Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Genetic and molecular analysis of the drosophila melanogaster polycomb group gene additional sex combs Milne, Thomas Arthur 1998

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_1998-0135.pdf [ 9.07MB ]
Metadata
JSON: 831-1.0088433.json
JSON-LD: 831-1.0088433-ld.json
RDF/XML (Pretty): 831-1.0088433-rdf.xml
RDF/JSON: 831-1.0088433-rdf.json
Turtle: 831-1.0088433-turtle.txt
N-Triples: 831-1.0088433-rdf-ntriples.txt
Original Record: 831-1.0088433-source.json
Full Text
831-1.0088433-fulltext.txt
Citation
831-1.0088433.ris

Full Text

Genetic and Molecular Analysis of the Drosophila melanogaster Polycomb Group Gene Additional Sex Combs by Thomas Arthur Milne B . S c , The University of British Columbia, 1994 A thesis submitted in partial fulfillment of the requirements for the degree Master of Science in The Faculty of Graduate Studies Department of Zoology We accept this thesis as conforming to the required standard The University of British Columbia Apr i l 1998 © Thomas Arthur Milne, 1998 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of ~2.ao \ oqy The University of British Columbia Vancouver, Canada Date DE-6 (2/88) A B S T R A C T Proteins of the trithorax group (trxG) and the Polycomb group (PcG) maintain the spatially restricted active and repressed states respectively of the homeotic genes throughout development. Mutations in PcG genes cause homeotic genes to be ectopically expressed which generally results in posterior homeotic transformations. Mutations in trxG genes have the opposite effect. They reduce homeotic gene expression and generally cause anterior transformations. Mutations in trxG and P c G genes mutually suppress each others homeotic transformations, suggesting that the two groups of proteins either act antagonistically or that they have opposite and independent functions. Additional sex combs (Asx) can enhance both trxG and P c G homeotic mutations, suggesting that Asx is important for both activation and repression of the homeotic loci. In addition, Asx has both tissue and stage specific effects on homeotic gene regulation.To gain further insight into Asx function, I used antibodies to the Asx protein to examine its expression pattern in embryos and to show that Asx protein binds to sites on polytene chromosomes containing other P c G proteins, but that it also binds to some unique target sites. Asx protein is probably part of a large P c G complex that binds to many target sites but the fact that it can bind to some unique target sites also indicates that it may be a member of other non-PcG protein complexes. To identify and clone genes important for Asx activity, interactor c D N A s from a yeast 2 hybrid screen were mapped to specific chromosomal sites. A large deletion (Df[3L)ZN4T) that removes the locus of the z40 interactor genetically interacts with Asx and shows target specific homeotic regulatory defects. Also, the P c G gene super sex combs (sxc) shows a strong genetic interaction with Asx and an attempt was made to transposon tag and clone sxc. These results suggest that Asx is a component of both repression and activation and that its stage and tissue specific activities are modulated by interactions with specific protein subsets such as z40. 11 T A B L E O F C O N T E N T S Abstract i i Table of contents i i i List of Tables i v List of Figures v Acknowledgements v i i G E N E R A L I N T R O D U C T I O N 1 M A T E R I A L S A N D M E T H O D S 14 C H A P T E R O N E Introduction 24 Results 25 Discussion 49 C H A P T E R T W O Introduction 53 Results 54 Discussion 73 C H A P T E R T H R E E Introduction 76 Results 79 Discussion 91 C H A P T E R F O U R Introduction 95 Results 98 Discussion 110 G E N E R A L D I S C U S S I O N 112 R E F E R E N C E S 117 ii i LIST OF TABLES Table 1: Comparison of Asx with Ph/Pc protein binding sites on polytene chromosomes.45 Table 2: Enhancement of trxG and P c G phenotypes by Asx?! 60 Table 3: Lack of enhancement of trxG and P x G phenotypes by the Asx?! reveitant A s x R 1 1 61 Table 4: Enhancement of P c G phenotypes in Asx;Pc^ double heterozygotes 63 Table 5a: Asx crosses that enhance the penetrance of trx phenotypes 64 Table 5b: Crosses that do not enhance the penetrance of trx phenotypes 64 Table 6a: Enhancement of Df(3L)ZN47 trxG phenotypes by heterozygous Asx^ mutat ions 84 Table 6b: Enhancement of Df(3L)ZN47 trxG phenotypes by Asxpl 84 Table 7: Results of two separate sxc local P screens 105 iv LIST OF FIGURES F i g . l : Comparison of the embryonic expression pattern of the homeotic gene Ubx with the corresponding adult fly segments 4 Fig.2: A comparison of the C N S derepression phenotype of the homeotic gene Ubx between Asx and the moderately strong P c G gene Pel 28 Fig.3: The regulation of the homeotic genes Ubx and AbdB and the bxdl4 maintenance element in Asx mutant embryos 30 Fig.4: Sites of antibody generation on the Asx protein 34 Fig.5: Developmental expression of Asx protein in embryos 37 Fig.6: C N S and brain development in Asx mutant embryos 39 Fig.7: Binding of Asx and Pc proteins to polytene chromosomes 42 Fig.8: Rabbit anti-Asx H R P staining on polytene chromosomes 44 Fig.9: Binding of Asx to Asx?! mutant chromosomes 48 Fig . 10: Example of anterior and posterior transformations seen in Asxpl homozygotes 57 F i g . l 1: Enhancement of trx phenotypes by Asx?! homozygotes 59 Fig.12: T I mutant phenotype in Pc4;Df(2R)trix double heterozygotes 67 Fig.13: Scanning electron microscope (SEM) image of T I mutant phenotype in Pc^;Df(2R)trix double heterozygotes 69 F ig . 14: Binding of Pc protein to Asx?! mutant chromosomes 72 F ig . 15: Mapping of interactor c D N A ' s to polytene chromosomes 81 Fig. 16: The large genomic deletion DJ(3L)ZN47 removes the z40 locus which produces a transcript of 1.8 kb in length 83 Fig.17: Expression of various homeotic genes in Df(3L)ZN47 homozygous mutants ... 87 F i g . 18: Diagram and restriction digest of the P Z construct 100 F i g . 19: L o c a l P screen for sxc insertions 102 Fig.20: Southern blots of individual putative lines 104 Fig .21: P enrichment screen for sxc insertions 109 vi A C K N O W L E D G M E N T S I would like to sincerely acknowledge the contributions made by my supervisor Dr. Hugh Brock to my intellectual and scientific development. I would also like to acknowledge past and present lab members for support and for fruitful scientific discussions, Greg Mullen for teaching me how to use the confocal microscope, and members of the Grigliatti lab for their interest and help. I would also like to acknowledge Dr. Jacob Hodgson for his interest in my project and for many late night scientific discussions. I would like to dedicate this thesis to my wife Isabella. vii General Introduction Overview The significance of the Polycomb group (PcG) of genes in Drosophila melanogaster was recognized by Lewis in 1978 when he suggested that the segmental transformations observed in Polycomb (Pc) mutants were the result of the misregulation of homeotic genes. Although at the time the Polycomb 'group' was not yet a group as there were only two clearly identified members, this initial observation was important in recognizing that separate genes existed which did not contribute to the final morphological differentiation of the adult fly except by ensuring that the homeotic genes were expressed within their proper expression boundaries. Drosophila development occurs in three basic stages; i) embryogenesis, ii) the larval stages and iii) pupation. The process of pupation eventually produces the adult fly. During pupation, most adult tissues are derived from a small set of disc-like tissues termed the imaginal discs that in embryogenesis segregate away from cells that give rise to the larval tissues. P c G proteins are important components of a system that maintains cell memory throughout these developmental stages by regulating developmentally important target genes (Struhl and Akam 1985a; Wedeen et al. 1986b; Riley et al. 1987a; Glicksman and Brower 1988a; Dura and Ingham 1988b; Glicksman and Brower 1990b; M c K e o n and Brock 1991a; Simon et al. 1992b; Moazed and O'Farrell 1992; Pelegri and Lehman 1994; Serrano et al. 1995). P c G proteins appear to be required for the regulation of multiple target genes but the homeotic loci are their best characterized targets (Struhl and Akam 1985b; Wedeen et al. 1986a; Riley et al. 1987b; Glicksman and Brower 1988b; Dura and Ingham 1988a; Glicksman and Brower 1990a; McKeon and Brock 1991b; Simon et al. 1992a; Soto et al. 1995), and are perhaps of the most importance. The homeotic loci determine the anterior-posterior morphological polarity of the developing organism. The homeotic genes 1 The 8 homeotic genes are arranged in two gene complexes (the H O M - C ) : the Antennapedia complex (ANT-C) , which determines the proper development of the anterior regions of the body including the anterior of the thorax, and the bithorax complex ( B X - C ) , which determines the proper development of the posterior thorax and the abdomen (Lewis 1978; Struhl 1982; Bender et al. 1983; Sanchez-Herrero et al. 1985; McGinnis and Krumlauf 1992). The A N T - C contains the homeotic genes labial (lb), proboscipedia (pb), Deformed (Dfd), Sex combs reduced (Scr) and Antennapedia (Antp) (McGinnis and Krumlauf 1992), and the B X - C contains the genes Ultrabithorax (Ubx), abdominalA (abdA), and AbdominalB (AbdB) (McGinnis and Krumlauf 1992). The order in which the genes appear in the H O M - C reflects the order in which the genes are expressed in the developing embryo. Starting from the 3' end of the complex, the H O M - C genes are activated with progressively more posterior domains of expression moving in a 5' direction through the complex (McGinnis and Krumlauf 1992). This incredible correlation between the position of a H O M - C gene in the complex and its domain of expression in the embryo is also highly conserved in the mammalian homeotic (or Hox) gene complexes (Duboule and Dolle 1989; Graham et al. 1989; Krumlauf 1994). A developing Drosophila embryo is divided into discrete units termed parasegments. Embryonic parasegments are slightly offset from but roughly correlate with the segments of the adult fly (see Fig. 1). Each of the homeotic genes listed has a unique embryonic expression pattern that is restricted to a specific set of parasegments (see Fig . 1 for the Ubx expression pattern). The parasegmental expression pattern of a particular homeotic gene correlates with the morphological determination of the corresponding set of segments in the adult fly (Lewis 1978; Bender et al. 1983; Sanchez-Herrero et al. 1985; Beachy et al. 1985; Carroll et al. 1986; Akam 1987; McGinnis and Krumlauf 1992). A n alteration in the parasegmental expression domain of a particular homeotic gene can result in the morphological transformation of one segment into another (Lewis 1978; Struhl 1982; 2 Fig. l Comparison of the embryonic expression pattern of the homeotic gene Ubx with the corresponding adult fly segments. Parasegments 1-14 in the embryo roughly correlate with the three head segments (C1-C3), three thoracic segments (T1-T3), and eight abdominal segments (A1-A8) of the adult fly. Ubx embryonic expression in parasegment (PS) 5 is faint and the protein accumulates in the anterior region of PS 5 which corresponds to the posterior of the second thoracic segment (T2). The anterior boundary of PS5 is marked by an arrow. Ubx protein accumulates to high levels throughout PS6 (the PS6 anterior boundary is marked by an arrowhead) and at moderate levels in parasegments 7 to 12. P c G proteins maintain repression of the Ubx gene in PS 1-4. A P c G mutation can cause the PS5 restricted boundary of Ubx expression to break down, thus causing Ubx expression to expand into more anterior (para)segments. Ubx expression in more anterior segments causes transformations of anterior segments into more posterior ones such as the transformation of the wing (a T2 structure) towards the haltere (a T3 structure). The embryo is positioned laterally and oriented with anterior to the left, posterior to the right, ventral down and dorsal up. The fly drawing is taken from page 4 in Lawrence 1992. 3 4 Bender et al. 1983; Beachy et al. 1985; Carroll et al. 1986; Duncan 1987; McGinnis and Krumlauf 1992). For example, i f Ubx is ectopically expressed in parasegment 4, the second thoracic segment (T2) can be transformed towards the third (T3). This would be visible as a transformation of the T2 wing towards the small wing-like structure known as a haltere found on T3 (Lewis 1978). A T2 to T3 transformation is an example of a posterior homeotic transformation, where a more anterior structure is transformed towards a more posterior one. Posterior transformations are seen with gain of function (GOF) regulatory mutations of the homeotic genes where homeotic genes are ectopically expressed beyond their normal expression boundaries (Struhl 1981; Duncan 1987; McGinnis and Krumlauf 1992; Shimell et al. 1994). Mutations in P c G genes also produce G O F homeotic transformations (Slifer 1942; Duncan 1982; Duncan and Lewis 1982; Ingham 1984; Jurgens 1985; Dura et al. 1985). This indicates that the P c G are negative regulators required for the repression of the homeotic genes outside of their normal expression domains. Loss of function (LOF) regulatory mutations of the homeotic genes have the opposite effect, they cause anterior transformations (Struhl 1981; Duncan 1987; McGinnis and Krumlauf 1992). Another set of proteins, the trithorax group, (trxG) are required to maintain normal levels of homeotic gene expression within their spatial expression domains (Kennison 1993). Thus trxG mutations reduce expression and generally cause L O F or anterior homeotic transformations (Kennison 1993). Early versus late regulation of the homeotic genes Homeotic gene regulation is divided into an early initiation stage and a late maintenance stage. Early developmental regulation of the homeotic genes is achieved by the segmentation genes. Products of the segmentation genes set up an early regulatory heirarchy in the embryo that reflects the clear metameric development of later stages (Akam 1987; Ingham 1988). Disruption of homeotic gene regulation during early embryogenesis 5 has an eventual effect on imaginal disc morphogenesis and thus direcdy effects the determination and final morphology of the adult structures. Since segmentation gene expression patterns are transitory and fade away early in development, other factors are required to maintain homeotic gene expression domains in later stages of the developing embryo and in the imaginal discs. In PcG mutants, early expression of the homeotic genes is normal, it is only at later embryonic stages (stage 10 or 11) that homeotic gene regulation breaks down resulting in the ectopic expression of the homeotic genes (Struhl and Akam 1985a; Wedeen et al. 1986b; Riley et al. 1987a; Glicksman and Brower 1988a; Dura and Ingham 1988b; Glicksman and Brower 1990b; McKeon and Brock 1991a; Simon et al. 1992b; Soto et al. 1995). This indicates that the PcG are not required for the initial repression of the homeotic genes but are required for the maintenance of homeotic gene repression. The trxG gene products are also required at this stage for maintenance of homeotic gene activation within their normal expression domains (Kennison 1993). Thus the activity of these two groups of proteins maintains the tight homeotic gene expression patterns that are set up by the early acting segmentation genes. P c G Homeo t i c Phenotypes Most homeotic genes are expressed with restricted anterior expression boundaries. In P c G mutants, these restricted boundaries break down and homeotic genes are ectopically expressed in more anterior regions (Struhl and Akam 1985a; Wedeen et al. 1986b; Riley et al. 1987a; Glicksman and Brower 1988a; Dura and Ingham 1988b; Glicksman and Brower 1990b; M c K e o n and Brock 1991a; Simon et al. 1992b; Soto et al. 1995). The ectopic expression of homeotic genes in P c G mutants is also referred to as derepression. Derepression of the homeotic genes produces distinctive G O F homeotic transformations such as the transformation of the wings to the halteres (small 'pseudowing' structures located posterior to the wing), the transformation of antenna to leg, the transformation of the 4th abdominal segment (A4) to the 5th (A5, most easily seen in males), and the 6 transformation of the first abdominal segment ( A l ) into more posterior segments (Slifer 1942; Duncan 1982; Duncan and Lewis 1982; Ingham 1984; Jurgens 1985; Dura et al. 1985). One exception to the above is the extra sex combs phenotype which results from a break down in the posterior expression boundary of the A N T - C gene Sex combs reduced (Scr) (Glicksman and Brower 1988a). In P c G mutants, Scr is ectopically expressed in the second and third legs transforming them towards the first leg. The first legs of male Drosophila melanogaster have unique structures about midway down the leg termed sex combs. Transformation of the second and third legs into the first results in the appearance of sex combs on the second and third legs. Many of the PcG members were originally identified due to this distinctive phenotype, and it is a characteristic of many PcG genes that they are haplo-insufficient with regards to the sex comb phenotype. That is, heterozygous P c G mutants display sex combs on the second and sometimes the third legs of male flies, indicating that a loss of one copy of the gene is enough to disrupt its normal function. Double heterozygous combinations between different PcG mutations tend to enhance the severity of the extra sex combs phenotype and other P c G homeotic phenotypes (Hannah-Alava 1958; Kennison and Russell 1987; Kennison and Tamkun 1988; Adler et al. 1989; Cheng et al. 1994; M c K e o n et al. 1994; Campbell et al. 1995) The P o l y c o m b group genes A l l the P c G members identified so far maintain repression of the homeotic genes outside their normal expression domains. To date, about 15 P c G members have been identified; extra sex combs (esc), Polycomb (Pc), polyhomeotic (ph), Posterior sex combs (Psc), Enhancer ofzeste (E(z)), Polycomb-like (Pel), Additional sex combs (Asx), Sex combs on midleg (Scm), Sex combs extra (See), super sex combs (sxc), pleiohomeotic (pho), multi sex combs (mxc), Suppressor ofzeste 2 (Su(z)2), Enhancer of Polycomb (E(Pc)) (Simon 1995) and cramped (crm) (Yamamoto et al. 1997). Most members have 7 been identified either by their strong adult homeotic transformations (Slifer 1942; Lewis 1978), cuticle defects (Nusslein-Volhard et al. 1984) or enhancement of P c G or G O F homeotic phenotypes (Botas et al. 1982; Kennison and Russell 1987). It is likely that most of the stronger members of the group have already been identified. However, deletion analysis of the genome looking for enhancers of PcG homeotic phenotypes has produced estimates that there may be up to 40 PcG genes in total (Jurgens 1985; Landecker et al. 1994). There is strong evidence that PcG proteins form large functional complexes with one another. Polycomb (Pc) and polyhomeotic (ph) proteins co-immunoprecipitate as members of a large M W complex (Franke et al. 1992). The Sex combs on midleg (Scm) and ph proteins physically interact in vitro (Peterson et al. 1997) and domains from Pc, ph and Posterior sex combs (Psc) proteins also interact in vitro (Strutt and Paro 1997; Kyba and Brock 1998). Additionally, antibody binding indicates that Pc, ph, and Polycomblike (Pel) proteins completely overlap at about 100 sites on the large polytenized chromosomes from the larval salivary glands (Franke et al. 1992; Lonie et al. 1994). Scm overlaps with these proteins at all sites examined (Peterson et al. 1997, a complete list of Scm binding sites is not yet published) while Psc, Suppressor of zeste 2 (Su(z)2), and Enhancer of zeste (E(z)) overlap at a subset of these polytene binding sites as well as binding to a number of unique sites (Rastelli et al. 1993; Cairington and Jones 1996). A l l these P c G proteins bind to the homeotic loci . This suggests that many of the PcG proteins form at least one functional complex in vivo and complex formation is probably required at a large number of chromosomal target sites. P c G proteins act on specific target loci through regulatory elements termed Polycomb Group Response Elements (PREs). PREs were originally identified as regulatory maintenance elements that were required to maintain homeotic gene expression patterns late in embryogenesis (Simon et al. 1990; Zink et al. 1991; Muller and Bienz 1991). Enhancer regions from various B X - C and A N T - C genes have been shown to drive 8 the expression of reporter constructs in a homeotic gene specific expression pattern (Simon et al. 1990; Zink et al. 1991; Muller and Bienz 1991). However, these highly regulated expression patterns begin to break down at about 3-6 hrs., producing ectopic expression of the reporter constructs. Maintenance of spatially restricted expression patterns late in embryogenesis only occurs when a P R E is present in conjunction with the gene enhancer elements (Zink et al. 1991; Muller and Bienz 1991; Busturia and Bienz 1993; Simon et al. 1993; Pirrotta et al. 1995; Chan et al. 1994; Chiang et al. 1995). Maintenance of these spatially restricted expression domains is abolished in P c G mutants (Zink et al. 1991; Muller and Bienz 1991; Busturia and Bienz 1993; Simon et al. 1993; Pirrotta et al. 1995; Chan et al. 1994; Chiang et al. 1995) indicating that these P R E maintenance elements are dependent on functional P c G activity . PREs have also been identified by their P c G dependent variegated repressive effects on a white reporter gene (Fauvarque and Dura 1993; Kassis 1994; Gindhart and Kaufman 1995; Hagstrom et al. 1997). Not all PREs necessarily act as maintenance elements, but all maintenance elements that have been identified are PREs. For the sake of simplicity, P R E maintenance elements wi l l simply be referred to as PREs. Transgenes with a P R E present produce an ectopic P c G binding site on polytene chromosomes (Chan et al. 1994; Chiang et al. 1995), indicating that PREs probably contain binding sites for P c G proteins. In vitro analysis of the bxd5.1 element, a minimal P R E from the regulatory region of the homeotic gene Ubx, has identified several high and low mobility complexes that contain the PcG protein ph and bind to distinct subelements within the 5.1 element (Hodgson and Brock 1998). Thus, at least in the case of the bxd P R E , maintenance elements bind large complexes containing P c G proteins that function in maintaining the repression of target loci late in development. M o d e l s of P c G funct ion 9 The specific mechanism of PcG mediated repression of target loci is unknown. Several models of P c G function have been suggested; 1) that P c G proteins repress target loci by compacting the D N A into a heterochromatin-like structure (Paro 1990), 