STUDY OF THE ROLE OF EGL-38 PAX IN THE DEVELOPING EGG-LAYING SYSTEM AND GERMLINE CELL SURVIVAL IN CAENORHABDITIS ELEGANS DISSERTATION

Transcription

1 STUDY OF THE ROLE OF EGL-38 PAX IN THE DEVELOPING EGG-LAYING SYSTEM AND GERMLINE CELL SURVIVAL IN CAENORHABDITIS ELEGANS DISSERTATION Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy in the Graduate School of The Ohio State University by Vandana Rajakumar * * * * * The Ohio State University 2007 Dissertation Committee: Name: Dr. Helen M. Chamberlin, Advisor Name: Dr. Amanda S. Simcox Name: Dr. Paul K. Herman Approved by Adviser Department of Molecular Genetics Name: Dr. Stephen A. Osmani

2

3 ABSTRACT Understanding how genes coordinate the organization of cells into organs and how mutations in these genes upset this organization are fundamental questions in both cancer research and developmental biology. PAX factors play important roles to coordinate the development of mammalian organs like the kidney, eye and thyroid. The disruption of PAX activity leads to diseases like renal-coloboma syndrome and cancers in mammals. Dissecting the different functions for these factors has been difficult due to the complexity of organ development. To better understand these processes at a cellular level, I studied the role of a Pax2/5/8 transcription factor, EGL-38 in the developing egglaying system and in the survival of germline cells of the nematode, Caenorhabditis elegans. egl-38(n578) mutant animals exhibit of high levels of Programmed Cell Death (PCD) in the germline and the soma. Wild-type EGL-38 can transcriptionally regulate the bcl-2 gene, ced-9 in C. elegans. In this dissertation, I performed mosaic analysis for EGL-38 and found that its activity is required in both the germline and soma to promote germline cell survival. EGL-38 is also involved in the development of the C. elegans egg-laying system. The egl-38(n578) mutants are egg-laying defective, so the eggs hatch within the hermaphrodite parent. The latter then bursts to release viable progeny. The C. elegans ii

4 egg-laying system has two tissues, the somatic gonad and the vulva. The Pax2/5/8 transcription factor, EGL-38 is required for the normal development of the egg-laying system. Previously it was hypothesized to function in one tissue, the vulva to affect the development of both the vulval and somatic gonad tissues of the egg-laying system. While its mammalian counterpart, Pax 2 has independent functions in both the tissues of the metanephric kidney and thereby coordinates development between them. However, our studies involving various mutants and mosaic analysis demonstrate that egl-38 does have distinct functions in both the tissues of the egg-laying system of C. elegans. Furthermore, this influences both the autonomous and non-autonomous functions during the development of this system. Therefore, egl-38 behaves like the mammalian Pax2 by acting in the various tissues of an organ to promote its coordinated development. I also performed a genetic screen to isolate mutations in genes that suppress the egl-38(n578) egg-laying defect. Thereby I isolated four strong suppressor mutations. Two of these were mapped to separate locations on chromosome IV. Both these suppressor mutations suppress both the egg-laying defect and the elevated levels of cell death associated with the egl-38(n578) mutation. These studies have demonstrated that EGL-38 acts in multiple tissues to coordinate development like its mammalian counterpart Pax2 in the developing metanephric kidney. Thus there is conservation of the development logical by which PAX factors coordinate organogenesis. Finally, identifying the genes affected by the suppressor mutations could have implications in understanding development and survival of not only C. elegans but also of higher animals if these genes are evolutionarily conserved. iii

5 DEDICATION To the four people who have guided, supported and inspired me my parents, Ambika and Lt Col. C. Rajakumar, my advisor, Dr Helen M. Chamberlin and my husband, Ashwin Uttam, I am deeply grateful. iv

6 ACKNOWLEDGMENTS I would like to acknowledge my advisor, Dr. Helen Chamberlin, for her patience, guidance and support throughout my graduate education. I joined her Lab when I first came to the USA. Through the years, she has stood by me through all my failures and been an invaluable mentor in my professional career. She has also silently helped me develop confidence in myself as a person and most important taught me to be become an independent thinker. I am also deeply grateful to the members of my Ph.D. committee, Dr. Helen Chamberlin, Dr. Mandy Simcox, Dr. Paul Herman and Dr. Stephen Osmani for their valuable suggestions, guidance and generosity with their time. I have also grown professionally and benefited from the guidance and suggestions of other professors in the Department of Molecular Genetics and the Program in Molecular, Cellular and Developmental Biology (MCDB). My appreciation extends especially to Dr. Russell Hill for useful suggestions and stimulating discussions. I would also like to acknowledge Dr. David Bisaro for his guidance as the Chair of the MCDB Program. This Research was supported by grants from the National Science Foundation, the National Institutes of Health. My gratefulness also extends to my current and former colleagues for their personal friendship and professional guidance and suggestions, including Rong-Jeng Tseng, Dr. Xiaodong Wang, Dr. DonHa Park, Sama Sleiman, Ryan Johnson, Kristen v

7 Armstrong, Hongtao Jia, Xin Li, Steven Sewell, Jesse Sloan, Dr. Ihab Younis, Dr. Nesrine Affara, Aysha Osmani, Litty Paul and Steven Justiano. I am also deeply grateful to Jan Zinaich, Jessica Siegman, Debbie Dotter, Marsha Hronek and other administrative staff for their cheerful and valuable help. My appreciation also extends to my family, especially my parents, Ambika and Lt. Col. Rajakumar for their unconditional support and encouragement. They allowed me to move to another country to pursue my dreams and at every step thereafter gave me the freedom to make my own decisions. My thanks also extend to my sisters Bhagya and Anandita for their support. Finally I would like to acknowledge my husband, Ashwin for his patience, support and help through our years together. He has lifted my spirits and directing my focus to the goal at hand when the going got tough. vi

8 VITA February 26, Born Bangalore, Karnataka, INDIA B. S. Zoology, Chemistry and Physics St. Josephs College of Arts and Sciences, Bangalore, INDIA M.S. Biochemistry M. S. University, Baroda, INDIA Teaching and Research Associate, The Ohio State University Columbus, Ohio. PUBLICATIONS Rajakumar V and Chamberlin HM. The Pax2/5/8 gene egl-38 coordinates organogenesis of the C. elegans egg-laying system. Developmental Biology 2006 Sep 6 Park D, Jia H, Rajakumar V, Chamberlin HM. Pax2/5/8 proteins promote cell survival in C. elegans. Development 2006, 133(21): Zhang G, Sleiman SF, Tseng RJ, Rajakumar V, Wang X, Chamberlin HM. Alteration of the DNA binding domain disrupts distinct functions of the C. elegans Pax protein EGL- 38. Mechanisms of Development (7-8): FIELDS OF STUDY Major Field: Molecular, Cellular and Developmental Biology vii

9 TABLE OF CONTENTS Page ABSTRACT...ii DEDICATION...iv ACKNOWLEDGEMENTS... v VITA...viii LIST OF FIGURES... x CHAPTERS: 1. Introduction Characterization of the role of EGL-38 in the egg-laying system using molecular markers Introduction Materials and Methods Results Discussion Site of action study for EGL-38 in germline cell death and the egg-laying system development Introduction Materials and Methods Results Discussion A screen for genetic suppressors of the egl-38(n578) egg-laying defect and the identification of rig-4 as a possible candidate suppressor Introduction viii

10 Materials and Methods Results Discussion Discussion List of References ix

11 LIST OF FIGURES Figure Page 1.1 Amino acid sequence alignment of the paired domain of Pax2/5/8 proteins Different hypomorphic alleles of egl-38 disrupt different functions The egg-laying system in C. elegans Expression of molecular markers in the vulva of egl-38(n578) mutants Expression of molecular markers in the uterus of egl-38(n578) mutants lin-3 expression, uv1 cell specification, vulval separation and egg-laying ability of the different egl-38 mutant animals Study of germline cell death Lineage diagram of the early embryonic divisions in C. elegans Summary of observations made during mosaic analysis for tissue source egl-38(+) activity promoting germline cell death egl-38::gfp is expressed in both the uterus and in the vulva A cell lineage chart of the early cell divisions in C. elegans Mosaic analysis for egl-38 function An activated allele of let-23 (sa62) can bypass the uv1 cell defect of egl-38 (n578) A model for EGL-38 functions in the egg-laying system F2 suppressor screen using EMS...89 x

12 4.2 Summary of the isolated suppressors Complementation test Three-factor mapping of the suppressor mutation, gu Three-factor mapping of the suppressor mutation, gu DNA clones (cosmids and Yeast Artificial Chromosome or YAC) tested in rescue analysis for gu68 and gu Phenotypic analysis of the suppressor with respect to the vulval tissue in the egglaying system Phenotypic analysis of the suppressor with respect to the uterine tissue in the egglaying system gu68 and gu51 suppress elevated levels of germline death seen in egl-38(n578) animals Phenotypic analysis of the suppressor, gu68 with respect to the germline cell death Phenotypic analysis of the suppressor, gu51 with respect to the germline cell death Phenotypic analysis of the suppressors with respect to LIN-48::GFP expression the hindgut cells Summary of the phenotypic analysis of suppressors, gu68 and gu Analysis of rig-4 as a candidate gene for gu xi

13 CHAPTER 1 INTRODUCTION Pax genes play important roles during development in most species. They influence the proliferation of cells and their subsequent assembly into organs. In addition, altering their expression or activity results in various developmental disorders. In order to better understand how these genes control development, I focused on the Pax2/5/8 subclass gene in the worm Caenorhabditis elegans (C. elegans) called egl-38. My work pays specific attention to the n578 mutation in the egl-38 gene which preferentially causes defects in the development of the egg-laying system and in cell survival. Thereby I intend to elaborate on the mechanism by which Pax genes in general and egl-38 in particular can affect development. This knowledge may have further implications in the understanding and treatment of diseases. This Chapter offers an introduction to the Pax gene, egl-38, establishing the context for the experiments I have performed as part of my dissertation. 1

14 Pax genes and development The Pax genes encode a class of evolutionarily conserved transcription factors called the Pax factors. These factors are characterized by the presence of a 128-amino acid DNA binding domain called the Paired Domain (PD) (Fig1.1). Study of the Drosophila and the human Pax6 PD crystal structures revealed that the carboxy and the amino ends of this protein each have helix-turn-helix motifs (Xu et al., 1995; Xu et al., 1999). Each half-site binds to the major groove of DNA and both these sites contribute to the DNA-binding specificity (Czerny et al., 1993). Pax genes were initially identified in Drosophila melanogaster (D. melanogaster) and subsequently nine members have been isolated in vertebrates. Some Pax proteins also have an additional Homeodomain and/or Octapeptide domain. To add to the complexity of the Pax proteins, they undergo extensive alternative splicing, use alternative promoters and interact with multiple proteins. Based on genomic sequence, structural similarity and conserved function, the Pax gene family in mammals is classified into four sub-groups. For example the Pax2/5/8 subgroup of proteins has similar PD sequence, an octapeptide domain and a partial homeodomain (Reviewed in Chi and Epstein, 2002). The Pax proteins play crucial roles during the development of organisms ranging from the Coelenterate Hydra to the worm C. elegans and the mammals (Reviewed in Noll, 1993). Mutations in the PD are functionally relevant and result in various developmental and proliferation defects. For example, mutations in Pax1, Pax3 and Pax6 are responsible for the undulated, Splotch and Small eye developmental mutants of mice (Balling et al., 1988; Epstein et al., 1991 and Hill et al., 1991). In humans, Pax6 2

15 mutations affect eye development and lead to aniridia, while mutations in Pax3 are associated with Waardenburg syndrome (Chi and Epstein, 2002). The members of the Pax2/5/8 gene family have been classified together as a subfamily because they have similar PDs, an octapeptide domain and (in some species) a partial homeodomain. The expression pattern and mutant phenotypes for these genes have revealed that they contribute to the development of various tissues. The Pax2, Pax5 and Pax8 genes all affect the Central Nervous System and individually play roles in the development of the kidney, B cell and the thyroid respectively (Chi and Epstein, 2002). Pax5 gene mutants lack mature B lymphocytes (Urbanek et al., 2004). Pax2 mutants are defective for multiple steps in kidney development, leading to the development of the renal and genital tract, epithelial transformation of mesenchymal tissue and other events (Torres et al., 1995, 1996). To better dissect the roles for these factors at a cellular level during development, I am using the simple nematode C. elegans Pax2/5/8 genes in C. elegans C. elegans was proposed as a model system by Sydney Brenner. It has a simple anatomy and a transparent body making it amenable to microscopy. It also has a short life-span and can undergo self-fertilization. The development of C. elegans is invariant wherein all animals have a fixed number of somatic cells (959), each of which can be traced back to the zygote with a distinct lineage pattern (Sulston and Horvitz, 1977). Several tools have been developed to study development in C. elegans. Among these, 3

16 classical genetics has elucidated many signaling pathways and identified many molecular mechanisms. Completion of the genome sequence and the development of reverse genetics using techniques like RNA interference have further contributed to the same (Fire et al., 1998). Further extrachromosomal transgenes of a wild-type gene can be used to rescue animals mutant for that gene. These transgenes can be randomly lost during cell division generating genetically mosaic animals. The site of action of this gene s function can be easily determined from these mosaic animals because C. elegans has a reproducible cell lineage (Herman; 1984; Yochem et al., 1998). There are two related Pax2/5/8 genes in C. elegans: egl-38 and pax-2 (Wang et al., 2004). Both are almost identical in their PD domains and the two genes are proposed to be generated by a gene duplication event (Fig1.1). egl-38 is an essential gene affecting the development of various tissues and promoting cell survival. It was identified in different genetic screens (Trent et al., 1983; Chamberlin et al., 1997; Zhang et al., 2005). The other Pax 2/5/8 gene, pax-2 is not fully redundant with egl-38 because egl-38 mutants exhibit defects even though they have a normal pax-2. Deletion mutations were generated in pax-2 by the C. elegans Gene Knockout Project ( Although pax-2 has not been genetically characterized to a great extent, a recent study has shown that egl-38 and pax-2 contribute to cell survival in the germline and the somatic tissues (Park et al., 2006). Structurally, EGL-38 and PAX-2 include a PD and an octapeptide domain. Their PD is 74-75% identical to the human PAX proteins (Fig 1.1). egl-38 can produce two classes of transcripts. The larger one can be translated to a protein with 289 amino acids 4

17 and seven exons. The smaller one is produced when the C. elegans trans splice leader SL1 is trans-spliced to the second exon. This smaller transcript produces a protein that lacks the amino terminus of the larger protein including 28-amino acids of the PD (Chamberlin et al., 1997). The pax-2 gene can alternatively incorporate two more exons at its 5 end (Wang et al., 2004). The null allele of egl-38, s1775 is lethal (Chamberlin et al., 1997). Different viable alleles of egl-38 have been recovered and all are missense mutations in the paired DNA binding domain (Fig 1.1). Three of them, sy287 (G9S), sy294 (G33V) and n578 (G69E) have mutations that lead to substitution of Glycine (G) to other amino acids, Serine (S), Valine (V) and Glutamic Acid (E) respectively. The fourth gu22(m29i) has an Isoleucine in place of Methionine (M). To understand how egl-38 affects multiple tissues during development, high affinity binding sites were identified using in vitro selection. However these sites did not function effectively as EGL-38 response elements in vivo. Therefore, it is likely that both sequence features and additional factors play important roles in the selection and function of PAX sites in different tissues in vivo (Zhang et al., 2005). In this study, I will be focusing on the egl-38(n578) mutation, examining its role in the developing egg-laying system. Further I will investigate the site of action of EGL- 38 activity that promotes germline cell survival. In egl-38(n578) mutants there is elevated level of cell death in both the germline and the soma. 5

18 egl-38 and the egg-laying system The egg-laying system in C. elegans is located in middle of the body on the ventral side of hermaphrodites. It mainly comprises two tissues: the uterus which is a part of the somatic gonad, and the vulva. Fertilized eggs are initially stored in the uterus and then subsequently extruded outside through the vulva. Sperm can also enter the hermaphrodite through the vulva. The vulva consists of a toroid of cells which form a tube, the internal end of which opens into the uterus, while the external end opens outside. The process of egg-laying depends on the structure of the vulva and the uterus, and also on the muscles and the neurons that promote it (reviewed by Greenwald, 1997 ). During development of the vulva there is reciprocal signaling between the vulva and the somatic gonad (Fig. 1.3.). Initially, in the third larval (L3) stage, the Anchor Cell (AC) of the somatic gonad sends an Epidermal Growth Factor (EGF) signal, LIN-3, to its receptor LET-23 in three of the six Vulval Precursor Cells (VPCs) that lie along the ventral cuticle. LIN-3 induces the VPCs to adopt the vulval fate (reviewed by Greenwald, 1997). The VPC closest to the AC (P6.p) adopts the 1 fate while the other two (P5.p and P7.p) adopt the 2 fate. P6.p divides to produce eight descendents, whereas P5.p and P7.p each produce seven. These cells undergo invagination to form a stack of vulval cells. During the mid fourth larval (ml4) stage the eight P6.p descendants lie at the apex of this stack of cells. The four P6.p descendants closest to the somatic gonad are called the vulf cells, and they express the EGF signal LIN-3 and reciprocally signal to the somatic gonad to specify a subset of π uterine cells that lie above the vulf cells called uv1 cells (Chang et al, 1999). Following specification, the uv1 cells migrate to flank the forming vulva. The other π uterine cells, along with the AC, move away from the vulva 6

