. Because tm3886 deletion introduces a frame shift followed by a stop codon before transcript still exists. Because tm3886 deletion introduces a frame shift followed by a stop codon before the catalytic site, the this transcript presumably does not code a functional ULP-4fat protein. (B-E) Worms’byfatOil-Olevel thiscatalytic transcriptsite, presumably does not code for a functional ULP-4for protein. (B-E) Worms’ level was determined was by ulp-4(tm3688) Oil-O-Red staining. ulp-4(tm3688) worms have control less fat(B). than(D) theThe wild-type Reddetermined staining. (C) mutant (C) worms have less fatmutant than the wild-type level ofcontrol fat is marginally worms over-expressing ulp-4::gfp . (E) ulp-4::gfp construct ulp-4::gfp. partially rescues fat loss ofconstruct ulp-4(-) (B). (D) Theincreased level ofinfat is marginally increased in worms over-expressing (E) the ulp-4::gfp mutants rescues indicating ulp-4::gfp construct is functional bar,ulp-4::gfp 50 µm. construct is functional. Scale bar, 50 partially thethat fatthe loss of ulp-4(-) mutants indicating. Scale that the µm.

159

Figure S7

Figure S7. Opposite effects of ulp-4 loss and ulp-4::gfp over expression on C. elegans life span

(A) Life span analyses type, ulp-4 and worms over expressing own promoter. Figure S7. Opposite effectsofofwild ulp-4 loss andmutants, ulp-4::gfp over expression on ulp-4::gfp C. elegansunder life its span ulp-4 mutants have a longer median and maximal life span than wild type control (B) whereas ulp-4 over-expression (A) Life span analyses of wild type, ulp-4 mutants, and worms over expressing ulp-4::gfp under its own promoter. ulp-4 opposite effect. The effect of ulp-4 on C. elegans life span is significant based on multiple statistical mutantsexhibits have a anlonger median and maximal life span than wild type control (B) whereas ulp-4 over-expression exhibits an analyses (C). All life span experiments were held at 20째C on plates seeded with OP50 at the same time. opposite effect. The effect of ulp-4 on C. elegans life span is significant based on multiple statistical analyses (C). All life span experiments were held at 20째C on plates seeded with OP50 at the same time

160

Sumoylation of HMGS-1 regulates metabolism during aging

Figure S8 Figure S8

FigureS8. S8.Proteomic Proteomicanalysis analysis HMGS-1 interactors Figure ofof HMGS-1 interactors (A)AAdiagram diagramofofthe themetabolic metabolicnetwork networkdiscovered discoveredbybythe thesurvey surveyofofHMGS-1 HMGS-1intercators. intercators.This Thisanalysis analysissuggests suggeststhat (A) that acetyl-CoA and acetoacetyl-CoA synthesis is physically coupled to condensation by HMGS-1, presumably in a acetyl-CoA and acetoacetyl-CoA synthesis is physically coupled to condensation by HMGS-1, presumably in a hand-off hand-off type of mechanism. (B)ofStatistics of the KEGG (Kyoto Encyclopedia of Genomes) Genes andcomponents Genomes) components type of mechanism. (B) Statistics the KEGG (Kyoto Encyclopedia of Genes and enriched in the immunoprecipitated fraction. (C) The fraction. list of ubiquitin-proteasome related proteins interacting with HMGS-1. enriched in the immunoprecipitated (C) The list of ubiquitin-proteasome related proteins interactingHMGS-1 with itself, labeledHMGS-1 in red, was found as theinmost enriched the enriched immunoprecipitated fraction, demonstrating fraction, the potency HMGS-1. itself, labeled red, was foundprotein as the in most protein in the immunoprecipitated ofdemonstrating our assay. The is comprised fifteen identified by the KOG (Eukaryotic Orthologous Groups of proteins) thelistpotency of our of assay. Theproteins list is comprised of fifteen proteins identified by the KOG (Eukaryotic analysis and five proteins that were identified manually. Orthologous Groups of proteins) analysis and five proteins that were identified manually.

161

Figure S9 Figure S9

Figure S9. ClustalW sequence alignment between C. elegans ULP-4 SUMO protease and Human SENP-3 Figure S8. ClustalW sequence alignment between C. elegans ULP-4 SUMO protease and Human SENP-3 and and SENP-6 SENP-6 Specific acidic are labeled labeledininred, red,and and acidic regions (rich in acidic residues) are underlined inAcidic blue. Acidic Specific acidic residues residues are acidic regions (rich in acidic residues) are underlined in blue. regions 2. regions not present in otherSENPs. humanThe SENPs. The was alignment wasusing annotated usingsoftware the JalView software2每每每. are not are present in other human alignment annotated the JalView

Supplemental References Supplemental References: 1.

162

Kelley LA & Sternberg MJ (2009) Protein structure prediction on the Web: a case study using the Phyre server.

Kelley LA & 4(3):363-371. Sternberg MJ (2009) Protein structure prediction on the Web: a case study using the Nat Protoc Waterhouse AM, Protoc Procter JB, Martin DM, Clamp M, & Barton GJ (2009) Jalview Version 2--a multiple sequence Phyre server. Nat 4(3):363-371. alignment editor and analysis workbench. Bioinformatics Waterhouse AM, Procter JB, Martin DM, Clamp M, &25(9):1189-1191 Barton GJ (2009) Jalview Version 2--a multiple sequence alignment editor and analysis workbench. Bioinformatics 25(9):1189-1191

Sumoylation of HMGS-1 regulates metabolism during aging

163

Chapter 5 General discussion and concluding remarks

Thijs Koorman

166

General discussion and concluding remarks

ell polarity is an essential property of many cells, contributing to cell-fate specification, differentiation, signal transduction, asymmetric cell division, and cell migration both during development and in adult life. In the past decades, genetic and biochemical studies in a variety of model systems uncovered a series of proteins involved in the establishment and maintenance of polarity1. Nevertheless, it remains unclear how the unequal distribution of molecules within a cell is arranged. Moreover, the roles of known polarity regulators differ depending on the cellular and developmental context, making the study of mechanisms underlying cell polarity even more complicated. Three key modules in cell polarity, described in Chapter 1, are the conserved PAR-complex, the Crumbs-complex, and the Scribble group of proteins. Kemphues and colleagues identified and characterized Partitioning defective mutants of Caenorhabditis elegans (C. elegans)2. Mutations in the par genes resulted in defects in the symmetric positioning of the mitotic spindle and resulted in an equal distribution of cell fate-determinants. Forward genetic screens in Drosophila uncovered proteins of the Crumbs and Scribble groups, which play a major role in polarity in epithelial cells3. The Crumbs-complex promotes the establishment of an apical membrane identity, whereas the Scribble group of proteins is essential for establishing basolateral identity. In our studies, we aimed to study the network of molecular interactions that underlies the establishment of cell polarity. We combined a fragment-based high-throughput yeast two-hybrid (Y2H) screen with in vivo studies using existing models of polarity in C. elegans (described in Chapter 2). The Y2H system has proven to be a powerful system to identify protein interactions, at both small and large scale4. One of the reasons why the Y2H system is such an attractive technique is that screens can be performed relatively quickly and inexpensively. When prey libraries are available, the cloning of bait proteins and culturing of yeast strains is the most time consuming step. Much effort has been put into optimization of the system such that screens can be performed with high quality and with high throughput5,6,7. A common source of false positives that used to be present in screens is the auto-activation of bait proteins. Auto-activating proteins are capable of activating the reporter genes without the presence of an interaction. Auto-activating bait proteins can be identified using a counter-selection step that eliminates the prey protein, based on sensitivity to cycloheximide resulting from expression of the CYH2 gene present on the prey plasmid8. Another method to eliminate false-positive protein interactions is the retesting of the interaction by transforming the prey encoding plasmid into a fresh bait yeast population. This eliminates false-positives due to mutations in the bait plasmid or yeast that may occur during the screening procedure and allow the yeast to grow in the absence of an interaction. When identified, false positives should be removed from the network. 167

A major recent improvement which we applied to all our Y2H screens is the use of multiple fragments per full length protein. Full length proteins may sometimes not be capable of interacting in yeast because the proteins can e.g. be folded in a closed conformation. The use of protein fragments can circumvent this problem, increasing the chance of identifying interactors for a particular protein. Furthermore, the use of different fragments enables the identification of the region responsible for an interaction. This approach was successfully demonstrated using a C. elegans prey library, containing fragments and full length proteins of 750 genes that are essential in the first two embryonic divisions9. We used this approach in every Chapter of this thesis. In Chapter 2, we established a C. elegans polarity protein interaction network (CePIN) with Y2H screens using 69 polarity regulators as baits. A specific improvement we made in these screens was the use of fragments on the bait side. In total, 65 full length proteins and 375 protein fragments were used as baits. The design of the fragments was based on gene length and protein domain composition, and aimed to keep known protein interaction domains intact. All full length and fragment baits were screened against two C. elegans prey libraries: a cDNA library designed to contain mostly full-length genes and the C. elegans embryonic fragment library described above. Screens using full-length bait proteins resulted in the identification of 80 protein-protein interactions. Through screens using protein fragments as baits, we identified an additional 359 interactions. This result demonstrates that the use of fragments increases the amount of interactions per full length bait protein. The interactions found were validated with both bio-informatic and experimental approaches. The use of protein fragments for both bait and prey fusions allowed us to examine the minimal regions of interaction (MRIs) between identified proteins. The MRI is defined as the smallest region shared by all fragments of a protein that are able to interact. To evaluate the accuracy of the identified MRIs, we compared the MRIs identified with the literature and found 29 interactions from our list for which a binding site was previously described for either C. elegans or homologous proteins. In all but 2 cases the MRI we identified matched the literature-described domain, validating the accuracy of Y2H-based interaction domain mapping. The novel MRIs identified by Y2H may provide a starting point for domain based analysis of protein interactions. Bio-informatic tools and literature enquiries can provide an indication that an interaction network is of high quality. Methods generally used is the analysis of GeneOntology terms as well as enrichment of similar mRNA expression profiles during development. The CePIN was enriched for similar GO terms, though enrichment scores identified with the AD-fragment library were lower probably due to intrinsic