2) that P c G proteins target D N A segments to repressive "compartments" in the nucleus (Schlossherr et al. 1994), 3) that P c G proteins inhibit transcription initiation (Bienz 1992), or 4) that P c G proteins interfere with the interaction between enhancer elements and the promoter by "looping" out the D N A (Pirrotta 1995). The main problem with all of these models is that none of them have any strong data to support them. The heterochromatin model was originally suggested because the P c G protein Pc and the chromatin protein HP1 share a functional domain in common. HP1 is encoded by a Suppressor of position effect variegation (Su(var)), Su(var)205 . Su(var)s appeal- to be modifiers of heterochromatin (Henikoff 1990). However, most P c G mutations that have been tested are not modifiers of P E V and of those that are, only E(Pc) acts as a Su(var) (Sinclair et al. 1998a). A s well, alterations in regulation of the homeotic loci do not affect restriction enzyme accessibility suggesting that the D N A is not in a compacted state (Schlossherr et al. 1994; M c C a l l and Bender 1996). The compartment idea has not been disproven but it does not seem likely to be true. P c G proteins, in general, bind to about 100 sites on both polytene and mitotic chromosomes (Franke et al. 1992; Buchenau et al. 1998) and these sites are not localized to a specific region of the nucleus, so one would have to argue that there are at least 100 separate repressive compartments in the nucleus. P c G proteins do not seem to interfere with basal transcription since transgenes inserted within the homeotic loci are still transcribed even in regions where the endogenous homeotic gene is repressed (McCal l and Bender 1996). Thus i f they do interfere with transcriptional initiation, it must be through promoter specific elements and not through the general transcriptional machinery. Finally, there is no evidence for the existence of looping structures. Presumably, the looping model would predict that PREs isolate enhancer elements and make them 10 unavailable for enhancer-promoter interactions. Constructs that are flanked by PREs are still subject to position specific alterations in reporter gene expression (Zink and Paro 1995) suggesting that the enhancer and promoter elements have not been isolated from interacting with external elements. However, (Muller 1995) has shown that stable P c G repressive complexes can be assembled at artificial target sites i f at least one P c G protein is targeted to the site and i f there is an enhancer element and promoter from the endogenous Ubx locus present. This indicates that there is specificity in P c G mediated repression of enhancer-promoter interactions and suggests that PcG activity is somehow dependent on P R E -enhancer-promoter cross talk. That PREs can crosstalk with each other is supported by the observation that transgenes containing a P R E show stronger repression of the reporter when they are paired in trans (Fauvarque and Dura 1993; Kassis 1994; Chan et al. 1994; Zink and Paro 1995; Hagstrom et al. 1996; Hagstrom et al. 1997) than when they are present as single elements. One possibility is that weak P c G binding sites in enhancer regions somehow crosstalk with strong P c G binding sites located on PREs and interfere with the function of activation complexes. Double mutations of PcG and trxG genes generally lead to the suppression of the homeotic transformations seen with mutations in either group alone (Ingham 1983; Capdevila et al. 1986; Kennison and Tamkun 1988). This has led to the suggestion that P c G and trxG proteins either have opposite and independent functions (Ingham 1983) or they function antagonistically to one another, possibly involved in a stoichiometric competition for repression or activation at specific target loci (Jones and Gelbart 1993). Thus another model of PcG mediated repression is that binding of P c G proteins to multiple regulatory sites within a locus directly interferes with the activity of trxG proteins and prevents them from activating loci in specific embryonic domains. The Additional sex combs (Asx) gene 11 The P c G gene Additional sex combs (Asx) shares many general attributes with other P c G genes, but it has some distinctive aspects that may shed light on P c G repressive activity. First of all, most P c G mutations show the same extent of homeotic gene derepression in all embryonic tissues. Asx, however, has some very strong tissue specific effects. It is required for homeotic gene regulation in the epidermis but has only a minor effect on regulation in the central nervous system (CNS) (McKeon and Brock 1991a; Soto et al. 1995). This suggests that unlike other P c G genes, Asx activity is targeted in both a parasegment specific and a tissue specific manner in the embryo. Secondly, Asx mutations generally display weak P c G homeotic transformations, but the allele Asx?! strikingly displays both trxG and P c G homeotic transformations (Sinclair et al. 1992). This suggests the possibility that Asx could have a role in mediating both P c G repression and trxG activation of the homeotic loci. Finally, many Asx mutations display segmentation defects (Sinclair et al. 1992), a phenotype that is usually associated with mutations in early patterning events. This possibly implicates Asx as having a possible role in initiation as well as maintenance. Taken together, an analysis of Asx could tell us 1) possible ways of targeting P c G repressive activity in a tissue specific manner, 2) the relationship between P c G mediated repression and trxG mediated activation at target loci, and 3) possible implications for P c G function in initiation as well as maintenance. Both molecular and genetic approaches were used to analyze basic Asx function in relation to the activity of other PcG genes. Since different Asx mutations may disrupt different aspects of Asx function, the regulation of two different homeotic loci and a Ubx P R E were examined for allele specific requirements for Asx activity in the embryo. Antibodies raised to the Asx protein were used to look for possible relationships between the distribution of Asx protein both in the embryo and on polytene chromosomes and the tissue and target specific requirements for Asx activity that were observed. To examine the possibility that Asx could function as both a member of the P c G and the trxG, Asx mutations were tested for their ability to enhance both P c G and trxG homeotic mutant 12 phenotypes. A n d finally, two different approaches were used to try to identify interacting proteins that were important for Asx function. First of all, interacting proteins identified in a yeast-2 hybrid screen were analyzed for possible effects on homeotic gene regulation. Secondly, an attempt was made to clone the P c G gene super sex combs (sxc), a P c G gene that shows a strong genetic interaction with Asx. The results of these experiments are discussed within the context of Asx function in particular and P c G function as a whole. 13 Materials and Methods Fly strains The fly strains used in this study are described in Lindsley and Zimm 1992 and on Flybase (http://flybase.bio.indiana.edu/). Some additional specific references are also given below. Asxl, Asx^, Asx$, Asx$, and Asx^ are all Asx homozygous lethal alleles that are described in Jurgens (1985) or Sinclair et al. (1992). Because these Asx mutations have stronger phenotypes than a deletion of Asx, they are probably gain of function Asx alleles. The Asx deficiency Dfi2R)trix is described in Breen and Duncan 1986 and is probably the only known null allele of Asx. Df(3R)red is a large deletion the removes the trx locus and is therefore a null. The trx^^ allele has a deletion in the coding region of trx and is probably a null and trx^l is a point mutation in the SET domain, a protein domain conserved with some other regulatory proteins (Stassen et al. 1995). Pc4 is a strong, probably null allele. Pcfi is a strong Pel hypomorph (Jurgens 1985). The Asx?! allele is homozygous viable and is described in Sinclair et al., 1992. The Asx?! chromosome carries the recessive eye mutations cn and bw so homozygous Asx?! flies are identifiable by their white eyes.The Asx^^cn bw/Asx^^cn bw ; Ly/TM3 stock was constructed in the following manner. Asx?! cn bw/Asx^l cn bw; +/+ female flies were crossed to a +/+ ; Ly/TM3 stock. Asx?! cn bw/+ ; Ly/+ and Asx?! cn bw/+ ; TM3/+ male flies were collected from the F l and separately back crossed to the Asx?! cn bw/Asx^l cn bw ; +/+ stock. Asx?! cn bw/Asx^l cn bw ; Ly/+ and Asx?! cn bw/Asx^l cn bw; TM3/+ flies were collected from the F2 and crossed to each other. Asx?! cn bw/Asx^l cn bw; Ly/TM3 flies were collected from the F3 and used to make a stock. Df(3R)ZN47 and all other interactor deficiency strains were acquired from the Bloomington stock center. Transformed lines containing the bxd!4 element were obtained 14 from W . Bender and are described in Simon et al 1990 and 1993. A l l sxc alleles are described in Ingham, 1984. Immunostaining of mutant and wild type embryos Embryos were collected by allowing flies of the appropriate genotype to lay overnight at 25° on 2% agar plates that were coated with a thin film of 5% ethanol, 5% acetic acid and 5% yeast. Plates were also supplemented with a sprinkle of dry bakers yeast. Embryos were fixed as follows. Overnight collections were rinsed with tap water and strained through a sieve. Embryos in the sieve were placed in a 50% bleach solution for 3 min. to remove the cuticle. Embryos were then placed in an eppendorf tube containing 700LI1 of heptane, 630iil of P B S (pH 7.0) and 70ul of paraformaldehyde and rotated for 20 min. at room temperature. The PBS/paraformaldehyde layer was removed and 700pJ of methanol was added. The embryos were then vortexed for 2 min. and any embryos that did not sink to the bottom of the tube were removed along with all liquid. Embryos were then rinsed 3 X with methanol and stored in methanol at -20° for up to two months or until needed. Immunostaining was performed by washing fixed embryos of the desired genotype 3 X with a P B T solution (0.1% TritonX, 0.2% B S A in PBS) , 1 X 30 min. with P B T , 1 X 30 min. with P B T + 2% normal serum, and then incubating with the primary antibody at the appropriate dilution rotating overnight at 4°. The next day, embryos were again washed 3 X with P B T , then 4 X 30 min. with P B T , 1 X 30 min. with P B T + 2% normal serum, and then incubated with a secondary antibody. For H R P (Horse Radish Peroxidase) staining, the secondary antibody was either conjugated directly to H R P or conjugated to biotin. Secondary antibody was added at the appropriate dilution (1/200-1/500 for an H R P conjugated secondary and 1/5000 for a biotinylated secondary) and the embryos were incubated for 2-4 hrs. at room temperature or at 4° overnight. After incubation with the secondary, embryos were washed as above 15 omitting the P B T + 2% normal serum wash. If amplification was being used, avidin conjugated to H R P was incubated at a dilution of 1/1000 at room temperature for one hour. Washes were then repeated as above. The H R P signal was detected by incubating the embryos with 0.3 mg/ml D A B (Diamnobenzidine) in P B T for 10 min. 30% H2O2 was added at a concentration of 1/1000 and the reaction was allowed to proceed for 10 min. After the reaction was complete, embryos were washed with PBS then washed with a dilution series of ethanol (30%,70% then 95%). A l l ethanol was removed and the embryos were then mounted in Gary's Magic Media ( G M M ) , a 1:1 mixture of Canada Balsam and methyl salicylate. Embryos were viewed using Nomarski optics and photographed using Kodak Ektachrome 160T slide film (exp. +1). For immunofluorescence, the secondary was conjugated to Texas Red and incubated in the dark at a concentration of 1/200 for 2 hrs. at room temperature. P B T washes were carried out as before, embryos were rinsed in P B S and then 1:1 D A B C O : P B S was added ( D A B C O = 2.5mg/ml l,4-diazobicyclo-(2,2,2)-octane in 90% glycerol, used to retard photobleaching). Embryos were allowed to settle, the D A B C O : P B S was removed and D A B C O was added. The embryos were mounted on slides and fluorescent images were collected using a Bio-Rad 600 confocal microscope. Confocal images were projected using N I H image and the projected images were arranged using Photoshop 4.0. The following primary antibodies were used. The monoclonal antibody FP3.38 was used at 1/10,000 with amplification (1/1000 without) to detect Ubx protein. The monoclonal antibody 8C11 was also used at 1/10,000 with amplification to detect AntP protein. A monoclonal antibody to Scr protein was used at 1/100 and a monoclonal antibody to AbdB protein was used at 1/10 with no amplification. The axon specific monoclonal antibody BP102 was used at 1/100 to detect the C N S of Asx mutant embryos. A n anti-6 gal antibody was used at a dilution of 1/2,000 without amplification to detect lacZ staining in embryos. Mouse and Sheep anti-Asx antibodies were used at a dilution of 1/200. 16 A l l embryos were staged according to Campos-Ortega and Hartenstein 1985. Homozygous Asx and other PcG mutant embryos were identified by their homeotic misexpression phenotypes, except for Dfi2R)trix embryos stained with the axon specific antibody BP102 which were identified by double staining a Df(2R)trix/CyOj3elav-lacZ stock with BP102 and an anti-Bgal antibody. Homozygous Df(2R)trix embryos stained only with BP102 and did not show the fielav lacZ expression pattern. This was also used to identify Df(2R)trix homozygotes that were double stained with anti-Bgal and a mouse anti-Asx antibody. Asx^ mutants were also stained with B P 102 and all embryos were scanned equally for possible C N S defects. Df(3L)ZN47 homozygotes were identified by their morphological defects. Immunostaining of polytene chromosomes 20 pairs of wi ld type or Asx^cn bw/Asx^^cn bw flies were allowed to lay for two days on standard cornmeal sucrose medium containing tegosept as a mold inhibitor. Dry yeast was added periodically and the larvae were raised at 17°. Late stage third instar larvae were picked off the sides of the bottles and salivary glands were dissected out of them in a 0.1% Tri ton-X/PBS solution. Glands were then transferred to a droplet of pre-fixative for 6-20 sec. for Asx and 15-20 sec. for Pc. The glands were then transferred to a droplet of fixative on a siliconized coverslip and left for 2-3 min. Two different pre-fixative and fixative solutions were used. A 1% paraformaldehyde(Pf)/l% Triton-X in P B S pre-fix with a 1% Pf/45% acetic acid fix or similar solutions with an increased Pf concentration of 3.7%. It was found that the lower (1%) Pf concentration improved chromosome morphology without reducing the amount of signal. The coverslips were picked up with poly-L-lysine treated slides and the glands were broken up with a pencil tip and squashed using thumb pressure. Coverslips were removed using liquid N2 and a razor blade and the slides were stored in P B S . Chromosomes with good morphology were used immediately for immunostaining. 17 Chromosomes were washed in a blocking buffer (10% nonfat dry milk, 1% B S A , 0.2% NP40 and 0.2% Tween) for 30 min. at room temperature. The primary antibody was added at the appropriate dilution (1/20 for sheep and rabbit Asx antibodies and 1/50 for the rabbit Pc antibody, diluted in blocking buffer) and incubated in a moist chamber at 4° overnight. The next day, slides were washed 2 X 1 5 min. in wash buffer (300mM N a C l , 0.2% NP40,0.2% Tween 20 in PBS) and incubated with the secondary antibody. For immunofluorescence, the slides were incubated at room temperature in the dark in a moist chamber for 1-4 hrs. The secondaries used were either conjugated to FITC (used at a dilution of 1/100) or Texas Red (used at a dilution of 1/50). Slides were then washed as before and mounted in D A B C O . Fluorescent images were collected using a Bio-Rad 600 confocal microscope, except for Pc binding to Asx?! chromosomes where photographs were taken using Kodak 400ASA color slide film on a Zeiss Axiophot Photomicroscope. For H R P staining, slides were incubated for 1-2 hrs at room temperature with a biotin conjugated secondary. Washes were performed as before and preincubated biotin-avidin complexes from the Vectastain kit were added to the slides. The complexes were allowed to incubate with the slides for 40 min. at room temperature and then the slides were washed in 2 X P B S . The H R P reaction was allowed to proceed by adding 0.5 mg/ml D A B and 1/1000 H2O2. Chromosomes were either counterstained with 5% Giemsa which stains them a blue-gray or with 0.1 mg/ml D A P I , a flourescent dye that is quenched by H R P staining. Photos were taken using both bright field and phase contrast optics with Kodak 160T color slide film (exp. +1). Crosses A l l crosses were performed at 25°. Flies were raised on standard cornmeal sucrose medium containing tegosept as a mold inhibitor. 15-20 females were each crossed to 15-20 males of the appropriate genotypes. Crosses were turned over once after allowing them to 18 lay for four days. Parents were discarded after four days and the F I was allowed to eclose. F I flies were then scored at two day intervals over a 10 day period. Asx?! homozygous crosses were performed in the following manner. A stock of Asx^^cn bw/AsxPlcn bw; Ly/TM3 flies were crossed to balanced homozygous lethal mutants on the third chromosome (ie. Df(3L)ZN47, Pc^ and various trx alleles). Males of the genotype Asx^^cn bw/+ ; mutant/TM3 or Ly were collected from the F I . Since there is no crossing over in males, loss of the chromosomal markers at this stage was not a problem. Asx^^cn bw/+ ; mutant/TM3 or Ly flies were then crossed to a homozygous AsxPlcn bw/Asx^^cn bw; +/+ stock. The genotypes of the F2 progeny were identified in the following manner. Asx^^cn bw/Asx^^cn bw ; mutant/+ flies were white eyed (due to homozygosity of the recessive cn bw eye markers) but otherwise wild type. Asx^^cn bw/Asx^^cn bw; +/TM3 or Ly flies were white eyed but also carried a second dominant marker (either Ly or Sb). Asx^^cn bw/+ ; mutant/+ flies had wild type eyes and no dominant markers while Asx^^cn bw/+ ; +/TM3 or Ly flies had wild type eyes plus a dominant marker. Crosses with the Asx^^ revertant line Asx^^ were done the same way. To test for interactions between homozygous lethal Pc and Asx alleles, mutant balanced stocks were crossed to one another and the transheterozygous adults in the F I were scored for the penetrance of extra sex combs on the second and third legs and for the penetrance of posterior abdominal transformations (scored as patches of pigment in abdominal segments 2-4). The penetrance of P c G homeotic phenotypes was compared to sibling single Pc and Asx mutants. Single versus double mutants were easily distinguished via independant assortment of the mutant alleles in relation to dominant markers carried on the second and third chromosome balancers. To test for interactions between homozygous lethal trx and Asx alleles, mutant balanced stocks were crossed to one another and the transheterozygous adults in the F I were scored for overall penetrance of adult homeotic transformations of either the haltere to the wing or the fifth abdominal segment to a more anterior segment. Haltere to wing 19 transformations were scored as positive i f the haltere was bloated or i f triple row margin bristles were present. Anterior transformations in the abdomen were scored as patches lacking pigment in the 5th and 6th abdominal segments. Transheterozygous mutants were compared to their siblings that contained single mutant alleles for either trx or Asx . Single versus double mutants were easily distinguished via independant assortment of the mutant alleles in relation to dominant markers carried on the second and third chromosome balancers. A C h i square test was perfomed on all Asx/trx crosses to determine i f there was a statistically significant increase in the penetrance of trxG homeotic mutations in the transheterozygotes compared to the single trx mutants. In each case, the number of single trx mutant individuals with trxG homeotic mutations was used to calculate the number of individuals we would expect to see in the double heterozygotes i f there was no enhancement. For example, in the cross Asx^ females crossed to Df(3R)red males, 31/97 double heterozygotes and 16/113 single trx mutants had trxG phenotypes. 16/113=0.1416, 0.1416 * 97=13.7. Therefore, i f there was no enhancement, we would expect about 14/97 double mutants to have trxG phenotypes. The actual number seen was 31. The C h i square calculation was performed as usual to yied a p value of 20.6. The Chi square table p value for 5% at 1 df was 7.879 indicating that in this case, the hypothesis that there was no enhancement was rejected. The sxc screens were conducted at 25° on standard media as above. A l l crosses were en masse except for the final selection crosses which were single pair matings between one male and 5-10 females. Otherwise, screens were performed as described in the results section of Ch.4. M o u n t i n g and photographing flies F ly abdomens were cut dorsally with a razor blade and mounted on slides in G M M (1:1 methyl salicylate : Canada Balsam). Slides were cleared at 65° for 2-5 days. Thoraxes 20 were prepared by first boiling the flies for 10 min. in 10% K O H . The flies were then washed with distilled water and then a dilution series of ethanol (30%, 70%, 95%, 100%). The dorsal thorax was dissected using a pair of dissecting scissors and mounted in G M M . A l l fly parts mounted in G M M were photographed using bright field optics and Kodak Tech Pan film (exp. -1,0,-1-1). Whole flies were photographed using a dissecting scope and Kodak 160T slide film. Scanning electron microscope (SEM) images were collected with no sample preparation. Flies were mounted on tabs and the S E M image was collected immediately. DNA in situ hybridization to polytene chromosomes 20 pairs of wild type flies were allowed to lay for two days on standard cornmeal sucrose medium containing tegosept as a mold inhibitor. Dry yeast was added periodically and the larvae were raised at 17°. Late stage third instar larvae were placed in a droplet of 0.8% N a C l solution and then into a droplet of 45% acetic acid. Salivary glands were dissected and transferred to a droplet of fixative (3:2:1 Acetic acid:water:lactic acid) on a siliconized coverslip. Fixation was allowed to occur for 4-5 min. and then the glands were broken up by tapping the coverslip with the end of a pencil. The pencil was used to spread the chromosomes by trailing it across the coverslip in a zig zag motion and the slides were left at 4° overnight. This allows the fixative to evaporate and flattens the chromosomes. Coverslips were removed by plunging the slide into liquid nitrogen and flipping them off with a razor blade. Slides were plunged into 95% ethanol for 10 min., air dried, and stored at room temperature until needed. A standard lab protocol was followed for the labelling and hybridization of probes to polytene chromosomes. The general approach is outlined below. Interactor c D N A probes were labelled using a Boehringer-Mannheim digoxigenin (DIG) random prime labelling kit. lp.g of gel purified template D N A was used for each reaction and the protocol was followed as stated in the kit. Once the reaction was complete, 21 the labelled probe was precipitated along with 40|ig of carrier D N A (sonicated salmon sperm D N A , denatured) using NaOAc and ethanol. The D N A pellet was resuspended in water, precipitated again and then resuspended in 150 (il hybridization buffer. (0.6M N a C l , 0 .05M N a P 0 4 p H 7.0 buffer, l x Denhardt's reagent, 5% dextran sulphate and 50% formamide) Slides were heat treated at 65° for 30 min. in 2 X S S C , l X l O m i n . in 2 X S S C , and then rinsed in 70 % then 95% ethanol. The chromosomes were then treated with 0.07 N a O H for 2.5 min. and then rinsed as before. The probe was denatured by heating it to 70° for 10 min. and 10 p:l was added to 22mm^ coverslips. The coverslips were picked up by each slide and the edges were sealed with rubber cement. The slides were incubated in a moist chamber at 37° overnight. The next day, the rubber cement was removed and coverslips were rinsed off of the slides and slides were washed in 2 X S S C at 65° and followed by multiple washes at room temperature. The slides were then washed in a blocking buffer (lOOmM Tr i s -HCl , p H 7.5, 150 m M N a C l plus 1% block from B M kit) and an A P (alkaline phosphatase) conjugated anti-DIG antibody (from the B M kit) was added to 22X40 coverslips at a dilution of 1/5000. Incubation with the antibody was in a moist chamber for 60 min. Slides were rinsed twice in a wash buffer ( lOOmM Tr i s -HCl , p H 7.5,150 m M NaCl) and the A P reaction was detected using standard N B T and X -phosphate substrates. Slides were examined under the microscope using phase contrast optics. Black and white photos were taken with Kodak Tech Pan film (exp. +1) and color photos were taken with Kodak 160T slide film. Northern and Southern Blotting Genomic D N A was recovered from 6-10 adult flies by first homogenizing the flies in 200 i l l of a buffer containing 0 .1M Tris (pH 9.0), 0.5% SDS, 50mM E D T A , 5% sucrose and 100p,g/ml protease K and incubating them for one hour at 50°. 