19 and fuse to form a syncytium called the uterine seam cell (utse). Meanwhile, the vulf cell nuclei separate along the vulva-uterine border to leave only a thin membrane of tissue between the vulva and the uterus at the prospective opening. Then the vulval cells fuse to form torroidal rings that line the vulva cavity connecting the uterus and the external cuticle. Finally, the vulva undergoes eversion (Sharma-Kishore et al., 1999). Therefore proper organ development of the egg-laying system requires cell fate specification in both vulva and uterus, signaling between the two tissues, and cellular migration and morphogenesis. The role of egl-38 in the developing egg-laying system was first evident when the egl-38(n578) mutant was insolated by Trent et al., in 1983 as a part of a screen to isolate worms that were defective in egg-laying or Egl-. The mutants isolated in this screen were fertile but had abnormalities in the release of the progeny. egl-38(n578) was characterized as a very sluggish mutant with an abnormal vulva. The eggs hatched within the hermaphrodites to form a bag of worms, which then burst to release viable progeny. The n578 mutation was mapped to chromosome IV, between unc-8 and dpy-20 (Trent et al., 1983). Morphological studies on the developing egl-38(n578) animals revealed that the uterine-vulval junction in these animals was abnormal. The membrane between the uterus and the vulva is thick in these animals because the vulf cells are mispositioned. Also, the uterine uv1 cells migrate like the presumptive utse cells instead of lying next to the vulva (Chamberlin et al., 1997). Finally Chang et al, showed that egl-38(n578) mutants lack lin-3::gfp expression. Interestingly, a gain-of-function allele of the lin-3 receptor, let- 23(sa62) bypasses the need for egl-38 in specifying the uv1 cells separate from the 7

20 uterine utse cells (Chang et al., 1999). In conclusion, egl-38 was hypothesized to function in one tissue, the vulva to affect the development of both the vulval and somatic gonad tissues of the egg-laying system egl-38(n578) and apoptosis Programmed Cell Death, or Apoptosis, is an evolutionarily conserved process wherein metazoans undergo biologically controlled cell death. This process occurs to regulate cell number during development as well as to remove damaged cells. The core biological pathway in apoptosis is mediated by the activation of enzymes called caspases. These enzymes cause irreversible cellular damage which necessitates the elimination of the cell (Reviewed by Lockshin, 2004; Yan, 2005). The conservation of apoptotic machinery has allowed for the use of simple organisms to dissect the biology of PCD. Some of the major breakthroughs in the study of apoptosis have come in the nematode C. elegans. For example, in the 1980s, Horvitz and his colleagues identified genes specific to PCD in worms called the ced genes. These genes encode proteins similar to mammalian cell death proteins like Bcl-2 and apoptosis protease-activating factors (Apafs) (Reviewed in Lettre and Hengartner, 2006; McCarthy, 2003; Mirkes, 2002). In C. elegans PCD occurs in both the somatic and the germline cells. PCD in each of these tissues has both different goals and different regulatory inputs although the core biochemical pathways remain the same. During somatic development 131 cells of 1090 undergo cell death in a lineage-specific manner. In contrast, the germline does not have a fixed number or lineage of cells. About 300 germ cells undergo PCD using the 8

21 same execution machinery as the somatic cells to maintain germline homeostasis. Although somatic and germline apoptosis utilize the same execution machinery, the mechanisms for activating cell death differ in somatic and germline cells. For example, the RAS-MAPK pathway promotes apoptosis in germline but not in somatic cell death. It is proposed that germline cell death helps to eliminate excess germ cells which act as nurse cells to release cytoplasmic content for the maturing oocytes (Gumienny et al, 1999; Twomey and McCarthy, 2003) Goals of this study The main goal of my dissertation is to understand how PAX2/5/8 proteins promote the development. To this end I have studied the role of the PAX2/5/8 transcription factor, EGL-38 in the development of the C. elegans egg-laying system and in the process of PCD in the germline. In the egg-laying system, previous results have shown that EGL-38 is necessary for the development of the cells at the junction of the uterus and the vulva. In Chapter 2, I present data characterizing the defects associated with these cells in the egg-laying defective egl-38(n578) animals using molecular markers. Subsequently, I compared these defects among the different hypomorphic mutants of egl-38 to learn if these defects were influenced by each other. Then in Chapter 3, I examined the site-of-action of the EGL-38 in the egg-laying system. Finally, to isolate other molecules that may act with EGL-38 in the developing egg-laying system, I performed a genetic screen for suppressors of the 9

22 egl-38(n578) egg-laying defect. The rationale and results for this screen are presented in Chapter 4. In addition to learning more about the role of EGL-38 in the egg-laying system, I also examined the site-of- action of EGL-38 in promoting germline cell survival using mosaic analysis in Chapter 3. By learning how EGL-38 contributes to two development processes involving multiple tissues, I hope to gain a deeper understanding of the biology of both the egglaying system and PCD in the germline of C. elegans. Studies done in the egg-laying system can shed light on how a PAX transcription factor promotes organogenesis at a cellular level. While in the germline, learning where EGL-38 acts to promote germline cell survival may gives us a better understanding of the mechanism of germline cell death in C. elegans. EGL-38 belongs to the evolutionarily conserved PAX2/5/8 subclass of transcription factors. Therefore, the data presented in this dissertation may provide useful insight in understanding the role of PAX2/5/8 transcription factors in development and disease of higher organisms as well. 10

23 sy287 G9S gu22 M29I G33V sy294 Ce-EGL-38 Ce-PAX-2 Hl-PaxB Dm-sparkling Mm-Pax2 SHTGVNQLGGVFVNGRPLADTVRAQIVEMSQHGTRPCDISRQLKVSHGCVSKILGRYYSTGSVR SHTGVNQLGGVFVNGRPLPDTIRAQIVEMSQHGTRPCDISRQLKVSHGCVSKILGRYYSTGSVR NHGGINQLGGTFVNGRPLIEPVRRKIVELAHQGVRPCDISRQLRVSHGCVSKILSRFYETGSVR GHGGVNQLGGVFVNGRPLPDVVRQRIVELAHNGVRPCDISRQLRVSHGCVSKILSRYYETGSFK GHGGVNQLGGVFVNGRPLPDVVRQRIVELAHQGVRPCDISRQLRVSHGCVSKILGRYYETGSIK G69E n578 Ce-EGL-38 Ce-PAX-2 Hl-PaxB Dm-sparkling Mm-Pax2 PGVIGGSKPKVATPRVVECIAGYKRANPTMFAWEIRQKLIEDQICGEENVPSVSSINRIVRNKS PGVIGGSKPKVATPRVVECIAGYKRANPTMFAWEIRQKLIEDQICGEENVPSVSSINRIVRNKS PGVIGGSKPKVATPSVVAKIQEYKQHNPTMFAWEIRDKLLSEQICDSDSVPSVSSINRIVRNRL AGVIGGSKPKVATPPVVDAIANYKRENPTMFAWEIRDRLLAEAICSQDNVPSVSSINRIVRNKA PGVIGGSKPKVATPKVVDKIAEYKRQNPTMFAWEIRDRLLAEGICDNDTVPSVSSINRIIRTKV Figure 1.1. Amino acid sequence alignment of the paired domain of Pax2/5/8 proteins. C. elegans (Ce) EGL-38 (Chamberlin et al., 1997) and PAX2 (Wang et al., 2004), Hydra littoralis (Hl) PaxB (Sun et al., 1997), Drosophila sparkling (Fu and Noll, 1997) and the Mus musculus (Mm) or Mouse Pax2 (Dressler et al., 1990) proteins are shown. The different sequences were aligned using Clustal W 1.8 (Jeanmougin et al., 1998) and shaded using Boxshade 3.21 ( Identical amino acids are shaded in black while the conserved amino acids are in grey. Below the sequence are the structural features of the domain, with arrows representing β-sheets and boxes indicating α-helices. Above the amino acids sequence are the changes associated with egl-38 alleles

24 egl-38 ALLELES AFFECTED SYSTEM EGL-38 FUNCTION + sy287 gu22 sy294 n578 wt G9S M29I G33V G69E Egg-laying system egg-laying ability vulf morph male spicule Hindgut lin-48::gfp ex. duct Excretory system viability lin-48::gfp Figure 1.2. Different hypomorphic alleles of egl-38 disrupt different functions. All values for scoring functions are a percent of the wild type function. +++ indicates that function =/> 80% is retained. ++ means that 50-79% of the function is retained. + means that 10-49% of the function is retained. indicates that 0-9% of the function is present. Data are summarized from Chamberlin et al 1997, 1999 and Zhang et al

25 Figure 1.3. The egg-laying system in C. elegans. In the early third larval stage, six Vulval Precursor Cells (VPC; P3.p-P8.p) lie in the ventral mid-section of the body below the Anchor Cell (AC). The AC induces three of the six VPC to adopt 1 and 2 vulval fates. These cells undergo cell division resulting in cells that will form the vulva. In the mid-late L4 stage (ml4), four of the cells produced from the 1 lineage lie in each lateral plane. They include the dorsal vulf cells (F) and the ventral vule cells (E). The vulf cells separate in the ml4 stage to reveal and thin membrane between the uterus and the vulva. The AC also induces uterine cells dorsal to the vulva to adopt a π cell fate. The π cells closest to the vulva are specified as uv1 cells ( uv1 ). In the ml4 stage these cells move to either side of the vulva. Two uv1 cells can be seen in each lateral plane lying adjacent to vulf cells. Meanwhile the other π cells and the AC move away from the vulva and ultimately fuse to form the utse syncitium ( utse ). By the ll4 stage, there is a syncitium of four π nuclei in each lateral plane. In one of the lateral planes, however, the nuclei of the AC can also be seen. Therefore in this plane, there are five nuclei in the utse syncitium. 13

26 D Gonad A V P AC P3.p P4.p P5.p P6.p P7.p P8.p L3 stage E E F F F F E E Early L4 stage Lateral plane uv1 F F E E uv1 UTERUS VULVA Mid plane AC 2 2 π cells Mid-L4 stage uv1 E F E F uv1 utse utse Thin membrane 2 2 utse utse Late L4 stage uv1 F F E E uv

27 CHAPTER 2 CHARACTERIZATION OF THE ROLE OF EGL-38 IN THE EGG-LAYING SYSTEM USING MOLECULAR MARKERS 2.1. INTRODUCTION The first allele of the Pax2/5/8 gene in C. elegans, egl-38 was isolated in a screen for egg-laying defective (Egl-) mutants (Trent et al., 1983). This allele n578 was used to characterize the roles of egl-38 in the development of the cells in both the vulva and the uterine of the egg-laying system using morphological analysis and a limited number of molecular markers (Chamberlin et al., 1997; Chang et al., 1999). In this Chapter, a much broader characterization of the cells affected in the egl-38(n578) using molecular markers is described. Furthermore, the egg-laying system has been analyzed in other egl-38 mutants that retain different levels of egg-laying ability compared to egl-38(n578). These studies help us to understand the roles of egl-38 during the development of the egg-laying system. 15

28 EGL-38 and the developing egg-laying system in C. elegans The egg-laying system in C. elegans comprises of two tissues, the uterus where fertilized eggs are stored and the vulva which that forms the opening through which eggs and sperm can enter and exit the worm (reviewed by Greenwald, 1997). This system is a good model to study organogenesis. The pattern of cell fates in this system and many of the signaling pathways involved have been well studied (reviewed in Kornfeld 1997). The development and the final anatomy of the egg-laying system are described in detail in the Section of Chapter 1. EGL-38 plays an important role in the development of cells at the uterine-vulval junction during the fourth larval stage. EGL-38 is a member of the PAX 2/5/8 subfamily of transcription factors (Chi and Epstein, 2002; Refer Section of Chapter 1). The different alleles of egl-38 were isolated in different genetic screens (Trent et al., 1983). The s1775 allele is lethal. The other alleles of egl-38 disrupt different functions. For example, the n578 allele of egl-38 most severely affects the egg-laying system, while the sy287 allele preferentially disrupts the male tail development (Chamberlin et al., 1997; Fig 1.2). egl-38(n578) animals are egg-laying defective. The eggs hatch within the hermaphrodite which then bursts to release viable progeny (Trent et al., 1983). Studying the egg-laying system further in these animals has enabled us to understand developmental processes that are disrupted in the egl-38(n578) animals during the fourth larval stage to form a connection between the uterus and the vulva. Previous studies have shown that EGL-38 influences processes in specifically the 1 vulval vulf cells and the π 16

29 cells, uv1 that are found at the uterine-vulval (uv) junction (Chamberlin et al., 1997; Chang et al., 1999; Fig 1.3). In the vulval vulf cells, genetic, morphological and cell ablation studies have shown that EGL-38 is necessary for the transcription of the Epidermal Growth Factor (EGF) signal LIN-3 and for the separation of the vulf cells revealing a thin membrane at the uv-junction (Chamberlin et al., 1997; Chang et al., 1999). In egl-38(n578) animals there is no LIN-3 expression in the vulf cells during the fourth larval stage (Chang et al., 1999). Further, the vulf cells appear to be mispositioned close to each other. As a result, there is no thin membrane seen at the uv-junction (Chamberlin et al., 1999). In the uterus of egl-38(n578) mutants, the uv1 cells are not specified. Therefore they fail to remain adjacent to the vulva and move away like the other π cells, the utse (Chamberlin et al., 1997). Previous results suggested that EGL-38 plays an indirect role in the specification of the uv1 cells. Specifically, a gain-of-function of mutant of the LIN- 3 receptor gene, let-23(sa62), can bypass the need for egl-38 and specify uv1 cells in the egl-38(n578) mutant background (Chang et al., 1999). The main goal of the studies reported in chapter is to understand further the role of EGL-38 in the development of the uv-junction. In this regard, I wanted to examine three main questions. First, I wanted to learn if the vulf cells are merely abnormal in egl- 38(n578) or if they get transformed into the other 1 vulval cells, vule. Secondly in the uterus, do the uv1 cells physically move away from the vulva or do they adopt the utse cell fate and fuse with the utse cells to form a syncitium? Finally, I compared the defects seen in the egl-38(n578) egg-laying system to those of other egl-38 alleles. Thereby, by 17

30 studying the different roles of EGL-38 in sensitive backgrounds, I wanted to determine if these roles are dependent on each other or not. In order to answer these questions, I mainly used vulval and uterine molecular markers in different egl-38 backgrounds. Examining the various roles of EGL-38 in the egg-laying system will help in the understanding of how a major transcription factor can regulate the development of individual tissues like the vulva and the uterus. It will also aid in understanding how individual tissue development is coordinated to form an organ system. Finally, studying biology at the level of a single cell in C. elegans can extend understanding of organogenesis in general MATERIALS AND METHODS Strains used The C. elegans strains were cultured using standard techniques (Sulston and Hodgkin, 1988). Mutations are described in wormbase ( or as listed below. The following mutations were used: Linkage Group II (LGII): let-23(sa62) (Katz et al., 1996); LGIII: unc-119(e2498); LGIV: dpy-20(e1282), egl-38(sy294), egl- 38(sy287), egl-38(n578), egl-38(gu22) (Zhang et al., 2005, and references therein); LGV: him-5(e1490); LGX: lin-2(e1309). The following transgenes were used: inis179 (pida-1::gfp, Zahn et al, 2001), syex598 (lin-11::gfp, Gupta and Sternberg, 2002), syex724 (bam-2::gfp, Colavita et al 2003; Inoue et al., 2005), syis67 (zmp-1::pes-10::cfp; Inoue et al., 2002), syex234 (let- 18

31 23::gfp; Chang et al., 1999), kuis29 (egl-13::gfp, Hanna-Rose and Han, 1999), jcex44 (ajm-1::gfp, Koppen et al., 2001) and guex1000 (lin-3::gfp, Hwang and Sternberg, 2004) Phenotypic analyses Developmental stages: Hermaphrodite animals were observed at the adult (A) stage or at different fourth larval stages (L4): early L4 (el4), mid L4 (ml4) and late L4 (ll4). These developmental stages were defined on the basis of the vulval morphology in the mid-plane under a compound microscope (reviewed by Greenwald, 1997). During the el4 stage, the anchor cell has invaded the vulva between the vulf cells. During the ml4 stage, the vulf cells separate, leaving a thin membrane between them. During the ll4 stage, the vulva begins to evert, and the thin membrane is no longer seen between the vulf cells. Finally in an adult worm, the vulva has undergone complete eversion. Adherens junctions (AJs): AJs were visualized in the vulval tissue using an ajm- 1::gfp transgene (Koppen et al., 2001). The AJs between cells appear as spots at cell boundaries. Egg-laying ability: ml4 stage worms were placed on separate plates. These plates were checked for presence of eggs laid by the worm once a day for three days or until the worm burst. Worms which laid any eggs at all on the plate were categorized as Egl+ and those that did not were classified as Egl-. Vulval morphology: ml4 larvae were observed under Nomarski optics to determine whether the vulf nuclei are positioned anterior and posterior, leaving behind a thin laminar process between the vulva and uterus. In egl-38(n578) mutants the 19