168

General discussion and concluding remarks

functional similarity between early embryogenesis genes. Interactions identified from the AD-cDNA library screens were enriched for similar mRNA expression profiles. Early embryogenesis genes in the AD-fragment library are already similar in expression profiles which makes further enrichment hard to observe9. Literature curation of the CePIN showed that 53 molecular interactions were previously described (18 in pubmed, and 35 solely in Intact). With the above described methods we confirm the quality of the identified interactions and demonstrate the ability to identify biologically relevant interactions. Experimental procedures such as an orthogonal affinity purification assay of interaction proteins in mammalian cell culture or the use of MAPPIT (reconstituting a membrane-bound receptor complex in mammalian cells) can further increase the confidence in the quality of an interaction network. Such approaches generally reproduce 20-30% of the interactions10,11. We tested 33 interactions with affinity purification from transfected HEK293T tissue-culture cells and 16 out of 33 (48%) interactions tested positive. In this procedure we attempted to not only purify the full length prey protein but also the identified fragment. Using protein fragments, we were able to recover a greater number of interactions by co-affinity purification than by using full-length proteins alone (11 fragments vs. 5 full-length proteins of the 33 interactions). Thus, the use of fragments in addition to full-length clones provides an advantage in co-affinity purification. Since the fragment based approach worked well for C. elegans we expanded this to human genes in Chapter 3. By randomly fragmenting DNA encoding full length open reading frames (ORF) present in the human ORFeome 5.1 collection, we generated a human Y2H fragment library for 15,483 genes12. We tested a subset of baits involved in cell division and polarity establishment and were able to identify both known and unknown protein interactions and interaction domains. To assess the quality of the interactions identified, we validated all identified interactions by co-affinity purification from transfected tissue-culture cells. Of the 22 protein interactions identified, we were able to reproduce 12 (55%) interactions. Of the 12 interactions, 11 were detectable when using the prey fragment identified earlier in the Y2H screens. This validates the interaction domains identified by Y2H. While a physical interaction in yeast indicates a possible functional role in vivo, not all interactions identified are biologically relevant. The establishment of genome-wide RNAi libraries for C. elegans provides a tool to identify functional interactions15. Although a disadvantage of RNAi is the incompleteness of inactivation and variability throughout experiments, a robust interpretation of an RNAi phenotype can be achieved by examining a population of worms18. Integrating physical interaction networks with phenotypic data allows for the identification of

169

interactions that are likely to be relevant in vivo16. Therefore, we performed RNAi profiling of the polarity interaction network in Chapter 2. The use of multiple fluorescent marker lines increases the chance to find a similar phenotype following RNAi of the individual genes that encode two interacting proteins. We started with analysis of the 69 polarity regulators that were used as baits, by feeding RNAi of the marker strains. In total, we observed 44 defects and grouped the phenotypes based on the severity of the effect. To determine whether the phenotypes observed give an indication of the functional relationship between the two interactors, we performed unbiased hierarchical clustering of the genes based on the phenotypes observed. Indeed, genes which were known to function together clustered, such as the genes of the PAR complex. The accurate grouping of genes validates the quality of our RNAi screens. Combining Y2H screens with phenotypic analysis by RNAi profiling is useful for tissues that were previously not extensively studied such as axon versus dendrite specification in the PVD-neuron and the morphology of the excretory canal. Remarkably, inactivation of lgl-1 led to mislocalization of axon specific endosomes into the dendrites. This represents the first identification of a critical function for lgl-1 in C. elegans somatic cells as well as the first identification of LGL in neuronal polarity in the PVD neuron. Another striking observation was that the inactivation of par-2 and par-4 led to irregular excretory canal morphology, while other PAR components did not show this defect. For 100 protein pairs, RNAi of the genes encoding the interacting proteins resulted in a defect in the same marker strains. In some cases, the phenotypes of both genes were quite similar, giving a strong indication that the protein interaction is biologically relevant. In many cases however, while RNAi for both genes caused a phenotype in the same marker strain, the phenotypes were quite dissimilar. This was not entirely unexpected, since we aimed at identifying interactions that link cell polarity to other cellular processes. In such cases, it is not likely that RNAi of the identified prey protein results in an identical and specific polarity defect, and a more detailed analysis of the phenotype will be required to find the process in which the two phenotypes overlap14. One such interaction was investigated in more detail in Chapter 2, the interaction between the Rho GTPase activating protein PAC-1 and the polarity regulator PAR6. PAC-1 (PAR-6 at Cell-contacts 1) was identified in a genetic screen for mutations that cause a defect in PAR-6 radial polarization in the four-cell stage16. In wild-type embryos, the active Rho GTPase CDC-42 binds to PAR-6[ref. 17,18]. PAC-1 localizes basolaterally, thereby locally inactivating CDC-42 and restricting recruitment of PAR-6 by CDC-42 to apical contact-free cell surfaces.

170

General discussion and concluding remarks

Although PAC-1 and PAR-6 do not appear to co-localize, our data suggest that interaction between these proteins is essential for their function in vivo. We validated the interaction by affinity purification from transfected tissue culture cells, and showed that the proteins co-localize in the same punctated pattern in HeLa cells. To test whether the interaction had any function in vivo we attempted to rescue the PAC-1 phenotype by expressing a mutant PAC-1(ΔPBD) which lacks the PAR6 binding domain. However, GFP-PAC-1(ΔPBD) failed to restore the localization pattern of PAR-6. Thus, the direct interaction between PAC-1 and PAR-6 is required to establish radial polarity in the four cell stage. Earlier work showed that two RhoGEFs ECT-2 and CGEF-1 compete with PAC-1, and that the competition between these proteins is required to inactivate CDC-42 at cell contact sites16. In addition, the N-terminal region of ECT-2 binds preferentially to CDC-42 in vivo and not to RHO-1[ref. 18]. The N-terminal PAR-6 interaction domain of CDC-42 differs when compared with PAR-6 binding to PAC-1, which is via the PDZ domain of PAR-6 and an internal PAC-1 PDZ-motif. The primary localization of PAR-6 before radial polarization might be basolateral, while PAC-1 locally inactivates CDC-42 and thereby triggers the relocalization of PAR-6 to contact free cell surfaces. Why PAR-6 binding to PAC-1 is required for exclusion of PAR-6 from the inner membranes is unclear. One possibility is that PAC-1 only inactivates CDC42 in a complex together with PAR-6. In this model, PAR-6 functions as a scaffold, bringing together CDC-42 and PAC-1. Further studies are needed to elude whether how PAC-1, CDC-42 and PAR-6 are linked in the establishment of radial polarity.

Concluding remarks In this thesis, we aimed to find interactions of proteins involved in the control of cell polarity, by using a combinatorial approach of fragment-based high-throughput protein identification by Y2H, and RNAi analysis using C. elegans polarity marker strains. As addressed in several chapters, the ability to identify protein interactions with the Y2H system increases several-fold by using protein fragments. Furthermore, the use of protein fragments allows the identification of minimal interaction domains, and therefore may help to unravel molecular mechanisms. The phenotypic analysis of the polarity interactome we present is the beginning of functionally clarifying the network, with PAC-1 and PAR-6 as a prime example. It will be interesting to investigate the identified interactions in different model systems and tissues to identify additional interactions that affect specific aspects of cell polarity. 

171

References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18.