50pl of 5 M potassium acetate was added and mixed by inversion and the mixture was left on ice for 10 22 min. The mixture was centrifuged and the supernatant was removed and placed in a new tube. The supernatant was extracted with phenol/CIA (95% chloroform, 5% Iso amyl alcohol) and followed by a C I A extraction and D N A was ethanol precipitated out of the solution (no extra salt was added). The D N A pellet was redissolved and digested with the appropriate restriction enzyme and subjected to agarose gel electrophoresis for blotting. The gel was blotted to Hybond N filter paper in 10X SSC using a standard gravity blot set up. D N A probes were radioactively labelled with a Boehringer-Mannheim random prime kit to a specific activity of at least 1 x 10^ cpm/Lig. A standard protocol for hybridizing probes to genomic blots was used with the following alteration. A phosphate buffer containing 0.1M NaH2PC>4, 0.05M Na4P2C»7-10 H2O, I m M E D T A , 7% SDS and 100Lig/ml Salmon sperm D N A was used for pre-hybridization and hybridization. A blot containing embryonic, larval, pupal and adult poly ( A ) + m R N A s (Hugh Brock and Kryn Stankunas) was hybridized as described above for Southern blots. 23 Chapter 1 Introduction Additional sex combs (Asx) was originally identified independently in a screen for embryonic lethal mutations that caused pattern defects in the larval cuticle (Nusslein-Volhard et al. 1984; Jurgens 1985), and as a mutation that enhances the dominant sex combs phenotype of Pc mutants (Dura et al. 1985; Jurgens 1985). Homozygous Asx mutants die at the end of embryogenesis with severe head defects and mild posterior transformations in the cuticle of the abdomen (Jurgens 1985; Breen and Duncan 1986). Because the posterior transformations were less severe than those normally seen with other P c G mutations, Asx was categorized as a weak member of the P c G . The cuticle phenotype of a P c G gene reflects the state of homeotic gene regulation during embryogenesis. Specifically, posterior transformations are a consequence of ectopic expression of homeotic genes outside of their normal domains of expression. Consistent with the above observation, homeotic derepression phenotypes in the C N S (central nervous system) are less severe in Asx mutant embryos compared to most other P c G genes (McKeon and Brock 1991a; Simon et al. 1992b) even when the maternal contribution of Asx is removed (Soto et al. 1995). Most P c G mutations show equal levels of homeotic gene derepression in all tissues (Soto et al. 1995; M c K e o n and Brock 1991a; Simon et al. 1992b). Although Asx mutations have only weak effects in the C N S , they are unusual in that they have a much stronger homeotic derepression phenotype in the epidermis (Soto et al. 1995; M c K e o n and Brock 1991a; Simon et al. 1992b). Polycomblike (Pel) is the only other P c G member that has clear tissue specific effects, but it is the opposite of Asx in that homeotic genes are strongly derepressed in the C N S but only weakly derepressed in the epidermis. This tissue specific pattern of homeotic gene derepression in Asx mutants suggests that Asx may not be globally required like most of the other P c G proteins, and instead may have a highly 24 specific role in homeotic gene regulation. One possibility is that Asx mediates crosstalk between different PREs and tissue specific enhancers. A s well as its tissue specific effects, one allele of Asx, Asx^f is unusual in that it occasionally displays trxG mutant phenotypes as well as typical P c G mutations in the adult fly (Sinclair et al. 1992). This allele implicates Asx in both trxG and P c G activity and therefore a possible role in the maintenance of activation, as well as its role in the maintenance of repression. In this chapter, I set out to determine: 1) i f Asx would regulate endogenous homeotic loci and a reporter construct containing a Ubx specific P R E in the same tissue specific manner; 2) i f the tissue specific activity of Asx would be reflected in a tissue specific expression pattern of Asx protein in the embryo; 3) i f Asx protein bound to the same chromosomal target sites as other P c G proteins or i f the differences in Asx activity would also be reflected in the binding of Asx to unique target sites; and 4) possible mechanistic explanations for the effect of the Asx?! mutation. It was found that although there are stage and tissue specific requirements for Asx activity, Asx protein is expressed with no tissue or stage specificity at all. Asx binds to multiple target sites that contain other P c G proteins but also binds to some unique sites. These unique binding sites may reflect some of the unique aspects of Asx activity. Finally, the binding of Asx to AsxPl mutant chromosomes was examined but found to be completely wild type. This suggests that the Asx?! mutation does not reduce the accumulation of Asx protein or affect its ability to bind to chromosome target sites. Results Asx mutations have tissue specific effects on homeotic gene regulation in the embryo In the developing Drosophila embryo, each homeotic gene is expressed with a unique parasegmental register. Fig.2 (A-C) compares the expression of the homeotic gene 25 Ultrabithorax (Ubx) in the Asx mutant Df(2R)trix (a deletion that removes the entire Asx locus) with the P c G mutant Pcfi. Pel is considered to be only a moderately strong member of the P c G , but the derepression of Ubx in the C N S is far more extensive than that seen in an Asx mutant (compare Fig.2 B and C) . Previous studies with Asx have shown that there are tissue specific differences in Asx dependent homeotic gene regulation. Asx is required to maintain the proper anterior expression boundaries of the homeotic genes AbdB and Ubx in the epidermis of the developing embryo (McKeon and Brock 1991a; Simon et al. 1992b; Soto et al. 1995). However, the Asx mutation Df(2R)trix only slightly affects these boundaries in the C N S (McKeon and Brock 1991a; Simon et al. 1992b) and the Asx mutation Asx^ does not affect these boundaries in the C N S at all (Soto et al. 1995). To determine i f these differences in the literature reflect an actual difference in the effect of these two Asx mutations on homeotic gene regulation in the C N S , the Df(2R)trix and Asx? alleles were re-examined for alterations in the proper regulation of the homeotic genes Ubx and AbdB. Fig.3 C and D confirms that both the Df(2R)trix and Asx^ mutations show a break down in regulation of AbdB in the epidermis, but contrary to Simon et al (1992), there is no alteration in the pattern of AbdB expression in the C N S (compare to wild type, F ig . 3A) . Contrary to Soto et al (1995), Asx^ was found to cause a breakdown in the regulation of Ubx in the C N S which is similar to that seen in Df(2R)trix mutants (Fig.3 G and H). The extent of ectopic Ubx expression in the C N S of Asx mutants is variable and is sometimes seen extending into parasegment (PS) 4 only (see Fig.2 B) , into both PS3 and PS4 (Fig.3 H) , or sometimes as far as PS2 (Fig.3 G) . These results show that Asx function is required for Ubx regulation in the C N S , although it is not as important as other PcG genes such as Pel (see Fig.2 A - C for a comparison). Asx function is not required in the C N S for the regulation of Abd B but it is essential for both Abd-B and Ubx regulation in the epidermis. No other P c G gene shows this particular pattern of tissue specific regulation of the homeotic genes. 26 F i g . 2 A comparison of the C N S derepression phenotype of the homeotic gene Ubx between Asx and the moderately strong PcG gene Pel. Embryos are positioned laterally and oriented with anterior to the left, posterior to the right, ventral down and dorsal up. In Fig. 1 A , the C N S is situated ventral to the midgut (a) and the hindgut (long thin tube-like structure, b). In wi ld type embryos (Fig. 1 A ) , Ubx is normally expressed in a very strong, spatially restricted pattern in the C N S . Ubx expression is faint in parasegment (PS) 5 (the PS 5 anterior boundary is marked by an arrow), strong in PS 6 (the PS 6 anterior boundary is marked by an arrowhead) and at moderate levels in parasegments 7 to 12 (Fig. 1 A ) . In an Asx mutant embryo, the PS 5 anterior boundary breaks down and misexpression of Ubx is seen in PS4 (Fig. 1 B) . Pel mutant embryos show a much stronger breakdown of Ubx regulation and Ubx expression is seen all the way into the posterior region of the brain (Fig. 1 C) . 27 28 F i g . 3 The regulation of the homeotic genes Ubx and AbdB and the bxdl4 maintenance element in Asx mutant embryos. The embryos in A - H are late stage 15 (~ 12-13 his. old) and are oriented with anterior to the left, posterior to the right and with the ventral region facing the viewer so as to make comparisons between the C N S and epidermal expression patterns easier. They are stained with either an anti-AbdB (A-D) or an anti-Ubx (E-H) primary antibody which is then detected with an H R P conjugated secondary antibody. The embryos in I-L are positioned laterally with anterior to the left and posterior to the right. They are stained with an anti-(3gal antibody which is then detected with an H R P conjugated secondary. The embryos in I -K are at late stage 9 (~ 4 hrs. old) and the embryo in L is at stage 11 (~ 7 hrs. old). A , wild type Abd-B expression. Abd B is normally expressed only in the extreme posterior of the embryo with a C N S expression boundary at PS 10 (marked with a straight line in the C N S ) . B , Abd-B expression in Asx?* mutants is wild type. C and D , Abd-B expression in an Asx^ and a Df(2R)trix mutant respectively. The PS 10 boundary in the C N S is maintained but ectopic expression in the epidermis is apparent (large arrow points to ectopic expression in the epidermis of PS 11). The Df(2R)trix mutant has been dissected dorsally in order to show the epidermal expression of Abd-B more clearly. E , wi ld type expression of Ubx, described in detail in F i g . l . F , Ubx expression in AsxPl mutants is wild type. G and H , Ubx expression in an Asx^ and a Df(2R)trix mutant respectively. Ectopic Ubx expression is seen in PS 4 and 3 (G and H) as well as occasionally as far as PS2 (G). The PS 6 boundary is marked with an arrowhead and PS 3 and PS 4 boundaries are marked with small arrows. I, lacZ expression of the bxdl4 reporter construct in a wild type embryo. It has a boundary of expression at PS6 (arrowhead). J - L , expression of the bxd l4 transgene in an Asx?! , an Asx^ and a Df(2R)trix mutant respectively. In all cases the PS 6 boundary breaks down and the transgene is expressed up to PS 2. 29 30 The Asx?! mutation does not affect endogenous homeotic gene expression in the embryo but is required for regulation of a Ubx transgene AsxPl displays both trxG mutant phenotypes in the adult fly (abdominal segment 5 and 6 transformed towards a more anterior segment and transformation of the haltere towards the wing) as well as typical PcG mutations (anterior abdominal segments transformed towards more posterior segments). These adult homeotic phenotypes are a direct consequence of alterations in expression of the homeotic genes Ubx, AbdA and AbdB. T r x G homeotic transformations are caused by a reduction in the expression of the homeotic genes within their normal expression domains. To see i f the adult phenotypes observed in Asx^^ mutants were a consequence of homeotic misexpression phenotypes in the embryo, Asx^3^ homozygous mutant embryos were examined for alterations in the regulation of the homeotic genes Ubx and Abd-B. 500-800 homozygous Asx?! embryos were examined for homeotic misexpression phenotypes. A l l embryos were completely wi ld type for AbdB (Fig. 3 B) and Ubx (Fig.3 F) expression both in the C N S and the epidermis. Therefore either the Asx?! mutation does not affect Ubx or AbdB regulation in the embryo and the effect of the Asx?! mutation must occur during imaginal disc development, or Asx?* has only subtle effects on homeotic gene regulation in the embryo that are not detectable with this assay. The bxd l4 transgene is a 14.5 kb regulatory element from the Ubx gene that is sufficient for the maintenance of lacZ expression from PS6-13 for up to 12 hrs during embryogenesis. It has been shown to contain a minimal 5.1 kb P c G response element (PRE) based on the fact that it binds P c G proteins both in vivo and in vitro and it fails to maintain its proper expression boundaries in P c G mutants (Simon et al. 1990; Simon et al. 1993; Chan et al. 1994; Chiang et al. 1995; Hodgson and Brock 1998). Derepression of the bxd l4 in specific P c G mutants is usually comparable to the level of derepression seen at the endogenous Ubx locus in those same mutants. Despite the fact that the endogenous Ubx gene has only a minor requirement for Asx activity in the C N S , the Asx^ mutation 31 causes a very strong breakdown in the bxdl4-lacZ expression boundary relatively early in embryogenesis (Soto et al, 1995 and Fig.3 K ) . Df(2R)trix and Asx?! were both tested for their effects on maintenance of the bxdl4 element and they also cause a breakdown of the bxdl4- lacZ expression boundary in the C N S (Fig.3 J and L) . The ectopic staining in the Asx?! mutants (Fig.3 J) is less intense than the other two Asx alleles (Fig. 3 K and L ) , but no reduction of bxdl4-lacZ expression in PS6-13 was seen in the Asx^^ mutants. The derepression of the bxdl4 construct in Asx?! mutants indicates that the Asx?l mutation does have some embryonic activity. The Asx^^ mutation acts like a typical Asx allele in the embryo in that it causes derepression of the bxdl4 element but it does not reduce expression of the element in PS6-13. It is possible then that the trxG adult homeotic phenotypes seen in Asx^ mutants are not the result of misregulation of the homeotic genes in the embryo; but are the result of a mutant effect at the level of imaginal disc development. The fact that Asx mutations have a much stronger effect on bxdl4 regulation in the C N S than the endogenous Ubx locus also indicates two possibilities: i) other response elements must exist in the Ubx regulatory region that can direct proper maintenance of the Ubx gene or ii) the bxdl4 element interacts with other Ubx regulatory sequences to stabilize its activity. Asx is expressed ubiquitously in all stages of embryogenesis One possible mechanism for the tissue specific regulatory effect of Asx is that it has a tissue specific expression pattern, for example, strong expression in the epidermis and reduced expression in the C N S . The Asx gene has recently been cloned and sequenced and three different antibodies were raised to Asx (Sinclair et al. 1998b). A diagram of the Asx protein showing most of the major identifiable domains is shown in Fig.4. Asx contains a region of 22 repeating alanines, multiple glutamine repeats and two domains homologous to a putative human Asx protein, including a putative pair of zinc fingers at the carboxy terminus. A sheep antibody was raised to a peptide sequence in the amino terminus of Asx 32 Fig.4 Sites of antibody generation on the Asx protein. A sheep antibody was raised to a sequence in the amino terminus of Asx while a rabbit and a mouse antibody were raised to the carboxy terminus. 33 \ 2 2 Q1Q2Q3Q4Q5Q6 Q7 0 75 94 Sheep anti-Asx — 1 1 1 1 1 1 1 C cluster 1668 1590 1668 Rabbit and Mouse anti-Asx n peptide sequences used rorantibody generation mammalian homology regions Alanine repeat Glutarnine repeats 34 and a mouse and a rabbit antibody were raised to the carboxy terminus of the Asx protein (see Fig. 4). A l l three antibodies either recognize a band the size of Asx on a western blot or can immunoprecipitate in vitro translated Asx (Sinclair et al. 1998b). The mouse antibody was used to examine the stage specific expression pattern of Asx in wi ld type embryos (Fig. 5 A - G ) . From very early (1-2 hrs, F ig 5 A ) to very late (-12 hrs, F ig 5 G) stages, Asx protein is distributed ubiquitously with a slightly stronger concentration in the C N S of later stages compared to other regions of the embryo. Staging details are given in the figure legend. To show that the mouse antibody was specific for the Asx protein, late stage homozygous Df(2R)trix embryos that lack Asx protein were stained with the mouse antibody and did not produce a signal (Fig. 5 H). For comparison, wi ld type mid to late stage embryos were stained with affinity purified sheep anti-Asx antibodies and produced an identical ubiquitous staining pattern (Fig. 5 I-L). Despite its expression in the C N S and its late stage head defects, Asx is not required for normal C N S or brain development as Asx mutant embryos have a morphologically wild type C N S and brain (see Fig .6) . The ubiquitous pattern of Asx protein expression in embryos is typical of all P c G genes studied so far (Paro and Zink 1992; Martin and Adler 1993; Lonie et al. 1994; DeCamillis and Brock 1994; Gutjahr et al. 1995). If Asx is expressed in all tissues at all stages of embryonic development, this raises the question of how Asx recognizes its target genes only in those tissues and those spatial domains where its activity is required. Asx protein binds to unique target sites on polytene chromosomes A l l P c G proteins tested so far are chromatin proteins that bind to discrete euchromatic sites. Polytene chromosomes are large cytologically visible chromosomes found in the salivary glands of late larval stages and are useful for examining the chromosome binding sites of specific proteins. Staining of polytene chromosomes with antibodies to specific P c G proteins has shown that many P c G proteins (Pc, ph, Pel, and 35 F i g . 5 Developmental expression of Asx protein in embryos. Embryos are oriented with anterior to the left, posterior to the right, dorsal up and ventral down. Asx is expressed ubiquitously at all stages and in all tissues. Embryos in A - H are stained with the mouse anti-Asx antibody. Embryos in I-L are stained with the affinity purified sheep anti-Asx antibody. A stage 3, ~1 hr. B stage 5, - 2 1/2 hrs., C stage 8 , - 3 1/2 hrs., D stage 12, ~ 8 hrs., E stage 13, -10 hrs., F stage 1 4 - 1 1 hrs., G stage 15 - 12 hrs. Embryos in I - K approximately match the staging of those in E - G respectively. L shows a stage 16 (~ 15 hrs) embryo. H is a stage 15 Df(2R)trix homozygous mutant embryo that lacks Asx protein. Df(2R)trix was balanced over a chromosome that contained a lac-Z reporter gene so that homozygous mutants could be identified with the absence of lac-Z staining. 36 F i g . 6 C N S and brain development in Asx mutant embryos. Embryos were stained with the axon specific monoclonal antibody BP102 and detected with a fluorescent conjugated secondary. Fluorescent images were collected with a confocal microscope. A , wild type C N S and brain in a stage 16 embryo. Embryo is oriented with anterior to the left. The image is collected at a slight angle so that the brain is facing towards the page while the ventral portion of the C N S is facing away. B, C N S in a Df(2R)trix homozygous mutant embryo. The image is collected so that the ventral portion of the C N S is facing towards the page. C , close up of a brain in a Df(2R)trix homozygous mutant embryo. Similar results were seen with the Asx$ mutation. 38 39 probably also Scm) bind to about 100 target loci in a completely overlapping pattern (Franke et al. 1992; Lonie et al. 1994; Peterson et al. 1997) and other P c G proteins (Psc, E(z), Su(z)2) overlap with a large number of those 100 sites but also bind at some additional unique sites (Rastelli et al. 1993; Carrington and Jones 1996). This large degree of overlap between P c G binding sites has been used to argue for at least one discrete P c G protein complex containing all the various P c G members. Asx has unique phenotypic characteristics that indicate it may be functionally distinct from other P c G proteins, so polytene chromosomes were stained with the antibodies described above to determine i f the Asx polytene staining pattern was different than other P c G proteins. Polytene chromosomes were double labelled with the sheep anti-Asx antibody and a rabbit anti-Pc antibody to see i f Asx protein would overlap with any P c G proteins at specific target sites (Fig. 7 A - C ) . Surprisingly, these double labelled chromosomes showed very little overlap between Asx and Pc (F ig . 