32 vulval nuclei separate to a lesser extent, and the tissue between the vulva and uterus remains thick (Chamberlin et al., 1997). Uterine morphology: To assess uterine cell morphology, individual ml4 animals were selected using Nomarski optics, recovered for three hours and observed again during the ll4 stage. The uterine morphology was scored according to previous studies (Newman et al., 1996). In wild-type animals during the ml4 stage, the π cell progeny that become the uv1 cells (VT4 and VT8) move to flank the vulf cells of the vulva (Fig. 1. 3). The other π cell progeny migrate with the Anchor Cell (AC) to either the anterior or the posterior end of the uterus. In the ll4 stage, the uv1 cells continue to remain close the vulva and now lie to the anterior and posterior side of the vulf cells. At this stage the other π cell progeny move near the extreme ends of the uterus and fuse to become the syncytial uterine seam cell (utse). In egl-38(n578) mutants, the uv1 cells lie above the vulf cells in the el4 stage as in wild-type animals. However, in the ml4 stage, the uv1 cells migrate distally with the other π cell progeny and the AC (Chamberlin et al., 1997). Assay for uv1 specification using egl-13::gfp: egl-13::gfp is localized to the nucleus of all π cells in the uterus (Hanna-Rose and Han, 1999). Thus it serves as a useful marker to determine the position and thus the fate of the different π cells during the mid and late L4 stages. In wild-type animals, two uv1 cells per side (left/right) are positioned next to the vulf cells, while the other four π cells position away from the vulva (Newman et al., 1996). In one of the lateral planes, the Anchor Cell (AC) also migrates with the four π cells that move away from the vulva. Thus, per side, wild-type animals have two proximal and four or five distal egl-13::gfp-positive cells, whereas in egl-38(n578) mutants all cells are distal. 20

33 2.3. RESULTS Molecular markers reveal defects in the egg-laying system of egl-38(n578) animals Previous studies showed that egl-38 mutants fail to specify the uv1 cells of the uterus, and are abnormal for morphogenesis and expression of LIN-3 in the vulf cells of the vulva (Chamberlin et al., 1997; Chang et al., 1999). To further characterize the functions of egl-38, we have assessed the expression of additional molecular markers for the vulval and uterine cells of the egg-laying system in mutants. We find that although egl-38 mutants have defects in the uv1 and the vulf cells, these cells are not completely abnormal. The vulval cells derive from the ventral hypodermal cells, P5.p P7.p (Fig. 1.3). These cells divide to produce 22 cells that invaginate during the L4 stage and form an inverted U shaped structure (Sulston and Horvitz, 1977). The cells at the top of this inverted U (the vulf cells) move apart to form an open "apex" in the mid L4 stage, leaving behind only a thin membrane necessary to form a functional uterine-vulval connection (Hanna-Rose and Han, 2002). Subsequently, cells of similar type fuse, forming rings that line the vulval cavity (Sharma-Kishore et., 1999). At this stage the number of cells and the architecture of the uterine-vulval connection can be visualized using the adherens junction marker AJM-1::GFP (Koppen et al., 2001). When observed in the mid-plane, AJM-1::GFP appears as dots between cells. In both egl-38(n578) animals and their wild type counterparts, there are eight dots corresponding to seven 21

34 distinct vulval cells (Fig 2.1, line 1). This result suggests that although vulf is abnormal, there are no fusions between the vule and vulf cells. However, the four vulval nuclei closest to the uterus are mispositioned (Chamberlin et al., 1997), and this leads to an abnormal arrangement of the AJM-1::GFP dots. Consequently the apex of the vulval cavity near the uterus is essentially blocked. There are two vulval cell types that derive from P6.p: the vulf cells, which lie closest to the uterus and the vule cells which are found ventral to the vulf cells and thus away from the uterus (Fig. 1.3). These two cell types are related by lineage, and also overlap in developmental potential, since if the vulf precursors are ablated then the vule cells can adopt features of the vulf cell fate (Chang et al., 1999). vulf and vule cells differ with respect to expression of several cell markers (Inoue et al., 2002). In particular, LIN-3/EGF is expressed by vulf to signal to the somatic gonad, but it also serves as a vulf-specific cell marker which is not expressed in egl-38(n578) animals (Fig 2.1, line 2; Chang et al., 1999). We wanted to determine whether the presumptive vulf cells in egl-38 mutants are simply abnormal, or instead adopt a fate similar to vule. Our experiments with AJM-1::GFP (above) showed that the vule and vulf cells do not fuse together in egl-38 mutants. This provides evidence that the two cell types retain some differences, and that the presumptive vulf cells are not transformed into vule cells. We used additional reporter transgenes to better characterize the vulf defects in egl-38 mutants. The marker bam-2::gfp is normally expressed brightly in vulf cells and dimly in vule cells in wild-type animals (Inoue et al., 2005) and this expression remains unaltered in egl-38(n578) mutants (Fig 2.1, line 3). Similarly, zmp-1::gfp is normally restricted to vule cells (Inoue et al., 2002), and we find that this pattern is retained in egl- 22

35 38(n578) mutants (Fig. 2.1, line 4). This result is in contrast to that obtained by Inoue et al, who report that zmp-1::gfp is expressed ectopically in the vulf cells of egl-38(n578) mutants (Inoue et al., 2005). However, our experiments utilize transgenes containing a different extent of gene regulatory sequences. One interpretation of the different results is that the transgene utilized by Inoue et al. (syis49) includes negative regulatory elements absent from the transgene used in this study (syis67). The two transgenes dissect the behavior of zmp-1::gfp in response to egl-38, but the result with syis67 indicates that vulf and vule cells are different in egl-38 mutants. Altogether, we find that the vulf cells in egl-38(n578) animals are distinct from vule cells, and that all cellspecific gene expression is not altered. Thus egl-38 affects a subset of genes important for normal vulf cell fate. We next utilized molecular markers to examine the role of egl-38 in the other tissue of the egg-laying system, the uterus. In the mid-l4 stage, the π cells in the uterus that lie closest to the vulva become the uv1 cells. The remaining π cells migrate away and fuse to form the utse syncytium. In egl-38(n578) animals, the uv1 nuclei fail to remain proximal to the vulva. Instead they lie close to the distal π cell nuclei, and have been interpreted to adopt the utse fate (Chamberlin et al., 1997). To confirm this interpretation, we used lin-11::gfp to study π cell migration and fusion (Newman et al., 1999). lin-11:gfp is expressed in all π cells, but expression in uv1 cells persists longer and more intensely than that in the utse. Using this marker, we found that the presumptive uv1 cells still express lin-11::gfp as in wild type, but migrate and physically fuse with the distal utse syncytium resulting in a fluorescent arch in the egl-38(n578) mutant animals (Fig. 2.2). We also confirmed that π cell nuclei are present in their normal 23

36 numbers in egl-38(n578) mutants using a different π cell fate marker, egl-13::gfp (Fig. 2.2; Cinar et al., 2003). Finally, we found that expression of a uv1 cell specific marker, ida-1:: gfp, is essentially absent in egl-38 mutants (Fig. 2.2; Zahn et al., 2001). These studies confirm that π cell number and fate is maintained in egl-38(n578) mutants, but that there are specific defects in the uv1 cells. As a consequence, the presumptive uv1 cells fuse with the other π cell nuclei to form the utse Effects of hypomorphic egl-38 alleles on the egg-laying system The function of egl-38 in the egg-laying system has been characterized primarily using a single allele, egl-38(n578). egl-38(n578) was originally identified in a screen for egg-laying defective or Egl- animals (Trent et al., 1983). Subsequently, several additional alleles of egl-38 have been identified based on other defects (Fig. 1.2). Genetic studies show that the different homozygous viable alleles disrupt different egl-38 functions in a tissue-preferential manner (Zhang et al., 2005). Of all the egl-38 alleles, the n578 allele was found to confer the greatest egg-laying defect, and the functions of egl-38 in the egglaying system have been interpreted based on analysis of egl-38(n578) mutants. However, egl-38(n578) retains significant levels of activity in other tissues that require egl-38 function. Therefore, we wanted to assess the egg-laying system defects associated with different egl-38 mutant alleles to determine whether the egl-38(n578) phenotypes reflect the full range of egl-38 functions in this organ. In addition, these studies enabled us to examine whether these different functions are genetically linked to each other or separable, and to what extent they contribute to the egg-laying ability of the animal. 24

37 In egl-38(n578) mutants the vulf cells are defective for two functions: they neither separate properly nor express a vulf marker, LIN-3 (Chamberlin at al., 1997; Chang et al., 1999). In contrast, the other hypomorphic alleles confer varying levels of defect for each (Fig.2.3). Notably, we found that in egl-38(sy294) mutants, the vulf cells do not express detectable lin-3::gfp. However, in a majority of animals the vulf cells do separate as in the wild type. This result suggests that different egl-38 alleles can preferentially disrupt different functions within the same cells of a tissue. However, since no mutant shows the reciprocal defect (significant disruption of the vulval separation but not the lin-3::gfp expression), it is possible that the phenotype reflects differential sensitivity of the vulf functions to overall egl-38 activity. egl-38(n578) mutants fail to express LIN-3 in the vulf cells in the vulva and lack uv1 cells in the uterus. Several previous results suggested that the uterine cell defect in egl-38 mutants is a secondary effect of the LIN-3 expression defect. In particular, killing the vulf precursor cells results in failure of uv1 cells to be specified, and mutants in the lin-3/let-23 pathway are disrupted for uv1 cell fate specification (Chang et al., 1999). A prediction of this model is that LIN-3 expression should correlate with uv1 cell specification in other egl-38 mutants with greater levels of egg-laying ability than egl- 38(n578). To test this prediction, we assessed lin-3::gfp (to report LIN-3 expression) and lin-11::gfp (to track uv1 cell fate) in the different egl-38 mutants. Our results show that although there is a correlation between these two functions in most of the egl-38 mutants, it is not universal. Interestingly, in egl-38(sy294) mutants, almost all the animals have normal uv1 cell specification even though no lin-3::gfp expression is detected in the vulf cells (Fig 2.3). This result may reflect the possibility that the reporter gene does not 25

38 sufficiently report lin-3 expression, or that limited amounts of LIN-3 are adequate for normal uv1 fate (Dutt et al., 2004). However, lin-3::gfp expression and uv1 cell fate are well correlated in egl-38(sy287) and egl-38(gu22) animals. For example, egl-38(sy287) mutants exhibit a greater defect in uv1 cell specification than egl-38(sy294), but have detectable levels of lin-3::gfp. These results suggested that egl-38 may have direct functions in the uterus, a possibility we tested in Chapter 3 using genetic mosaic analysis. egl-38(n578) mutants have a defective connection between the vulva and the uterus, and therefore fertilized eggs hatch within the animal. We examined both the uterine and vulval functions among the egl-38 mutants to learn which of these functions is critical to form a functional egg-laying system and thereby confer egg-laying ability. Based on the correlation of different egg-laying system phenotypes (Fig. 2.3), we suggest that vulval morphogenesis leading to vulf cell separation plays a major role in enabling an animal to lay eggs. For example, in the egl-38(sy287) mutants all the animals examined lay eggs, but only about 60% of them express LIN-3 or specify uterine uv1 cells. Almost 90% of these animals do show vulf cell separation. We interpret from these data that the separation of the vulf cells may relieve a physical block in the formation of the uterine vulval connection and enable eggs to pass through the vulva DISCUSSION In this Chapter, I have tried to broaden the understanding of the role that EGL-38 plays in the egg-laying system. First using molecular markers, I have shown that the number of vulval cells remains unaltered in the egl-38(n578) Egl- mutant. However 26

39 subsets of vulf cell functions like LIN-3 expression are altered in egl-38(n578) mutants. For example, these animals fail to express lin-3 in the vulf cells and the vulf cells are mispositioned. Next, in the uterine tissue, the uv1 cells fail to be specified at their wild type position next to the vulva. Rather they fuse with the other π cells to form the utse as shown using the marker, lin-11::gfp. Finally studying the system in various egl-38 mutants suggests that different egl-38 activities contribute differently to the egg-laying function The vulf and the uv1 cells During the development of the uterine-vulval connection in C. elegans, the vulf cells of the vulva and the uv1 cells of the uterus form the connection between the two tissues of the organ, and cell signaling establishes the fate of each. Our studies of the development of these two cells in egl-38 mutant animals dissect features that underlie each cell fate. In the egl-38(n578) mutants, the correct number of presumptive vulf cells is produced and they lie in their appropriate positions. They also acquire a cell fate distinct from those of the other cell type derived from the 1 VPC (the vule cells), as they express bam-2::gfp at higher levels than vule cells, as do wild-type vulf cells. However, the presumptive vulf cells are abnormal in egl-38(n578) mutants, as they fail to express lin-3::gfp and fail to undergo proper morphogenesis. Therefore, only subsets of features that define the vulf cell fate are dependent on EGL-38. In the uterus, the uv1 cells are defined both by the expression of markers and cell position. We have found that in egl-38 mutants, the presumptive uv1 cell expresses π cell markers like egl-13::gfp and lin-11::gfp, but not the uv1-specific marker, ida-1::gfp. 27

40 The presumptive uv1 cells also migrate near the other π cells, and fuse with them to form the utse. We speculate that normal uv1 cell fate involves cell-cell adhesion or another process that anchors the cells at the vulva, and that egl-38 is normally required for this function Hypomorphic alleles of egl-38 expand the understanding of the role of egl-38 in the egg-laying system The hypomorphic alleles of egl-38 preferentially disrupt functions in different tissues (Chamberlin et al., 1997; Zhang et al., 2005). In this Chapter, the ability of these hypomorphic egl-38 alleles to affect the different aspects of egg-laying system development was examined. Interestingly, these studies led to the understanding that besides merely causing tissue-preferential defects, these alleles also differentially influence subsets of egl-38 functions within a single tissue. For example, in the vulf cells of the vulva, egl-38(sy294) mutants have no LIN-3 expression, but do show vulf cell separation. This allows us to learn about multiple independent roles for egl-38 within the cells of a single developing tissue. To understand tissue-specific phenotypes of a transcription factor like EGL-38 Zhang et al, identified high affinity binding sites for the hypomorphic egl-38 mutant proteins using in vitro selection. The high affinity binding sites discovered by in vitro selection however did not function properly in vivo (Zhang et al., 2005). Therefore, two models were proposed for such tissue-preferential behavior. The first suggested that different mutant alleles could differentially bind to the target DNA sequences, while the second invoked the presence of other proteins to influence DNA target selection. 28

41 In such context, it would be interesting to learn the method by which egl-38 affects different activities within a single cell of a tissue. Perhaps either or both the mechanisms mentioned above might be involved.. Preliminary studies done so far by us have not lead to the identification of specific DNA sequences on the lin-3 promoter which are sensitive to the egl-38(n578) mutation. For the second activity of vulf separation, no genes downstream of egl-38 are yet known. Perhaps, the suppressor genes isolated in Chapter 4 could be involved in specifically mediating vulf separation and not affecting LIN-3 expression (Refer Chapter 4). Also, within the same cell, egl-38 could mediate direct and indirect roles. 29

42 TRANSGENE wild type GENOTYPE egl-38(n578) 1. ajm-1::gfp ape ape 8 4 ape ape F F F F F F F F 2. lin-3::gfp 3. bam-2::gfp F E F E F E F E F E F E F E F E 4. zmp-1::gfp F E F E F E F E F E F E F E F E Figure 2.1. Expression of molecular markers in the vulva of egl-38(n578) mutants. Each image set has Nomarski images to the left and its corresponding epi-fluorescence image to the right. At least twenty animals were observed for each genotype and transgene. 1) Adherens junctions (AJs) at the uterine-vulval junction were observed in ml4 animals as dots between cells in the mid-plane using AJM-1::GFP. The AJs in the vulva are numbered 1 to 8. The first AJ is between the vula cell and hyp7, the next six are AJs between vulval cells, while the eighth lies between the vulf cell and the uterine uv1 cell. The number of AJs is normal in egl-38(n578) mutants. The opening between AJs of the vulf cell and the uv1 cell is called the apex (denoted by the dotted line). The apex is wide in wild-type animals, but narrow in egl-38(n578) animals. 2) lin-3::gfp is expressed in vulf cells of wild-type animals and absent in egl-38(n578) mutants. 3) bam-2::gfp is expressed brightly in the vulf cells and dimly in the vule cells in both wild type and egl-38(n578) mutants. 4) zmp-1::gfp is expressed in the vule cells in both wild type and egl-38(n578) mutants. 30

43 Figure 2.2. Expression of molecular markers in the uterus of egl-38(n578) mutants. Each image set has Nomarski images to the left and its corresponding epi-fluorescence image to the right. lin-11::gfp is expressed brightly in uv1 cells, and less brightly in utse cells. In the el4 stage, all the π cells are present close to and above the vulva in both the wild type and the egl-38(n578) animals. In the wild type animals, the uv1 cells are close to the vulva (within the brackets) throughout the L4 stage. The other utse cells migrate distally and fuse to form a syncitium. In egl-38(n578) animals the uv1 cells are close to vulva in el4. By the el4 stage they have moved away from the vulva. In ll4 stage, uv1 cells are not seen near the vulva. Instead the utse syncitium appears brighter, suggesting fusion of the presumptive uv1 cells to the utse. Fifteen animals were examined in each stage. ida-1::gfp is expressed in uv1 cells. In wild-type animals, the uv1 cells express ida-1::gfp and lie close to the vulva (within the brackets) in both the ml4 and the ll4 stages. In egl-38(n578) animals there is no detected ida-1::gfp expression in the ml4 stage. In the ll4 stage, faint GFP is detected in the utse. Ten animals were examined in each stage. 31

44 TRANSGENE STAGE Wild type GENOTYPE n578 ut ut early-l4 ut ut ut ut lin-11::gfp mid-l4 ut ut ut ut late-l4 ut ut mid-l4 ida-1::gfp ut late-l4 ut 32