172

Rodriguez-Boulan, E. & Macara, I. G. Organization and execution of the epithelial polarity programme. Nat. Rev. Mol. Cell Biol. 15, 225–42 (2014). Kemphues, K. J., Priess, J. R., Morton, D. G. & Cheng, N. S. Identification of genes required for cytoplasmic localization in early C. elegans embryos. Cell 52, 311–320 (1988). Bilder, D., Li, M. & Perrimon, N. Cooperative regulation of cell polarity and growth by Drosophila tumor suppressors. Science 289, 113–116 (2000). Vidal, M., Cusick, M. E. & Barabási, A.-L. Interactome networks and human disease. Cell 144, 986– 98 (2011). Caufield, J. H., Sakhawalkar, N. & Uetz, P. A comparison and optimization of yeast two-hybrid systems. Methods San Diego Calif 58, 317–24 (2012). Venkatesan, K. et al. An empirical framework for binary interactome mapping. 6, 83–90 (2009). Rolland, T. et al. A Proteome-Scale Map of the Human Interactome Network. Cell 159, 1212–1226 (2014). Vidalain, P.-O., Boxem, M., Ge, H., Li, S. & Vidal, M. Increasing specificity in high-throughput yeast two-hybrid experiments. Methods San Diego Calif 32, 363–370 (2004). Boxem, M. et al. A protein domain-based interactome network for C. elegans early embryogenesis. Cell 134, 534–45 (2008). Braun, P. et al. An experimentally derived confidence score for binary protein-protein interactions. 6, 91–97 (2009). Lievens, S. et al. Array MAPPIT: high-throughput interactome analysis in mammalian cells. J. Proteome Res. 8, 877–886 (2009). Lamesch, P. et al. hORFeome v3.1: a resource of human open reading frames representing over 10,000 human genes. Genomics 89, 307–315 (2007). Rual, J.-F. et al. Toward improving Caenorhabditis elegans phenome mapping with an ORFeomebased RNAi library. Genome Res. 14, 2162–2168 (2004). Walhout, A. J. M. et al. Integrating interactome, phenome, and transcriptome mapping data for the C. elegans germline. Curr. Biol. CB 12, 1952–1958 (2002). Wang, Z. & Sherwood, D. R. Dissection of genetic pathways in C. elegans. Methods Cell Biol. 106, 113–157 (2011). Anderson, D. C., Gill, J. S., Cinalli, R. M. & Nance, J. Polarization of the C. elegans embryo by RhoGAP-mediated exclusion of PAR-6 from cell contacts. Science 320, 1771–4 (2008). Aceto, D., Beers, M. & Kemphues, K. J. Interaction of PAR-6 with CDC-42 is required for maintenance but not establishment of PAR asymmetry in C. elegans. Dev. Biol. 299, 386–397 (2006). Chan, E. & Nance, J. Mechanisms of CDC-42 activation during contact-induced cell polarization. J. Cell Sci. 126, 1692–702 (2013).

General discussion and concluding remarks

173

Addendum Summary Nederlandse samenvatting voor niet-ingewijden Curriculum vitae; in het Engels en Nederlands Scientific Publications, talks, fellowships and posters Dankwoord

Mens, wat ben je mooi als je doorgaat, tegen de tegenslag jezelf verslaat. En daar dan oprecht zelf verstelt van staat. Mens, wat ben je mooi als je de onmacht eventjes ontkracht.

176

Summary

Summary Networking for proteins A yeast two-hybrid and RNAi approach to uncover C. elegans cell polarity regulators

ell polarity is a near universal trait of life and guides many aspects of animal development: from the creation of cells with different developmental potential through asymmetric divisions, to proper functioning and morphogenesis of multicellular tissues and organisms. Although a number of key polarity proteins have been identified (summarized and discussed in chapter 1), many interactions with proteins acting downstream likely remain to be elucidated. In addition, there is extensive crosstalk between different polarity pathways, but the mechanistic details underlying this crosstalk are not well known. Mutations in polarity proteins or deregulation of polarity pathways are causative for various diseases including cancer. To identify new proteins or new links between polarity regulators, we combined a fragment based high-throughput yeast two-hybrid approach with in-vivo RNAi studies using existing models of polarity in the small nematode Caenorhabditis elegans, which I describe in chapter 2. We started with 69 polarity regulators and designed an additional 375 protein fragments, which we designed based on gene size and protein domain composition. In total we identified 459 protein interactions between 296 proteins, which greatly expand the polarity interaction network. Computational and experimental approaches were used to validate the quality of our network. The 69 polarity regulators were characterized in several C. elegans polarity marker strains by RNAi. Furthermore, we analysed the interacting preys identified in the Y2H screens in C. elegans strains in which the original polarity regulator exhibited a phenotype. This approach led to 100 protein pairs that physically interact and for which RNAi of the corresponding genes causes a defect in the same polarity-related process. One interaction was investigated in more detail, the interaction between the RhoGAP protein PAC-1 and the polarity protein PAR-6. Our experiments show that PAC-1 lacking the PAR-6 binding domain expressed in a pac-1 mutant is unable to rescue the PAR-6 localization defect, even though mutant PAC-1 localizes as wildtype. The network of candidate polarity-related interactions we identified will be a valuable resource for the research community. In chapter 3 we expand the concept of Y2H fragment based interaction mapping to a human genome wide scale. Protein function is mostly dictated by its structure and by its interactions. Full-length proteins are often not capable to perform in screening methods because of e.g. miss-folding, the active domain being masked or

177

because of lack of post-translational modifications. We generated a fragment based human Y2H prey library by mechanically fragmenting the ORFeome 5.1 library. The new library was experimentally validated by screening 44 bait proteins involved in cell polarity and cell division with well described interactions. We identified 22 protein interactions of which 8 were described previously. Furthermore, we are able to identify new minimal interaction domains. These 22 interactions were validated by protein affinity purification and 12 (55%) tested positive. Moreover, we validated the newly identified minimal interaction domains for these 12 protein pairs by affinity purification. In chapter 4 we collaborated with the Broday and Sternberg laboratoryâ&#x20AC;&#x2122;s and performed Y2H screens to identity novel interactions. We identified the protein HMGS-1 (HMG-CoA synthase), the key regulator of the Mevalonate pathway, interacting with the small ubiquitin modifier protease ULP-4. HMGS-1 is the first step of the Mevalonate pathway which facilitates the syntheses of steroid hormones and cholesterol precursors. Perturbations of this pathway are causative to cardiovascular diseases and cancer. During the lifespan of C. elegans the expression and localisation pattern of ULP-4 changes from cytoplasmic in young animals and localizing to mitochondria in old animals. When ULP-4 is in the cytoplasm, the protein sumoylates HMGS-1, by which metabolism of the worm increases. When worms age, ULP-4 is relocated to the mitochondria, upon which HMGS-1 will be desumoylated and degraded by the proteasome, decreasing metabolism specifically altering the Mevalonate pathway. This conserved regulatory rewiring mechanism may be a reference for future therapies. In summary, we used the yeast two-hybrid technique to generate a comprehensive map of polarity-related interactions, which we further analysed and clarified in vivo in C. elegans. We expanded the use of fragment-based yeast two-hybrid mapping to human proteins, and identified a novel protein interaction between HMGS-1 and ULP-4. ď&#x201A;ž

178

Nederlandse samenvatting voor niet-ingewijden

Nederlandse samenvatting voor niet-ingewijden Netwerken voor eiwitten Een yeast two-hybrid en RNAi methode om C. elegans cel polariteit regulatoren te ontdekken

e cel is de kleinste functionele eenheid van het levend wezen. Ter illustratie: bacteriën bestaan uit één enkele cel, de worm waarmee wij werken ‒ Caenorhabditis elegans ‒ heeft 958 cellen en de mens heeft er miljarden. Cellen zijn dus de bouwstenen van het leven, van ons. Wat er in ons lichaam op grote schaal gebeurt ‒ organen die samenwerken ‒ vindt ook in de cel plaats, maar dan in het klein. Iedere cel bevat een kern met DNA, waarin een compleet blauwdruk voor de bouw en de ontwikkeling van het hele lichaam is opgeslagen. DNA is een lange keten van vier verschillende moleculen, aangeduid met A, T, G en C. Een bepaalde volgorde van deze letters vormen de code voor een eiwit. Eiwitten zijn de bouwstoffen en werkpaarden van de cel, en werken alleen of samengebonden. Vergelijk het DNA met een cassetteband die afspeelt. Maar in plaats van muziek worden er eiwitten geproduceerd. Daarbij heeft een cel net als wij ‘organen’ die het nodig heeft om te functioneren. Een orgaan is bijvoorbeeld het mitochondrium, vergelijkbaar met een lever, dat dient als energieleverancier en ervoor zorgt dat de cellulaire motor blijft draaien. Een cel heeft tevens fabrieken waarin de eiwitten uit het DNA worden bewerkt, zodat ze hun functie kunnen uitvoeren. Deze worden het endoplasmatisch reticulum en het golgi-apparaat genoemd. Nog veel indrukwekkender: een cel heeft zelfs een skelet dat tevens dient als een transportsysteem, met verschillende ‘banen’ genaamd het cytoskelet. Deze kun je in grote schaal vergelijken met het openbaar vervoer en de snelweg. Zoals mensen met elkaar communiceren, zo kunnen eiwitten en cellen dat ook. Dit wordt met een moeilijke term signaaltransductie genoemd. In de cel vinden honderden, zo niet duizenden, processen plaats die er samen voor zorgen dat een cel en zo ook een levend wezen gezond blijft. Gelijke en ongelijke delingen van cellen Laten we uitzoomen. Een groepje cellen noemen we een weefsel. Dit kan bijvoorbeeld spierweefsel of zenuwweefsel zijn. Gaan we een stapje verder, dan komen we bij organen zoals de darmen of de lever. Nog verder uitgezoomd komen we tot systemen, zoals het centraal zenuwstelsel of het ademhalingsstelsel. Nu vraagt u zich misschien af hoe de diversiteit tussen cellen ontstaat. Hoe weet een darmcel dat het een darmcel moet zijn en een spiercel dat het een spiercel is? De grondslag hiervan ligt bij celdeling of celvermenigvuldiging. Een cel kan namelijk gelijk en ongelijk delen. Bij gelijke delingen ontstaan twee dezelfde cellen, en neemt