7 C) and each antibody detected only about 40- 50 individual sites (Fig. 7 A and B) . However, when each antibody was tested alone, both the anti-Pc antibody and the anti-Asx antibody reliably detected between 70-80 sites (Fig. 7D and E respectively). This result suggests that one or both antibodies are interfering with each others ability to recognize their antigens at certain loci. This raises two possibilities: i) either Asx and Pc proteins completely overlap at all binding sites but the antibodies interfere with detecting this overlap, or ii) they only overlap at a subset of binding sites and Asx binds to some completely unique sites. To distinguish between the above two possibilities, Asx sites were directly mapped to polytene chromosomes using H R P staining (Fig. 8). 90 sites were mapped to polytene chromosomes using the rabbit anti-Asx antibody. Similar results were seen with the sheep and mouse anti-Asx antibodies. To show that the signal on polytene chromosomes was specifically due to anti-Asx reactivity, rabbit anti-serum was depleted for anti-Asx reactivity by incubation with the carboxy terminus of Asx. When this immune-depleted serum was used to stain polytene chromosomes, no signal was apparent on the chromosomes (Fig. 40 F i g . 7 Binding of Asx and Pc proteins to polytene chromosomes. Primary antibodies were detected with Texas Red (red) and FITC (green) fluorescent secondaries. A - C , chromosomes double labelled with sheep anti-Asx (red) and rabbit anti-Pc (green) antibodies. Regions of overlap are yellow. D, chromosomes singly labelled with anti-Pc only. E , chromosomes singly labelled with anti-Asx only. F , same nucleus as in E , stained with the fluorescent dye D A P I to show the banding pattern of the chromosomes. 41 A PC D < ' PC B ASX F ASX C PC/ASX ' w °A , >' 4} Fig . 8 Rabbit anti-Asx H R P staining on polytene chromosomes. Chromosomes were counterstained with a blue-gray Giemsa stain to detect chromosome banding patterns and allow mapping of antibody binding sites. A , rabbit anti-Asx binding sites. The two homeotic complexes, the bithorax complex (BX-C) and the Antennapedia complex (ANT-C) are shown as well as a strong Asx binding site, 3 5 A B . B , polytene chromosomes stained with immune depleted rabbit anti-Asx antibody. C and D , comparison of binding sites between Asx and ph, respectively, on the distal part of the X chromosome. Lines connect shared binding sites and arrowheads mark unique sites. 43 44 Table 1: Comparison of Asx with Ph/Pc protein binding sites on polytene chromosomes ASX PH/PC ASX PH/PC A S X binding PH/PC binding binding binding binding sites binding sites sites sites sites sites X 2L 3L - IA 21A 21A - 61A ID - 22A 22A 61C 61C IF - 22B 22B - 61D 2D (ph) 2D - 22C 61F 61F 4C 4C 24A 24A 62F 62F 5A 5A 25 E F 25EF 63A -5B - 26F 26F - 63F-64A 5D 5D 27B - 64C -7A - 28A 28A - 65D 7B 7B 28D - 66A -7D - - 29E 66C -8A 8A 30AB 3 OB - 66E - 8B - 30C 67CD 67D 9A 9A 32EF 32EF 67E (E(z)) 67E 10A - - 33B - 67F 10B - 33F 33F - 68A 12D 12D - 34C 69C 69C 13E 13E - 34D 69D 69D 14B 14B 35AB 35AB 70AC 70AB - 16D 35CD 35D 70DE 70DE 17A 17A - 36A 75D -- 17E - 36B - 76C - 17F 36CD - 77A -18CD - - 37B - 77E 19D 19D 37D - - 78EF 19F - 38C - 79B 79B - 38F 39F-40A 39EF 2R 3R 41C (sxc) 41CD - 82E - 43C - 83C 44A 44A 8 4 B ( A N T - C ) 84AB - 45C 84DE 84D 46A - 84F 84EF 46CD 46C 85D -47A - 85EF (Scm) 85E 48A (en) 48A 86C 86C 49F (Psc) 49EF 87B 87BC 50A - 88A 88A 5 1 A (Asx) 51A 89B 89B - 51D 89C 89C 56C 56C 89D -- 57A 89E ( B X - C ) 89E 57B 57B 90E 90E - 58CD 93E 93E - 58F - 94DE - 59A 96A -- 59C 96CD 96BC 59F 59F - 96F-97A 60E 60E 97D -- 60F 98BD 98CD 99A 99 A D 99F 99F 100AB 100A 100F 100F 45 8B). F ig . 8A highlights examples of some Asx binding sites. Table 1 lists the cytological location of all 90 sites as well as highlighting the position of several genes and loci of interest. 63 of the 90 Asx binding sites overlap with Pc/ph/Pcl sites including the homeotic loci (Fig. 8 A and table 1) and the Asx locus itself (table l).Thus 37 Pc/ph/Pcl sites do not contain Asx protein. 27 of the Asx sites were completely unique to Asx and didn't overlap with any other P c G binding sites. To illustrate this, a comparison of Asx (Fig. 8C) and ph (Fig. 8D) binding sites is shown for part of the X chromosome. In addition to the differences in actual binding sites, there are also differences in staining intensity at specific sites. Sites 48A, 49EF and 100A all stain very intensely with antibodies to Ph or Pc but stain very weakly with Asx , whereas sites 35AB (see Fig . 8A), 56C and 93E stain very intensely for Asx but weakly for Ph or Pc. This argues that the accumulation of these proteins can vary at different target sites and this may relate to differences in activity or differences in complex formation at these target sites. The AsxPl mutation does not alter the distribution of Asx protein on polytene chromosomes The AsxPl mutation is caused by the insertion of a P transposable element in the 5' U T R of the Asx gene, thus the P mutation does not directly affect the coding sequence of Asx. Northern blots with Asx^l mutants are essentially normal although the amount of transcript present looks slightly reduced compared to wild type (D. Sinclair, H . Brock unpublished data). To test the possibility that the molecular basis of theAsx^^ mutation may be to slightly reduce the dose of Asx protein or disrupt the ability of Asx to bind to its target sites, polytene chromosomes from Asx?! homozygous mutants were stained with Asx antibody. A n antibody to Asx still recognizes 90 sites on homozygous Asx?! mutant polytene chromosomes (Fig.9). This indicates that the binding and accumulation of Asx protein at its polytene target sites is normal in Asx?! mutants which strongly suggests that the molecular basis of the Asx?! mutation does not involve reducing the dose of Asx 46 F i g . 9 Binding of Asx to Asx?! mutant chromosomes. A , 90 binding sites were detected with H R P staining. B , Chromosomes were counterstained with the fluorescent dye D A P I to help in identification of the number of binding sites. D A P I fluorescence is effectively quenched by H R P staining so every strong H R P binding site in A can be correlated with a quenched site in B (see arrowhead for an example). 47 48 protein or altering the binding capabilities of the Asx protein. However, it is still possible that Asx binding is slightly reduced at a level that is not detectable using H R P staining. Discussion Targeting PcG activity P c G proteins are only required in specific regions and specific tissues of the embryo for regulation of individual homeotic genes. A l l the PcG proteins that have been examined are expressed ubiquitously and thus it is unknown how they recognize their targets only in those regions that require their activity. The activity of Asx is even more specific than other P c G proteins but despite its tissue and stage specific effects, Asx protein is expressed with no temporal or spatial specificity. One possibility is that the tissue specific activity of Asx is regulated by interactions between factors bound to tissue specific enhancers and Asx protein bound to PREs. No such factors have been identified but they could have only subtle mutant phenotypes that would make them refractory to traditional genetic screens and they may only be identified with the purification and identification of factors bound to P c G protein complexes. Asx may have imaginal disc specific as well as tissue specific requirements for its activity Although the Asx?! allele has trxG as well as P c G phenotypes in the adult, it has no effect at all on regulation of the endogenous Ubx and AbdB loci in the embryo. The bxdl4 reporter element creates a new trx binding site on polytene chromosomes and its expression is slightly reduced in a trx mutant background indicating that it is responsive to trxG activity. Asx?! causes derepression of the bxdl4 construct, indicating that it does have a mutant P c G phenotype in the embryo. However, it does not cause reduced expression of the construct and thus does not have a trxG embryonic phenotype. It could 49 be that the trxG effect is too subtle to be detected with this assay, or it may be that the AsxPl mutation may have its effects at the level of imaginal disc development. This argues that there may be a different requirement for Asx activity during imaginal disc development than there is during embryogenesis. The bxdl4 maintenance element contains a functional P R E (Simon et al. 1990; Simon et al. 1993) that binds multiple high molecular weight P c G complexes (Hodgson and Brock 1998). However, a deletion, pbx^, that removes the entire endogenous bxdl4 element as well as some flanking sequences has no effect on the regulation of the endogenous Ubx gene in embryos (J. Hodgson, unpublished observation; Duncan 1987). This argues that the bxdl4 regulatory sequence is not required for the maintenance of endogenous Ubx expression during embryogenesis. Since almost all P c G mutations have a very strong effect on the embryonic regulation of Ubx, this argues that the bxd l4 is not the only P R E at the Ubx locus or that other regulatory sequences at the Ubx locus can maintain restricted expression throughout embryogenesis. Muller and Bienz (1991) have shown that some combinations of regulatory elements from the Ubx locus can act as P R E maintenance elements although these regulatory elements act as simple enhancers when present by themselves. Additionally, there is also evidence that there may be at least one other P R E at Ubx, the b x l 7 (Chiang et al. 1995), but it has not been as thoroughly characterized as the bxd 14 . Although the embryonic phenotype of the pbx? deletion argues that the bxd 14 is not required or is redundant for the embryonic regulation of Ubx, the pbx? deletion has an obvious adult homeotic phenotype (Bender et al. 1983; Duncan 1987) suggesting that the bxd l4 element may be required for Ubx regulation in the imaginal discs. Next to the bxd l4 and within the pbx^ deletion, are a set of enhancers that are only expressed in the imaginal discs, but by themselves they are expressed with a non restricted pattern (Pirrotta et al. 1995). Pirrotta et al (1995) have argued that the bxdl4 P R E functions by transferring 50 spatially restricted expression information from embryogenesis to the imaginal disc enhancers during larval growth and pupation. Asx mutations have only a subtle effect on the C N S regulation of Ubx, but have a relatively strong effect on regulation of the bxdl4 element. If the bxdl4 element is crucial for the transfer of spatially restricted information from the embryo to the imaginal discs, then the central role of Asx in bxdl4 regulation in the embryo may implicate Asx activity in this process. If the Asx?! mutation alters imaginal disc regulation of homeotic genes, it may do this by disrupting Asx mediated transfer of spatially restricted information from the bxd l4 to the imaginal disc enhancers. Asx is a member of a PcG complex and binds close to Pc protein at multiple overlapping euchromatic sites Asx binding overlaps with a large number of other P c G sites which is consistent with the possibility that Asx is part of a large P c G complex at these sites. The observation that chromosomes double labelled with antibodies to Asx and Pc do not show any overlap suggests that the two antibodies are sterically hindering each others ability to bind to specific sites. The fact that each antibody on its own can recognize many more loci than when they are present together supports this idea. If we assume that the Pc antibody can recognize up to 80 sites on its own and the Asx antibody should recognize about 27 sites that are completely unique to Asx, this gives a total of about 107 sites that should be recognized by both or one of the antibodies. On double labelled chromosomes, the total number of sites recognized by one or the other antibody is about 100 sites which is about the same as the 107 expected. However, the number of sites recognized by each antibody individually is only about 50 sites (compared to about 70 or 80 sites on individually labelled chromosomes) which indicates that at some sites of overlap, Asx antibody is preventing Pc from binding while at other sites of overlap, Pc antibody is preventing Asx from binding. There may be no significance as to which specific sites-preferentially bind 51 which specific antibody as this could be a random event depending on particular local binding conditions. Steric hindrance between the two antibodies argues that the respective antigens they each recognize are in close proximity to one another. This suggests the possibility that Asx is either within a P c G complex bound very close to Pc protein or Asx is contacting the complex very near the Pc protein. Using the yeast two hybrid system, Asx does not directly bind to Pc or any other P c G protein tested ( M . Kyba and H . Brock, unpublished observation) but this does not rule out the possibility that Asx recognizes an intermediary protein that allows it to interact with Pc. However, the fact that Asx binds many sites where no Pc protein is present also argues that Asx does not require the presence of a Pc containing P c G complex in order to bind to chromosomal target sites. In the salivary gland tissues, the homeotic genes are repressed and so the presence of all known P c G proteins bound to the homeotic loci strongly suggests that these proteins are in a functional complex together at these loci. However, P c G proteins also bind many loci that are obviously not repressed in salivary gland tissues such as the ph locus or the Asx locus itself. This suggests then that either i) P c G proteins are not global repressors and their specific function can vary at particular loci, perhaps by interacting with enhancer specific factors or ii) a bound P c G complex is not necessarily a functional complex. The fact that Asx and some other PcG proteins can bind to unique sites also suggests that there are different requirements for PcG activity at different sites at thus possibly different P c G complexes. 52 Chapter 2 Introduction Asx was originally isolated as a lethal mutant that died at the end of embryogenesis and displayed mild posterior transformations of the cuticle (Nusslein-Volhard et al. 1984; Jurgens 1985). Posterior transformations are a hallmark of P c G mutations and are caused by the ectopic expression of homeotic genes. Most homeotic genes are expressed with restricted anterior boundaries so that when expression boundaries break down, homeotic gene expression expands in an anterior direction resulting in posterior transformations. A n exception to this is the expression of the homeotic gene Sex combs reduced (Scr) that expands in both an anterior and a posterior direction (Glicksman and Brower 1988a), producing extra sex combs on the second and third legs of male flies. This dominant extra sex combs phenotype has been used to identify many members of the P c G as well as modifiers of P c G activity. Double mutant combinations between P c G genes tend to enhance the penetrance of this phenotype. The number of legs that display extra sex combs is a measure of the strength of the mutant interaction and has been used in the past in an attempt to categorize the P c G into specific functional subsets (Cheng et al. 1994; Campbell et al. 1995). Mutations in members of the trithorax group (trxG) of genes reduce homeotic gene expression within their normal domains and cause posterior structures to be transformed into more anterior ones (Kennison 1993). A n exception to this is the transformation of the first and second legs towards the third leg, suppressing the formation of sex combs on the first legs of male flies presumably due to a reduction in Scr expression (Ingham and Whittie 1980). Double mutant combinations between two trxG genes enhance the penetrance of trxG homeotic transformations (Kennison and Russell 1987; Kennison and Tamkun 1988). Double mutant combinations between a P c G gene and a trxG gene suppress each others homeotic mutations to produce an essentially wild type fly (Kennison and Russell 1987; Kennison and Tamkun 1988; Campbell et al. 1995). This mutual suppression of homeotic 53 phenotypes in PcG/trxG double mutants has led to the suggestion that the two groups of proteins either act antagonistically or have opposite and independent functions (Ingham 1983; Jones and Gelbart 1993). The extra sex combs phenotype is rarely seen in adult heterozygous Asx mutant flies (Sinclair et al. 1992), but Asx can strongly enhance the extra sex combs phenotype of other P c G genes. In particular, the Asx allele Df(2R)trix strongly interacts with mutations in Pc to produce flies with multiple sex combs (Campbell et al. 1995). However, the fact that the Asx^ allele has both trxG and PcG homeotic transformations in adult flies (Sinclair et al. 1992) raises the possibility that Asx may have a role in both repression and activation. The major objective of this chapter was to determine i f Asx could function as both a member of the P c G and the trxG. To test this possibility, Asx?! homozygotes were crossed to both Pc and alleles of the trxG gene trithorax (trx). Asx?! homozygotes showed strong enhancement of P c G phenotypes when combined with a Pc heterozygote and strong enhancement of trxG phenotypes when crossed to a trx heterozygotes. Lethal Asx alleles were crossed to Pc and trx alleles and found to enhance both PcG and trxG homeotic transformations. Df(2R)trix, the only known true null allele of Asx, showed an extra strong interaction with Pc not seen with other PcG mutations. This indicates that there may be a specific in vivo functional interaction between Asx and Pc . Although Asx^ homozygotes strongly enhance Pc homeotic phenotypes, Pc binding is normal on Asx^ homozygous polytene chromosomes. Thus the functional interaction between Pc and Asx may occur after the proteins are already bound to their target sites. Results Asx?! homozygotes strongly enhance both PcG and trxG phenotypes 54 Unlike other Asx alleles, Asx^^ survives as a homozygote and occasionally displays trxG mutant phenotypes (abdominal segments 5 or 6 transformed towards a more anterior segment or transformation of the haltere towards the wing, see Fig. IOC) as well as some typical P c G mutations (anterior abdominal segments transformed towards more posterior segments, see Fig. lOD). The penetrance of these homeotic transformations is rather low and only occurs in about 10 - 20% of adult flies (Sinclair et al 1992, and Table 2). Double mutant combinations between a PcG gene and a trxG gene usually suppress homeotic transformations to produce a nearly wild-type fly. However, homozygous Asx?! mutations strongly enhance both the penetrance and expressivity of homeotic transformations seen with two heterozygous trx null alleles, Df(3R)red and trx^^ (Table 2 and Fig . 11). Heterozygous trx alleles almost never show transformations of the haltere towards the wing, and transformations in the abdomen are much weaker than when enhanced by Asx?! (compare F i g . l l C and D to E and F). A s well, the transformation of the second thoracic segment (T2) towards the third (T3) is a very strong anterior transformation (see Fig . H E ) that is never seen in Asx^ or trx mutants alone (Table 2). The trx allele trx^!! is a point mutation in a specific protein motif termed the SET domain (Stassen et al. 1995) and is probably not a complete null mutant. Compared to Dfi3R)red and trx^^, trx^!! is relatively weak and shows a low penetrance of abdominal homeotic transformations (see Table 2, row 3). Homozygous Asx?! mutations enhance both the penetrance and the expressivity of homeotic transformations in trxZU mutants to the same level as with Dj(3R)red and tnfi^ (see Table 2). This indicates that the trx^^ SET mutation alone is sufficient for the strong genetic interaction between Asx?! and trx. Double mutant combinations between PcG genes enhance P c G homeotic transformations such as the extra sex combs phenotype (Campbell et al. 1995). In particular, the Asx null allele Dfi2R)trix strongly interacts with mutations in Pc to produce flies with multiple sex combs (Cambell et al, 1995). Asx?! can behave like a typical P c G 55 Fig. 10 Example of anterior and posterior transformations seen in Asx?! homozygotes. A , haltere on a wi ld type fly indicated by the black arrowhead. B, abdomen of a wild type male fly showing abdominal segments 3-6. Note that A5 and A 6 are both heavily pigmented. C, an example of a haltere to wing transformation in an Asx?! homozygote. The haltere (indicated by the black arrowhead) is much larger and has triple row margin bristles and the overall structure of a small wing. D, an example of both posterior and anterior transformations in the abdomen of an Asx^l homozygote. Note the patches of extra pigmentation in A 3 and A 4 (white arrowheads, facing left) indicative of posterior transformations and the large unpigmented patch in A 5 (white arrowhead, facing right) indicative of an anterior transformation. 56 57 F i g . 11 Enhancement of trx phenotypes by Asx^^ homozygotes. A , thorax and haltere (arrowhead) of a wi ld type fly. B abdomen of a wild type fly. Note the pigmentation in abdominal segments 5 and 6. C , typical thorax and haltere of a heterozygous trx mutant. In this genetic background, single trx mutants have a wild type haltere (arrowhead) and thorax. D , example of a strong abdominal transformation in a single trx heterozygote. There are patches of light pigment in abdominal segment 5 (small arrow) which is indicative of the transformation of A 5 towards a more anterior segment. E , thorax and haltere of a mutant that is homozygous for Asx?! and heterozygous for trx. The haltere is much enlarged (arrowhead) indicating a partial transformation of the haltere towards the wing. Additionally, the posterior of the third thoracic segment, T3 (large arrow) is partially transformed towards the second thoracic (T2) segment, an anterior transformation that is not seen in Asx^^ or trx mutants alone. F , example of a strong abdominal transformation in a mutant that is homozygous for Asx?! and heterozygous for trx . Abdominal segment 5 (small arrow) has very little dark pigmentation which is indicative of an almost complete transformation of A 5 towards a more anterior segment. 