45 PERCENT WILD-TYPE lin-3 uv1 vulval egg-laying 0 + sy2 gu sy2 n5 egl-38 GENOTYPE Figure 2.3. lin-3 expression, uv1 cell specification, vulval separation and egg-laying ability of the different egl-38 mutant animals. In this graph, the different phenotypes of various non-null egl-38 mutants are plotted against the percentage of animals exhibiting wild-type levels of each phenotype. lin-3 expression in the vulf cells during ml4 was assessed using a lin-3::gfp reporter transgene (see Fig. 2). uv1 cell specification was assessed using a lin-11::gfp reporter in the ll4 stage. Vulval separation was scored based on vulf cell separation observed in the mid-plane. At least twenty animals were observed for each genotype and phenotype. 33

46 CHAPTER 3 SITE OF ACTION STUDY FOR EGL-38 IN GERMLINE CELL DEATH AND THE EGG-LAYING SYSTEM DEVELOPMENT 3.1. INTRODUCTION Studies done previously and those in Chapter 2 have demonstrated that the viable egl-38 alleles disrupt functions in different tissues (Chamberlin et al., 1997; Zhang et al., 2005; Rajakumar and Chamberlin, 2006). Perhaps each of these alleles either work in a combinatorial manner with other transcription factors and co-factors or have different DNA binding affinities to effect their tissue-preferential activity. Learning the site of action of the EGL-38 activity might help in understanding in which cells it acts to influence a biological process and in delineating the signaling pathway involved therein. In this Chapter, mosaic analysis is employed to learn where EGL-38 acts to impact two processes programmed cell death in the germline and the development of the egglaying system egl-38(n578) mutants are defective in multiple biological processes EGL-38 belongs to the PAX2/5/8 class of transcription factors. This 34

47 class is characterized by the presence of the Paired Domain (PD). This domain is bipartite in structure. Each domain can bind to two half-sites on two major grooves of the DNA respectively. The affinity of this interaction is dependent on both domains of the protein (Czerny et al., 1993; reviewed in Chi and Epstein, 2002). The viable alleles of egl-38 each have missense mutations. The hypomorphic alleles of egl-38 disrupt functions in different tissues. Interestingly, egl-38(n578) affects different functions in different tissues. Recent studies done by Park et al, have revealed that egl-38(n578) disrupts both somatic and germline cell death (Park et al., 2006). Additionally, egl- 38(n578) mutants have a defective egg-laying system. In the germline and the soma, the two Pax2/5/8-related genes in C. elegans, egl- 38 and pax-2 influence Programmed Cell Death (PCD or Apoptosis). Both egl-38(n578) and pax-2(ok935) mutants show elevated levels of both germline and somatic cell death. Further experiments have uncovered that these Pax genes mediate their activity by transcriptionally regulating the bcl-2 gene in C. elegans called ced-9. ced-9 is a member of the core apoptotic pathway (Reviewed in Twomey and McCarthy, 2005). These studies show that the function of Pax genes in promoting cell survival is evolutionarily conserved (Refer Chapter 1 for the role of Pax genes in cell survival in higher systems). Furthermore, they extend the understanding of how cell survival signals are integrated into the core apoptotic pathway through the Pax genes. However, in order to understand this mechanism in its entirety, it is necessary to understand the site of action of EGL-38. The other biological process disrupted in the egl-38(n578) mutants is the development of the egg-laying system. As discussed in Chapter 2, these mutants have a defective uterine-vulval (uv) connection. In the vulva, the vulf cells are mispositioned 35

48 and fail to express the EGF signal, LIN-3, while in the uterus, the uv1 cells fail to be specified and fuse with the lateral utse cells rather than lying adjacent to the vulva (Chamberlin et al., 1997; Chang et al., 1999; Zhang et al., 2005; Rajakumar and Chamberlin, 2006; Refer to Chapter 1 and Chapter 2 for further information on the role of egl-38 in the egg-laying system). During the development of the egg-laying system, studies done by Chang et al suggested that EGL-38 in the vulva indirectly regulates the specification of the uterine uv1 cells by regulating LIN-3. Therefore, a gain-of-function mutation in the LIN-3 receptor, LET-23, could bypass the uv1 specification defect in the egl-38(n578) mutants (Chang et al. 1999). However, while analyzing various egg-laying system defects in the hypomorphic alleles of egl-38, most animals with the sy294 allele had no LIN-3 expression, but did show uv1 cell specification. This result suggested that perhaps there was a LIN-3-independent role for EGL-38 in specifying the uv1 cells in the uterus (Rajakumar and Chamberlin, 2006). To resolve whether EGL-38 specifies the uv1 cells in a vulva-dependent or independent manner, it is necessary to learn the tissue source for EGL-38 activity in the egg-laying system Genetic mosaic analysis dissect the site for action of egl-38 To learn the site of action of EGL-38 activity both in the process of PCD and in the development of the egg-laying system, I used genetic mosaic analysis. To generate mosaic animals, I used a strain homozygous for egl-38(n578) and bearing an extra chromosomal transgene with egl-38(+) rescuing DNA and a broadly expressed marker (sur-5::gfp) (Yochem et al., 1998). This transgene can be lost spontaneously in cell division, resulting in mosaic animals. In these animals cells retaining the transgene have 36

49 functional egl-38 and express SUR-5::GFP, whereas cells that lack the transgene are mutant for egl-38 and lack SUR-5::GFP. Therefore, it is possible to determine in what cells egl-38 functions. In this Chapter, I want to determine the site of action of egl-38 gene activity in the biological processes which are disrupted in egl-38(n578) mutant animals germline celldeath and the development of the uv-connection in the egg-laying system. Firstly in the process of PCD, I was interested in learning which tissues were the sources of the germline cell survival signals promoted by the egl-38 gene. Studies done by Park et al suggested that the Pax2/5/8 genes, egl-38 and pax-2, contribute to germline cell survival by regulating the transcription of the bcl-2 gene, ced-9, a member of the core apoptotic pathway (Park et al., 2006). I performed genetic mosaic analysis to determine whether egl-38acts within the germline to promote germline cell survival, as would be predicted by the work of Park et al. (2006). Similarly, I am interested in learning if egl-38 acts merely in the vulval vulf cells to trigger EGF/LIN-3 signal expression to promote uv1 cell specification in the uterus, or if egl-38 also has an autonomous role in the uv1 cells. Studies done in mammals have shown that PAX transcription factors play a role in promoting cell survival and thereby influence cell death. During kidney development in mice, renal hypoplasia is seen when there is a high Pax2 gene expression (Dressler et al., 1993). Further, altering the expression of these genes raises levels of apoptosis (Muratovska et al., 2003; Buttiglieri et al., 2004; Wu et al., 2005). In both primary cancer cells and cancer cell lines, there is inappropriate expression of Pax genes. Uncovering the tissue-source of egl-38 activity will enable us to extend the understanding of the mechanism by which PCD occurs in the C. elegans germline. This could lend 37

50 insight into whether the cell survival promoted by Pax genes in mammals occurs by an autonomous or non-autonomous manner. Pax2/5/8 genes have been shown to play an important role in the development of organ systems. For example, egl-38 plays important roles during the development of the egg-laying system in C. elegans. Also during the development of the metanephric kidney in mouse, Pax2 is expressed in both the metanephric mesenchyme and the uteric bud. It is also required for reciprocal inductive interactions in both these tissues that promote and coordinate organogenesis (Majumdar et al., 2003; Torres et al., 1995; reviewed by Dressler, 1996). In addition, within the metanephric mesenchyme Pax2 promotes the mesenchymal-to-epithelial transition and subsequently promotes formation of the nephrons (Dressler et al., 1993; Grote et al., 2006; Rothenpieler and Dressler, 1993; reviewed by Schedl and Hastie, 2000). These results suggest that Pax proteins can play multiple roles in different tissues of an organ to coordinate organogenesis. Mosaic analysis for egl-38 activity in the egg-laying system will help us learn to what extent developmental logic is conserved among the Pax2/5/8 genes MATERIALS AND METHODS Strains The C. elegans strains used were cultured using standard techniques (Sulston and Hodgkin, 1988). Mutations are described in wormbase ( or are listed below. The following mutations were used: 38

51 Linkage Group II (LGII): let-23(sa62) (Katz et al., 1996); LGIV: dpy-20(e1282), egl- 38(n578) (Trent et al., 1983; Chamberlin et al., 1997). The following transgenes were used: syex234 (let-23::gfp, (Chang et al., 1999), kuis29 (egl-13::gfp, Hanna-Rose and Han, 1999) and the transgene within strain UL1251 (egl-38::gfp, John Reece-Hoyes and Ian Hope, personal communication; Dupuy et al., 2004) Phenotypic analyses Developmental stages: Hermaphrodite animals were observed during the Adult (A) and at two different points of the fourth larval stages (L4): mid L4 (ml4) and late L4 (ll4). These developmental stages were defined on the basis of the vulval morphology in the mid-plane under a compound microscope (reviewed by Greenwald, 1997). During the ml4 stage, the vulf cells separate, leaving a thin membrane between them. During the ll4 stage, the vulva begins to evert, and the thin membrane is no longer seen between the vulf cells. Finally in an adult worm, the vulva has undergone complete eversion. Vulval morphology: ml4 larvae were observed under Nomarski optics to determine whether the vulf nuclei are positioned anterior and posterior, leaving behind a thin laminar process between the vulva and uterus. In egl-38(n578) mutants the vulval nuclei separate to a lesser extent, and the tissue between the vulva and uterus remains thick (Chamberlin et al., 1997). Uterine morphology: To assess uterine cell morphology, individual ml4 animals were selected using Nomarski optics, recovered for three hours and observed again during the ll4 stage. The uterine morphology was scored according to previous studies 39

52 (Newman et al., 1996). In wild-type animals during the ml4 stage, the π cell progeny that become the uv1 cells move to flank the vulf cells of the vulva (Fig. 1). The other π cell progeny migrate with the Anchor Cell (AC) to either the anterior or the posterior end of the uterus. In the ll4 stage, the uv1 cells continue to remain close the vulva and now lie to the anterior and posterior side of the vulf cells. At this stage the other π cell progeny move near the extreme ends of the uterus and fuse to become the syncytial uterine seam cell (utse). In egl-38(n578) mutants, the uv1 cells lie above the vulf cells in the el4 stage as in wild-type animals. However, in the ml4 stage, the uv1 cells migrate distally with the other π cell progeny and the AC (Chamberlin et al., 1997). Assay for uv1 specification using egl-13::gfp: egl-13::gfp is localized to the nucleus of all π cells in the uterus (Hanna-Rose and Han, 1999). Thus it serves as a useful marker to determine the position and thus the fate of the different π cells during the mid and late L4 stages. In wild-type animals, two uv1 cells per side (left/right) are positioned next to the vulf cells, while the other four π cells position away from the vulva (Newman et al., 1996). In one of the lateral planes, the Anchor Cell (AC) also migrates with the four π cells that move away from the vulva. Thus, per side, wild-type animals have two proximal and 4 or 5 distal egl-13::gfp-positive cells, whereas in egl- 38(n578) mutants all cells are distal. LET-23 expression pattern in the uterine cells: LET-23 expression was examined using a LET-23::GFP reporter transgene (Chang et al., 1999). Animals were observed at three time points in the L4 (early, mid and late), and in the adult. Assay for Programmed Cell Death (PCD)/ Apoptosis in the hermaphrodite germline: The assay for relative numbers of dying cells per gonad arm was performed as 40

53 described by Gumienny et al. (1999) using a vital stain that preferentially stains germ cell undergoing apoptosis. For this assay ml4 stage hermaphrodites were incubated at 20 C for 24 hours and then submerged in 33µM aqueous solution of SYTO12 for 4 hours. The worms were then recovered for 30 minutes on Agar plates. The germline of the worms was then observed using epifluorescence microscopy. In wild-type animals usually twoto-three apoptotic cells are seen (Gumienny et al., 1999; Park et al., 2006). In egl- 38(n578) animals the level of apoptosis is elevated (Park et al., 2006) Analysis of egl-38::gfp expression Two reporter genes were used to visualize egl-38::gfp expression : pxw124 which expresses in the vulval cells, and the unnamed reporter gene contained in strain UL1251, which expresses in the uterus. pxw124 includes ~ 5 kb sequence upstream of egl-38 cloned into the GFP reporter vector ppd95.69 (a gift from A. Fire). Transgenic strains of the pxw124 plasmid were generated using standard methods for germline transformation (Mello et al., 1991). The plasmid pxw124 (76 ng/ul) was injected using unc-119(+) (15ng/ul) as a transformation marker into unc-119(2498); him-5(e1490) animals. The UL1251 strain was obtained from Ian Hope (J. Reece-Hoyes and I. Hope, personal communication). This clone includes 457 bp of sequences from the upstream region of egl-38. It is notable that all of the sequences contained within the UL1251 reporter transgene are present within pxw124. Thus it is not clear why pxw124 promotes expression only in the vulva, and not both in the vulva and the uterus. We only conclude that sequences upstream of egl-38 include enhancers that can promote expression in the uterus, and in the vulva. 41

54 Mosaic analysis for egl-38 activity in the germline egl-38(n578) animals were injected with a mixture of 10ng/ul of C04G2 cosmid (egl-38 rescuing cosmid; Chamberlin et al., 1997), 15ng/ul of a cytoplasmically-localized transformation marker (myo-2:: gfp; Okkema et al., 1993) and 100ng/ul of a ubiquitously expressed nuclear marker (sur-5::gfp, Gu et al., 1998; Yochem et al., 1998). Mosaic analysis for egl-38 was done by selecting mid-l4 stage animals and scoring a panel of cells derived from different lineage precursors for presence or absence of sur-5::gfp (see Fig 3.5). These animals were recovered, allowed to mature 24 hours, and then were stained with SYTO12 and analyzed for apoptosis as described above. The offspring of each mosaic animal were subsequently assayed for GFP to determine whether the transgene was present in the germline Mosaic analysis for egl-38 activity in the egg-laying system egl-38(n578) animals were injected with a mixture of 10ng/ul of C04G2 cosmid (egl-38 rescuing cosmid; Chamberlin et al., 1997), 15ng/ul of a cytoplasmically localized transformation marker (myo-2::gfp, Okkema et al., 1993) and 100ng/ul of a ubiquitously expressed nuclear marker (sur-5::gfp, Yochem et al., 1998; Gu et al., 1998). Mosaic analysis for egl-38 was done by selecting ml4 stage transgenic animals that express myo- 2::gfp in some cells of the pharynx under the dissecting microscope. These animals were then assessed for mosaicism by examining for presence or absence of sur-5:: gfp expression in vulval nuclei using a compound microscope (Carl Zeiss). Simultaneously the vulval morphology was examined under Nomarski optics for vulf cell separation. All 42

55 the animals listed in Fig 3.6 as vulf(+) contained the transgene in all vulval cells (derived from P5.p-P7.p). Thirty-five of 37 animals listed as vulf(-) contained the transgene in no vulval cells. Two of 37 retained the transgene in the P5.p and P7.p-derived cells, but lacked the transgene in P6.p-derived cells (vulf and vule). These were both in mosaic category 2 (vulf(-) and uv1(-)), and were phenotypically the same in the egg-laying system as other category 2 animals. As Pn cells derive alternately from ABpl and ABpr, the most parsimonious interpretation of these animals is that that transgene was lost in ABp or ABa, although two separate losses are also possible. This scoring strategy does confirm that the transgene was absent in the P6.p precursor to the vulf cells. Disruption of the vulf separation function of egl-38 in vulf(-) animals also supports the idea that wild-type egl-38 activity is absent from these cells. These animals were recovered and examined again during ll4 stage for either presence or absence of sur-5::gfp in uterine nuclei and the specification of uv1 cells. All the animals listed in Fig 3.6 as uv1(+) contained the transgene in both uv1 cells, and those listed as uv1(-) lacked the transgene in both cells RESULTS EGL-38 acts both in the germline and the soma to promote germline cell survival Studies done in Park et al. (2006) report that EGL-38 and its other PAX2/5/8 counterpart in C. elegans, PAX2 both contribute to germline cell survival by regulating the transcription of the bcl-2 gene in C. elegans, ced-9. In order to learn the site of action of egl-38 activity that promotes germline cell survival, I completed genetic mosaic 43

56 analysis. Here, a transgene containing egl-38(+) rescuing clone, a ubiquitous nuclear marker, sur-5::gfp, and a transformation marker, myo-2::gfp, was injected into egl- 38(n578) mutants. These mutants are hypomorphs for egl-38 activity and have elevated levels of germline cell death (5.64 +/ corpses per gonad arm) compared to wild type animals (3.69 +/ corpses per gonad arm) (Fig 3.1). The caveat of using egl- 38(n578) mutants is that there will be wild type PAX2 activity and some EGL-38 activity in these animals. However, a benefit of the approach is that it provides a sensitized genetic background wherein I can specifically learn the tissue source for egl-38 activity required to promote germline cell survival. The transgenes in the egl-38(n578) mutants can be randomly lost during cell division resulting in genetically mosaic animals. Therefore the presence or absence of sur-5::gfp expression in a certain cell can give information of whether it has egl-38(+) rescuing activity or not. A range of somatic cells were thereby examined during the mid fourth larval stage for presence of the transgene. After 24 hours, levels of germline cell death were examined in the meiotic region of the gonad arm and recorded as number of corpses per gonad arm (Fig 3.1.A). Since germline expression of sur-5::gfp cannot be detected in the germline, the presence of the transgene was examined in the progeny of these animals. The entire panel of tissues and cells examined therein are listed in Figure 3.2. The results of the mosaic analysis are summarized in Figure 3.3. Animals that had the transgene in the germline and the soma had levels of cell death similar to that seen in wild-type animals (3.04 +/ corpses per gonad arm). Conversely, those that lacked the transgene in the entire animal had germline cell death that was comparable to that 44