179

alleen het aantal toe. Met ongelijke delingen ontstaan twee verschillende cellen, waarbij de diversiteit toeneemt. Vergelijk het delen van cellen als het snijden van een taart. Als u een taart in twee gelijke stukken moet delen door één keer te snijden, hoe zou u dat doen? Gelijke stukken krijgen we als we de taart in het midden verticaal doorsnijden. Maar als ik de taart horizontaal in het midden doorsnijd, dan zit de slagroom boven en de cake onder. Ik heb de taart dus in twee ‘ongelijke’ stukken verdeeld. De taart illustreert nog een eigenschap die veel cellen hebben: polariteit. De taart is van boven anders dan aan de onderkant. Daarom kan hij in twee verschillende helften worden gesneden. Ook cellen hebben vaak verschillende helften. Bij het delen (doormidden snijden) krijgt de ene afstammeling een andere set eiwitten dan de andere. We spreken dan van een ongelijke deling. Voornamelijk stamcellen hebben deze eigenschap: de moederstamcel verdeelt haar componenten ongelijk of anders gezegd asymmetrisch. Het gevolg is dat de moederstamcel een stamcel blijft en de dochtercel zich kan specialiseren tot bijvoorbeeld een darmcel. Een mechanisme dat asymmetrische delingen kan reguleren noemen we polariteit. Maar, polariteit is niet alleen belangrijk bij asymmetrische delingen, want veel cellen bevatten verschillende kanten met verschillende functies. Hoe weet een darmcel wat de boven- en onderkant is? En hoe weet een huidcel wat de boven- en onderkant is? Een cel heeft dus een richtingsgevoel, dat heet ook polariteit. Een ziekte waarbij cellen geen deling of richtingsgevoel meer hebben kennen we allemaal: kanker. Kankercellen delen ongecontroleerd, gelijk en ongelijk, en weten niet goed wat de boven- en onderkant is. Cellulaire polariteit Mijn promotieonderzoek stond grotendeels in het teken van de (on)gelijke deling en het richtingsgevoel van cellen, polariteit. In het bewerkstelligen van cellulaire polariteit spelen eiwitten een rol. Specifieke eiwitten kunnen een zijde van de cel markeren, waardoor deze zich op de juiste manier kan ontwikkelen. Zo moeten darmhaartjes in de darm alleen aan de bovenkant zitten om voedingsstoffen op te nemen. Maar kennen wij alle eiwitten die dit kunnen? Nee. De afgelopen decennia is hier veel onderzoek naar verricht. Door experimenten in de fruitvlieg en in de worm zijn de belangrijkste eiwitten echter ontdekt. Hoe? Door de genen ‘aan’ en ‘uit’ te zetten wordt inzichtelijk welke eiwitten voor deze verkeerde delingen zorgen. Het weghalen van deze belangrijke eiwitten heeft vaak verkeerde delingen met uiteindelijk kanker tot gevolg. In hoofdstuk 1 worden de specifieke eiwitten uitgebreid beschreven. Deze polariteitseiwitten beïnvloeden bovendien een groot aantal processen binnen de cel, bijvoorbeeld de organisatie van het transportsysteem van eiwitten of van het cytoskelet. Hoe dit exact in zijn werking gaat is tot vandaag de dag nog niet bekend. 180

Nederlandse samenvatting voor niet-ingewijden

Wegenkaart In ons onderzoek zoeken we naar andere eiwitten waarmee de al bekende polariteitseiwitten samenwerken. Hiermee maken we een ‘wegenkaart’ met routes die deze eiwitten afleggen: hoe polariseert hij, hoe onderhoudt een cel dit en hoe controleert hij gelijke en ongelijke delingen. Al deze gegevens staan beschreven in hoofdstuk 2. We onderzoeken polariteit in de eerder genoemde worm Caenorhabditis elegans (C. elegans). Dit diertje heeft weinig geheimen voor ons. Het DNA is bekend, we kennen alle genen en ook de meeste eiwitten in de worm. Het diertje is transparant en daardoor goed met de microscoop te bekijken. Ten slotte is het goed te houden in een laboratorium. Als u nou nog niet overtuigd bent hoe wetenschappelijk mooi de worm eigenlijk is, moet u mij maar even aan de jas trekken. Wat we niet weten is hoe alle eiwitten samenwerken en welke functie ze in welke organen van de worm hebben. Dit weten we tevens niet over de polariteitseiwitten. Naast de worm hebben we gist gebruikt. Gist is een eencellig organisme, net als een bacterie, maar is wel een stuk complexer. We hebben gist gebruikt om C. elegans eiwitverbindingen (interacties) te vinden. Zoals gezegd kunnen eiwitten aan elkaar binden, maar zijn niet alle eiwitinteracties bekend. Voor ons onderzoek hebben we 69 bekende C. elegans polariteitseiwitten uitgekozen. We hebben deze elk in verschillende stukken verdeeld en zijn daarmee op grote schaal op zoek gegaan naar interacties met andere eiwitten. In totaal hebben we er 459 gevonden. Nu we wisten welke eiwitten met elkaar interacteren, zijn we op zoek gegaan in C. elegans. De coderende genen van de polariteiteiwitten hebben we uitgeschakeld en de gevolgen daarvan in verschillende gepolariseerde weefsels en organen van de worm geanalyseerd. Ook van defecte eiwitten hebben we de interacties één voor één geremd en deze geanalyseerd. Door het onderzoek naar deze nieuwe eiwitten heeft de lijst met interacties een extra dimensie gekregen. Vervolgens besloten we om één specifieke eiwit interactie verder uit te diepen. Hiermee hebben we een nog onbeschreven nieuwe rol van polariteit laten zien tussen bekende eiwitten. Deze studie is ondertussen bijna klaar, we zijn bezig om het te publiceren in een vooraanstaand wetenschappelijk tijdschrift. Een nieuwe eiwitbibliotheek In hoofdstuk 3 hebben we een soortgelijke studie gedaan, maar met een nieuwe dimensie: een vertaalslag naar de mens, met eiwitfragmenten van het complete DNA. Om het functioneren van een eiwit beter te begrijpen is het belangrijk om te achterhalen welke fragmenten of domeinen binnen een eiwit deze functies uitvoeren. Dit bepaalt namelijk vaak de functie van een eiwit en verschilt per type cel, of zelfs binnen bepaalde regio’s in de cel. Om snel door functies te kunnen ‘bladeren’

181

hebben we een bibliotheek van genfragmenten gegenereerd waarin verschillende eiwitdomeinen zijn opgeslagen. Deze hebben wij in gist geplaatst en vervolgens getest op interacties met bekende humane polariteits- en celdelingseiwitten. We hebben 22 eiwitinteracties gevonden, waarvan acht al eerder beschreven en veertien nieuwe. Deze hebben we uitgebreid gecontroleerd. Uit ons onderzoek blijkt dat het middels onze nieuwe eiwitbibliotheek mogelijk is om eiwitinteracties en eiwitdomeinen te identificeren. Doordat deze methode succesvol werkt hebben we deze kunnen publiceren in het wetenschappelijk tijdschrift Journal of Proteome Research. Expertise Vanwege onze techniek zijn van een van de experts op gebied van eiwitinteracties. Het komt dan ook regelmatig voor dat andere wetenschappers in het C. elegans-veld ons vragen om samen experimenten uit te voeren. Internationale samenwerkingen met multidisciplinaire groepen staan voor ons hoog in het vaandel. In hoofdstuk 4 staat beschreven hoe wij met het eiwit ULP-4 hebben gezocht zoeken naar nieuwe eiwitinteracties. ULP-4 is een eiwit dat bindt aan andere eiwitten, om ze een andere functie mee te geven. Wij vonden het eiwit HMGS-1. Dit eiwit functioneert in de Mevalonaat-route, waar cholesterol en hormonen worden geproduceerd die er onder andere voor zorgen dat cellen kunnen groeien. Als de humane versie van HMGS-1 niet wordt gereguleerd door andere eiwitten kan het uiteindelijk kanker tot gevolg hebben. Wat dit betekent voor de worm was nog niet eerder bekend. Samen met de collega’s uit het andere lab laten we zien dat ULP-4 HMGS-1 kan reguleren afhankelijk van de leeftijd van de worm. Bij jonge wormen bindt ULP-4 aan HMGS1 in het cytoplasma van de cel, wat ervoor zorgt dat de mevalonaat-route ‘aan’ staat. Als de worm ouder wordt en niet meer hoeft te groeien, verplaatst ULP-4 naar het mitochondrium waardoor HMGS-1 wordt afgebroken en de Mevalonaat-route minder actief wordt. Deze modificatie van HMGS-1 door ULP-4 blijkt geconserveerd naar de mens, waar mee het op lange termijn mogelijk is een therapie te ontwikkelen voor het remmen van de Mevalonaat-route. Deze bevindingen zijn gepubliceerd in het wetenschappelijke tijdschrift The Proceedings of the National Academy of Sciences, wij noemen het makkelijk PNAS. In het kort: in dit proefschrift beschrijven we een methode om op grote schaal eiwitinteracties te vinden. We hebben een netwerk van polariteit eiwitinteracties gegenereerd en deze geanalyseerd in verschillende gepolariseerde weefsels in C. elegans. Hiermee hebben wij een naslagwerk ‒ een bibliotheek ‒ gecreëerd dat voor het onderzoeksveld een kostbare bron voor vele jaren studie is. 