58 59 Table 2: Enhancement of trxG and P c G phenotypes by Asx?! mutant allele Asxpl/Asxpl-y mutant / + Asxpl/Asxpl • + / + Asxpl/+; mutant / + Asxpl/+ + / + Df(3R)red 94.4 (54) t 19.1 (115)* 33.6 (110) § 0(130) trxBU 100 (82)f 15.8 (82) * 14.3 (147) § 0(151) trxZ11 97.1 (104) f 26.3 (76) X 3.0 (99) § 0(95) Pc4 * 40.0 (45) * 4.0 sex combs 4.2(118) * 2.0 sex combs 12.2 (148) * 3.0 sex combs 0(125) 2.0 sex combs The first number is % penetrance of either anterior (trxG) transformations (f ,$,§) or posterior (PcG) transformations (*), the number in parentheses is the total number of flies scored for each genotype. t scored for transformations of abdominal segment 5 (A5) or abdominal segment 6 (A6) towards more anterior segments, transformations of the posterior of the third thoracic segment (T3) towards the second (T2) or transformations of the haltere towards the wing $ only abdominal and haltere to wing transformations were seen § only abdominal transformations were seen abdominal transformations of A 2 - A 4 towards A 5 or A6 plus average number of legs with sex combs (wild type is 2.0) were scored 60 Table 3: Lack of enhancement of trxG and P x G phenotypes by the Asx?! revertant mutant allele AsxR11/AsxRU; mutant / + AsxR11/ AsxRn; + / + AsxR11/+; mutant / + AsxR11/+; + / + Dff3R)red 16.1 (62)§ 0(38) 15.9 (63)§ 0(40) trxB11 20.3 (79)§ 0(35) 17.4 (69)§ 0(61) Pc4 * 0(80)* 3.5 sex combs 0(85)* 2.0 sex combs 0(98)* 3.5 sex combs 0(110)* 2.0 sex combs The first number is % penetrance of either anterior (trxG) transformations (§) or posterior (PcG) transformations (*), the number in parentheses is the total number of flies scored for each genotype. § only transformations of abdominal segment 5 (A5) or abdominal segment 6 (A6) towards more anterior segments were seen * no abdominal transformations were seen, average number of legs with sex combs (wild type is 2.0) were scored 61 mutation in that homozygous Asxpl mutations can also strongly enhance homeotic transformations of the P c G gene Pc (Table 2). Asxpl is homozygous viable but it is semi-lethal with strong Asx alleles. Asx^^ is a revertant line derived from Asxp^ in which loss of the P element insertion at the Asx locus is also associated with loss of semi-lethality with strong Asx alleles (Sinclair et al. 1992). AsxRH homozygotes fail to enhance homeotic transformations of either Pc or trx (Table 3) indicating that the above described genetic interactions are specific for the Asxpl insertion allele and are not due to a second site mutation on the Asx^l chromosome. Other Asx mutations can also enhance trxG phenotypes P c G mutations enhance one anothers homeotic transformations. Consistent with this, all Asx alleles tested strongly enhance the extra sex combs phenotype of Pc and some also show enhanced penetrance of anterior to posterior abdominal transformations (Table 4). The genetic interaction between Asxpl and trx could be due to neomorphic activity of the Asxpl allele or it could reflect a more general interaction between Asx and trx. To distinguish between these possibilities, other Asx alleles were crossed to trx mutant alleles. P c G and trxG mutations mutually suppress each others homeotic phenotypes. Double heterozygous combinations between Pc4 and the trx null allele Df(3R)red completely suppresses trx abdominal homeotic transformations (Table 5b, cross 10). Double mutant combinations between two trxG genes generally show enhancement of trxG phenotypes. Surprisingly, when introduced from females, Asx^, Df(2R)trix, Asx*?, and Asx^3 heterozygotes all enhance the penetrance of trxG homeotic transformations in Dfi3R)red heterozygotes (Table 5a, crosses 1-4), although not as dramatically as Asxpl homozygotes do. Where tested, these Asx alleles can also enhance the penetrance of trxG homeotic transformations in trx^^ and trx^! heterozygotes (Table 5a, crosses 5-9). 62 Table 4: Enhancement of P c G phenotypes in Asx;Pc^ double heterozygotes Asx allele AsxiPc4 D O U B L E H E T E R O Z Y G O T E S Df[2R)trix 2.2 (93) 3.4 sex combs Asx 1 0(34) 4.9 sex combs Asx 3 69.0 (42) 4.7 sex combs Asx & 0(30) 4.7 sex combs Asx 9 0(36) 5.8 sex combs Asx 13 0 (55) 6.0 sex combs Asx PI 1.7 (60) 3.1 sex combs The first number is % penetrance of abdominal posterior transformations.The number in parentheses is the total number of flies scored. Flies were also scored for the average number of legs with sex combs, the wild type number is 2.0. 6.0 is the maximum number of legs that can have sex combs, seen i f all double heterozygotes show complete transformation towards the first leg. Only double heterozygotes are shown in the table because all other genotypic classes (including single Pc mutants alone) showed 0% abdominal transformations and the wild type average of 2.0 legs with sex combs. 63 Table 5a: Asx crosses that enhance the penetrance of trx phenotypes C R O S S F E M A L E S X M A L E S Asx;rrx D O U B L E H E T E R O Z Y G O T E S trx S I N G L E H E T E R O Z Y G O T E S 1) Df(2R)trix X Df(3R)red 6.1 (279) 2.5 (278) 2) Asx1 X Dfi3R)red 32.0 (97) 14.2(113) 3) Asx9 X Df(3R)red 49.2 (122) 28.3 (152) 4) Asx13 X Df(3R)red 41.3 (167) 26.2 (130) 5) Df(2R)trix X trxBn 6.1 (359) 2.4 (338) 6) Asx9 X ? r x 5 i i 40.2(117) 22.2 (81) 7) Asx13 X f r x 5 i i 44.1 (102) 26.6 (94) 8) Df(2R)trix X f r x Z i i 4.8 (230) 1.9 (210) 9) A s x 9 X trxzn 25.5 (153) 10.6(141) Table 5b : Crosses that do not enhance the penetrance of trx phenotypes C R O S S F E M A L E S X M A L E S D O U B L E H E T E R O Z Y G O T E S trx S I N G L E H E T E R O Z Y G O T E S 10) Pc4 X Df(3R)red 0 (114) 23.9 (67) 11) Asx8 X Df(3R)red 37.9 (153) 31.4(169) 12) Asx3 X Df(3R)red 11.5(156) 12.7 (165) 13) A s x P i X Df(3R)red 21.0 (57) 18.4 (38) 14) Df(3R)red X Df(2R)trix 17.7 (237) 13.7 (117) 15) Df(3R)red X A s x ; 25.9 (139) 38.9(113) 16) Df(3R)red X A s x 9 32.1 (53) 27.5 (40) 17) Df(3R)red X A s x s 25.7 (101) 34.7 (101) 18) Df(3R)red X A s x P i 20.6 (102) 19.7 (71) 19) Df(3R)red X A s x 7 - 3 13.9(101) 27.2 (81) 20) Df(3R)red X A s x 3 31.1 (106) 36.9 (130) 21) * r x 5 i i X Df(2R)trix 10.7 (187) 14.0 (157) 22) ? r x 5 J i X A s x 9 7.7 (39) 16.7 (36) 23) fr-x5ii X A s x 7 5 30.1 (103) 18.9 (53) 24) r r x Z i i X Df(2R)trix 7.3 (218) 9.4 (245) 25) * r x Z i i X A s x 9 29.8 (47) 15.9 (63) Flies were scored for transformations of abdominal segment 5 (A5) towards more anterior segments. The first number is % penetrance of trxG transformations.The number in parentheses is the total number of flies scored for each genotype. Differences in penetrance between the double heterozygotes and the single trx heterozygotes were tested for significance using a Chi-square test (see Materials and Methods). 64 Asx^, Asx$ and Asxpl heterozygotes do not significantly enhance Df(3R)red homeotic transformations, but they also do not suppress them the way Pc4 does (Table 5b, crosses 11-13 ). For all crosses, when the trx allele is introduced from the females, the double heterozygotes fail to show significant enhancement of trx phenotypes (Table 5b, crosses 14-25). This indicates that the genetic interaction between Asx and trx has a maternal effect relative to the Asx mutant alleles. The genetics shows that Asx behaves like a member of the P c G with respect to interactions with Pc , and can act as a member of the trxG with respect to interactions with trx, providing a different possible interpretation of the intermediate homeotic phenotype seen in the cuticle of Asx homozygous mutants. Double homozygous mutations between Pc and trx produce a weak PcG cuticle phenotype that looks very similar to homozygous single Asx mutants alone (Capdevila et al. 1986; Jurgens 1985; Breen and Duncan 1986). Thus, instead of viewing the weak cuticle phenotype of Asx mutants as resulting from weak P c G activity, this intermediate phenotype could be the result of abolishing both Pc and trx activity. Asx could therefore be a component of the system that integrates the activation signal of the trxG with the repression signal of the PcG. The Asx mutation Df(2R)trix shows an extra strong PcG phenotype with Pc4 A l l Asx mutations tested show strong enhancement of PcG phenotypes when crossed to Pc4. The Asx null mutant Df(2R)trix enhances both the extra sex combs phenotype and the penetrance of abdominal posterior transformations. In addition, Df(2R)trix;Pc4 double heterozygotes also show an unusual darkly pigmented outgrowth in the first thoracic (TI) segment (F ig . l2A and B) . The penetrance of T I outgrowths is quite variable and can range from 7% (from a total of 94 flies) to 24% (from a total of 116 flies) in individual crosses. The expressivity of the phenotype can range from a small, condensed outgrowth (Fig. 12 A , B and Fig . 13 C) to producing an actual pseudo-wing on T I (Fig. 13 65 Fig.12 TI mutant phenotype in Pc4;Dfi2R)trix double heterozygotes. A-B, darkly pigmented outgrowth on the first thoracic segment (TI) of Pc4;Df(2R)trix double heterozygotes. C, darkly pigmented internal mass in the first thoracic segment of Pc4;Df(2R)trix double heterozygotes. D , close up of the mass in (C). Upon close examination, the dark pigmentation is actually the result of condensed wing tissue and closely packed bristles. 6 6 67 Fig.13 Scanning electron microscope (SEM) image of TI mutant phenotype in Pc4;Df(2R)trix double heterozygotes. A, stereo image of a scanning electron microscope (SEM) image of a pseudo-wing outgrowth on TI . B, SEM image of the same fly in (A) showing the size of the pseudo-wing relative to the rest of the fly. C, SEM image showing a small, condensed outgrowth on TI . 68 A and B) . In a few cases, instead of an actual outgrowth, there is a darkly pigmented internal mass (Fig. 12 C) . Upon closer examination, the darkly pigmented internal and external growths are both made up of wing-like tissue (Fig. 12 D). The internal growths are likely due to an unsuccessful imaginal disc evagination of a pseudo-wing from T I . Overall, this phenotype is interpreted to be a strong, but rarely seen, posterior transformation of T I towards T2 . This T I to T2 transformation is not seen when Df(2R)trix is crossed to mutations in the P c G genes Sex combs on midleg (Scm) or Polycomblike (Pel) indicating that this interaction is probably not a generalized PcG-Asx interaction. The fact that crosses between Dfi2R)trix, the only known Asx null allele, and the strong Pc allele Pc4 can produce this unusually strong posterior transformation may indicate a specific, strong in vivo functional interaction between Asx and Pc . The b ind ing of P c is no rmal on Asx^1 mutant chromosomes The AsxRl mutation can strongly enhance Pc homeotic transformations both as a heterozygote and as a homozygote. PcG proteins are found bound to multiple loci on polytene chromosomes, including genes that are known to be active in the salivary glands. The trx protein is found bound along with Pc at multiple loci, including the homeotic loci, that are known to be repressed in the salivary glands. One possible interpretation of this data is that binding of P c G and trxG complexes is necessary but not sufficient for repression or activation respectively. The enhancement of P c G phenotypes seen with AsxR1 mutants could be due to i) a reduction in the ability of P c G complexes to bind at target sites or ii) an alteration in functional activity of the complex once it is bound to its target site. To distinguish between these two possibilities, homozygous mutant AsxR1 polytene chromosomes were stained with an antibody to Pc to look for gross alterations in binding activity. Binding of Pc to homozygous AsxB1 polytene chromosomes was normal (Fig. 14) indicating that the effect 70 F i g . 1 4 Binding of Pc protein to Asxpl mutant chromosomes. Pc protein was detected using a secondary conjugated to the fluorescent molecule FITC. A , W T chromosomes. B , homozygous Asxpl mutant chromosomes. The image in (A) was collected using a confocal microscope and the image in (B) was photographed using a zeiss axiophot microscope. 71 A r 1 j V / s W T B i • / i • * < • 1 f • f i P1 A S X 72 of the AsxRl mutation probably occurs at the level of functional activity of the complex after it has bound to its target site. Discussion Asx mutations enhance both PcG and trxG homeotic phenotypes. AsxPl homozygous mutants strongly enhance both P c G and trxG mutations. A l l Asx alleles tested strongly enhance P c G phenotypes. Unexpectedly, some Asx alleles can also enhance trxG phenotypes. This enhancement shows a maternal effect with regards to the Asx mutant allele and could reflect an early developmental requirement for Asx activity in regards to trx function. B y themselves, AsxR1 homozygotes have a low penetrance of both P c G and trxG adult phenotypes with a higher penetrance of trxG phenotypes compared to P c G phenotypes. This suggests that i f Asx has both a trxG and a P c G function, the Asx^1 mutation has a stronger affect on the frx-dependent function of Asx. The molecular basis for the Asx^1 mutation is unknown. The P element does not interrupt the open reading frame and although northern blots show a slight decrease in the amount of transcript present (D. Sinclair, H . Brock unpublished), the level of Asx protein accumulation and binding to polytene target sites appears to be unaffected. The AsxPl mutation may alter Asx expression in specific tissues such as the imaginal discs or it may somehow alter the translation of Asx. The fact that some Asx alleles can enhance both P c G and trxG phenotypes implicates Asx in both P c G and the trxG activity. If Asx is required for the proper functioning of both the P c G and the trxG, Asx mutations would be expected to produce a phenotype similar to one seen in PcG/trxG double mutants. Consistent with this idea, Pc;trx and esc;trx homozygous double mutants suppress the strong trxG and strong P c G phenotypes seen with each gene individually to produce a more wi ld type embryo that has a slight P c G phenotype (Ingham 1983; Capdevila et al. 1986). The weak P c G phenotype 73 seen in these PcG;trxG double mutants is very similar to the weak P c G cuticle phenotype seen in Asx homozygous lethal mutants (Jurgens 1985; Breen and Duncan 1986). The P c G gene Enhancer ofzeste (E(z)) can also enhance the phenotypes of some trxG mutations. E(z) mutants have a strong P c G phenotype in the embryo but temperature sensitive studies have indicated that they have a strong trxG phenotype in the imaginal discs (LaJeunesse and Shearn 1996). Recently, it has also been shown that the G A G A protein, the product of the trithoraxlike (trl) locus (a trxG gene), functions at PREs and is required for the binding of P c G complexes to target sites (Hagstrom et al. 1997; Hodgson and Brock 1998). This suggests that there may be a subset of genes that either are required for the function of both groups or are involved in mediating the establishment of repression versus activation at target loci. It also suggests that the relationship between P c G mediated repression and trxG mediated activation is not entirely antagonistic and that there is a functional interdependance. Asx may interact specifically with both trx and Pc proteins The trxZH allele is a point mutation in the SET domain of trx (Stassen et al. 1995). The SET domain has an unknown function but it is found in other regulatory proteins such as the trxG protein ash-1, the P c G protein E(z), and in Su(var)3-9, a chromatin protein. Interestingly, a specific E(z) SET domain point mutation alters the E(z) S E T domain and makes it more similar to the trx SET domain (L. Sipus, R. Jones, H . Gyurkovics, unpublished data). This E(z) point mutant has strong trxG phenotypes, suggesting that the SET domain itself plays an important role in mediating activation and repression of the homeotic loci. The trxZH allele is enhanced by Asxpl homozygotes to the same extent as null alleles of trx, suggesting that the SET domain mutation alone is sufficient for the strong interaction seen between Asxp^ and trx. In vitro binding assays show that the SET domain of trx can bind to a carboxy terminal region of Asx (Kyba et al. 1998). Polytene staining shows that these two proteins overlap at multiple sites (Kyba et al. 74 1998), suggesting that the two proteins may also interact in vivo. These data suggest that Asx and trx proteins directly interact with one another and that Asx probably modifies trx activity. Asx genetically interacts much more strongly with Pc than with most other P c G genes suggesting that the two proteins may have a specific in vivo functional interaction. Antibodies to Pc protein and an antibody to the amino terminus of Asx sterically hinder one another from binding to specific sites on polytene chromosomes suggesting that the antigens are close together. Although they have not been shown to directly physically interact in vitro, Asx and Pc protein are probably bound quite close to one another at specific target sites. Antibodies to trx have indicated that trx protein can be found bound to multiple sites overlapping with Pc protein. It is possible that trx binds to the carboxy terminus of Asx while the amino terminus of Asx binds to a small intermediary protein that binds to Pc. Asx could therefore mediate activation versus repression of the homeotic loci by directly regulating the activity of both trx and Pc proteins. 75 Chapter 3 Introduction P c G genes were originally categorized as members of the P c G based on two criteria: i) they either display homeotic phenotypes which result from the ectopic expression of homeotic genes, or ii) they enhance the homeotic phenotypes of other P c G mutations. Ideally, a P c G gene should do both but this has not always been the case. Some P c G genes (such as Asx and Psc) were originally identified by their failure to undergo head involution and had only very weak homeotic mutant phenotypes (Nusslein-Volhard et al. 1984; Jurgens 1985). Their role in homeotic gene regulation was not as apparent until they were combined with other P c G mutations (Jurgens 1985). Double and triple mutant combinations between weak P c G mutations can produce an overall strong mutant phenotype (Jurgens 1985). A s well, some P c G genes show strong enhancement of homeotic phenotypes in other PcG mutants but by themselves have no obvious homeotic phenotypes (Sato et al. 1983; Adler et al. 1989). Enhancer of Polycomb (E(Pc)) and Suppressor ofzeste 2 (Su(z)2) are examples of two P c G mutations that can strongly enhance the homeotic phenotypes of some other PcG mutations, but on their own do not show adult homeotic transformations or homeotic misexpression phenotypes in the embryo (Sato et al. 1983; Adler et al. 1989; McKeon and Brock 1991a; Soto et al. 1995). Only when the maternal component of the E(Pc) gene is Completely removed do mutant embryos exhibit a very slight ectopic expression of the homeotic gene AbdB in the embryo (Soto et al. 1995). Using the above criteria, most strong members of the P c G (ie. those that display dose sensitive dominant adult homeotic transformations) have probably already been identified. However, using genomic deletions to enhance the extra sex combs phenotype of other P c G mutations, it has been estimated that the genome may contain up to 25 unidentified P c G genes (Jurgens 1985; Landecker et al. 1994). There are two major problems with identifying new P c G mutations using genetic means. 76 First of all, not all mutant PcG combinations show enhancement of P c G homeotic phenotypes. Some double or triple mutant PcG combinations produce very strong homeotic mutant phenotypes and some do not show any genetic interaction at all (Campbell et al. 1995; Cheng et al. 1994). This has been used to argue for specific functional subsets within the P c G that may reflect the existence of distinct functional complexes (Campbell et al. 1995; Cheng et al. 1994). Most of the traditional genetic screens have looked for P c G interactors using Pc (and in one case Pel) alleles so these screens may have missed mutants that only interact with other P c G genes (Kennison and Tamkun 1988; Landecker et al. 1994). Screening the entire genome looking for genetic interactors with the entire panel of known P c G mutants would be tedious. The second major problem with a genetic approach is that not all genes that contribute to P c G function may be dose sensitive. The most straight forward method of screening for new mutations is to look for enhancement of adult homeotic transformations in double heterozygous mutants. However, this is only effective i f a new mutation is dose sensitive and can enhance other P c G mutations as a heterozygote. A more direct way of identifying new P c G genes is to use the yeast 2 hybrid system to screen for proteins that interact with a target P c G protein of interest and then to try and establish a possible in vivo functional role for the new gene. This has two major advantages over a genetic approach; 1) it can be used to identify proteins that contribute to the function of only one P c G gene product and may be relatively dose insensitive and 2) it provides an immediate c D N A clone, avoiding the difficult process of cloning a gene of interest after a genetic interaction has been identified. Asx protein can be found bound to multiple chromosome sites containing other P c G proteins. A t some of these sites, it may bind extremely close to the P c G protein Pc, suggesting that Asx is part of a PcG complex at these sites. Asx does not interact directly with any P c G proteins in the yeast 2 hybrid system ( M . Kyba and H . Brock, unpublished) 77 so i f it is a member of a P c G complex, it must interact with the complex through proteins other than those tested. Unlike other P c G genes, Asx has tissue-specific effects on homeotic gene regulation, suggesting that its activity is required differentially in different tissues. Asx itself is expressed ubiqitously in all tissues, so it is unclear how the protein is targeted to its sites of activity only in those tissues where it is required. One possibility is that binding to its target sites is mediated by other proteins that are themselves expressed in a tissue specific pattern. The main objective in this chapter was to take four new proteins retrieved from a yeast 2 hybrid c D N A library screen that interact with Asx, and to try to determine i f they could contribute to the in vivo function of Asx. To test the possibility that there may be genes that contribute to the function of Asx but do not themselves have any obvious homeotic phenotypes, a yeast-2 hybrid screen was initiated in our lab using the carboxy terminus of the Asx protein ( M . Kyba and H . Brock, unpublished data). Several interacting genes were recovered ( M . Kyba and H . Brock, unpublished data) and four of them were chosen for further analysis. Genomic deletions that uncover P c G loci tend to enhance the homeotic transformations of other P c G mutations, so genomic deletions that contain a dose-sensitive gene suspected of interacting with Asx should enhance the severity of homeotic transformations in double mutant combinations. To test this possibility, the chromosomal location of each interactor was mapped and large genomic deletions that remove each endogenous locus were tested for genetic interactions with Asx. A large genomic deletion (Df(3L)ZN47) that removes the locus of the interactor termed z40 shows weak enhancement of trxG phenotypes in an Asx mutant background indicating that this protein may contribute to one specific aspect of Asx function. The observed genetic interaction may be weak because z40 is not a very dose-sensitive gene, so homozygous mutant Df(3L)ZN47 embryos were examined for the regulation of several homeotic genes. Surprisingly, Df(3L)ZN47 mutant embryos have 78 highly target specific embryonic homeotic misexpression phenotypes, including extensive derepression of the homeotic gene Sex combs reduced (Scr). This particular pattern of target specific homeotic gene regulation has not been seen with any other PcG mutations. It is unknown what the relationship is between these embryonic phenotypes and Asx function in particular, but since Df(3L)ZN47 interacts with Asx mutations in the adult, one possible interpretation is that the z40 gene product mediates Asx target specificity. Results Mapping the interactors The genomic location of each interactor was mapped by hybridizing the labeled cDNA's to polytene chromosomes (Fig. 15). The z3, z34 and z40 genes are all on the third chromosome while the zll locus maps to the X chromosome. The z3 interactor maps to 85E (Fig. 15 A) , zll to 14C (Fig. 15 B), z34 to 100E (Fig. 15 C) and Z40 to 65A (Fig. 15 D) . A genomic deletion that uncovers the z40 locus displays enhanced trxG abdominal transformations in an Asx mutant background Large genomic deletions that remove each of the above loci were tested for their ability to enhance adult homeotic transformations of Asx mutations. Deletions of zll and z34 failed to have any noticeable effect. The z3 interactor maps to the same location as the PcG gene Sex combs on midleg (Scm) but it does not have a sequence that matches Scm and thus is a unique gene (Bornemann et al. 1996). Any deletions that remove z3 would also remove Scm and so this locus could not be genetically tested. Southern blotting was used to confirm that the large genomic deletion Df(3L)ZN47 completely removes the z40 locus (Fig. 16, A-D). Df(3L)ZN47 shows variable penetrance of a weak trxG homeotic transformation of abdominal segment 5 (A5) towards more anterior abdominal segments. In a heterozygous Asx^ mutant background, the penetrance 79 Fig. 15 Mapping of interactor c D N A ' s to polytene chromosomes. A , z3 at 85E. B , zll at 14C. C , z34 at 100E. D, z40 at 65A. 80 81 F i g . 