57 seen in egl-38(578) mutants (6.93 +/ corpses per gonad arm). The ability of the transgene in the germline versus in the surrounding somatic tissues arising from the MS and the E lineage to influence germline cell survival was compared. Here, the presence of the transgene in the germline but absence in either the MS lineage or the E lineage or both resulted in wild-type levels of germline cell death (2.75 +/ corpses per gonad arm). This suggests that egl-38(+) activity is required in the germline to support germline cell survival. Interestingly, upon examining animals with rescuing transgene in the MS and the E lineage but lacking in the germline also showed a low level of cell death (4.03 +/ corpses per gonad arm). This result implies that besides playing an autonomous role in the germline, egl-38(+) activity is also required non-autonomously in the soma for germline cell survival. To simplify analysis, the different mosaic categories were classified as wild type or mutant based on if they had five or lower corpses per gonad arm or if they had six or more corpses per gonad arm. The number six represents a value which is more than two standard deviations from the average number of corpses per gonad arm seen in animals having transgene in cells of all lineages. Based on this criterion, within the somatic tissues, all six animals which lacked egl-38(+) activity in the E lineage which contributes to the intestine has wild-type levels of germline cell death. In conclusion, egl- 38(+) activity is required both in the germline and the soma to promote germline cell survival. 45

58 EGL-38 is expressed and functions in both vulval and uterine cells Our analysis of different egl-38 mutants suggested that the gene may have roles both in the vulva and the somatic gonad to mediate development of the egg-laying system (Refer Chapter 2). If this is the case, then the gene should be expressed and should function in both tissues. To characterize the expression pattern of egl-38, we utilized egl- 38::gfp reporter transgenes. Consistent with this model, we found that these transgenes express in the vulf cells of the vulva, and the uv1 cells of the uterus (Fig. 3.4). To more directly test in which cells egl-38 functions, we performed genetic mosaic analysis. We utilized the mosaic strain used in Section (see materials and methods; Yochem et al., 1998). This transgene can be lost in cell division, resulting in genetically mosaic animals. For these animals, cells retaining the transgene have functional egl-38 and express SUR-5::GFP, whereas cells that lack the transgene are mutant for egl-38 and lack SUR-5::GFP. We selected mid-l4 animals that contained the transgene in at least some cells, and observed the vulva for presence or absence of SUR- 5::GFP. At the same time we verified whether vulf separation was normal or defective. These animals were then recovered and examined in the late L4 stage for the expression of SUR-5::GFP in their uterine cells, and for whether the uv1 cells remained at the vulva, or had migrated distally with the utse cells. We found that the mosaic animals fell into four categories based on the deduced presence or the absence of egl-38(+) in the vulval and the uterine tissue (Fig. 3.6). The first category comprised of animals that had wild-type egl-38 in both tissues. As expected, these animals were found to have wild-type development. In a second category, egl-38 was absent in both the tissues, and development was abnormal in both. The final 46

59 two categories comprised of animals that were mosaic for egl-38 in either the vulva or the uterus. When there was egl-38(+) rescuing activity in the vulva and not the uterus (category 3), then correspondingly there was wild-type vulf separation but improper uv1 cell specification. The reverse was true when egl-38 activity was found in the uterus and not in the vulva (category 4). Thus we find that even with functional egl-38 activity in the vulva, disruption of egl-38 in the uterus prevented uv1 cell fate. Our studies have therefore uncovered a vulva-independent function for egl-38 in the uterus. Furthermore, in the egl-38(n578) background, egl-38(+) activity is both necessary and sufficient in each of tissues to perform the corresponding functions EGL-38 and LET-23 in the uterus Our genetic mosaic results prompted us to revisit the functional relationship between EGL-38 and the LIN-3(EGF)/LET-23(EGFR) signaling pathway. Earlier studies by Chang et al. (1999) showed that expression of LIN-3 in the vulva requires egl-38 (see Fig. 2.1, line 2). In addition, a gain-of-function mutation, let- 23(sa62), suppresses the egl-38(n578) uv1 defect, indicating that activated let-23 can bypass the requirement for egl-38 in uv1 cell development. These data are consistent with a model in which egl-38 acts in the vulva to signal to the uterus to specify uv1 cell fate. In contrast, our genetic mosaic analyses suggest that egl-38 activity in the uterus is necessary and sufficient to specify the uv1 cells. To reconcile the results obtained from these different approaches, we carried out additional analysis of the relationship between let-23 and egl-38 in the uv1 cells. 47

60 In the uterus there are twelve π cells (Newman and Sternberg, 1996). These cells initially lie above the vulva with six cells in each lateral plane. Subsequently, the four central π cells that lie adjacent to the vulva are specified to be uv1 cells. The remaining cells distal to the vulva fuse to form a syncytium called utse. To replicate the analysis of Chang et al. (Chang et al., 1999), we used a reporter, egl-13::gfp, to visualize the π nuclei. In wild-type animals, the uv1 cells are specified and these animals have two proximal and 4-5 distal egl-13::gfp-positive nuclei on each side, corresponding to two proximal uv1 nuclei and four to five distal utse nuclei (Fig 3.7). In contrast, in egl- 38(n578) animals, all nuclei lie distal to the vulva because the presumptive uv1 cells fuse with the other π cells to form the utse cell. We constructed let-23(sa62);egl-38(n578) double mutants with the egl-13::gfp reporter and found that these animals exhibit the wild-type arrangement of π nuclei, indicating that the gain-of-function let-23 allele can bypass the egl-38 defect. These experiments using a molecular marker to replicate the morphological analysis of Chang et al. (1999), supporting the idea that activated let-23 can bypass the egl-38(n578) uv1 cell development defect. Bypass of egl-38(n578) also occurs in response to induced expression of lin-3, indicating that the result does not reflect a neomorphic property of let-23(sa62) (L. Huang and W. Hanna-Rose, 2006). Since activated let-23 can bypass egl-38 to promote uv1 cell fate, we wanted to test whether egl-38 might influence the abundance or localization of LET-23 in the uv1 cells. For example, our combined results might be obtained if egl-38 influences the expression levels of both the ligand, LIN-3, in the signaling cell and the receptor, LET- 23, in the responding cell. We used a LET-23::GFP reporter previously shown to rescue the loss-of-function defects of let-23 alleles (Chang et al., 1999). We found that normally 48

61 LET-23 expression in the uv1 cells begins in the third larval stage and persists until early adulthood (Fig. 3.7.B-E). Although during the fourth larval stage LET-23::GFP expression is found to surround the entire uv1 cell, it is concentrated between the uv1 cell and the vulf cell. In the egl-38(n578) mutants we found that the LET-23::GFP pattern is normal until the middle of the fourth larval (ml4) stage. At this stage LET-23::GFP is localized around the presumptive uv1 cell. However, instead of staying adjacent to the vulf cells, these uv1 cells begin to move away from the vulva. Finally during the late fourth larval stage, there is no LET-23::GFP expression around the presumptive uv1 cells. Overall, we do not observe obvious changes in the expression dynamics both with respect to localization and in terms of levels of expression of LET-23 between wild type and egl-38(n578) mutant animals. Consequently, we conclude that egl-38 does not significantly influence the expression or localization of LET DISCUSSION In this Chapter, I have focused on the role of egl-38 in the developing egglaying and germline cell survival. To this end, I used genetic mosaic analysis to determine in which cells and tissues egl-38 function is required. With respect to germline cell death, the results show that egl-38 activity is required both in the germline tissues as well as in some somatic tissues to promote germline cell survival. Similarly in the developing egg-laying system the results support a model in which egl-38 functions in both the somatic gonad and the vulval cells to coordinate egg-laying system development. In particular, I find that functional egl-38 in the vulva is not sufficient for 49

62 normal uterine development, indicating that it influences non-autonomous as well as autonomous developmental functions within the organ. I propose that a general property for Pax2/5/8 transcription factors may be to act broadly in the different tissues with roles in coordinating organogenesis and promoting cell survival in tissues, thereby contributing to the development of the animal EGL-38 acts both in germline and somatic tissues to influence germline cell death Mosaic analysis for egl-38 in germline cell survival revealed that egl-38 can act autonomously in the germline and non-autonomously in the soma to perform this function. A simple interpretation of these results is that egl-38 triggers signaling pathways both in the germline and the soma. Survival signal from the soma feeds into the germline tissue where along with the germline pathway, cells are prevented from dying. Park et al, have shown that the Pax2/5/8 genes in C. elegans, egl-38 and pax-2, both play a role in regulating germline cell death by regulating the transcription of the bcl-2 gene, ced-9 (Park et al., 2006). In the germline, CED-9::GFP was shown to be enriched in the region where apoptosis occurs although such a pattern is not seen for CED-9 protein (Chen et al., 2000; Park et al., 2006). Assuming that these differences could be due to variations in stability and post-transcriptional regulation among the two forms of CED-9 used, perhaps egl-38 could directly regulate ced-9 gene expression at times when high levels of apoptosis occurs to reduce the number of dying cells (Park et al., 2006). Understanding how egl-38 activity coming from the somatic cells contributes to germline cell survival requires further work. Some of the unanswered questions include 50

63 learning what is the immediate downstream target of EGL-38, what is the nature of the signaling pathway that feeds into the germline to promote cell survival, and does this pathway also regulate the transcription of ced-9 in the germline. Furthermore learning that egl-38 activity from both soma and the germline contribute germline cell survival prompts the question, could the activity of the other related Pax2/5/8 gene in C. elegans, pax-2 also arises from the soma and the germline EGL-38 plays multiple roles in the development of the egg-laying system EGL-38 plays a direct role in the normal development of both the vulva and uterus of the C. elegans egg-laying system. Previous experiments suggested a model in which EGL-38 mediates development of both tissues by acting solely in the vulval cells, where it ensures proper vulval morphogenesis and promotes expression of the EGF signal, LIN-3. LIN-3 then signals to the uterus to specify the uv1 cells (Chang et al., 1999). Our studies demonstrate a role for EGL-38 in both the vulva and the uterus in coordinating the development of the egg-laying system. We find that reporter genes for egl-38 are expressed in both uterine and vulval cells, and that in genetically mosaic animals egl-38 can act cell-autonomously. Thus we propose a model in which egl-38 plays a role in cells of both tissues to coordinate development of the organ (Fig 3.8). In the vulva, egl-38 acts to both promote normal morphogenesis and to promote expression of LIN-3, which signals to the uterus. egl-38 also acts in the uterus, and is necessary for prospective uv1 cells to respond to the LIN-3 signal. We favor this model as it integrates our results with earlier studies. We speculate that the uv1 cells develop normally in mosaic animals in which egl-38 is mutant in the vulva either because they can respond to 51

64 LIN-3 expressed elsewhere in the animal, or there is trace LIN-3 expressed by the vulva cells in egl-38(n578) mutants that is not detectable using the lin-3::gfp reporter transgene. Overall, we interpret egl-38 to have distinct roles in each tissue of the egg-laying organ, and to promote appropriate cellular development as well as signaling to coordinate development between the tissues EGL-38 influences both signaling and responding during egg-laying system organogenesis Our genetic mosaic results suggest that egl-38 acts autonomously within the two tissues of the egg-laying system to promote normal development. At the same time, genetic studies (Chang et al., 1999) indicate that egl-38 influences signaling between the vulva and the uterus, acting upstream of genes both for the signal (lin-3) and the receptor (let-23) required to specify uv1 cell fate. Functionally, egl-38 affects the transcription of the lin-3 gene in the vulval cells, whereas we find that let-23 expression is not similarly affected by egl-38 in the uterus. What might be the mechanism for how egl-38 affects uv1 cell fate specification? Our experiments replicate and extend those of Chang et al. (1999), confirming that an activated allele of let-23 can bypass the requirement for egl-38 to specify uv1 cells. This suggests that egl-38 does not act to mediate the response to let- 23. Although we favor the idea that egl-38 acts in parallel to let-23 within the uv1 cell, our results cannot rule out other possibilities. For example, egl-38 could act within the uv1 cells to somehow influence the activity of let-23. In either situation, our results demonstrate that the same Pax transcription factor can act in different cells to coordinate signaling and responding during development. 52

65 Pax transcription factors in development EGL-38 belongs to the Pax2/5/8 family of transcription factors. Members of this family play important roles in the development of different cell types and organs. Pax transcription factors have roles in promoting cell survival, proliferation and differentiation. For example, in mice and humans, Pax2 mutations cause renal hypoplasia (Sanyanusin et al., 1995; Torres et al., 1995). In contrast, primary cancers and various cancer cell-lines show increased levels of Pax gene expression (Muratovska et al., 2003; Wu et al., 2005). Mosaic analysis for egl-38 activity in germline cell survival shows that egl-38 is necessary in both the germline as well as the soma for this function. Therefore there are autonomous and non-autonomous roles for egl-38 in germline cell death. Pax transcription factors also play important roles to coordinate development of tissues and thereby promote organogenesis. During the development of the metanephric kidney in mouse, Pax2 is expressed multiple times, and performs different functions in the different tissues of the system to promote and coordinate organ development (reviewed by Schedl and Hastie, 2000). Our work demonstrates similarities between the role played by Pax2 in mammalian kidney formation and that of its orthologue EGL-38 during the development of the egg-laying system in C. elegans. EGL-38, like Pax2, is expressed in both epithelial and mesodermal cells of the organ. It also influences organ development in both autonomous and non-autonomous ways, as it affects the development of the cells in which it is expressed, as well as expression of signaling molecules that influence the development of other cells. We conclude that coordinating 53

66 the development of cells and tissues within an organ is a conserved function for Pax2/5/8 transcription factors. 54

67 A MEIOSIS I CELL CORPSES DISTAL ARM DISTAL TIP CELL PROXIMAL ARM SPERMATHECA UTERUS B C + N = 18 n578 N = 13 Figure 3.1. Study of germline cell death. A. Diagram of half of the adult hermaphrodite gonad. The two halves of the gonad end at the uterus on the ventral side of the animal. On the dorsal end lies the distal arm followed by a bend where meiosis I occurs. On the ventral side lie the proximal arm and spermatheca ending in the uterus. Germline cell death occurs in the meiosis I region close to the bend in the gonad arm (cell corpses are indicated by arrows).b-e. Level of germline cell death assayed using SYTO12 dye (see Materials and Methods). N indicates number of animals assessed for each genotype. Nomarski images (B, D) and the matching fluorescent images (C, E) of wild type and egl-38(n578) mutants stained using SYTO12 dye. The corpses are seen near the bend of the gonad arm and indicated by an arrow. In this experiment,level of germline cell death was found to be / corpses in wild-type animals, while in egl-38(n578) mutants it was much higher at / corpses per gonad arm. 55

68 Precursor: AB MS Cell scored: -m3vr - m4dl - Cell name (ABarapappp) (Lineage) (MSaaaaapp) -P5.p-P7.p (ABpx) -m4dr (MSpaaaapp) -uv1 (MSx) E -intestine cells (E) C -body muscle (Cappppv) -hyp11 (Cpappv) D P4 -body -germmuscle -line (Daaaa) Figure 3.2. Lineage diagram of the early embryonic divisions in C. elegans showing the panel of cells examined during mosaic analysis and their corresponding lineage (see Sulston et al., 1983). Here x indicates that the cell arises from more than one cell in the lineage. 56

69 Figure 3.3. Summary of mosaic analysis for egl-38(+) activity in germline cell survival. The mosaic categories examined are indicated. Animals with five or less corpses per gonad were designated to be wild type. Animals with six or more corpses per gonad arm were classified as mutant (six represents a value that is two standard deviations higher than the average number of corpses per gonad arm seen in wild type animals). Animals where the transgene was lost in the offspring were classified to lack the transgene in P4 lineage. A more stringent analysis wherein a loss in P4 was only scored when there was also a corresponding loss in the closely related D lineage was done. It resulted in a similar although slightly lower proportion of wild type animals (11 of 18 or 61%). Further analysis on these observations can be seen in Section of the Results. Mosaic Cat. AB MS E P4 </= 5 deaths >/=5 deaths # animals # wild type # mutant % wild type % wild type

70 A uv1 cells B uv1cells C vulf cells D vulf cells Figure 3.4. egl-38::gfp is expressed in both the uterus and in the vulva. Transgenes that include different enhancers from the egl-38 gene are expressed in the uv1 cells of the uterus (A and B) and the vulf cells of the vulva (C and D). At least twenty animals were observed for each transgene. 58

71 The vulva and uterus arise from different founder cells, AB and P1 ABa AB ABp zygote MS EMS E P1 P2 ABpl ABpr Uterus Vulva Figure 3.5. A cell lineage chart of the early cell divisions in C. elegans. The lineage chart indicates the embryonic source of the vulval and uterine cells (after Sulston et al., 1983).Genetically mosaic animals result when an extrachromosomal transgene is lost following cell division. Losses in the first division can result in animals that retain the transgene in vulval cells but not the uterus, or vice versa. 59