182

Curriculum vitae and publications

Curriculum Vitae Curriculum vitae Thijs Koorman wasThijs bornKoorman on June was 29th born 1986 on in Dedemsvaart, Netherlands. June 29th 1986 The in Dedemsvaart, The Netherlands. He attended in 1998 the ‘Middelbare

He attended in 1998 the De ‘Middelbare De Driemark’ in Winterswijk. He school Driemark’school in Winterswijk. He passed his MAVO passed his MAVO exam exam in in 2002 2002 and andstarted startedhis hismbo/mlo mbo/mlostudies studiesCytoCyto-Histo Histo and Pathology at the Rijn IJssel college in Arnhem. During

and Pathology at this thestudy Rijn he IJssel college two in Arnhem. During the thisfirst study he performed rotation projects, in the Pathology department of the Dutch Cancer Center Antonie

performed two rotation projects, the under first insupervision the Pathology of the the van Leeuwenhoek of P.department Wisman and second in the department the Academic Dutch Cancer Center Antonie vanPathology Leeuwenhoek under of supervision of P.Hospital Wisman of Maastricht under supervision of R. Tonk and R. Büchholz.

and the second inDuring the Pathology department of theinAcademic of both rotations he was trained techniquesHospital related to

cytology, histology and pathology. He passed his exam in 2006. In the same year Thijs

Maastricht under supervision of R. Tonk and R. Büchholz. Duringvan both rotations was trained in started his hbo/hlo studies at the Hogeschool Arnhem and he Nijmegen (HAN) on Biochemistry Signal Transduction. performed rotations, techniques relatedfocussing to cytology, histology andand pathology. He passed hisHe exam in 2006.two In the same the first in the research group of prof. G. Pruyn, PhD, under supervision of A. Lokate,

PhD, in the studies Radboud for Molecular Life Sciences (RIMLS),(HAN) working on the year Thijs started his hbo/hlo atInstitute the Hogeschool van Arnhem and Nijmegen focussing auto-immune response caused by Scleroderma disease. His next rotation was in the

on Biochemistry and Signal Transduction. rotations, the first the research group research group of prof.He H.performed Bos, PhD, two under supervision of F.inZwartkruis, PhD, at the Molecular Cancer Research University Medical Centre. Thijs of prof. PhD. G. Pruyn, under supervision of PhD. in A. Utrecht Lokate in the Radboud Institute forHere, Molecular worked on the small GTPase Rheb and generated a protein purification technique to

purify Exchange factors from small GTPases. studies, Thijs followed Life Sciences (RIMLS), working on the auto-immune responseDuring causedhis byhbo Scleroderma disease. His Bio-medical courses at the Vrije University in Amsterdam. He passed his hbo/hlo

next rotation was inexam the research group ofThijs prof.dr. H. Bos supervision of ‘Cancer dr. F. Zwartkruis atand the in 2009. In 2009, started hisunder bio-medical Master Genomics

Developmental Biology’ at Utrecht University, he completed first rotation Molecular Cancer Research in Utrecht University Medical Centre. Here, Thijshis worked on theproject small

in the research group of prof. H. Bos, PhD, under supervision of F. Zwartkruis, PhD, investigating the exchange factor technique PDZ-GEF. His secondExchange intershipfactors was performed in GTPase Rheb and generated a protein purification to purify from small the group of prof. Sander van den Heuvel, PhD, under supervision of M. Wildwater, GTPases. During his hbo studies, Thijs followed Bio-medical courses at the Vrije University in PhD, at the Developmental Biology department of Utrecht University. During this internship Thijs got familiar with C. elegans and investigated the tumour suppressor Amsterdam. He passed his hbo/hlo exam in 2009. In 2009, Thijs started his bio-medical Master protein Discs-Large. Before passing his exam in 2011, he his master thesis under supervision of P. vanBiology’ der Sluijs, on signalling thecompleted neuronal presynaptic cleft. ‘Cancer Genomics and Developmental at PhD, Utrecht University,inhe his first rotation In 2011 Thijs started working in the research group of Mike Boxem, PhD, in the project in the research group of prof.dr. Bos underofsupervision of dr. F.on Zwartkruis Developmental BiologyH. department Utrecht University the work investigating described in this thesis. Thijs presented parts of his work at national and international meetings the exchange factor PDZ-GEF. His second intership was performed in the group of prof.dr. Sander van and was awarded with several fellowships during his PhD training. Thijs had a board at ‘Vereniging Zonnebloem’ which he organized trips and outings. den Heuvel under position supervision of dr. M. de Wildwater at theatDevelopmental Biology department of He was actively bound to the University of Applied sciences Nijmegen as a guest Utrecht University.lecturer Duringand thisasinternship familiar with C. elegans investigated a memberThijs of thegot graduation committee. After and obtaining his PhDthe he was invited to join the University of Applied sciences for a short period as lecturer/ tumour suppressor protein Discs-Large. Before passing his exam in 2011, he his master thesis under researcher and took the opportunity to develop his soft skills extensively. Thijs will scientific career in abroad as a postdoc in the lab of prof. A. Thijs McClatchey, supervision of dr. P.continue van der his Sluijs on signalling the neuronal presynaptic cleft. In 2011 started PhD, at the Massachusetts General Hospital, Havard Medical School in Boston USA.

working in the research group of dr. Mike Boxem in the Developmental Biology department of Utrecht University on the work described in this thesis. Thijs presented parts of his work at national 183

and international meetings and was awarded with several fellowships during his PhD training. Thijs had a board position at ‘Vereniging de Zonnebloem’ at which he organized trips and outings. He was actively bound to the University of Applied sciences Nijmegen as a guest lecturer and as a member of

Curriculum vitae in het Nederlands Thijs Koorman werd geboren op 29 juni 1986 te Dedemsvaart. In 1998 ging hij naar de middelbare school De Driemark in Winterswijk. Het mavo-examen werd in 2006 afgerond, waarna hij in hetzelfde jaar begon met de mbo/mlo-opleiding CytoHisto- en Pathologie aan het Rijn IJssel college te Arnhem. Tijdens deze studie heeft Thijs twee stages gevolgd. De eerste op de Pathologie-afdeling van het Nederlands Kanker Instituut Antoni van Leeuwenhoek onder begeleiding van P. Wisman en de tweede stage op de Pathologie-afdeling in het Academisch Ziekenhuis Maastricht onder begeleiding van R. Tonk en R. Büchholz. Tijdens beide stages heeft Thijs ervaring opgedaan in technieken gerelateerd aan cytologie, histologie en pathologie. Zijn diploma haalde hij in 2006. Datzelfde jaar begon hij met zijn studie aan het hbo/ hlo aan de Hogeschool van Arnhem en Nijmegen met een focus op biochemie en signaaltransductie. Tijdens deze periode heeft Thijs twee stages gevolgd; de eerste in de onderzoeksgroep van prof.dr. G. Pruyn onder begeleiding van dr. A. Lokate in het Radboud Institute for Molecular Life Sciences (RIMLS). Tijdens deze stage heeft hij de auto-immuunziekte Scleroderma bestudeerd. Zijn tweede stage was in de onderzoeksgroep van prof.dr. H. Bos in het Universitair Medisch Centrum Utrecht (UMCU), onder begeleiding van dr. F. Zwartkruis. Hier heeft Thijs gewerkt aan een eiwitzuiveringsmethode om de exchangefactor van de kleine GTPase Rheb te identificeren. Tijdens het hbo/hlo heeft hij vrijekeuzevakken gevolgd op de Vrije Universiteit van Amsterdam uit de bachelor Biomedische Wetenschappen. Thijs haalde zijn hbo/hlo-diploma in 2009. Vervolgens is Thijs gestart met de universitaire masteropleiding Cancer Genomics and Developmental Biology aan de Universiteit Utrecht. Wederom zijn er twee onderzoeksstages gevolgd; de eerste in de onderzoeksgroep van prof.dr. H. Bos in het UMCU onder begeleiding van dr. F. Zwartkruis, waarin het moleculaire mechanisme van de exchangefactor PDZ-GEF werd onderzocht. Zijn tweede onderzoeksstage werd uitgevoerd in de groep van prof.dr. S. van den Heuvel aan de Universiteit Utrecht onder begeleiding van dr. M. Wildwater. Hier heeft Thijs zich bekwaamd in C. elegans en het tumorsupressoreiwit Discs-Large in C. elegans onderzocht. Thijs heeft zijn scriptie ‒ over signalering in de presynaptische ruimte van zenuwcellen ‒ geschreven onder begeleiding van dr. P. van der Sluijs. In 2011 is Thijs begonnen met zijn promotietraject in de onderzoeksgroep van dr. Mike Boxem op de Universiteit Utrecht en daar het in dit proefschrift beschreven werk uitgevoerd. Gedurende zijn promotietraject heeft hij op verscheidene nationale en internationale congressen mogen presenteren en heeft hij verschillende beurzen ontvangen. Thijs heeft in zijn vrije tijd een bestuurlijke functie bekleed bij Vereniging de Zonnebloem, waarin hij verschillende activiteiten heeft georganiseerd. Daarnaast was hij actief verbonden aan het hlo in Nijmegen als gastdocent en lid van de afstudeercommissie. Na zijn aioschap werd hij uitgenodigd om een korte periode als docent/onderzoeker op deze school te werken. Thijs zal zijn wetenschappelijke carrière voortzetten als postdoc in het lab van prof.dr. A. McClatchey aan de Harvard Universiteit in Boston.  184