16 The large genomic deletion Df(3L)ZN47 removes the z40 locus which produces a transcript of 1.8 kb in length. A , polytene chromosome showing the z40 1.0 kb c D N A fragment signal at 65A and the extent of the Df(3L)ZN47 deletion. The deletion extends from cytological position 64C to 65C. B , Genomic Southern blot with D N A prepared from Dfi3L)ZN47 heterozygous flies (lane 1) and wild type flies (lane 2) probed with the 1.0 kb z40 c D N A fragment. There should be half as much D N A from region 64C;65C in the Df(3L)ZN47 flies. The signal in lane 1 is about half as intense as in lane 2, indicating that the 1.0 kb z40 c D N A fragment is derived from the 64C;65C region. The M lane contains 1 kb ladder molecular weight marker D N A , sizes are given in bp. C, same lanes as in (B) probed with a D N A fragment from the 48A region, outside the boundaries of the Dfi3L)ZN47 deletion. Signal intensity is about equal in both lanes indicating the amount of genomic D N A loaded in each lane is about equal. D , gel of blot from (B) also showing the amount of genomic D N A loaded in lanes 1 and 2 is about equal. E , Northern blot probed with the 1.0 kb z40 c D N A fragment giving a signal at about 1.8 kb. Sizes are marked in bp. 82 83 Table 6a : Enhancement of Df(3L)ZN47 trxG phenotypes by heterozygous Asx3 mutations Females X males Df(3L)ZN47/Asx Df(3L)ZN47/+ Asx/+ Asx3 X Df(3L)ZN47 100 (36) 90.5 (42) 15.9 (44) Df(3L)ZN47 X Asx3 100 (25) 93.8 (32) 0(28) Table 6b : Enhancement of Df(3L)ZN47 trxG phenotypes by Asxpl mutant allele Asxpl/Asxpl-y Df(3L)ZN47/ + Asxpl/ Asxpl • y + / + Asxpl/+ ; Df(3L)ZN47/ + Asxpl/+ + / + Df(3L)ZN47 72.9 (59) 18.1 (238) 28.1 (235) 0(385) The first number in each column represents the percentage of flies of each genotype that had a transformation of abdominal segment 5 (A5) towards the anterior. The number in brackets represents the total number of flies scored for each genotype. Average number of sex combs per fly was also examined and found to be wild type. 84 of these weak trxG abdominal transformations is slighdy increased (Table 6a). No genetic interaction was seen between Df(3L)ZN47 and the Polycomb allele Pc4. Flies that are homozygous for the AsxR1 mutation and heterozygous for Dfi3L)ZN47 are semi-lethal as observed by the relative absence of this genotype compared to others in the table 6b cross. A l l the surviving Asxpl ;Df(3L)ZN47/+ flies showed a significant increase in the penetrance of strong transformations of A 5 towards the anterior, similar to the phenotypes seen when Asx alleles are crossed to trx mutations. The full length z40 transcript is 1.8 kb The partial z40 c D N A retrieved from the yeast 2 hybrid interactor screen is only about 1.0 kb in length. A Northern blot was probed with this c D N A to determine the size of the transcript produced by the z40 locus and the full length size of the z40 transcript was determined to be 1.8 kb (Fig. 16, E) . B y examining a Drosophila E S T database, a further 360 nucleotides of z40 were recovered and the resulting c D N A now includes the full z40 open reading frame ( M . Kyba, unpublished result). The genomic deletion Df(3L)ZN47 that uncovers the z40 locus has embryonic homeotic misexpression phenotypes The interactions of Df(3L)ZN47 with different Asx mutations is relatively weak. One reason could be that the z40 gene product is relatively dose-insensitive so that one wild type copy can fulfill most functions of the z40 gene. To examine the effect of complete removal of the z40 locus on homeotic gene regulation, Dfi3L)ZN47 homozygous mutant embryos were stained with antibodies to the homeotic proteins Ultrabithorax (Ubx), Abdominal B (AbdB), Antennapedia (Antp), and Sex combs reduced (Scr) (Fig. 17). The deletion that removes the z34 locus was also examined but no alterations in homeotic gene regulation were seen. 85 F i g . 17 Expression of various homeotic genes in Df(3L)ZN47 homozygous mutants ( A , C , E , G , I ) compared to wi ld type embryos ( B , D , F , H , J ) . The embryo in A is viewed from a ventral-lateral position to better highlight the Ubx expression in the ventral nerve cord, anterior is to the right and posterior to the left. Embryos in B - H are laterally viewed with anterior to the left, posterior to the right, dorsal up and ventral down. Embryos in I and J are ventrally viewed with anterior to the left and posterior to the right. A and B , expression of Ubx is increased in PS5 of Df(3L)ZN47 mutants (A) compared to wi ld type (B) but the PS5 boundary is maintained (white arrow). The parasegment 6 (PS6) boundary is marked with an arrowhead and the PS5 boundary is marked with a small white arrow in (A) and a small black arrow in (B). C and D , AbdB expression is normal in Df(3L)ZN47 mutants (C) compared to wild type (D). E and F , AntP expression is slightly reduced in Df(3L)ZN47 mutants (E) compared to wild type (F). G to J , Scr expression is ectopically expressed in Df(3L)ZN47 mutants (G and I) compared to wi ld type ( H and J) . G , Mutant Dfl3L)ZN47 embryos lack many of the typical morphological landmarks of wt embryos, but the patch of Scr expression in the head appears to be in the brain (arrowhead,^ ). Scr expression is also seen in the central nervous system (arrowheads,c), and in the posterior of the embryo (arrowhead,^. H , wild type Scr expression can be seen in PS2 and 3, in the foregut (small arrow,a), and very slightly in the hindgut (arrowhead,ci). There is no Scr expression in the brain (arrowhead,^) or the C N S (arrowheads,*;). I, the misexpression of Scr in the posterior of the embryo (arrowhead,tf) is more obvious in this ventral view where Scr is expressed in a large patch. Df(3L)ZN47 mutants show essentially wild type Scr expression in the foregut (small arrows, a). J , expression of Scr in the posterior of wild type embryos is limited to a small region in the hindgut (arrowhead,<i). 86 87 Df(3L)ZN47 homozygous mutant embryos show extensive morphological defects (Fig. 17, A ,C ,E ,G , I ) . The midgut is bloated and undefined and the embryos are shorter and have an overall stunted appearance. The C N S lacks definition and axon formation is disorganized. Anterior embryonic development appears to be slowed compared to posterior development. For example, in the embryo shown in Fig. 17 G , the embryonic head and brain (arrowhead,*?) has the features of a stage 11 embryo while the hindgut (arrowhead,^) has the features of a stage 13-14 embryo. Despite the lack of defining morphology, the PS5 expression boundary of the Ubx gene appears to be maintained in Dj\3L)ZN47 homozygous mutants (compare Fig . 17 A to B , small arrows mark the PS 5 expression boundary, arrowheads mark the PS 6 expression boundary). However, in wi ld type embryos, Ubx is expressed at much lower levels in PS5 relative to PS6 and only in the anterior compartment of PS5 (see Fig . 17 B) . In Df(3L)ZN47 homozygous mutants, Ubx expression is increased so that PS5 expression is now equal to that of PS6 and Ubx is also expressed ubiquitously throughout the PS5 compartment (see Fig . 17 A ) . There is no alteration in the expression of AbdB in Df(3L)ZN47 homozygous mutants (Fig. 17 C) compared to wild type embryos (Fig. 17 D) . Regulation of the Antp gene is altered in that expression of Amp is slightly reduced in Df(3L)ZN47 homozygous mutants (Fig. 17 E) compared-to wi ld type embryos (Fig. 17 F) . Reduction of homeotic gene expression is usually associated with mutations in members of the trxG. Since homeotic proteins expressed in posterior regions repress the expression of homeotic genes with more anterior expression boundaries, this reduction could be indirect and due to the slight increase of Ubx expression. However, Antp expression is reduced in regions where there is no Ubx expression which indicates that the reduction of Antp expression cannot be explained entirely by an increase in Ubx expression. 88 Mutations in most P c G genes normally cause suppression of Scr expression in the gut due to repression from expanded posterior homeotic gene expression boundaries. Surprisingly, Scr expression is widely derepressed in multiple tissues of Df(3L)ZN47 homozygous mutants (Fig. 17 G and I). In wi ld type embryos, Scr expression is normally restricted to PS2 and PS3 (Fig. 17 H and J) and the foregut (Fig. 17 H and J, short arrows, a) with a slight expression in the hindgut (Fig. 17 H and J, arrowhead,^). Because Dfi3L)ZN47 homozygous mutants lack morphological landmarks in the head, it is difficult to compare Scr expression in the head region with wild-type embryos. However, Scr appears to be ectopically expressed in the brain of Df(3L)ZN47 homozygous mutants compared to wild-type (compare Fig. 17 G , b arrowhead with Fig. 17 H , b arrowhead). In Dfi3L)ZN47 homozygous mutants Scr is also ectopically expressed throughout the C N S (Fig. 17 G compare to 17 H , c arrowheads), and in the posterior of the embryo (Fig. 17 G and I compare to 17 H and J, d arrowheads ). Ectopic staining of Scr protein in these regions is relatively weak, probably due to repression from other homeotic genes preventing full expression of the Scr gene in these regions. Despite its lack of morphology, expression of Scr in the foregut appears normal (Fig. 17, G and I compare to H and J, short arrows, a). In typical P c G mutants, regulation of all the homeotic loci are affected resulting in derepression of the homeotic genes. Df(3L)ZN47 is unusual because it displays a range of homeotic gene misexpression phenotypes. In Df(3L)ZN47 homozygous mutants, Scr expression is widely derepressed in many tissues. Ubx is only weakly derepressed in PS5, while Antp expression is actually suppressed. AbdB expression is unaltered and appears normal. This indicates that the z40 locus could regulate only some homeotic loci and possibly have a different specific role at each locus. A n attempt was made to make double homozygous Asx;Df(3L)ZN47 mutant embryos but no double mutant combinations were recovered suggesting that embryos of this genotype all died early in embryogenesis. 89 Df(3L)ZN47 does not affect the regulation of two PcG dependent Scr pairing sensitive regulatory elements Two putative PREs have been identified in the regulatory region of the Scr gene, a 10.0 kbXbal sequence and an 8.2 kb Xbal sequence. They were identified because they show P c G dependent variable repression (or variegation) of a miniwhite reporter gene. The miniwhite reporter gene is responsible for depositing pigments into the eye in a cell autonomous manner so i f the gene is on in some cells and off in others, a red and white mosaic or variegated pigment pattern results. Normally, when additional copies of the miniwhite reporter gene are introduced into the genome, eye pigment is increased. However, expression of the miniwhite reporter gene is reduced when a second copy of either of the above constructs is introduced on to the homologous chromosome. This has been interpreted to mean that the two elements are somehow pairing with one another in trans and increasing the repression of the miniwhite reporter gene. This trans enhancement of repression is termed pairing sensitivity. Although not all PREs that show pairing sensitive repression are necessarily maintenance elements, most P c G maintenance elements have a pairing sensitive effect. The variegation of the 10.0 kb and the 8.2 kb Xbal sequences is decreased in some, but not all, P c G mutant backgrounds, indicating that the repression is at least in some cases P c G dependent. Asx mutations do not affect variegation of either construct. Since it has such a strong effect on regulating the endogenous Scr gene, Df(3L)ZN47 was tested with lines containing these response elements to see i f it would alter the regulation of these transgenes. Df(3L)ZN47 was tested with one line containing the 8.2 kb Xbal construct and three separate lines containing the 10.0 kb Xbal construct. In all cases, Df(3L)ZN47 had no detectable effect on eye variegation. Thus, Df(3L)ZN47 probably does not regulate the 90 Scr gene through either of these regulatory sequences, indicating that there may be other important regulatory elements at the Scr locus that have not yet been identified. Discussion Df(3L)ZN47 genetically interacts with Asx mutations Df(3L)ZN47 is a large genomic deletion that removes the z40 locus as well as many other genetic loci. Thus, the results with this deletion are suspect at least to the extent that they do not rule out the possibility that another gene in the deletion other than z40 is responsible for the observed genetic interaction. However, the fact that this deletion removes the z40 locus and z40 interacts with Asx in the yeast 2 hybrid system favors the possibility that these results are due to the z40 locus itself. Dfi3L)ZN47 heterozygous mutants have a very weak trxG homeotic mutation in the abdomen. In an Asx3 mutant background, the double mutants show complete penetrance of this phenotype. However, since the penetrance of the phenotype in the single Dj\3L)ZN47 heterozygous mutants was high to start with, the significance of this increase is difficult to gauge. Dj\3L)ZN47 interacts much more strongly with Asxpl in that it appears to be semi-lethal with it. This semi-lethality could be indirect and simply due to the combined mutant effects of the Asxpl mutation and the Df(3L)ZN47 deletion, but the fact that all the survivors show a significant increase in the penetrance of strong trxG abdominal transformations argues for a specific genetic interaction between Asxpl and the Df(3L)ZN47 deletion. This enhancement of trxG phenotypes suggests that the Df(3L)ZN47 deletion (and therefore possibly the z40 gene itself) contributes to Asx activity in trxG mediated homeotic gene regulation. Anterior homeotic transformations in the abdomen are generally caused by reduced expression of the abdA gene. Unfortunately, no antibodies were available to look for possible alterations of abdA regulation in Df(3L)ZN47 homozygous mutant embryos. However, the fact that Df(3L)ZN47 homozygotes could suppress the 91 expression of the homeotic gene AntP (which is a trxG phenotype) at least argues for the possibility that z40 could also be required for activation of the abdA homeotic locus. Df(3L)ZN47 removes the z40 locus and has target specific effects on homeotic gene regulation Df(3L)ZN47 specifically causes the extensive derepression of the Scr gene. This is unique for two reasons. No P c G mutation examined to date causes extensive derepression of only one specific homeotic locus. Also, P c G mutations tend not to show ectopic Scr expression in the embryo. This second observation is due to the fact that the proteins of the B X - C (Ubx, abdA and AbdB) , all repress expression of Scr. When these genes are ectopically expressed in P c G mutants, the result is the downregulation of Scr expression throughout most of the embryo. However, even in P c G mutants, Scr expression is still maintained at relatively normal levels in its normal domain of expression, parasegments 2 and 3. This is probably due to the fact that moderate levels of Antp expression in these two parasegments maintains the expression of Scr. Both extra low levels and extra high levels of Antp expression tend to suppress the expression of Scr. In the Df(3L)ZN47 homozygous mutants, although ectopic Scr expression is spatially extensive, the levels of ectopic expression are relatively low. This is probably due both to repression from the presence of Ubx, abdA and AbdB expression, and to suppression of activity from the reduced levels of Antp expression observed (see below). In addition to its effect on the regulation of Scr, Df(3L)ZN47 also has a weak effect on the regulation of both Antp and Ubx. Ubx is not ectopically expressed outside its normal domain but parasegment 5 (PS5) expression is increased to levels normally only seen in parasegment 6 (PS6). This indicates that the product from Dj\3L)ZN47 is somehow required for proper Ubx expression in PS5, but not in other parasegments. Antp expression is reduced in Dfi3L)ZN47 mutants. Reduction of homeotic gene expression is normally considered to be a trxG phenotype. The only observable adult phenotype in 92 Dfi3L)ZN47 heterozygotes is a weak trxG transformation in the abdomen, which raises the possibility that expression of the homeotic gene abdA is also reduced in DJf 3L)ZN47 mutants. In vivo functional role for z40 One thing that might be expected from a screen designed to identify new P c G genes that are not easily identified using a traditional genetic approach is that these new genes might have non standard (in terms of P c G function) effects on the regulation of the homeotic loci. The z40 gene seems to fit this idea as it is relatively dose insensitive in terms of adult homeotic phenotypes and it displays an unusual array of target specific effects on homeotic gene regulation. Since double homozygous combinations between Asx and Dfi3L)ZN47 apparently did not produce viable embryos, it is difficult to determine the extent to which the embryonic phenotypes of Dj\3L)ZN47 contribute to the in vivo function of Asx. However, one unknown aspect of P c G function is how P c G proteins act on specific target sites only in domains where their activity is required. A l l P c G proteins are expressed ubiquitously throughout the embryo. Thus one possibility is that P c G proteins interact with other proteins that themselves have target specific activity. The z40 protein could be such a target specific factor and may contribute to Asx function by targeting Asx activity towards different specific loci. The combination of both P c G and trxG homeotic mutant phenotypes is also consistent with a possible in vivo interaction with Asx that itself has been implicated in both P c G and trxG activity. Thus z40 could be required for the Asx mediated activation of some target loci and the Asx mediated repression of others. Asx protein binds to some chromosomal target sites very near the Pc protein but it does not bind to Pc directly ( M . Kyba, unpublished). However, the z40 protein has been tested and it interacts directly with the Pc protein both in the yeast-2 hybrid system and in an in vitro G S T fusion assay ( M . Kyba, unpublished). This result provides a possible role 93 for z40 protein in the function of Asx. Perhaps z40 binds to Pc only at specific target loci and then the Asx protein binds to z40 and is able to interact with Pc. Asx;Pc heterozygous mutants show strong enhancement of the extra sex combs phenotype that results from the ectopic expression of Scr. The fact that Df[3L)ZN47 mutants show strong derepression of the Scr locus is consistent with the possibility that it interacts with Pc and Asx in such a manner. The fact \hatDf(3L)ZN47 heterozygous mutants fail to enhance the extra sex combs phenotype of Asx mutations may again reflect the relative dose-insensitivity of the z40 locus. 94 Chapter 4 Introduction Not all double heterozygous combinations between different P c G mutations show the same level of enhancement of dominant homeotic phenotypes. Some double heterozygous combinations produce flies with very strong P c G homeotic phenotypes while some mutant combinations only produce weak , or in some cases, no P c G homeotic phenotypes at all (Campbell et al. 1995). One of the reasons for this variability in the degree of enhancement seen is at least partly due to the fact that some P c G genes have a strong maternal component that can provide partial rescue of mutant phenotypes. However, the variations in phenotypic strength seen with different mutant combinations cannot be wholly explained by variations in maternal contribution since one would then expect that all mutant combinations between genes with a strong maternal component should interact only weakly. To some extent, strong genetic interactions probably reflect important in vivo functional interactions. Asx shows strong genetic interactions with only a subset of P c G genes which include Polycomb (Pc), Poly comblike,(Pel), Sex combs extra (See) and super sex combs (sxc) (Campbell et al. 1995). The strongest interaction is seen between Asx and sxc (Campbell et al. 1995). The sxc3 mutation is lethal or semi-lethal as a double heterozygote with all Asx alleles tested (D. Sinclair and H . Brock, unpublished data). In cases where a small proportion of double heterozygotes survive, they show multiple P c G homeotic phenotypes including extra sex combs, transformations of abdominal segment 4 towards more posterior segments and transformations of the wing towards the haltere ( D . Sinclair and H . Brock, unpublished data). Other sxc alleles are not lethal with Asx which argues that the sxc3 mutation is a gain of function mutant that is directly interfering with the function of Asx protein. This could indicate that there is an interaction between the Asx and sxc proteins in vivo. 95 There are only five alleles of sxc, all isolated in the original screen in which sxc was identified (Ingham 1984). The sxc gene has not been cloned, perhaps because it is situated at the base of 2R in the 4 IC region, very close to centric heterochromatin. Alleles of sxc display segmentation defects, a phenotype rarely seen in PcG mutants but one that is also seen with several Asx alleles. Homozygous sxc mutations survive through embryogenesis all the way through pupation and die as pharate adults just before eclosure (Ingham 1984). These dead pharate adults display typical PcG homeotic mutations such as extra sex combs and wing to haltere transformations as well as occasionally displaying the rare antenna to leg transformation (Ingham 1984) that is only seen in mutants of one other PcG gene, Pc. The lethal interaction between sxc3 and Asx and the fact that both Asx and sxc have segmentation defects suggests that Asx and sxc proteins may interact directly in vivo and that sxc may contribute to one or several aspects of Asx function. To start examining the possibility that the two proteins may interact, an attempt was made to clone the sxc gene using P element insertional mutagenesis. P element insertion mutagenesis has long been used as a method for cloning genes of interest in Drosophila. P elements are transposable elements that can exise and reinsert into different locations throughout the genome when they are in the presence of a source of transposase. B y screening for specific mutations, an insertion in a particular gene of interest can be selected for. A P insert line can then be used to make a library of genomic D N A from which D N A flanking the site of insertion can be recovered by screening with P element sequences. Alternatively, flanking D N A can be recovered by using inverse P C R methods or by using a P element construct that can be rescued as a plasmid. A rescuable P element is modified to contain an antibiotic resistance gene as a selectable marker and a plasmid origin of replication (see Fig. 18A). When cut with Xbal and religated, the selectable marker plus origin of replication are combined with flanking D N A sequences and can be selected for as a plasmid that confers antibiotic resistance on competent cells. 