72 Figure 3.6. Mosaic analysis for egl-38 function. The mosaic animals were divided into four categories based on transgene presence or absence and phenotype. In category 1, egl-38(+) activity is present both in the vulva and in the uterine uv1 cells, the vulval and uterine phenotypes are wild-type. Therefore, the vulf cells separate and reveal a thin membrane (shown within circular brackets) while in the uterus, the uterine uv1 cells are specified and lie close to the vulva (shown within square brackets). In category 2, egl- 38(+) activity is absent in both the vulf cells and the uv1 cells, and vulval and uterine functions of egl-38 are both mutant. As a result, the vulf cells fail to separate and the uv1 cells move away from the vulva and fuse with other uterine cells. In category 3, egl- 38(+) activity is present in the vulva but absent in the uv1 cells, and there is wild-type vulf cell separation but no uv1 cell specification. In category 4, egl-38(+) activity is absent from the vulva and present in the uterus, and the vulf separation is defective whereas uv1 cells are normal. We conclude that egl-38 has autonomous functions in the developing egg-laying system, and that its presence in the vulva is not sufficient to specify uv1 cell fate. N is the number of animals observed in each category. 60

73 Mosaic Cat. Transgene vulf uv1 vulf cell separation uv1 cell specification N WILD TYPE WILD TYPE thin membrane UTERUS uv uv VULVA uv uv 7 MUTANT MUTANT thick membrane UTERUS 29 VULVA WILD MUTANT thin membrane 13 MUTANT WILD TYPE thick membrane uv1 uv1 uv1 uv1 8 61

74 A GENOTYPE cell specification (Percent of animals) vulf cell separation (Percent of animals) N kuis sa62; kuis n578; kuis sa62; n578; kuis B uv1 uv1 D uv1 uv1 egl-38(+) egl-38(+) C uv1 uv1 E uv1 uv1 egl-38(n578) egl-38(n578) Figure 3.7. A. An activated allele of let-23 (sa62) can bypass the uv1 cell defect of egl-38(n578). uv1 cell fate was assayed using egl-13::gfp (see Materials and Methods). N indicates number of animals assessed for each genotype. Nomarski images (B, D) and the matching fluorescent images (C, E) of wild type and egl-38(n578) mutants expressing LET-23::GFP. LET-23::GFP expression can be seen surrounding the uv1 cells (indicated by a straight line) that lie adjacent to and slightly below the vulf cells. The GFP expression can also be seen above the vulva in the processes that are extended by the lateral uv1 cells. A similar pattern of expression is observed in egl-38(n578) mutants. Twenty animals were observed for each genotype. 62

75 uv1 cell specification Uterus EGL-38 LET-23 Vulva EGL-38 LIN-3 morphogenesis vulf cell Figure 3.8. A model for EGL-38 function in the egg-laying system. EGL-38 has autonomous and non-autonomous roles in the development of the C. elegans egg-laying system. In the vulva, EGL-38 has at least two distinct functions in the vulf cells (vulf): it is required for vulval morphogenesis and also for the expression of the EGF signal, LIN- 3, which signals to the uterus to promote uv1 cell fate. In the uterus, EGL-38 is also required for uv1 cell development. Since an activated allele of let-23 can bypass the requirement for egl-38, we interpret that either egl-38 acts in parallel to let-23 (as shown), or that it indirectly influences let-23 function. 63

76 CHAPTER 4 A SCREEN FOR GENETIC SUPPRESSORS OF THE egl-38(n578) EGG-LAYING DEFECT AND IDENTIFICATION OF rig-4 AS A POSSIBLE CANDIDATE SUPPRESSOR 4.1. INTRODUCTION Studies done by others and described in the previous chapters have characterized the multiple roles that egl-38 plays in the developing egg-laying system. egl-38(n578) mutant animals are egg-laying defective and at a cellular level they have been found to be defective for several features in the egg-laying system (This study; Chamberlin et al., 1997; Chang et al., 1999; Rajakumar and Chamberlin, 2006). Little is know about the molecules that function with egl-38 during egg-laying system development. To identify these molecules, I carried out a genetic screen to isolate suppressors of the egl-38(n578) egg-laying defect. I also characterized the functions of a candidate gene for one of the isolate suppressor mutations Genetic screen to isolate suppressors of the egl-38(n578) egg-laying defect egl-38(n578) was isolated in a screen for mutant animals that are 64

77 egg-laying defective (Egl-) (Trent et al., 1983). The hermaphrodites of this genotype cannot mate with males. However they produce both sperm and eggs and are capable of self-fertilization. Fertilized eggs hatch within the parent and the parent subsequently bursts to release the progeny (Trent et al., 1983; Chamberlin et al., 1997; Refer Section of Chapter1). Studies done by others and in Chapter 2 show that in the fourth larval (L4) stage, cells in the uterine-vulval (uv) junction of egl-38(n578) animals are defective for several functions (Refer to Chapter 1 Section to learn more about the structure of the vulva; Fig 1.3). In the vulva, the vulf cells at the uv junction are mispositioned, resulting in a thick separation at the uv junction. In addition, these cues fail to express the EGF/LIN-3 signal. In the uterus, there is no uv1 cell specification in these mutants (Chamberlin et al., 1997; Chang et al., 1999; Zhang et al., 2005; Rajakumar and Chamberlin, 2006). Little is know about the molecules that act along with egl-38 in the developing egg-laying system. Therefore, I decided to perform an F2 genetic screen to isolate suppressors of the egl-38(n578) egg-laying defect. Isolating suppressors of the egg-laying of egl-38(n578) animals may help to identify new components of the egl-38 pathway in this system. Isolating and characterizing the suppressors of a Pax2/5/8 transcription factor is significant because these Pax factors are evolutionary conserved. Further, the Pax2/5/8 factors have important roles during the development of the kidney, thyroid and the B lymphocytes. At a cellular level, they influence the proliferation, cell survival and the differentiation. When disrupted, these factors cause various developmental and cell survival disorders (Chi and Epstein, 2004). If the isolated suppressors are evolutionarily conserved, then novel genes could be identified or new functions in already identified 65

78 ones can be uncovered. In mammals, the putative orthologues of the suppressors could show interaction with the respective Pax2/5/8 genes. Thereby, identifying the suppressors could have impact in development and cell survival not only in C. elegans but also in higher animals MATERIALS AND METHODS Strains Nematode strains were cultured according to standard techniques (Brenner, 1974; Sulston and Hodgkin, 1988). The following mutations were used: Linkage group (LG) I: dpy-5(e61), unc-5(e53) LGII: dpy-10(e128), rol-6(e187), unc-4(e120). LGIII: unc-32(e189), unc-119(e2498). LGIV: dpy-13(e184), dpy-20(e1282), egl-38(n578), lin-33(n1043), bli-6(sc16), skn-1 (zu67), sqv-1(n2819), lin-49(s1198), rig-4(ok1160), unc-5(e53), unc-2(s7), unc-24(e138), unc-31(e169), edf19, nt1. LGV: dpy-11(e224), him-5(e1490). LGX: lon-2(e678). The following transgenes were used: sais14 (lin-48::gfp; Johnson et al., 2001), guex1000 (lin-3::gfp; Hwang and Sternberg, 2004) and syex598 (lin-11::gfp; Gupta and Sternberg, 2002). 66

79 Genetic screen for mutations that can suppress the egl-38(n578) egg-laying defect Mutations that can suppress the egl-38(n578) egg-laying defect were isolated following EMS mutagenesis. Mid fourth larval stage (ml4) P0 hermaphrodites from the strain CM607 (egl-38(n578) dpy-20(e1282)) were mutagenized using 50 mm EMS for four hours (Sulston and Hodgkin, 1988) and placed individually on agar plates and incubated at 15 C (permissive temperature). After allowing them to grow for two generations at this temperature, L2 or L3 animals from the F2 generation were transferred to 25 C (non-permissive temperature) for a day. Then the plates were returned to 15 C. The F3 generation was screened for their ability to lay eggs. From a screen of about 49,000 mutagenized gametes, 14 suppressor mutations were recovered. The worms were subjected to a temperature shift before the L4 stage to select for mutations in genes whose products would function specifically in the process during the L4 stage in forming the uterine-vulval (uv) connection. This shift would allow temperature-sensitive alleles of essential genes to be recovered. However, none of the suppressor alleles turned out to be significantly temperature sensitive. The penetrance of the suppression phenotype was determined by picking 20 animals to separate plates. They were allowed to grow at 25 C. The plates were observed for eggs every day for either four days or till the parent burst. Also included in these studies were four mutations previously isolated by Arelis Berrios-Figueroa using the above technique. Among all the suppressors examined, four suppressors (gu68, gu91, gu51, gu52) were found to highly penetrant and were further mapped and characterized. Of these four, gu51 and gu52 were isolated by Ms. Berrios-Figueroa, while gu68 and 67

80 gu91 were isolated in this study. As subsequent analysis at 20 C and 15 C yielded levels of suppression comparable to that done at 25 C, all further experiments were performed at the standard culture temperature of 20 C Genetic mapping to Linkage Groups (LGs) and complementation tests All mutations were backcrossed twice to N2 (wild-type) stocks and frozen. The mutations (gu51, gu52, gu68, gu91) were assigned to LGIV using the markers, dpy-5 I, rol-6 II, unc-32 III, egl-38 IV, dpy-20 IV, unc-5 IV, dpy-11 V and lon-2 X in the following crosses: (a) gu51 like the other suppressors has egl-38(n578) dpy-20(e1282) in the background. Previous studies by Ms. Berrios-Figueroa indicated that it might be linked to LGIV. gu51 was self crossed and since Egl Dpy progeny had egg-laying ability,i concluded that gu51 was linked to LGIV. (b) gu52, gu68 and gu91 were mapped to LGIV by crossing to mapping strains with markers for each chromosome: MT3751 (dpy-5(e61) I; rol-6(e187) II; unc- 32(e189) III) and MT464 (unc-5(e53) IV; dpy-11(e224) IV; lon-2(e678) X). The tested suppressors exhibited linkage to unc-5(e53) and not to the other markers. All four suppressors mapped to LGIV and were therefore tested for complementation with each other and to egl-38(n578). The suppressors were tested first for dominance by crossing them with the PS2210 strain (egl-38(n578) dpy-20(e1282) / DnT1). All egg-laying defective (Egl) and dumpy (Dpy) cross-progeny did not lay eggs, 68

81 indicating that the suppressor mutations are recessive and are not intragenic mutations in egl-38(n578). Complementation tests for the suppressors against each other were done by crossing the sais14 (lin-48::gfp) transgene into each suppressor strain (say suppressor A ). Then the gfp transgene-bearing males from this cross were mated with other suppressor strains (say suppressor B ). The Dpy cross progeny from this mating which look phenotypically wild type and bears the gfp transgene will have suppressor A in trans to suppressor B. On the other hand, the self-progeny in this cross would all be dumpy (Dpy) but would lack the gfp transgene. To examine complementation between suppressors, the cross progeny thus obtained were examined for egg-laying ability Three-factor mapping for gu51 and gu68 gu51 and gu68 were mapped to the left arm of chromosome IV by crossing to the strain JR1227 (dpy-13(e184) unc-24(e138)). From heterozygous F1 animals, Dpy nonuncoordinated (Non-Unc) F2 recombinants were picked and verified for segregation of appropriate markers. These F2 animals were then self-crossed to get Dpy Non-Unc animals. They were confirmed to be egl-38(n578) animals based on a characteristic tail defect. These animals were then checked for their egg-laying ability. Since the egg-laying ability was not suppressed in the Dpy non-unc but not Unc non-dpy recombinants, the two sup mutations were found to lie to the right of dpy-13(e184) (Refer to Fig 4.4 and 4.5). The suppressors mutations gu68 and gu51 were found to be mutations in two separate genes by mapping them using the strain JT9819 ([unc-24(e138) lin-49(s1198) unc-22(s7)]/ DnT1). The recombinants picked in this case were Unc non-lin (Unc non- 69

82 Let; the lin-49(s1198) mutation is recessive lethal). After assessing segregation and crossing to make the recombinant chromosome homozygous, the recombinants were tested for their egg-laying ability. gu51 mapped to the left of unc-24(e138), while gu68 mapped to its right (Refer to Fig 4.4 and 4.5). gu51 was mapped by crossing into the following strains: (a) SP1052 (dpy-13(e184) unc-5(e53))- the recombinants studied were Dpy Non-Unc (b) MT5240 (unc-5(e53) lin-33(n1043) bli-6(sc16)) the recombinants initially picked were Unc Non-Blistered (Non-Bli). Then among these recombinants, those that had lin- 33(n1043) (were vulvaless (Vul) were chosen to examine segregation of egg-laying ability. (c) EU84 ([unc-5(e53) skn-1(zu67)] IV / nt1 [let-?(m435)](iv;v)) skn-1(zu67) animals are maternal effect lethal. So the recombinants analyzed were Unc Non-Skn animals which produced viable progeny. Three-factor mapping of gu68 on chromosome IV was done using the strain MT8998 ([unc-24(e138) sqv-1(n2819)] IV / nt1[unc-?(n7540 let-?] (IV;V)). Unc animals were chosen to examine segregation of egg-laying ability in egl-38(n578) background Cosmid rescue as an attempt to molecularly identify the gene affected by the suppressor mutations gu68 was mapped between unc-24 and lin-49 on LGIV. Wormbase lists the cosmids in this region ( The following cosmids and a Yeast Artificial Chromosome (YAC) were then requested from Alan Coulson : F20D12, F57H12, F53A12, C55E8, F42G8, K07H8, F17E9, T09A12, C07G1, F25H8, F53H10, 70

83 F57A12, K07H8, F17E9, T09A12, C13H8, C25E7, T25B5 and Y42H9B (YAC). The cosmid clones were prepared and confirmed using restriction enzyme digestion. The suppressor mutation gu51 was mapped to the left of unc-24 on chromosome IV. Cosmids in this region were selected and obtained as in the case of gu68. They include: K02B2, H04MO3, H32C10, F33E7, C10G6, C08A10, B0F43, F17B8, C44A5, C02C3, H0M01, B0547, C45E7, T19E7, C10G6, H35B03, F44E8 and F26D12. The DNA clones were confirmed using restriction enzyme digestion. The DNA was injected into the mitotic germline of mutant hermaphrodites according to the method of Mello et al. (Mello et al., 1991). Injection mixture for cosmids contained one to four cosmids, each at 10ng/µl, 15ng/µl of transformation marker myo-2::gfp (Okkema et al., 1993) and pbskii (+) of the required concentration to make a final DNA concentration in the mixture to 100ng/µl. The injection mixture containing the YAC included digested or undigested yeast genomic DNA including the YAC, 15ng/µl of transformation marker myo-2::gfp (Okkema et al., 1993) and pbskii (+) of the required concentration to make a final DNA concentration in the mixture to 100ng/µl. The DNA was injected into the respective suppressor strain being tested in the cosmid rescue assay. The transformants were selected and heritable stable lines were maintained on basis of myo-2::gfp expression. The heritable lines were assayed for their ability to revert the wild-type vulf morphology seen in the suppressor to that of the mutant egl-38(n578) (Chamberlin et al., 1997) Phenotypic characterization of the suppressors Developmental stages: Hermaphrodite animals were observed at the adult 71

84 (A) stage or at different fourth larval stages (L4): early L4 (el4), mid L4 (ml4) and late L4 (ll4). These developmental stages were defined on the basis of the vulval morphology in the mid-plane under a compound microscope (Fig 1.3; reviewed by Greenwald, 1997). During the el4 stage, the anchor cell has invaded the vulva between the vulf cells. During the ml4 stage, the vulf cells separate, leaving a thin membrane between them. During the ll4 stage, the vulva begins to evert, and the thin membrane is no longer seen between the vulf cells. Finally in an adult worm, the vulva has undergone complete eversion. Egg-laying ability: ml4 stage worms were placed on separate plates. These plates were checked for presence of eggs laid by the worm once a day for three days or until the worm burst. Worms which laid any eggs at all on the plate were categorized as Egl+ and those that did not were classified as Egl-. Vulval fate and morphology: (a) vulf cell position and membrane: ml4 larvae were observed under Nomarski optics to determine whether the vulf nuclei are positioned anterior and posterior, leaving behind a thin laminar process between the vulva and uterus. In egl-38(n578) mutants the vulval nuclei separate to a lesser extent, and the tissue between the vulva and uterus remains thick (Chamberlin et al., 1997). The vulval morphology of the suppressors was observed to assess their ability to confer a wild-type phenotype in the egl-38(n578) background. (b)vulval protrusions: The mid-plane of the ml4 larvae was examined under Nomarski optics for the finger-like protrusions into the vulval lumen. Wild-type animals have three pointed protrusions between the ventral hypodermis and the thin 72