Curriculum vitae and publications

Scientific publications, talks, fellowships and posters Scientific publications Published: - Axon injury triggers EFA-6 mediated destabilization of axonal mictotubules via TACC and doublecortin like Kinase. Chen L., Chuan M., Koorman T., Boxem M., Jin Y., Chisholm A.D. E-Life. 2015 sep 4;4. doi: 10.755/elife -

The C. elegans Crumbs family contains a CRB3 homolog and is not essential for viability. Waaijers S., Ramalho J.J., Koorman T., Kruse E., Boxem M. Biol Open. 2015: Feb 6;4(3):276-84. doi: 10.1242/bio.201410744

Controlled sumoylation of the mevalonate pathway enzyme HMGS-1 regulates metabolism during aging. Sapir A., Tsur A., Koorman T., Ching K., Mishra P., Bardenheier A., Podolsky L., Bening-Abu-Shach U., Boxem M., Chou T.F., Broday L., Sternberg P.W. Proc Natl Acad Sci U S A. 2014 Sep 16;111(37):E3880-9

Identification of human protein interaction domains using an ORFeomebased yeast two-hybrid fragment library. Waaijers S.*, Koorman T.*, Kerver J., Boxem M. J Proteome Research. 2013 may 30; pp 3181–3192; doi: 10.1021/ pr400047p * shared first author

Rap2A links intestinal cell polarity to brush border formation. Gloerich M., ten Klooster J.P., Vliem M.J., Koorman T., Zwartkruis F.J., Clevers H., Bos J.L. Nat Cell Biol. 2012 June:14,793–801(2012) doi:10.1038/ ncb2537

Submitted: - A combined binary interaction and phenotypic map of C. elegans cell polarity proteins. Koorman T., van der Voet M., Klompstra D., Ramalho J.J., Nieuwehuize S., van den Heuvel S.J.L., Nance J., Boxem M. -

Tissue specific purification of protein complexes in Caenorhabditis elegans. Waaijers S. et al.,

Work selected for oral presentations: - A combined binary interaction and phenotypic map of C. elegans cell polarity proteins. The 6th EMBO meeting Birmingham, England (2015, 5 – 8 September) Presented by M. Boxem. 185

- - - -

A combined binary interaction and phenotypic map of C. elegans cell polarity proteins. International DGZ meeting, Germany Cologne (2015, 24 â&#x20AC;&#x201C; 27 March) Presented by M. Boxem. Mapping the C. elegans polarity protein interaction network. T. Koorman; Dutch C. elegans meeting Wageningen WUR (March 2014). Unraveling the C. elegans Interactome underlying Cell Polarity. T. Koorman; ESF-EMBO symposium Cell Polarity and Membrane Trafficking. Pultusk, Poland. Selected for short talk. (2014, May 10 - 15). Unraveling the C. elegans Interactome underlying Cell Polarity. T. Koorman; Dutch C. elegans meeting Groningen ERIBA (March 2013). Unraveling the C. elegans Interactome underlying Cell Polarity. T. Koorman; Dutch C. elegans meeting Utrecht Utrecht (April 2012).

Fellowships: - ESF-EMBO Fellowship to attend the Cell Polarity and Membrane Trafficking symposium: 2014 - Genetics Society of America (GSA) Travel-Fellowship to attend the 19th International C. elegans meeting UCLA: 2013 - ESF-EMBO Travel-Fellowship to attend the Cell Polarity and Membrane Trafficking symposium: 2012 Posters: - A combined binary interaction and phenotypic map of C. elegans cell polarity proteins. Koorman, T., van der Voet, M., Klompstra D., Nance J., van den Heuvel S., Boxem, M. 20th International C. elegans meeting UCLA Los Angeles, US (2015, June 24 â&#x20AC;&#x201C; 28) Presented by M. Boxem. - Unraveling the C. elegans protein interaction network underlying cell polarity. Koorman, T., van der Voet, M., Klompstra D., Nance J., van den Heuvel S., Boxem, M. ESF-EMBO symposium Cell Polarity and Membrane Trafficking. Pultusk, Poland (2014, May 10 - 15). - Unraveling the protein interaction network underlying cell polarity. Koorman, T., van der Voet, M., Klompstra, D. Ramalho, J. van den Heuvel, S. Nance, J. Boxem, M. European C. elegans meeting MDC-Berlin Germany (2014).

186

Curriculum vitae and publications

- - -

dpy-19 and mig-21 control the persistent directionality of migrating Q neuroblasts in Caenorhabditis elegans Middelkoop, T. C., Koorman, T., Boxem, M., & Korswagen, H. C. 19th International C. elegans meeting, UCLA Los Angeles, US (2013, June 26 – 30). ULP-4 SUMO protease controls HMGS-1 activity in cytosolic and mitochondrial metabolic networks during development, aging, and stress. Sapir, A., Tsur, A., Koorman, T., Boxem, M., Sternberg, P., & Broday, L. 19th International C. elegans meeting, UCLA Los Angeles, US (2013, June 26 – 30). Dissecting the role of CEBP-1 in axon regeneration. Kim, K. W., Ying, P., Koorman, T., Boxem, M., & Jin, Y. 19th International C. elegans meeting, UCLA Los Angeles, US (2013, June 26 – 30). Unraveling the protein interaction network underlying cell polarity. Koorman, T., van der Voet, M., Ramalho, J. van den Heuvel S., Boxem, M. 19th International C. elegans meeting, UCLA Los Angeles, US (2013, June 26 - 30) Unraveling the protein interaction network underlying cell polarity. Koorman, T., van der Voet, M., Boxem, M. ESF-EMBO symposium Cell Polarity and Membrane Trafficking. Pultusk, Poland (2012, March 31 – April 5). A biochemical approach to identify a Rheb-specific GEF. R. Weterings, H. Rehmann, T. Koorman, M. Brunnig, C. Berghaus, M. Schwarten, R. Stoll, A. Wittinghofer and F.J.T Zwartkruis. mTOR signaling symposium Spain (2007). 

The PhD graduation research school Cancer Stem cells & Developmental (CS&D) is part of the graduate school of life sciences in which Utrecht University, Utrecht Medical Centre and the Hubrecht Institute are involved. The following educational programme was completed:

Subject

Seminars: Introductory courses: Specialized courses: Masterclass: PhD student retreat: General courses: Conferences:

Attended Yes Yes Yes Yes Yes Yes Yes

Number

At least 20 per year Several were attended Several were attended Three were attended Two were attended Several were attended Several were attended