96 The problem with P insertional mutagenesis is that the chances of getting an insertion event in a target gene of interest is quite low. P elements do not insert randomly into D N A but the exact sequence preferences for P element insertions are unknown. For any particular locus, the chance of an getting an insertion can range from 1/1000 (or 0.1%) to greater than 1/100,000 (or 0.001%). In practical terms, it is usually too labour intensive to screen for more than 10,000 possible events, so multiple P elements are often mobibzed to increase the likelihood of a single P inserting into the target locus. This approach creates the problem of cleaning up multiple P elements left over throughout the genome after the screen is completed. A cleaner approach uses a "local hop" (Zhang and Spradling 1993; Tower et al. 1993). Almost 30% of all mobilization events wi l l result in the P element preferentially reinserting into the genome within 100 kb of its starting point (Tower et al. 1993). A particular P element only mobilizes about 7-10% of the time so i f you start within 100 kb of your target gene, the chances of getting a P element reinsertion event within this 100 kb can be as high as 2% (Tower et al. 1993; Zhang and Spradling 1993). Thus local hops require two things; i) the approximate location of the gene of interest, ii) a P element that is within 100-200 kb nearby. One problem with local hops is that P elements wi l l often make small deletions when they excise from the D N A (Tower et al. 1993) so any local P screen must differentiate between small deletions that remove the locus of interest and genuine P insertion events. A local P screen was attempted for the sxc locus using a P element construct ( a P[lacZ,ry+] element, also called PZ) that is homozygous lethal and inserted in the 41C region, but not in the sxc gene itself. Two separate P screens resulted in the generation of multiple small deletions in the region including five new sxc alleles but no insertion events were recovered. To enrich for insertion events and select against deletions, the original screen was redesigned using a small deletion that was lethal with the original insertion but was viable with sxc mutations. This screen was attempted by two undergraduate students and they successfully recovered two potential inserts in the sxc gene. 97 Results A local P screen for sxc resulted in the generation of new sxc alleles P insertional mutagenesis was attempted on the sxc locus using a P[lacZ,ry+] (or PZ) element (shown in Fig. 18A) that was inserted within the 41C region. The P Z element is inserted in an uncharacterized lethal site (1(2)02047) and is therefore homozygous lethal, but it is viable with sxc null alleles indicating that the uncharacterized lethal site is not sxc itself. For ease, the P Z element inserted at 1(2)02047 in 41C wi l l be referred to simply as P 4 1 C . Details of the screen are given in Fig. 19. The screen was initially done with the presumed sxc null allele sxc4 and the phenotype screened for was lethality with sxc4. Genuine P element insertion events can be distinguished from deletions because lethality due to a P element insertion is revertable by remobilizing the P element and causing it to excise out of the gene. A s well, i f the P element has inserted into a new location, it should produce a new band on a genomic Southern blot so this can be used as quick way of checking for possible insertion events. Out of 2,856 chromosomes (ie. individual pair matings), 14 mutants lethal with sxc4 were recovered. When these 14 mutants were retested by crossing them to the sxc null alleles sxc^ and sxc$ and the sxc semi-lethal mutant sxc?, only two of the putative sxc mutants were lethal or semi-lethal with these other sxc alleles, although all 14 mutants continued to be lethal with sxc4. The most likely explanation for this is that the sxc4 mutant has a second lethal mutation between the sxc mutation and the lethal insertion site of the P41C element. Thus, out of these 14 lethals, 12 affected this second lethal site but not sxc itself and only therefore 2 were genuine sxc alleles (see Table 7). The two new sxc alleles recovered from the screen were originally named sxc^'4^ and sxcIX-1. The exact nature of these mutations is unknown, they could be deletions extending from P41C to sxc , they could be insertions in sxc that immediately remobilized 98 F i g . 18 Diagram and restriction digest of the P Z construct. A , diagram of the P Z construct showing the approximate positions of EcoRl and Xbal restriction enzyme cut sites. There is only one internal Xbal site. The construct contains a copy of the wild type Drosophila rosy gene (rosy +) and a heat shock promoter (hs 70) linked to a lacZ gene. Cutting an inserted P Z element with Xbal and religating it produces a smaller functional plasmid containing flanking D N A and a Kanamycin resistance gene (Kan1 -) and origin of replication (ori). The P element ends that are required for transposition are indicated by boxed Ps. Each is approximately 500 bp's in size. B, A restriction digest of the entire circularized element shown in (A). Lane 1, 1 kb ladder D N A , sizes are marked in bp's. Lane 2, the P Z plasmid cut with Xbal produces a single -18 kb band. Lane 3, cutting with EcoRl produces a distinctive set of 3 internal bands (indicated by white arrows) and one flanking band. Starting from the smallest to the largest, the EcoRl fragments are 3.0 kb, 3.5 kb, 4.5 kb and 7.0 kb. The 3.0 kb fragment contains the lacZ gene, the 3.5 kb fragment contains the K a n r and the ori and the 4.5 kb band contains the bulk of the rosy gene. The 7.0 kb fragment contains the rest of the rosy gene, the P element ends, plus some extra sequences not shown in the diagram. 99 EcoR I 4 PZ P EcoR lacZ hs70 Kanr on - O Xba I , EcoR I EcoR I \ J) rosy + jj P 100 F i g . 19 Local P screen for sxc insertions. The P Z element in its original position (labelled P41C) was mobilized by crossing it to a source of transposase (P[ r y + A 2-3]). Individual putative transposition events (labelled P41C*) were then captured by balancing them over a C y R o i marked balancer chromosome. Individual events were screened by crossing single males to 5-10 females carrying an sxc mutation. The F3 was examined for the phenotypes described in the figure. 101 P P41C/CyO ; +/+ & X Sp/CyO ; P[ ry+ A 2-3] Sb/TM6 9 F1 P41C/CyO ; P[ ry+ A 2-3] Sb/+ & X Gla Bc/CyRoi ; +/+9 F2 P41C*/CyRoi ; +/+cf X sxcVSMS? single pair matings F3 score for the presence of homeotic transformations or for the absence or reduced frequency of the P41 C/sxc* genotypic class 102 F i g . 20 Southern blots of individual putative lines. A , genomic D N A was digested with Xbal and probed with a radioactively labelled lacZ probe. Lane 1, genomic D N A from the original P41C line produces a large -10 kb band. Lane 2, genomic D N A from sxc?. Lane 3, genomic D N A from the 11-78 line. Lane 4, genomic D N A from sxc^. B, genomic D N A was digested with EcoRl and probed with the full length P Z element shown in Fig . 18. Lane 1, genomic D N A from the original P41C line produces a specific pattern of bands. Three internal bands from the P Z element itself are produced (marked with black arrowheads, compare with EcoRl digestion shown in Fig. 18). One very large band that contains flanking D N A from the rosy side of the P Z construct is also produced (white arrowhead). Lane 2, sxc1®. Lane 3,11-78. Lane 4, sxc^. Lane 5, sxc8. Lane 6, 6-3 line, a derivative of P41C in which the P element has failed to move. Lane 7, sxc9. A l l of the genuine sxc alleles are missing the internal P Z bands, indicating that the P Z element has undergone an internal deletion in these lines. No new bands are present in the sxc lanes that are not present in either the P41C or 6-3 lanes. 103 104 Table 7: Results of two separate sxc local P screens Total number of chromosomes screened Total number of putatives recovered Total number of new sxc lethal alleles recovered Rate at which new sxc alleles were generated 2,856 14 2 0.07 % 4,757 6 3 0.06 % Totals 7,613 20 5 0.07 % 105 causing an internal s x c deletion, or they could be P element insertions. Because of these variables, these and all other s x c alleles that were recovered were named simply using standard genetic notation. Following standard genetic notation, these alleles w i l l now be known as sxcP and s x c 7 , respectively. Genomic D N A was prepared from sxc& and s x c 7 stocks, blotted and probed with D N A from the lacZ gene. Compared to the original P41C line, the P element in s x c 7 did not produce a band of a different size indicating that the P element had not re-inserted into a new location (see Fig 20A). The sxc& line did produce a new band when probed with lacZ, a larger 20 kb band. A reversion of lethality was attempted on sxcP by remobilizing the P element. Usually, when a P element transposes to a new location, it leaves a copy of itself behind or it causes a small deletion in its original location. Occasionally, when it moves to a new location it excises completely from the D N A and leaves no trace of itself behind. This is called a clean excision and in order to revert a mutant phenotype associated with a P element insertion a clean excision is required. The P41C element in its original position reverts its lethality at a rate of 20/832 or 2.4%. Reversion of lethality with s x c was attempted on the s x c ^ line but out of 1091 flies, no revertants were found. It was concluded that the sxc^ and s x c 7 lines were not P element insertions and were probably deletions based on the fact that they either; i) did not produce a new band on a Southern blot or ii) the lethality of the line was not revertable. The same screen was attempted again with two modifications. First, instead of s x c 4 , the null allele s x c 1 , which should contain no second site mutations, was used to screen for P insertions. Second, P element insertions often cause more subtle mutations rather than lethality, so every single cross was examined for semi-lethals and homeotic mutations as well as lethal interactions. With these two modifications, the same screen as outlined in Fig . 19 was attempted. 4,757 chromosomes were screened and three lethal lines were recovered and established (see Table 7). No semi-lethal lines were recovered and no crosses showed any homeotic transformations. The lethal lines recovered continued to be lethal with strong sxc alleles so they were named s x c 8 , s x c 9 and sxc^O. None of these 106 lethals produced new bands on a southern (Fig. 20B) and none of them were revertable for lethality with sxcf Again, none of these lethal lines appeared to be inserts and all were probably deletions. Out of a total of 7,613 chromosomes, five new sxc alleles, all probably deletions, were recovered giving a recovery rate of sxc alleles of about 0.07 % (see Table 7). No insertions were recovered indicating that P element insertion at the sxc locus occurs at a rate lower than 1/7,613. A P element insertion enrichment screen produced putative sxc insertions From the above screens, it is obvious that local P screens in the region of sxc generate deletions much more readily than insertions. One other problem with the local P screen is that most of the time, when a P element is exposed to a source of transposase, it doesn't move at all. To try and enrich for insertion events, the original screen was modified to select for mobilizations of the P element and to select against the generation of deletions (see Fig.21). The P41C element was mobilized as before by crossing it to a transposase source. P41C excision events were selected for in the germline of the F l flies by crossing them to AH-78/CyRoi flies. z\^-7S [ s a double mutant recovered from the very first screen. It is lethal with sxc4 and with the original P41C insert but not with other sxc alleles. It is presumed to be a deletion based on the fact that it does not produce a new band on a Southern (see Fig.20) and its lethality with sxc4 is not revertable. When crossed together as stable stocks, heterozygous P41C*/A^'^ flies die. Selecting for heterozygous P41C*/AH~78 survivors in the F2 selects for P41C elements that have hopped out of their original location (and are no longer lethal with z\^ ~7<3) and it also selects against deletions that would affect the P41C lethal site or the sxc4 second lethal site. Thus, for any P41C*/AN-78 survivors we know that the P element has been mobilized and we also know that it did not generate a deletion in the process. The P41C*/A^'^ flies can then be 107 F i g . 21 P enrichment screen for sxc insertions. Details are given in the text. 108 P P41C/CyO ;+/+cf X Sp/CyO ; P[ ry+ A 2-3] Sb/TM6 9 F1 P41C/CyO ; P[ ry+ A 2-3] Sb/+ & X AH-78/CyRoi ; +/+ 9 F2 P41CVA»-78 ; +/+ & X sxc 4 /SM5 9 single pair matings F3 score for the presence of homeotic transformations or for the absence or reduced frequency of the P41 C/sxc4 genotypic class 109 crossed to sxc4 to select for mutations in the sxc locus itself. The selection works as follows. A H ~ 7 8 / S X C 4 heterozygotes w i l l die ( A 1 1 ' 7 8 is lethal with sxc4). Both P41C and AjI-78 chromosomes w i l l be viable over SM5 which is marked with a dominant Curly marker. P41C should be viable over sxc4 and produce straight winged flies. If the P element has inserted in the sxc locus the P41C/sxc4 genotypic class should either be absent, reduced in frequency or exhibit homeotic transformations. The modified screen was attempted by two undergraduate students, Susan Leong-Sit and Ester O'Dor under my direction. Out of 261 A^'78/sxc4 individual pair matings, 17 putatives were found based on the fact that these flies showed tergite defects and weak homeotic transformations in the abdomen. Lines were established by individually crossing curly winged males to A^'78/CyRoi females. Crosses that produced straight winged survivors contained the P41C* putative. CyRoi siblings were collected and used to establish stocks of the putative insertion lines. Discussion The P element insertion enrichment screen The insertion enrichment screen provides a modification of the local P hop, allowing for selection against deletions. Unfortunately, the enrichment screen was not possible until the original screen had been done. Thus the enrichment screen is probably not generally applicable to other loci. A modification of the enrichment screen could have provided further enrichment for insertion events. The main problem with the screen as it was used is that it selects for the P element hopping out of the P41C locus, but it does not select for P element reinsertion. In many cases, the P41C element probably hopped out and then was lost. One way to ensure that you are testing only reinsertion events would be to select for the presence of ry+. If the entire screen was done in a ry mutant background, then only those flies that had the P element would be ry+. Thus at the F2 stage, you would be selecting for flies that were P41C*/A^'78 and were also ry+ so that you would know 110 that they all contained an insert. The problem with this approach is that it would have been too time consuming to make all the needed stocks ry", so the screen was attempted without selecting for re-insertions at the F2 stage. Future work Now that we have acquired putative inserts in the sxc locus, work has continued on cloning the sxc gene. Once the gene is cloned, it would be interesting to see i f the sxc protein can interact with Asx. One of the odd things about Asx is that in the yeast two hybrid system, Asx does not interact with any PcG proteins, although it does interact with trx, a member of the trxG. Other genetic and molecular data are consistent with the likelihood that A s x is, however, somehow part of a P c G complex. Mutations in Asx show strong enhancement of the mutant phenotypes of other PcG genes and Asx protein appears to be bound close to Pc protein at many chromosomal target sites. If Asx protein actually does interact directly with sxc protein, perhaps this is how Asx participates in a P c G complex. Alternatively, Asx binds to many target sites where there are no other P c G proteins bound. It would be interesting to see i f antibodies to sxc overlapped with Asx at these unique target sites. The fact that both sxc and Asx mutations can show segmentation defects, a phenotype not seen in most other PcG mutations, may indicate a specialized functional role for these two proteins that would be reflected in binding to unique target sites. Ill General Discussion Asx may be a component of the system that mediates activation versus repression at target loci Asx mutations can enhance both trxG and PcG phenotypes. If a gene is important for the activity of both the P c G and the trxG, we would expect that a mutation in such a gene should have the same phenotype as double mutations between P c G and trxG genes. Pc;trx and esc; trx double homozygous mutations to a large degree cancel out each others P c G and trxG homeotic mutations and produce an embryo that is more wild type than with either mutation individually, but still has a slight overall PcG phenotype (Ingham 1983; Capdevila et al. 1986). Homozygous Asx mutants have a weak P c G phenotype in the cuticle that is very similar in appearance to the phenotype of these PcG;trxG double mutants (Jurgens et al. 1984; Breen and Duncan 1986). This supports the possibility that a mutation in Asx is disrupting the activity of both the PcG and the trxG, or at least the activity of the Pc and trx genes. The P c G geneE(z) also displays both P c G and trxG phenotypes (LaJeunesse and Shearn 1996). The trxG protein G A G A is required for transcriptional activation from chromatin templates in vitro (Tsukiyama et al. 1994) but it has also recently been found associated with P c G complexes and PREs (Strutt et al. 1997; Hodgson and Brock 1998). Taken together, all of this suggests that the traditional view of two distinct, separate and antagonistic groups of genes with one set of activators and one set of repressors is probably inaccurate and there is much more functional interdependance between the trxG and the P c G than previously thought. Pc and trx proteins overlap at a large number of target sites on polytene chromosomes (Chinwalla et al. 1995). This is surprising because one would not necessesarily expect both an activating complex and a repressive complex to be bound to the same targets. Some of these target loci are active in the salivary glands (such as the P c G genes themselves) and some of them are repressed (such as the homeotic loci). This 112 suggests that binding of either a P c G or a trxG complex is not itself sufficient for either repression or activation and that there is some other step downstream from binding that determines i f a particular locus is repressed or activated. Asx may be a component of such downstream activity. Asx shows strong genetic interactions with both Pc and trx and is thus required for the activity of both genes. Asx binds directly to the trx SET domain in vitro (Kyba et al. 1998) and it binds very near the Pc protein in vivo. Pc and Asx do not bind to each other directly in vitro, but they both bind to z40 protein and thus may interact with each other through this z40 interaction. The binding sites for trx protein overlap with about 66 of the total of 90 Asx binding sites (Kyba et al. 1998), suggesting that the two proteins may also interact in vivo. A n overlap between trx and Asx of 66 sites is very close to the 63 sites of overlap between Asx and Pc. Although it has not been shown directly, it is possible that Asx, Pc and trx protein all overlap at the same set of sites. It would be interesting to see i f there are two separate complexes, one with Asx-trx the other with Asx-Pc, or i f there is just one specific complex containing all three proteins. One possibility is that Asx protein binds to both Pc and trx at the same time and mediates crosstalk between two separate complexes, although it is also possible that Asx is a member of two separate and distinct complexes; a P c G complex and a trxG complex. Asx has tissue specific effects and binds to unique target sites The fact thatAsx may be important for the activity of both the P c G and the trxG complicates the interpretation of homeotic derepression phenotypes seen in Asx mutant embryos. To some extent, i f the weak P c G cuticle phenotype of Asx mutants results from disrupting both P c G and trxG activity, weak P c G derepression phenotypes in Asx mutant embryos probably results from the same thing. However, this does not explain tissue specific differences in regulation and the observation that although Asx mutants produce a weak Ubx derepression phenotype in the C N S , there is no effect on the regulation of Abd 113 B in the C N S at all. This suggests that Asx activity is not uniform and it does not regulate all targets in an equal manner. Most of the genetic and molecular data favors the idea that there are one or more P c G complexes involved in regulating target loci . The fact that Psc, Su(z)2, and E(z) proteins as well as Asx can bind to some polytene sites in the absence of other