85 laminar process bordering the vulva and the uterus. In egl-38(n578) and doubles with the suppressors, the protrusions could be abnormal in shape and number. (c) lin-3::gfp expression: lin-3::gfp is expressed in the L4 stage in vulf cells of wild-type animals and is lost in egl-38(n578) animals (Chang et al., 1999). The suppressors were analyzed for their ability to suppress this loss in expression. Assay for uv1 specification using lin-11::gfp: lin-11 gene is expressed in the π cells of the uterus (Newman et al., 1996; Gupta and Sternberg, 2002). Thus it serves as a useful marker to determine the position and thus the fate of the π cells during the mid and late L4 stages. In wild-type animals, two uv1 cells per side (left/right) are positioned next to the vulf cells, while the other four π cells position away from the vulva and fuse to form a syncitium called utse (Newman et al., 1996) Therefore during the ll4 stage in wild-type animals expressing lin-11::gfp, there are two bright proximal uv1 cells next to the vulva and a dim distal utse cell. In contrast egl-38(n578) mutants only a brighter utse cell is seen. The sups were examined for their ability to restore uv1 cell fate specification. Assay for Programmed Cell Death (PCD)/ Apoptosis in the hermaphrodite germline: The assay for relative numbers of dying cells per gonad arm was performed as described in Gumienny et al., 1999 using a vital stain that preferentially stains germ cell undergoing apoptosis (Gumienny et al., 1999). For this assay ml4 stage hermaphrodites were incubated at 20 C for 24 hours and then submerged in 33µM aqueous solution of SYTO12 for 4 hours. Next, the worms were recovered for 30 minutes on Agar plates. The gonads of the worms were subsequently observed using epifluorescence microscopy. In wild-type animals usually two-to-three apoptotic cells are seen in the gonad arms 73

86 (Gumienny et al., 1999; Park et al., 2006). In egl-38(n578) animals the level of apoptosis is elevated (Park et al., 2006). The sups were examined for their ability to suppress this elevated level of cell death in egl-38(n578) animals. Somatic Cell Death: 1.5 stage embryos were observed under Nomarski optics for refractile, elevated button shaped cells which are undergoing apoptosis (Sulston and Horvitz, 1977). Assay for lin-48::gfp expression: lin-48 promoter has two redundant elements called lin-48-response elements 1 and 2 (lre-1 and lre-2), that are responsible for its expression in four hindgut cells, F, U, K and K (Johnson et al., 2001). Clone-pAJ59 (with lre-1 intact but with mutations in lre2) and clone-paj148 (with wild-type lre-2 and mutant lre-1) driving GFP expression were injected into wild-type animals, egl-38(n578) mutants and the two strains with suppressor mutations (Johnson et al., 2001). Subsequently I examined the four hindgut cells, F, U, K and K for lin-48::gfp expression using epifluorescence Phenotypic analysis of rig-4(ok1160) Three-factor mapping led to rig-4 gene as a possible locus for the gu68 mutation on chromosome IV. Therefore several phenotypic studies were done with the strain RB1140 rig-4(ok1160): (a)vulval morphology: The mutants were observed for finger-like protrusions in the ml4 stage as described in (b) Construction of rig-4(ok1160) egl-38(n578) mutants 74

87 To construct rig-4(ok1160) egl-38(n578), I crossed rig-4(ok1160) animals to [unc- 24(e138) rig-4(ok1160) egl-38(n578)] / + heterozygotes. Subsequently I picked up Dpy- 20 non-unc-24 recombinant animals to separate plates. I examined these plates for recombinants that segregated Dumpy animals that were egg-laying defective. These animals would therefore be homozygous for rig-4(ok1160) egl-38(n578) dpy-20(e1282). I then genotyped these animals using PCR analysis. (c) Phenotypic analysis of rig-4(ok1160) egl-38(n578) mutants The embryos inside the rig-4(ok1160)/+ egl-38(n578)/egl-38(n578) heterozygous hermaphrodites were examined for levels of somatic cell death (Refer to Section of Materials and Methods) Sequencing of rig-4 gene in CM390 strain Wild type sequence of rig-4 was obtained from the wormbase ( wormbase.org/db/gene/gene?name=rig-4). To determine whether rig-4 is altered by the gu68 mutation, primers were designed to amplify the exons of the rig-4 gene. These primers were used to amplify the gene from the CM390 strain which carries the gu68 mutation using the Polymerase Chain Reaction (PCR) and Taq polymerase enzyme. The PCR products were sequenced by the OSU Plant-Microbe Genomics Facility RESULTS Genetic screen for suppressors of egl-38(n578) egg-laying defect To identify additional genes that function with egl-38 in egg-laying 75

88 system development, I performed an F2 suppressor screen on egl-38(n578) Egl- mutants. Thereby I wanted to isolate mutations in genes which could suppress the Egl- phenotype of the egl-38(n578) mutation and render the animals egg-laying competent (Egl+). About 49,000 haploid genomes were mutagenized and examined for suppressor mutations conferring egg-laying ability in egl-38(n578) background. From this screen, 14 suppressor mutations were isolated. Previously, Arelis Berrios-Figueroa had similarly isolated 4 mutations. All of these 18 suppressor mutations were backcrossed and verified for egg-laying ability (Fig 4.2). Of the 18, four mutations (gu68, gu91, gu51 and gu52) exhibited a suppression phenotype of at least 4% or more and were chosen for further analysis (Fig 4.2; green boxes in Fig 4.3) Mapping to Linkage Groups (LGs) and complementation tests The four suppressor mutations were all mapped to Chromosome IV after crossing with different mapping strains (see Materials and Methods, Fig 4.3, 4.4 and 4.5). Therefore, I performed a complementation test to examine if any of these mutations were alleles of the same gene. To do so, I first checked if these mutations were recessive or dominant (Fig 4.3). When the wild type chromosome was put in trans to any of suppressor mutations, the animals failed to suppress the egg-laying defect. Therefore all the suppressor mutation are recessive making them suitable for a complementation test. In order to perform a complementation test between the suppressors, inter se transheterozygous animals were constructed and identified as described in Materials and Methods and tested for their egg-laying ability (Refer Section of Materials and Methods). A similar cross was carried out between animals of similar genotype, to 76

89 determine the penetrance of the suppression phenotype associated with the different alleles. The complementation test results are shown in Figure 4.3. The penetrance of the different suppressors in shown in the green boxes. It ranges from 51.6% in gu51 homozygous mutants to 4.8% in gu91/gu91 animals. Upon examining transheterozygous animals I found that none of them exhibited a level of suppression equal to that of either homozygous animal. Although each of the mutations is recessive, I did observe a degree of non-complementation among all of the alleles. Given that at least two of the alleles map to different loci, I conclude that the results suggest that there is a genetic interaction between these alleles, rather than noncomplementation due to allelism. The results indicate possible allelism for gu68 and gu91 because the level of suppression in the gu91 and gu68 transheterozygotes is midway between that of the single mutant homozygotes. As an outcome of these studies, the two suppressor mutations with the highest penetrance, gu68 and gu51 were chosen for further mapping Three-factor mapping for gu68 and gu51 (Refer to Fig 3.4 and 3.5 for a summary of the results and the final map positions) Three-factor mapping crosses as described in Section and were employed to genetically map the two suppressor mutations. The results indicate that the gu68 and the gu51 mutation map to chromosome IV. gu68 was mapped to the right of unc-24 while gu51 maps to the left of it (Refer Fig 4.4 and 4.5). Therefore, I have isolated mutations in at atleast two different genes that act as suppressors of the egl- 38(n578) egg-laying defect. 77

90 Attempts at cosmid rescue for gu68 and gu51 In an attempt to clone the genes affected by the two mutations, gu68 and gu51, I decided to use transformation rescue experiments. Here DNA clones, either cosmids or Yeast Artificial Chromosomes (YACs), that cover the genetic intervals to which either gu68 or gu51 were mapped were chosen based on information in the Wormbase (Fig 4.6; Subsequently injection mixes with four DNA clones and a transformation marker, myo-2::gfp were injected into the germline mutant animals (Mello et al., 1991; Okkema et al., 1993; Refer Material and Methods Section 4.2.5). The transformants were selected and heritable stable lines were maintained on basis of myo-2::gfp expression. The heritable lines were assayed for their ability to revert the wild-type vulf morphology seen in the suppressor (either 90% in gu68 or 60% in gu51 animals; Fig 4.7.B) to that of the mutant egl-38(n578) (4%; Fig 3.7.B; Chamberlin et al., 1997). However, none of rescue analysis resulted in clear reversion from the suppressor wild type vulf phenotype to that closer to the level seen in egl-38(n578) animals. Therefore, I could not isolate the DNA clone that had a candidate gene for either of the suppressor mutations. Possible reasons as to why this approach failed further discussed in the Discussion Section of this Chapter (Refer Section 4.4.3) Phenotypic analysis of suppressors with respect to the egg-laying system The suppressor mutations were isolated in a screen for mutations in genes 78

91 that suppresses the egg-laying defect of egl-38(n578) animals (Trent et al. 1983). The egg-laying defect seen in the egl-38(n578) mutants arises from several defects in cells present at the uterine-vulval (uv) junction. Studies done by others and us have elaborated the defects in the egl-38(n578) egg-laying system at a cellular level (Chamberlin et al., 1997; Chang et al 1999; Chapter 2 and 3 of this study). In the vulva, the vulf cells bordering the uv-junction are defective in two independent processes. They are mispositioned leading to a thick membrane at this junction and they fail to express the EGF/LIN-3 signal (Chamberlin et al., 1997; Chang et al., 1999; Zhang et al., 2005; Rajakumar and Chamberlin, 2006). In the uterus, the uv1 cells are not specified in these animals (Chamberlin et al., 1997; Rajakumar and Chamberlin, in press). In order to characterize the isolated suppressor mutations, we hypothesized that since they suppress the egl-38(n578) egg-laying defect, they must do so by suppressing some of the cellular defects in the egg-laying system. To determine whether all or only a subset of these defects is suppressed, I analyzed the egg-laying system in the suppressor egl-38(n578) double mutants. The suppressors were tested for their ability to restore the EGF/LIN-3 expression lost in vulf cells of egl-38(n578) mutants (Fig 4.7.A). guex1000 (lin-3::gfp) was used to study LIN-3 expression in the vulf cells (Hwang and Sternberg, 2004). All wild-type animals examined express LIN-3::GFP, while it is completely lost in egl-38(n578) animals (Chang et al., 1999). I find that the gu68 and gu51 fail to restore LIN-3 expression in egl-38(n578) and thus this defect is not suppressed. Next, I examined vulf cell position in the suppressors (Fig 4.7.B). In wild-type animals, the vulf cells move apart to reveal a thin membrane at the uv junction. In 79

92 contrast, this is seen only in four percent of the egl-38(n578) mutants. Instead, the vulf cells generally stay together and there is a thick membrane at the uv junction. I find that close to 90 percent of gu68 and 60 percent of gu51 animals are suppressed for this defect and show wild-type vulf cell position. Thus this defect is significantly suppressed. Finally I examined the specification of the uv1 cells at the uv junction. syex598 (lin-11::gfp) was used to visualize the position uv1 cells (Gupta and Sternberg, 2002). Wild-type animals show uv1 cell specification wherein the uv1 cells lie close to the vulva. However, in egl-38(n578), none of the animals show uv1 cell specification. Instead, the cells move away from vulva and fuse with the utse cells (This study; Chapter 2; Chamberlin et al., 1997; Rajakumar and Chamberlin, 2006). I find that uv1 specification is present in the suppressor-bearing strains, with 12.5 percent of gu68 animals and 40.0 percent of gu51 animals exhibiting normal placement of uv1 cells. In conclusion, in the egg-laying system, gu68 and gu51 suppress the vulf cell position and the uv1 cell specification defect in egl-38(n578) mutant background. However, they do not alter the vulf LIN-3::GFP expression defect. The phenotypes that these suppressor mutations affect at a cellular level enable the double mutants to be egglaying competent even in egl-38(n578) background Phenotypic analysis of suppressors with respect to Programmed Cell Death (PCD/Apoptosis) The two related Pax2/5/8 genes in C. elegans, egl-38 and pax-2 have been shown to promote cell survival. In hypomorphic egl-38(n578) mutants, there are elevated levels of PCD (Park et al., 2006). Since the suppressors exhibit an ability to suppress the egg- 80

93 laying defect in egl-38(n578), I hypothesized that they could function to suppress another function disrupted by the n578 allele, PCD in the germline. Under Nomarski optics, dying germline cells appear irregular and raised and can also be seen using an epifluorescence microscope when stained with the SYTO12 dye (Fig 4.9; Gumienny et al., 1999). In order to study PCD levels, each suppressor was examined along with the wild type and egl-38(n578) animals in separate experiments. First, I examined gu68 animals along with the controls (Fig 4.10). Here, the wild-type animals, both N2 and dpy-20(e1282), had a low level of PCD whereas PCD in egl- 38(n578) mutants however was elevated. Interestingly, gu68 and gu51 were able to suppress this high level of PCD back to wild-type levels (Fig 4.10 and 4.11). Thus, both suppressors affect the egl-38(n578) defect in germline cell survival as well as the egglaying system Phenotypic analysis of the suppressors with respect to lin-48::gfp expression in the hindgut egl-38 plays an important role in the development many different tissues like the egg-laying system and the hindgut (Chamberlin et al., 1997). The target for EGL-38 in the hindgut is lin-48, a gene that encodes a C2H2 zinc finger protein. lin-48 is expressed in four cells in the hindgut, F, U, K and K. EGL-38 binds to one of two redundant elements in the lin-48 promoter that is necessary for its expression. Mutations in these elements called lin-48 response elements 1 and 2 (lre1 and lre2) can genetically interact with various egl-38 alleles. For example, when both lre1 and 2 are intact, LIN-48 expression in the hindgut cells visualized using GFP reporter transgene (sais14) is similar 81

94 in both egl-38(n578) and wild type animals. However, when either lre1 or lre2 is mutated, the LIN-48::GFP expression is reduced in egl-38(n578) mutant animals while remaining less affected in their wild-type counterparts (Johnson et al., 2001). I hypothesized that since the isolated suppressors can suppress defects associated with the egl-38(n578) allele in the egg-laying system and in the PCD process, they could also suppress the LIN-48::GFP expression defect seen in animals with mutant lre1 or lre2. Therefore, I examined the expression of LIN-48::GFP in the hindgut cells of suppressor-bearing strains (Fig 4.12; Fig 4.13). egl-38(n578) animals with transgenes mutant for either lre1 or 2 but intact for the other show reduced levels of LIN-48::GFP expression compared to wild type animals. The suppressors are however unable to restore wild type levels of LIN-48::GFP expression in the hindgut cells. Thus, gu68 and gu51 suppress only a subset of the defects associated with egl-38(n578) Phenotypic analysis of rig-4(ok1160) as a candidate gene for gu68 Three-factor mapping results led to the conclusion that gu68 lies between unc-24 and sqv-1 and closer to unc-24 on chromosome IV. I examined this interval for genes that might affect the phenotypes exhibited by gu68. One gene, rig-4, mapped to this region but was not covered by any existing cosmids. However, there was a preexisting deletion allele (ok1160) for rig-4. Upon examining the vulva of the rig-4 animals in the ml4 stage, I found that it had abnormal number and/or protrusions into the midplane of the vulval lumen. In wild-type animals and egl-38(n578) animals six protrusions representing three adherens junctions (AJs) can be seen. I classified any vulva with different number and shape of these protrusions as abnormal. In this way, I examined the 82

95 vulva of animals homozygous for egl-38(n578), gu68 or rig-4(ok1160). I found that both gu68 and rig-4(ok1160) animals show abnormal vulval protrusions (20 and 70 percentage respectively). Interestingly in an inter se complementation test between ok1160 and gu68, I observed 45.9 percent of animals with abnormal vulval protrusions. This represents a value that lies between that seen in the two homozygotes, and is consistent with noncomplementation. Next, I tried to construct a strain having both the rig-4(ok1160) and the egl- 38(n578) mutation in cis to each other. I recombined the egl-38(n578) mutation onto the rig-4(ok1160) chromosome (Refer to Section of Materials and Methods). This generated animals of genotype ok1160 n578/+ n578. I subjected these animals to a selfcross in an attempt to construct double homozygotes. However I was unable to recover these double homozygotes and establish a line. Instead, on the plate I saw thin, stationery L1 or L2 stage worms. I genotyped these thin worms using PCR and found them to be rig-4(ok1160) egl-38(n578). Since, these animals died young, I examined the commastage embryos within the hermaphrodites of the genotype ok1160/+ n578/n578 for levels of somatic cell death (Fig 4.14). There could be some embryos with high levels of somatic death, since rig-4(ok1160) egl-38(n578) doubles are synthetically lethal. The levels somatic cell death in rig-4(ok1160) single mutants is similar to that in wild-type animals, while the egl-38(n578) animals show higher levels. Interestingly, in the ok1160/+ n578/n578 heterozygotes, some embryos show somatic cell death levels similar to egl-38(n578) mutants while others have levels double that seen in the egl-38(n578) animals. As a result the average level of somatic cell death is much higher than either 83

96 single mutant. Therefore, I concluded that rig-49(ok1160) egl-38(n578) animals were inviable, and the two alleles are synthetically lethal DNA sequencing analysis of rig-4(ok1160) To examine if rig-4 is the candidate gene for the gu68 mutation, I sequenced the exons of this gene in gu68 homozygotes. rig-4 produces a 12.5 kb transcript from 12 exons. Using PCR, I amplified exon sequences from the CM390 strain which has the gu68 mutation. Subsequently I sequenced this DNA and compared the sequence obtained with that of rig-4 from the wormbase ( After sequencing all the exons, none of them were found to have a mutation in a region of the exon that has evolutionary conserved bases DISCUSSION In this Chapter, I describe a genetic screen to isolate mutations that suppress the egl-38(n578) egg-laying defect. Thereby I wanted to identify new molecules that function with egl-38 in egg-laying system development. Four suppressor mutations were phenotypically characterized. Two of these mutations were genetically mapped to two separate loci on chromosome IV. Cosmids and Yeast Artificial Chromosomes (YACs) were then used in an attempt to rescue the suppression phenotype and thereby clone the suppressors. However, I was unable to get rescue. Finally, I examined a candidate gene, rig-4 for the suppressor gu68 since it is present in the genetic interval to which gu68 had been mapped. Upon sequencing the exons of this gene in the gu68 strain I did not find 84