187

Dankwoord De jaren zijn werkelijk voorbij gevlogen. Vele jaren van studie en werken op vele laboratoria hebben mij uiteindelijk gebracht waar ik nu ben; van mbo tot aio. Als er één ding duidelijk wordt tijdens dit traject, is het wel dat de mensen buiten het lab net zo belangrijk zijn als de mensen binnen het lab. Wetenschap, je bedrijft het niet alleen. Ik heb veel geleerd en ben enorm gegroeid. Niet alleen als wetenschapper, maar ook als mens. Bij dezen neem ik de gelegenheid om de mensen te bedanken die hieraan een bijdrage hebben geleverd. Allereerst mijn copromotor en dagelijkse begeleider Mike of ook wel Dr. Boxem, Mieke of toch maar Baas B? Ik laat het bij Mike. Mike, sorry dat ik je lang aan het lijntje heb gehouden of ik bij je zou komen promoveren. Gelukkig had jij geduld en gaf je mij de ruimte om te wikken, wegen en uiteindelijk de juiste keuze te maken. Mike, bedankt voor alles! We hebben elkaar omarmd en vervloekt, hadden hoogtepunten en dieptepunten maar waar het om gaat is dat we mooie wetenschap hebben bedreven. Met als resultaat een boekje waar ik enorm trots op ben. Ik hoop dat in de toekomst je deur, net als de afgelopen jaren, open blijft staan voor gezelligheid maar ook voor wetenschap. Mijn promotor Sander, ik ben je nog steeds dankbaar dat ik ruim vijf jaar geleden stage bij je mocht komen lopen, om vervolgens verliefd te worden op de worm. Bedankt voor de afgelopen jaren. Je begeleiding en advies zowel professioneel als privé, het was goud! Toch maar Boston hè. Dankjewel! Hierbij wil ik ook mijn aio-commissie bedanken voor alle nuttige input en de leescommissie voor het beoordelen van het manuscript. Alle collega’s van de vakgroep Ontwikkelingsbiologie wil ik bedanken voor de betrokkenheid, gezelligheid en input. Het is een komen en gaan van collega’s, maar gelukkig is er gedurende je aioschap een ‘harde kern’. Adri bedankt voor je eerlijken oprechtheid en je nuttige commentaar. Ben nog steeds onder de indruk van de manier hoe jij je colleges geeft. Al knijp ik hem wel als jij met de rode pen door de tekst gaat. Inge, drukke bezige bij. Experimenten, onderwijs geven en dan ook een gezin onderhouden. Laat ik je gastvrijheid bij de gezellige Thanksgiving-feesten en barbecues niet vergeten. Je zoete aardappels zijn werkelijk goddelijk! Vincent, mijn labmaat en goede vriend buiten het lab! Vince! Paranimf! Hadden we in het begin verwacht dat we zo op één lijn zouden zitten? Samen naar concerten, foto’s maken, naar musea, dagjes uit, slechte grappen maken en laten we het belangrijkste niet vergeten: shoarma eten en de wereld redden! Poe, als de mensheid nou eens wist hoe vaak wij ze wel niet gered hebben… Serieus, Vince jij bent één van de onmisbare schakels tijdens mijn promotie geweest. Ik wens je het allerbeste met jouw promotie, dat komt vast goed. Selma, hoe kan het toch dat jij altijd overkomt als een oase van rust? Beheerst, voorbereid en berekenend. Selma, wij samen waren 188

Dankwoord

de aio’s van de Boxem-groep en ik had mij geen fijnere collega kunnen wensen. Wij hebben bijzonder prettig samengewerkt en daar is een prachtig artikel uitgekomen. Jana, bedankt voor al je praktische en technische hulp en ook de mentale steun. Zonder jou zou het allemaal een stuk lastiger geweest zijn. Ik heb erg prettig met je samengewerkt. En bedankt voor de tips over Praag! Suzanne, Truus, ik denk dat jij misschien wel één van de grootste vakidioten bent. Je vak- en theoretische kennis zijn echt ongelooflijk! Je hebt een pittig project, maar je gaat die cellen pakken! Als iemand het kan, ben jij het. Twijfel niet aan jezelf, je kunt het! Bedankt voor alle gezelligheid en dat we konden functioneren als elkaars klankkast. En anders, als echt niemand anders meer wil, wil ik wel met je mixen. Suzan, Suus, wat een doorzettingsvermogen! Je had wat pech in het begin, maar wat een prachtig onderzoek heb jij toch gedaan. En wat doe jij veel tegelijk, daarop ben ik nog steeds jaloers. We liepen behoorlijk parallel tijdens de promotie, het was altijd prettig om samen te overleggen. Heel veel succes met je volgende carrièrestap. Het komt vast allemaal goed. Lars, beugelbuddy. En ja ik ga het toch zeggen… kakker! Echt, sinds jij dat mailtje rondstuurde in de werkmail, dat de ‘Appie een doosje excellente wijn in de aanbieding had’, voldoe je voor mij aan dat stigma. Maar daar zit je vast niet mee toch? Lars, je verhalen zijn bijzonder en je bijnaam Lassie of Lightning Lars doe je onbetwist eer aan. Bedankt voor de gezelligheid en het allerbeste voor je promotie. Ruben, het is maar goed dat we allebei dezelfde slechte, echt slechte, humor hebben. Je werkt voor twee groepsleiders en dat valt niet mee. Doe je best en ga zo door! Je komt er vast! Misschien een keer samen muziek maken? Zouden ze die dwerg al hebben opgehaald? Aniek, heel veel succes met je postdoc bij Sander. Je start met een prachtig fundament. Tim, beste maat, ik heb nog steeds enorm veel respect voor je keuze. Maar eerlijk is eerlijk, het lab was niet hetzelfde zonder jou. Nooit meer ‘Thijs wat ben je weer lelijk vandaag’ en ‘Thijs waarom zou je aan iets beginnen wat je toch nooit gaat lukken?’ of de mooiste: ‘Thijs, kom je online op Facebook? Kunnen we chatten’. En dat terwijl je gewoon achter mij zat op het kantoor. Het allerbeste Tim. Geert, promoveren in het buitenland. Goede keus! Had ik een reden om naar Amerika te gaan! En Jurassic Park is natuurlijk veel leuker om in Berkeley in de bioscoop te zien dan in Nederland. Geert, zo ver weg en toch dichtbij. Ideaal dat Skype en online gaming. Volgens mij ben je helemaal niet weggeweest. Kom je voor je postdoc weer terug in Nederland? Nee, blijf maar daar! Kun je weer een meeting komen crashen (jouw badge was mooier dan die van ons) en hebben we een reden om Universal Studios te bezoeken. Misschien hebben ze ondertussen mijn telefoon wel gevonden… De oude garde Martine, Monique, Matilde, Christian. Allemaal dank! Monique, ik mocht verder gaan waar jij ophield. Niets is lastiger dan een steengoede aio/postdoc opvolgen. Ik hoop dat ik je wetenschapskindje goed heb laten opgroeien. Veel succes met je postdoc, ADHD-vliegen moeten niet worden 189

onderschat. Matilde, ik heb je maar even mogen meemaken maar dat was genoeg om een verpletterende indruk achter te laten. Succes met je carrière. Christian, niets is mooier dan iemand die zijn passie volgt. Voor jou is dat muziek en dat doe je geweldig. Toch mis ik af en toe wel onze discussies bij de lunch en borrels. I also would like to thank all the people from the Cell Biology department for the great borrels, nice retraits, chats and input during work discussions. Especially the nights out in the city and going to A State of Trance! That was great. We should definitely do that again. I would also like to thank our collaborators Diana Klompstra and Jeremy Nance for the discussions and valuable work on PAC-1 and Amir Sapir from the Sternberg/Broday labs for the collaboration on the ULP-4 project. It was also very nice meeting you people on the C. elegans meeting in Los Angeles. Gedurende je professionele carrière ontmoet je een heleboel mensen die een bijdrage leveren aan je ontwikkeling. Je wordt gevormd door de mensen die je ontmoet. De volgende mensen zijn voor mij onmisbaar. Mvr. Maria van Zon, u hebt mij op het mbo enthousiast gemaakt voor de medische richting van het laboratoriumonderzoek. Uw suggesties voor mijn stages waren van onschatbare waarde! Mijn ouders vonden het toen nogal wat. Een jochie van 18 die vanuit de Achterhoek naar het centrum van Amsterdam verhuist en stage gaat lopen in het Nederlands Kankerinstituut (bedankt Peter) en vervolgens naar het Academisch ziekenhuis van Maastricht (bedankt Ronald en Ruud)! U had er vertrouwen in, en het is helemaal goed gekomen. Als we het toch over het mbo hebben: Bjorn, al sinds die tijd houden we contact. Samen hebben we ook het hbo doorlopen. Jij hebt nu een prachtige baan in Nijmegen en bent een meer dan uitstekende analist! Ik geniet er altijd enorm van als we samen sushi eten en biertjes drinken. Remko, studiebegeleider op het hbo, wat fijn dat ook wij na al die jaren nog steeds zo goed contact hebben. Nu zijn we zelfs even collega’s. Terugkomen op de HAN voelt voor mij als een warm bad, als thuiskomen, en dat nog steeds na al die jaren. Ook jouw advies heeft mij een betere wetenschapper en mens gemaakt. Het scheelt dat we allebei wild worden van signaaltransductie. Onze ‘duo’-colleges waren echt leuk! Je voelt goed aan wat studenten nodig hebben en je haalt het beste uit ze. Een ongekend talent. Ik hoop dat ik nog vaak mag terugkomen, les mag blijven geven en onderdeel mag blijven van de HAN. Marjolein, nee ik maak er lieve drukke chaotische Marjolein van, ik zet je bij de HAN, oké? Jij was mijn allerlaatste stagebegeleider en op dat moment precies wat ik nodig had! Een super enthousiast ongeleid projectiel met allemaal geweldige ideeën! Als ik je zie is het echt onmogelijk om niet vrolijk te worden! Je haalt het beste uit de mensen om je heen, geeft ze het vertrouwen, herkent de talenten en laat ze groeien! Wat was het leuk om met jou de studenten les te geven. Je hebt ‘Soft skills for hard results’ nodig, was het niet?

190

Dankwoord

Han en Sam, matties vanaf het hbo. Leuk dat we nog steeds samen dingen ondernemen. De tripjes naar Schotland en Bristol waren onvergetelijk. Hanniebal, onze Brabander, onze Erpenees! Hoeveel uren wij wel niet in de kroeg en in gamen hebben gestopt? We hebben de hele periode op de VU in de trein MarioKart zitten spelen en nog kan je er niks van! Goed blijven oefenen Han, misschien win je nog wel een keer. Veel succes met het afronden van je promotie. Sammie, samen naar het hbo, dezelfde master gedaan en beiden bijna klaar met promoveren, in naastgelegen labs. Misschien moeten we die navelstreng toch echt eens doorknippen. De stapavonden, filmavonden, gameavonden en al het andere gezellige waren onvergetelijk! Ik ga het allemaal erg missen Sam. Je komt ons snel bezoeken, toch? Heel veel succes met het afronden van je promotie. Fried, stagebegeleider, mentor en vriend! Bij jou is het voor mij eigenlijk echt begonnen, wat werd wetenschap ineens leuk! Mijn hbo-afstudeerstage was onze eerste kennismaking en tot de dag van vandaag ben jij voor mij een hele belangrijke schakel. Het is mij een eer dat je de laatste jaren in mijn aio-commissie zat en dat je straks ook aan de ‘hoge heren tafel’ zit bij mijn promotie. Niet doorvertellen: Rheb en mTor blijven toch echt wel mijn favoFried! De afgelopen jaren heb ik een aantal studenten mogen begeleiden. Mijn wetenschappelijke erfstuk is bij jullie begonnen. En ja, ik heb ook een hele hoop van jullie geleerd! João, you were my first student, and I could not wish for a better one! Our collaboration went great and you did your work really well. I’m also really proud that you started you PhD in our group. Yes, ERM is cool! But, you still need to explain me one thing. How come that a Portuguese person does not know Piri Piri? Susan, Jij bent bij mij begonnen als bachelorstudent en we hadden meteen een klik. Je hebt je stage goed gedaan en een mooie scriptie geschreven. Wat nog veel mooier is; je kwam terug bij mij voor je eerste masterstage! Ik was daar erg blij om en wist meteen dat dat goed zou komen! Zo goed zelfs dat je een belangrijke bijdrage hebt kunnen leveren aan het artikel en dan ga je ook nog eens promoveren in Zwitserland! Trots, trots! Tijdens je stage startte er ook nog een andere student bij mij. Kaylee, jullie waren twee handen op één buik op het lab en samen waren we Groep Koorman. Kaylee, wat ben jij gegroeid tijdens je stage. Je kwam onzeker binnen en sloot het op de HAN af met een dijk van een presentatie! Je hebt ook mooi werk afgeleverd. Ondertussen ben je aan je master begonnen, wat goed, ik mag wel op je promotie komen he? Ik wens jullie allebei het allerbeste en ben erg trots jullie begeleid te mogen hebben. En dan last, but not least, Machiel. Je moest even opstarten bij ons, maar toen zag je het licht en alles kwam goed. Jouw meningen zijn prachtig, je woordgrappen even slecht en dan ook nog mooi werk afleveren. Succes met je verdere carrière. Hierbij wil ik ook alle korte-stagestudenten en vwoleerlingen ook bedanken voor de leuke tijd.

191

Zoals ik al zei, zowel binnen als buiten het lab zijn de mensen om je heen van onschatbare waarde. De mensen buiten het lab houden je in evenwicht en relativeren als je zelf met je hoofd te veel in de wolken zit. Daarom wil ik graag de volgende mensen bedanken. Martijn, Lars en Willem-Jan. In jullie dankwoord hebben jullie het vooral over mijn frommel-goals met FIFA. Laten we dat maar achter ons laten. Martijn, paranimf, ik weet niet of wij nou het beste of het slechtste in elkaar naar boven halen. Maar het heeft in ieder geval geleid tot prachtige experimenten, urenlang microscopie, de betere stapavondjes en barbecues (met spekfakkels). Ik heb er enorm veel respect voor hoe snel je denkt, doet, flexibel bent en al tien stappen vooruit kan denken. Uren en uren filosoferen over mooie proeven en ideeën. Zullen we anders samen een lab beginnen? Je bent een wetenschappelijke held en een goede vriend. Het was geweldig om jullie verschillende keren in de States te mogen bezoeken, al heb ik die Golden Gate nog steeds niet gezien. En echt, sorry van die sleutel, ik begrijp nog steeds niet het verschil tussen gooien en mikken. Wel jammer dat we elkaar weer mislopen de komende jaren, maar iemand moet toch jouw taak overnemen in Amerika. Lars, nog steeds ben ik vereerd met het feit dat ik je paranimf mocht zijn! Het is maar goed dat Lianne mij er even aan herinnerde dat jij die dag moest promoveren. Het scheelt tien jaar tussen ons, maar daar is alles ook mee gezegd. Oh, ik kan nou eindelijk zeggen “What about sample 1.13”, dat was toch het epje waar de Rheb-GEF inzit dat we kwijt waren geraakt toen we samen aan de MassSpec zaten? Misschien moeten we die maar een keer teruggeven aan Fried, het grapje is nu wel leuk geweest. Lars, ik hoop dat we nog lang vrienden mogen blijven en dat we in de toekomst vaak bij elkaar mogen komen voor een gezellige avond. Ik beloof hierbij ook geen panfluit-grappen meer te maken. Willem-Jan, zullen we binnenkort maar weer eens een potje gaan schaken onder het genot van een whiskey? Ik ben ook nog steeds onder de indruk van je labjournaal. Matties uit de Achterhoek Roy, Ivo en Jeffrey. Hebben we vroeger nou wel of niet naast elkaar in de couveuse gelegen? Dat moet wel! Jongens, jullie horen ook in dit rijtje. Ook al zijn we uit elkaar gegroeid en leiden we allemaal een eigen leven in andere plaatsen, toch blijven we elkaar opzoeken. En wat voelt het dan altijd vertrouwd. Binnen een fractie is het weer als vanouds. Zo moet het en laten we dat vooral zo houden! Bedankt voor alle gezelligheid. Oh ja en Roy, bedankt voor alle kansen die je mij hebt gegeven :D John, “Thijs, als jij nou ing. voor je naam hebt staan he, betekent dat dan dat je geen bankrekening meer mag bij de Rabo?” en “Ga jij nou bij de Halfords werken in Boston?”, dat is John! Dat maakt je geweldig! De Winterswijkse vriendengroep, bedankt voor de gezellige avonden en volksfeesten. Heerlijk om naar terug te komen en kijk er altijd weer naar uit.

192

Dankwoord

Mijn schoonfamilie, Clemens, Herma, Jeroen en Pascal. Clemens en Herma bedankt voor het ‘thuiskomen-gevoel’ en het relativeren. Af en toe hebben wij het nodig om met beide benen op de grond te worden gezet. Jeroen, wanneer kom je weer eens langs met de vrachtwagen? Ik zou er nog steeds wel een keer in willen rijden hoor! Pascal, of is het nou ‘Chef tosti’ onder werktijd? Enorm knap hoe jij Afghanistan hebt doorstaan. Wanneer gaan we een keer biertjes doen in Utrecht? We hebben ze lang nog niet allemaal gehad. Gerard, Carola en Anne. Lieve papa, mama en zusje. Het was een lange rit, en we denderen nog even door. Bedankt voor jullie onvoorwaardelijke steun, een luisterend oor, advies en dat jullie mee zijn geweest naar al die opendagen met al die aparte mensen en moeilijke woorden. Zonder jullie steun was ik nooit zo ver gekomen. Voor mij is het onbetaalbaar om thuis te komen, mamasoep met koekjespudding te eten, ‘even bij Anne kijken’ en in de avond gezellig op de bank samen Paul de Leeuw te kijken en ondertussen te onderhouden met de vogels. Op zulke momenten realiseer ik mij altijd weer wat echt belangrijk is. Komen jullie straks gezellig langs in Amerika? Lieve, lieve Lianne. Als we het hebben over onvoorwaardelijke steun, liefde, lief en rust. Hoe vaak jij wel niet trouw mee bent geweest naar het lab. Even dit doen, even dat doen. Het klinkt als een cliché maar zonder jou was het gewoon echt niet gelukt. Wat ben ik gelukkig dat een ontmoeting in de discobus tien jaar geleden tot dit heeft geleid! Laten we eerlijk zijn, jij bent het voor mij en wij zijn het voor elkaar. De laatste jaren waren niet altijd makkelijk voor ons. Maar wij zijn er sterker en verliefder dan ooit uitgekomen. Binnenkort word je mijn vrouw, nu ik dit typ voel ik mij zo gelukkig. Dat jij vandaag naast mij staat is uiteindelijk het enige wat echt telt. Hierna gaan we er samen op uit. Laten we verder gaan. Jij en ik, samen, op naar het heden. Kus X.

D 193

“Then I’ll be going now. I’ll come back when it is all over”

Networking for proteins

Uitnodiging

Networking for proteins A yeast two-hybrid and RNAi profiling approach to uncover C. elegans cell polarity regulators

Thijs Koorman

Voor het bijwonen van de openbare verdediging van het proefschrift

2016

Heeft u vragen? Neem a.u.b. contact op met mijn paranimfen: Vincent Portegijs v.c.portegijs@uu.nl Martijn Gloerich gloerich@stanford.edu