97 any mutations. However, rig-4 is an interesting gene as a deletion allele is synthetically lethal with egl-38(n578) Functional relationship between the suppressors and egl-38(n578) Of four mutations isolated in a screen for suppressors of egl-38(n578), two (gu68 and gu51) were examined further. They were tested for their ability to suppress defects associated with the egl-38(n578) mutation in various tissues. In the egg-laying system, they suppress the vulf cell separation and the uv1 cell specification. However, they fail to restore LIN-3 expression analyzed using a transgene. In the process of Programmed Cell Death (PCD) where egl-38(n578) animals show elevated levels of germline PCD (Park et al., 2006), the suppressors were able to restore wild type levels of PCD. Finally, in the hindgut, they were unable to restore wild-type levels of LIN-48 expression as assayed by transgenes defective in either lin-48 response element 1 or 2. One common theme among the phenotypes of the suppressors is that they suppress all tested defects associated with egl-38(n578) except those that examine gene expression using an extrachromosomal reporter transgene. Further experiments are required to determine if the effects reflect differences between endogenous genes, extrachromosomal transgenes or the suppressors preferentially affect certain egl-38 functions and not others. Although we do not know what genes are represented by the suppressors, several scenarios are possible. One possibility is that the suppressor molecules are bound to cell membranes or lay in the intercellular space affecting cell adhesion. This way these molecules would be present at the uterine vulval (uv) junction and affect both vulf cell position and uv1 cell specification. When mutated, these molecules would thus defective 85

98 in positioning the vulf cells apart. Also, they would be unable to maintain the uv1 cells close to the vulva and thereby allow these cells to move distal and follow the default pathway of fusing to become the utse syncitium (Newman and Sternberg, 1996; Chamberlin et al., 1997; Rajakumar and Chamberlin, in press). However, one caveat in the above speculation is that it is difficult to envisage a suppressor molecule with functions in cell adhesion, also playing a role in PCD in the germline. The cells in the germline that undergo apoptosis lie in the meiotic zone of the gonad. Another possibility for the nature of a suppressor molecule is that it functions with EGL-38 in transcription. For example it might work with mutated EGL-38 to strengthen its binding to DNA. EGL-38 protein has a Paired DNA Binding Domain (PD) where each half of the domain binds to DNA sequences in the two major grooves respectively. The affinity of DNA binding is dependent on both halves of the PD. The n578 mutation lies in the linker region between the two half domains, and it does affect the DNA binding (Zhang et al., 2005). As co-factors for EGL-38, the suppressors could strengthen the DNA binding affinity of the mutated EGL-38 protein. However these suppressor co-factors maybe unable to bind or regulate EGL-38 binding in the context of extrachromosomally located transgenes. This could explain why the suppressors are able to act in multiple tissues to suppress all egl-38(n578) defects except those assays measured using reporter transgenes Interaction of rig-4 with egl-38 The rig-4 (neuronal Ig) gene produces a cell adhesion molecule with 13 fibronectin and 30 immunoglobulin (Ig) domains. It has earlier been reported to be 86

99 expressed in some neuronal and non-neuronal tissues ( rig-4 is located in the region where the gu68 was genetically mapped (This study). There is a deletion allele, ok1160 for rig-4 ( Interestingly, rig-4(ok1160) egl-38(n578) animals are synthetic lethals. In contrast, gu68 suppresses elevated levels of cell death in egl-38(n578) animals (This study). Therefore, we hypothesize that gu68 might be a gain-of-function mutation in rig-4. We did not find a mutation in the rig-4 exon region upon sequencing the gu68 strain. If gu68 is indeed an allele of allele of rig-4, then the mutation could lie in the non-coding region. It is also possible that gu68 is not an allele of rig-4. RNAi against rig-4 was done in gu68 background without see any effect. These negative results could, among other possibilities, mean that the procedure did not work. In the future, overexpression studies using a heat shock promoter to drive rig-4 expression in gu68 animals can be done. Further the expression pattern of rig-4 can also be studied. Even if rig-4 is not the candidate gene for the gu68 mutation it is interesting to further explore the relationship between egl-38 and rig-4. Since my results suggest the rig-4(ok1160) egl-38(n578) animals are synthetically lethal, perhaps rig-4 could act as an effector of the EGL-38 pathway in various processes. It would be interesting to examine if the other Pax2/5/8 gene in C. elegans, pax-2 also genetically interacts with rig Molecular cloning of the suppressors The suppressors, gu68 and gu51 were mapped to different genetic intervals on chromosome IV (This study). Subsequently I attempted transformation rescue experiments by injecting DNA clones into the suppressor strains. Thereby I wanted to 87

100 examine if any DNA clone could revert the suppression phenotype back to that of egl- 38(n578) mutants. However I was unsuccessful in identifying any clone that did so. There could be a number of reasons why this approach was unsuccessful. Some of the technical reasons include that the clones used for the injections did not provide complete coverage for the genetic intervals examined. The injection mixtures could have insufficient concentration of the clones either due to adding less of it or due to degradation. Also, there was no positive confirmation carried out that the YAC used was intact and was able to form extrachromosomal array after injections. Further, although the cosmid clones were examined for their integrity using restriction analysis, they could still be harboring some deletions that were not detected. Besides these technical reasons, the nature of the suppressor molecules itself could determine if they would amenable to rescue analysis. If the suppressors have mutations which are neither loss-of-function nor null in nature, then it may not be possible to rescue those mutation using additional copies of wild-type gene. 88

101 Parent strain: egl-38(n578)... Egg-laying defective / Egl (-) Mutagen: 50mM EMS, 20 C, 4 hours EMS P0 (+/+) Egl (-) ~49,000 haploid genome screened F1 (m/+) F2 (m/m) Egl(+) 4 strong suppressors identified Figure 4.1. F2 suppressor screen using EMS. egl-38(n578) egg-laying defective (Egl-) animals were mutagenised using EMS to induce random mutations in the germ cells of P0 hermaphrodites. Since hermaphrodites are self-fertilizing, the sperm and eggs it produce will carry a copy of the mutation in the F1 generation. In the F2 generation, a quarter of animals from a mutation-bearing F1will be homozygous for the mutation. Therefore, F2 generation animals were screened for presence of egg-laying competent (Egl+) animals. The Egl+ animals thus obtained were self-crossed for one more generation to confirm the presence of the mutation. 89

102 ORIGINAL NAME STRAIN MUTATION M-5-11 CM355 gu56 M-5-12 CM356 gu57 M-9-25 CM367 gu59 M CM364 gu58 M CM371 gu63 M CM390 gu68 M CM372 gu64 M CM369 gu61 M CM387 gu66 M CM388 gu67 M CM497 gu80 M CM994 gu91 M CM396 gu69 M CM498 gu91 A9* CM252 gu51 C36* CM253 gu52 H25* CM271 gu55 H27* CM258 gu54 Figure 4.2. Summary of the isolated suppressors. Listed are the suppressors isolated with their respective original name and subsequently designated strain and allele names. The first fourteen mutations were isolated in this study, while the last four (original names with an * ) were isolated previously by Arelis Berrios-Figueroa.The mutations shown in bold letters were chosen for further analysis due to the high penetrance of the suppression phenotype. 90

103 COMPLEMENTATION TEST (% suppression) gu68 gu51 gu52 gu gu gu gu gu N=20 Figure 4.3. Complementation test. The suppressor mutations were subjected to a complementation test to determine if they were mutations in separate genes or different alleles of one gene. The columns represent the chromosome coming from the male parent while the rows represent chromosomes coming from the hermaphrodite. The numbers indicate percentage of cross progeny capable of laying eggs. The total number, N of animals examined for each cross is 20. Based on the results, the following conclusions can be made: (i) In the first row (colored grey) where each suppressor mutation is placed in trans to the wild-type copy, the mutant alleles are unable to suppress any egg-laying defect indicating that these mutations are all recessive to the wild-type allele. (ii) When a suppressor mutation is in trans to itself (green boxes), the penetrance of each mutation is revealed. The penetrance of the suppressors ranges from 52% suppression in gu51 to 5% suppression in gu91 animals. (iii) Examining one suppressor mutation in trans to another (in white boxes), all the mutations show some interaction suggesting non-complementation. However, the transheterozygotes generally exhibit weaker suppression than either homozygote. In addition, at least two of the alleles (gu68 and gu51) map to different loci indicating that the suppression phenotypes in the transheterozygotes results from allele interaction rather than intragenic non-complementation. However, it is possible that the gu68 and gu91 mutations are alleles of the same gene. The level of suppression seen in the transheterozygous animals (21%) lies midway between that of gu68 (45.5%) and gu91 (4.8%) homozygotes respectively. 91

104 A Suppressor mutation in egl- 38(n578) background gu68 Mapping strain JR1227 JT9819 MT8998 Genotype of heterozygote dpy-13(e184) unc- 24(e138) IV / + + gu68 egl-38(n578) dpy-20(e1282) IV unc-24(e138) lin- 49(s1198) unc-22(s7) IV /+ gu68 + egl- 8(n578)dpy-20(e1282)IV unc-24(e138) sqv- 1(n2819) IV / + gu68 + egl-38(n578) dpy-20(e1282) IV Phenotype of selected recombinant Dpy-13 non-unc Unc-24 non-let Unc non- Sqv Egg-laying ability of homozygosed recombinant 8/8 Suppressed 5/30 Suppressed 5/16 Suppressed B CHR IV GENES dpy-13 skn-1 unc-24 lin-49 unc-5 lin-33 sqv-1 egl-38 dpy-20 MAP UNITS gu68 Figure 4.4. Three-factor mapping of the suppressor mutation, gu68. A. Summary of results of three-factor mapping of the suppressor mutation, gu68. Listed in the second column for the different mapping strains used. In each case, transheterozygous animals were constructed of the generic genotype ab/c (column three). Here a and b represent mutations in the mapping strain used to pick recombinants. While c represents the suppressor mutation gu68 with egl- 38(n578) and dpy-20(e1282) in its background. From the progeny of these transheterozygote animals, A non B recombinants were picked (phenotypes are indicate in column four) and scored for the suppressor, C (egg-laying ability represented in column five). The frequency of recombinants in each class carrying the trans marker, c represents the relative distance of c from the a and b. B. Position of the gu68 mutation inferred from the three-factor mapping data. gu68 is present on the right arm of chromosome IV between unc-24 and sqv-1. All the map positions are approximated to the nearest tenths position and written as Map units or M. u. 92

105 A Suppress or mutation in egl- 38(n578) gu51 Mapping strain JR1227 JT9819 SP1052 MT5240 EU84 Genotype of heterozygote dpy-13(e184) unc- 24(e138) IV / + gu51 + egl- 38(n578) dpyunc-24(e138) lin- 49(s1198) unc-22(s7) IV / gu egl- 38(n578) dpydpy-13(e184) unc- 5(e53) IV / + gu51 + egl-38(n578) dpy-20(e1282) IV unc-5(e53) lin- 33(n1043) bli- 6(sc16)IV / + gu51 + egl-38(n578) dpyunc-5(e53) skn- 1(zu67) IV / + gu51 + egl-38(n578) dpy- 20(e1282) IV Phenotype of selected recombinant Dpy-13 non-unc Unc-24 non-let Dpy-13 non-unc Unc-5 non- Lin Unc-5 non- Skn Egg-laying ability of homozygosed recombinant 16/27 Suppressed 0/20 Suppressed 10/10 Suppressed 6/18 Suppressed 12/12 Suppressed B CHR IV GENES dpy-13 skn-1 unc-24 lin-49 unc-5 lin-33 sqv-1 egl-38 dpy-20 MAP UNITS gu51 Figure 4.5. Three-factor mapping of suppressor mutation, gu51. A. Summary of results of three-factor mapping of suppressor mutation, gu51. Data presented as in Fig 4.4 B. Position of the gu51 mutation inferred from the three-factor mapping data. gu51 is present on the right arm of chromosome IV between unc-5 and lin

106 Suppressor mutation Name of injection mix DNA clones present in the injection mix gu68 I-VR-A F20D12, F57H12, F53A12 I-VR-B C55E8, F42G8 I-gu68-2A (each cosmid K07H8, F17E9, T09A12 also injected separately) I-gu68-2B C07G1, F25H8, F53H10, F57A12 I-gu68-3A C13H8, C25E7, T25B5 YAC Y42H9B I-gu51-A gu51 (each cosmid K02B2, H04M03, H32C10 also injected separately) I-gu51-2A F33E7, C10G6, C08A10 I-gu51-2B B04B3, F17B8, C44A5, C02C3 I-gu51-3A H0M01, B0547, C45E7 I-gu51-3B T19E7, C10G6, H35B03 I-gu51-3C F44E8, F26D12 Figure 4.6. DNA clones (cosmids and Yeast Artificial Chromosome or YAC) tested in rescue analysis for gu68 and gu51. To clone the suppressor mutations I injected various available wild-type DNA clones and assess whether any of them was able to reverse the suppression phenotype back to that of egl-38(n578) phenotype with respect to vulf separation (See Fig 4.7.B). Each rescue injection was made using one to four DNA cosmids or a YAC, Y42H9B. 94

107 Figure 4.7. Phenotypic analysis of the suppressor with respect to the vulval tissue in the egg-laying system. A. gu68 and gu51 suppress the vulval morphology defect in egl-38(n578). The four images are Nomarski images of the vulf cell separation positions (shown by a red arrow) in the mid-fourth larval stage of the vulva of wild-type, egl-38(n578) and the two suppressor animals. Below are the observations of this separation phenotype for the different genotypes represented as percentage of animals showing wild-type vulf cell morphology. In wild-type animals, all the vulf cells separate revealing a thin membrane between the uterus and vulva (shown by a red arrow). In egl-38(n578) mutants, the vulf cells are mispositioned close to each other and have a thick membrane (shown by red arrow). Only four percent of the animals have wild-type vulf morphology. The suppressors are able to suppress this egl-38(n578) defect and either 90% (in gu68 animals) or 60% (in gu51 animals) of them have a wild-type vulf cell separation phenotype. The total number of animals, N examined in this study were 30 for each case respectively. B. gu68 and gu51 do not restore lin-3::gfp expression in vulf cells in egl-38(n578). guex1000 (lin-3::gfp) transgene was used to observe LIN-3::GFP expression in wildtype, egl-38(n578) and the suppressor animals. Shown here are Nomarski and the corresponding epifluorescence images of the vulf cells (indicated by black arrows) in wild type and egl-38(n578) animals. Below is the tabulation of results showing percentage of animals expressing LIN-3::GFP in the vulf cells. In wild-type LIN-3::GFP is expressed in the vulf cells of all fourth larval stage animals. This expression is absent in egl-38(n578) animals. The suppressors, gu68 and gu51, fail to restore the GFP expression. The total number of animals, N examined in this study was 20 for each case. 95

108 A. vulf cell position + egl-38(n578) gu68 n578 gu51 n578 UTERUS UTERUS VULVA VULVA vulf cell position (% wildtype) (N=30) B. LIN-3 expression + egl-38(n578) UTERUS vulf cells LIN-3::GFP vulf cells LIN-3::GFP GENOTYPE LIN-3:: GFP (% animals showing expression) egl-38 (n578) 0 gu68 n578 0 gu51 n578 0 (N=20) 96

109 uv1 cell specification + egl-38(n578) uv1 cells LIN-11::GFP uv1 cells LIN-11::GFP GENOTYPE + egl-38 (n578) gu68 n578 gu51 n578 LIN-11::GFP (% animals) (N=25) Figure 4.8. Phenotypic analysis of the suppressor with respect to the uterine tissue in the egg-laying system. The images show the Nomarski and the corresponding epifluorescence images of the uterus of wild type and egl-38(n578) mutant animals. syex598 (lin-11::gfp) transgene is expressed in the uterine uv1 cells (Gupta and Sternberg, 2002). uv1 cells in wild-type animals lie close to the vulva, within the brackets. In egl-38(n578) mutants, the uv1 cells are not specified resulting in their fusion with the flanking utse syncitium away from the vulva and outside the orange brackets. Below the images is a summary of uv1 specification assessed using the lin-11::gfp transgene as described above. In wild-type animals, 97.2 percent of the animals show uv1 cell specification where the four uterine cells above the vulva (two seen on each plane) lie close to the vulva and within the brackets. However in egl-38(n578) mutant animals there is no uv1 cell specification. These cells move away from the vulva and lie outside the brackets. The suppressor mutations are able to suppress this defect in egl-38(n578) animals to 12.5 percent in gu68 animals and 40.0 percent in gu51 animals. The total number of animals, N examined in this study was 25 for each case. 97

110 Figure 4.9. gu68 and gu51 suppress elevated levels of germline death seen in egl- 38(n578) animals Shown are pairs of Nomarski and corresponding epifluorescence images of posterior gonad arms of worms of different genotypes stained with SYTO12 dye (labels dying cells) and seen on one plane. The cells undergoing germline cell death appear as raised blobs in the Nomarski images (black arrows) and as green spots (white arrows) in the epifluorescence images. Low levels of dying cells are seen in wild-type animals. This level is much higher in the egl-389(n578) mutant animals. The suppressors, gu68 and gu51 suppress high levels of germline cell death in egl-38(n578) animals. 98

111 (+) n578 gu68 n578 gu51 n578 99