P c G proteins (Rastelli et al. 1993; Carrington and Jones 1996) suggests that these proteins participate in some non-PcG complexes or that they are able to bind and be active at some target sites on their own. Some Asx mutations have segmentation defects, mutant phenotypes that are usually associated with mutations in the early acting segmentation genes. Asx protein is present in early embryos so it is possible that it has a role in these early regulation events. O f the P c G , only sxc and pleiohomeotic (pho) also have segmentation defects (Bieen and Duncan 1986; Ingham 1984). Asx and sxc show an unusually strong, G O F genetic interaction with each other (D. Sinclair, unpublished) suggesting the possibility that the two proteins may interact in vivo. One possibility is that Asx and sxc proteins interact and are required for the regulation of some early acting segmentation genes. It would be interesting to see i f sxc protein overlaps with Asx at all of its unique binding sites. Future work There are several important questions that remain: 1) what is/are the mechanism(s) of PcG/trxG requirements for Asx activity?, 2) what is the basis of the tissue specific requirements for Asx activity?, and 3) is Asx a member of multiple functional complexes at different target sites? First of all, to determine i f Asx is a member of both Pc and trx containing complexes in vivo, antibodies to Asx, trx and Pc, could be used for three separate co-immunoprecipitation experiments. Co-IP experiments have been done with ph and Pc antibodies (Franke et al. 1992) and an immunoprecipitation with a ph antibody purifies a 114 soluble complex with a band the size of Asx (J. Hodgson, H , Brock unpublished). This suggest that P c G proteins can interact as soluble complexes. If Asx is a member of two separate complexes, then one would expect that Asx would copurify with both Pc and trx but that Pc would not be present in the trx co-IP and vice versa. If Pc, trx and Asx were to all co-purify together this would be strong evidence that these three proteins interact in vivo and that Pc and trx are members of complexes that interact directly with each other. However, i f Asx protein interacts with both Pc and trx containing complexes only after they have bound to D N A , it may not copurify with soluble complexes. To deal with this possibility, one could map the position of Asx containing complexes at a regulatory target and compare the position of Asx containing complexes with other P c G and trxG proteins. A n antibody to ph has been used in a supershift assay to identify ph containing complexes bound to the bxd5.1 minimal P R E (Hodgson and Brock 1998). It seems likely that Pc wi l l be present in most of these ph containing complexes, but it would be interesting to see i f Asx and trx proteins overlap with any or all of these protein complexes. If Asx containing complexes are completely separate from Pc containing complexes but are found present in trx containing complexes (or vice versa), this might suggest that in vivo , the bound complexes directly interact with each other and thus are able to alter chromatin structure at a target locus. The fact that Asx can bind to multiple target sites without other PcG proteins present suggests that it may be a member of different functional complexes. The supershift assay could be used to screen for Asx complexes bound to different-PREs and then specific complexes could be purified using D N A affinity chromatography. It would be interesting to purify separate Asx complexes from the PREs of different homeotic gene targets and determine i f there are different specific constituents at each target. This would be one way to identify possible factors involved in target specific activity. The purified proteins could be used to raise antibodies and then screen an expression library to identify new genes 115 involved in the regulation of specific targets. It would be interesting to see i f z40 was present only at specific targets. The tissue specific requirements for Asx activity suggests that Asx may interact with tissue specific enhancer elements at target loci. A n enhancer element that has a C N S specific Ubx expression pattern was identified (Christen and Bienz 1992) but no other tissue specific enhancers are known. Thus another important use of mapping Asx containing complexes would be to see i f they map to any specific regulatory sites that do not contain other P c G protein complexes. This may identify tissue specific elements and purification of any bound constituents may reveal what factors account for the tissue specific activity of Asx. C o n c l u s i o n Traditional models of PcG function have depended largely on the fact that P c G proteins were distinct components of repressive complexes and that trxG proteins were distinct components of activation complexes. Both genetic and molecular data now suggest that this simple view is inaccurate and that repression and activation may require some level of cooperative activity between the two sets of proteins. 116 References Adler, P. N., Charlton, J. , and Brunk, B. (1989). Genetic interactions of the Suppressor 2 of zeste region genes. Dev. Genet. 10, 249-260. Akam, M . E. (1987). The molecular basis for metameric pattern in the Drosophila embryo. Development 101, 1-22. Beachy, P. A., Helfand, S. L . , and Hogness, D. S. (1985). Segmental distribution of bifhorax complex proteins during Drosophila development. Nature 313, 545-551. Bender, W., Akam, ML, Karch, F., Beachy, P. A., Peifer, M. , Spierer, P., Lewis, E. B., and Hogness, D. S. (1983). Molecular genetics of the Bithorax complex in Drosophila melanogaster. Science 221, 23-29. Bienz, M . (1992). Molecular mechanisms of determination in Drosophila. Current Op. Cell Biol. 4, 955-961. Bornemann, D., Miller, E. , and Simon, J. (1996). The Drosophila Polycomb group gene Sex comb on midleg (Scm) encodes a zinc finger protein with similarity to polyhomeotic protein. Development 122, 1621-1630. Botas, J. , prado, J . M . d., and Garcia-Bellido, A. (1982). Gene-dose titration analysis in the search of trans-regulatory genes in Drosophila. EMBO J. 1, 307-310. Breen, T. R., and Duncan, I. M. (1986). Maternal expression of genes that regulate the bithorax complex of Drosophila melanogaster. Dev. Biol. 118, 442-456. Buchenau, P., Hodgson, J . , Strutt, H., and Arndt-Jovin, D. J. (1998). The distribution of Polycomb-group proteins during cell division and development in Drosophila embryos: impact on models for silencing. J. Cell. Biol, in press. Busturia, A., and Bienz, M . (1993). Silencers in Abdominal-B, a homeotic Drosophila gene. EMBO J. 12, 1415-1425. Campbell, R. B., Sinclair, D. A. R., Couling, M. , and Brock, H. W . (1995). Genetic interactions and dosage effects of Polycomb group genes of Drosophila. Mol. Gen. Genet. 246, 291-300. Campos-Ortega, J. A., and Hartenstein, V. (1985). "The embryonic development of Drosophila melanogaster." Springer- Verlag, Berlin Heidelberg. Capdevila, M . P., Botas, J. , and Garcia-Bellido, A. (1986). Genetic interactions between the Polycomb locus and the Antennapedia and Bithorax complexes of Drosophila. Roux's Arch. Dev. Biol. 195, 417-432. Carrington, E. A., and Jones, R. S. (1996). The Drosophila Enhancer of zeste gene encodes a chromosomal protein: examination of wild-type and mutant protein distribution. Development 122, 4073-4083. Carroll, S. B., Laymon, R. A., McCutcheon, M . A., Riley, P. D., and Scott, M . P. (1986). The localization and regulation of Antennapedia expression in Drosophila embryos. Cell 47, 113-122. 117 Chan, C.-S., Rastelli, L . , and Pirotta, V. (1994). A Polycomb response element in the Ubx gene that determines an epigenetically inherited state of repression. EMBO J. 13, 2553-2564. Cheng, N. N., Sinclair, D. A. R., Campbell, R. B., and Brock, H. W. (1994). Interactions of polyhomeotic with Polycomb group genes oi Drosophila melanogaster. Genetics 138, 1151-1162. Chiang, A., O'connor, M . B., Paro, R., Simon, J. , and Bender, W. (1995) . Discrete Polycomb-binding sites in each parasegmental domain of the bithorax complex. Development 121, 1681-1689. Chinwalla, V., Jane, E . P., and Harte, P. J . (1995). The Drosophila trithorax protein binds to specific chromosomal sites and is co-localized with Polycomb at many sites. EMBO J. 14, 2056-2065. Christen, B., and Bienz, M . (1992). A c/s-element mediating Ultrabithorax autoregulation in the central nervous system. Mech. Dev. 39, 73-80. DeCamillis, M. , and Brock, H. W. (1994). Expression of the polyhomeotic locus in development of Drosophila melanogaster. Roux's Arch. Dev. Biol. 203, 429-438. Duboule, D., and Dolle, P. (1989). The structural and functional organization of the murine H O X gene family resembles that of Drosophila homeotic genes. EMBO J. 8, 1497-1505. Duncan, I. (1987). The bithorax complex. Ann. Rev. Genet. 21, 285-319. Duncan, I. M . (1982). Poly comblike: a gene that appears to be required for the normal expression of the bithorax and Antennapedia complexes of Drosophila melanogaster. Genetics 102, 49-70. Duncan, I. M. , and Lewis, E . B. (1982). Genetic control of body segment differentiation in Drosophila. In "Developmental order: its origin and regulation" (S. Subtelny and P. B . Green, Eds.). Alan R. Liss, New York. Dura, J.-M., Brock, H. W., and Santamaria, P. (1985). Polyhomeotic: a gene in Drosophila melanogaster required for correct expression of segment identity. Mol. Gen. Genet. 198, 213-220. Dura, J.-M., and Ingham, P. W. (1988b). Tissue- and stage-specific control of homeotic and segmentation gene expression in Drosophila embryos by the polyhomeotic gene. Development 103, 733-741. Dura, J.-M., and Ingham, P. W. (1988a). Tissue- and stage-specific control of homeotic and segmentation gene expression in Drosophila embryos by the polyhomeotic gene. Development 103, 733-741. Fauvarque, M. -M. , and Dura, J.-M. (1993). polyhomeotic regulatory sequences induce developmental regulator-dependent variegation and targeted P-element insertions in Drosophila. Genes Dev. 7, 1508-1520. 118 Franke, A., DeCamiUis, M . A., Zink, D., Cheng, N., H.W. Brock, and Paro, R. (1992). Polycomb and polyhomeotic are constituents of a multimeric protein complex in chromatin of Drosophila melanogaster. EMBO J. 11, 2941-2950. Gindhart, J. G., and Kaufman, T. C. (1995). Identification of Polycomb and trithorax group responsive elements in the regulatory region of the Drosophila homeotic gene Sex combs reduced. Genetics 139, 797-814-Glicksman, M . A., and Brower, D. L. (1988a). Misregulation of homeotic gene expression in Drosophila larvae resulting from mutations at the extra sex combs locus. Dev. Biol. 126, 219-227. Glicksman, M . A., and Brower, D. L. (1988b). Misregulation of homeotic gene expression in Drosophila larvae resulting from mutations at the extra sex combs locus. Dev. Biol. 126, 219-227. Glicksman, M . A., and Brower, D. L . (1990a). Persistent ectopic expression of Drosophila homeotic genes resulting from maternal deficiency of the extra sex combs gene product. Dev. Biol. 142, 422-431. Glicksman, M . A., and Brower, D. L. (1990b). Persistent ectopic expression of Drosophila homeotic genes resulting from maternal deficiency of the extra sex combs gene product. Dev. Biol. 142, 422-431. Graham, A., Papalopulu, N., and Krumlauf, R. (1989). The murine and Drosophila homeobox gene complexes have common features of organization and expression. Cell 57, 367-378. Gutjahr, T., Frei, E . , Spicer, C , Baumgartner, S., White, R. A. EL, and Noll, M . (1995). The Polycomb-group gene, extra sex combs encodes a nuclear member of the W D - 4 0 repeat family. EMBO J. 14, 4296-4306. Hagstrom, K., Muller, M. , and Schedl, P. (1996). Fab-7 functions as a chromatin domain boundary to ensure proper segment specification by the Drosophila bithorax complex. Genes Dev. 10, 3202-3215. Hagstrom, K., Muller, M. , and Schedl, P. (1997). A Polycomb and GAGA dependant silencer adjoins the Fab-7 boundary in the Drosophila bithorax complex. Genetics 146, 1365-1380. Hannah-Alava, A. (1958). Developmental genetics of the posterior legs in Drosophila melanogaster. Genetics 43, 878-905. Henikoff, S. (1990). Position-effect variegation after sixty years. Trends Genet. 6, 422-426. Hodgson, J., and Brock, H. (1998). Site specific interaction of Polyhomeotic with subelements of bithoraxoid mediates maintenance of parasegmental silencing of Ubx. Submitted. Ingham, P., and Whittle, R. (1980). Trithorax: a new homeotic mutation of Drosophila melanogaster causing transformations of abdominla and thoracic imaginal segments. Mol. Gen. Genet. 179, 607-614. 119 Ingham, P. W. (1983). Differential expression of bithorax complex genes in the absence of the extra sex combs and trithorax genes. Nature 306, 591-593. Ingham, P. W. (1984). A gene that regulates the bithorax complex differently in larval and adult cells oi Drosophila. Cell 37, 815-823. Ingham, P. W. (1988). The molecular genetics of embryonic pattern formation in Drosophila. Nature 335, 25-31. Jones, R. S., and Gelbart, W. M. (1993). The Drosophila Polycomb-group gene Enhancer ofzeste shares a domain of sequence similarity with trithorax. Mol. Cell. Biol. 13, 6357-66. Jurgens, G. (1985). A group of genes controlling the spatial expression of the bithorax complex in Drosophila. Nature 316, 153-155. Jurgens, G., Wieschaus, E . , Nusslein-Volhard, C., and Kluding, H. (1984). Mutations affecting the pattern of the larval cuticle in Drosophila melanogaster. Roux's Arch. Dev. Biol. 193, 283-295. Kassis, J. A. (1994). Unusual properties of regulatory D N A from the Drosophila engraile d gene: three "pairing-sensitive" sites within a 1.6kb region. Genetics 136, 1025-1038. Kennison, J. A. (1993). Transcriptional activation of Drosophila homeotic genes from distant regulatory elements. Trends Genet. 9, 75-79. Kennison, J . A., and Russell, M . A. (1987). Dosage-dependent modifiers of homeotic mutations in Drosophila melanogaster. Genetics 116,75-86. Kennison, J . A., and Tamkun, J . W. (1988). Dosage-dependent modifiers of Polycomb and Antennapedia mutations in Drosophila. Proc. Natl. Acad. Sci. USA 85, 8136-8140. Krumlauf, R. (1994). Hox genes in vertebrate development. CelllS, 191-201. Kyba, M. , and Brock, H. W. (1998). The Drosophila polycomb group protein Psc contacts ph and Pc through specific conserved domains. Mol. Cell. Biol, in press. Kyba, M . , Milne, T. A., Jane, E . P., O'Grady, M. , Dorn, R., Reuter, G., Grigliatti, T. A., Harte, P. J. , and Brock, H. W. (1998). A s x binds the trx S E T domain and mediates activation and repression of homeotic loci, submitted. Lajeunesse, D., and Shearn, A. (1996). E(z): a polycomb group gene or a trithorax group gene? Development 122, 2189-2197. Landecker, H. L . , Sinclair, D. A. R., and Brock, H. W. (1994). A screen for enhancers of Polycomb and Polycomblike in Drosophila melanogaster. Dev. Genet. 15, 425-434. Lawrence, P. A. (1992). "The making of a fly." Blackwell Scientific Publications, London. 120 Lewis, E. B. (1978). A gene complex controlling segmentation in Drosophila. Nature 276, 565-570. Lindsley, D. L. , and Zimm, G. G. (1992). "The genome of Drosophila melanogaster." Academic Press, San Diego New York . Lonie, A., D'Andrea, R., Paro, R., and Saint, R. (1994). Molecu la r characterisation of the Polycomblike gene of Drosophila melanogaster, a trans-acting negative regulator of homeotic gene expression. Development 120, 2629-2636. Martin, E . C , and Adler, P. N. (1993). The Polycomb group gene Posterior sex combs encodes a chromosomal protein. Development 111, 641-655. McCall, K., and Bender, W. (1996). Probes for chromatin accessibility in the Drosophila bithorax complex respond differently to Polycomb-mediated repression. EMBO J. 15, 569-580. McGinnis, W., and Krumlauf, R. (1992). Homeobox genes and axial patterning. Cell 68, 283-302. McKeon, J. , and Brock, H. W. (1991a). Interactions of the Polycomb group of genes with homeotic loci of Drosophila. Roux's Arch. Dev. Biol. 199, 387-396. McKeon, J. , and Brock, H. W. (1991b). Interactions of the Polycomb group of genes with homeotic loci of Drosophila. Roux's Arch. Dev. Biol. 199, 387-396. McKeon, J . , Slade, E . , Sinclair, D. A. R., Cheng, N., Couling, M . , and Brock, H. W. (1994). Mutations in some Polycomb group genes of Drosophila interfere with regulation of segmentation genes. Mol. Gen. Genet. 244, 474-483. Moazed, D., and O'Farrell, P. H. (1992). Maintenance of the engrailed expression pattern by Polycomb group genes in Drosophila. Development 116, 805-810. Muller, J . (1995). Transcriptional silencing by the Polycomb protein in Drosophila embryos. EMBO J. 14, 1209-1220. Muller, J. , and Bienz, M . (1991). Long range repression conferring boundaries of Ultrabithorax expression in the Drosophila embryo. EMBO J. 10, 3147-3156. Nusslein-Volhard, C , Weischaus, E. , and Kluding, H. (1984). Mutations affecting the pattern of the larval cuticle in Drosophila melanogaster. Roux's Arch. Dev. Biol. 193, 267-282. Paro, R. (1990). Imprinting a determined state into the chromatin of Drosophila. Trends Genet. 6, 416-421. Paro, R., and Zink, B. (1992). The Polycomb gene is differentially regulated during oogenesis and embryogensis of Drosophila melanogaster. Mech. Dev. 40, 37-46. Pelegri, F., and Lehman, R. (1994). A role of polycomb group genes in the regulation of gap gene expression in Drosophila. Genetics 136, 1341-1353. 121 Peterson, A., Kyba, M. , Bornemann, D., Morgan, K., Brock, H. , and Simon, J. (1997). A domain shared by the Polycomb group proteins Scm and ph mediates heterotypic and homotypic interactions. Mol. Cell. Biol. 17, 6683-6692. Pirrotta, V. (1995). Chromatin complexes regulating gene expression in Drosophila. Curr. Opin. Genet. Dev. 5, 466-472. Pirrotta, V., Chan, C. S., McCabe, D., and Qian, S. (1995). Dist inct parasegmental and imaginal enhancers and the establishment of the expression pattern of the Ubx gene. Genetics 141, 1439-1450. Rastelli, L . , Chan, C. S., and Pirotta, V. (1993). Related chromosome binding sites for zeste, suppressor of zeste and Polycomb group protein and their dependence on Enhancer of zeste function. EMBO J. 12, 1513-1522. Riley, P. D., Carroll, S. B., and Scott, M . P. (1987a). The expression and regulation of sex combs reduced protein in Drosophila embryos. Genes Dev. 1,716-730. Riley, P. D., Carroll, S. B., and Scott, M . P. (1987b). The expression and regulation of sex combs reduced protein in Drosophila embryos. Genes Dev. 1, 716-730. Sanchez-Herrero, E . , Vernos, I., Marco, R., and Morata, G . (1985). Genetic organization of Drosophila bithorax complex. Nature 313,108-113. Sato, T., Russell, M . A., and Denell, R. E. (1983). Homoeosis in Drosophila: a new enhancer of Polycomb and related homoeotic mutations. Genetics 105, 357-370. Schlossherr, J . , Eggert, H. , Paro, R., Cremer, S., and Jack, R. S. (1994) . Gene inactivation in Drosophila mediated by the Polycomb gene product or by position-effect variegation does not involve major changes in the accessibility of the chromatin fibre. Mol. Gen. Genet. 243, 453-462. Serrano, N., Brock, H. W., Demeret, C , Dura, J . -M. , Randsholt, N. B., Kornberg, T. B., and Maschat, F. (1995). polyhomeotic appears to be a target of Engrailed regulation in Drosophila. Development 121, 1691-1703. Shimell, M . J . , Simon, J . , Bender, W., and O'Connor, M . B. (1994) . Enhancer point mutation results in a homeotic transformation in Drosophila. Science 264, 968-971. Simon, J. (1995). Locking in stable states of gene expression: transcriptional control during Drosophila development. Curr. Opin. Cell Biol. 7, 376-385. Simon, J. , Chiang, A., and Bender, W. (1992b). Ten different Polycomb group genes are required for spatial control of the abd-A and Abd-B homeotic products. Development 114, 493-505. Simon, J. , Chiang, A., and Bender, W. (1992a). Ten different Polycomb group genes are required for spatial control of the abd-A and Abd-B homeotic products. Development 114, 493-505. Simon, J . , Chiang, A. C , Bender, W., Shimell, M . J . , and O'connor, M . (1993). Elements of the Drosophila bithorax complex that mediate repression by Polycomb group products. Dev. Biol. 158, 131-144. 122 Simon, J. , Peifer, M. , Bender, W., and O'Connor, M . (1990). Regulatory elements of the bithorax complex that control expression along the anterior - posterior axis. EMBO J. 9, 3945-3956. Sinclair, D. A. R., Campbell, R. B., Nicholls, F., Slade, E . , and Brock, H. W. (1992). Genetic analysis of the Additional sex combs locus of Drosophila melanogaster. Genetics 130, 817-825. Sinclair, D. A. R., Clegg, N. J. , Antonchuk, J . , Milne, T. A., Stankunas, K., Ruse, C , Grigliatti, T. A., Kassis, J. A . , and Brock, H. W. (1998a) . Enhancer of Polycomb is a suppressor of position-effect variegation in Drosophila melanogaster. Genetics 148, 211-220. Sinclair, D. A. R., Milne, T. A., Hodgson, J. W., Shellard, J. , Salinas, C. A., Kyba, M. , Randazzo, F., and Brock, H. W. (1998b). The Additional sex combs gene of Drosophila encodes a chromatin protein that binds to shared and unique Polycomb group sites on polytene chromosomes. Development 125, 1207-1216. Slifer, E. H. (1942). A mutant stock of Drosophila with extra sex-combs. J. Exp. Zool. 90, 31-40. Soto, M . C , Chou, T.-B., and Bender, W. (1995). Comparison of germ-line mosaics of genes in the Polycomb group of Drosophila melanogaster. Genetics 140,231-243. Stassen, M . J . , Bailey, D., Fox, S., Chinwalla, V., and Harte, P. J. (1995). The Drosophila trithorax proteins contain a novel variant of the nuclear receptor type D N A binding domain and an ancient conserved motif found in other chromosomal proteins. Mech. Dev. 52, 209-223. Struhl, G. (1981). A homeotic mutation transforming leg to antenna in Drosophila. Nature 292, 635-638. Struhl, G. (1982). Genes controlling segmental specification in the Drosophila thorax. Proc. Natl. Acad. Sci. USA 79, 7380-7384. Struhl, G., and Akam, M . E. (1985a). Altered distribution of Ultrabithorax transcripts in extra sex combs mutant embryos of Drosophila. EMBO J. 4, 3259-3264. Struhl, G., and Akam, M . E. (1985b). Altered distribution of Ultrabithorax transcripts in extra sex combs mutant embryos of Drosophila. EMBO J. 4, 3259-3264. Strutt, H., Cavalli, G., and Paro, R. (1997). Co-localization of Polycomb protein and G A G A factor on regulatory elements responsible for the maintenance of homeotic gene expression. Embo J. 16, 3621-3632. Strutt, H., and Paro, R. (1997). The polycomb group protein complex of Drosophila melanogaster has different compositions at different target genes. Mol. Cell. Biol. 17, 6773-6783. Tower, J . , Karpen, G. H. , Craig, N., and Spradling, A. C. (1993) . Preferential transposition of Drosophila P elements to nearby chromosomal sites. Genetics 133, 347-359. 123 Tsukiyama, T., Becker, P. B., and Wu, C. (1994). ATP-dependent nucleosome disruption at a heat-shock promoter mediated by binding of G A G A transcription factor. Nature 367, 525-532. Wedeen, C., Harding, K., and Levine, M . (1986a). Spatial regulation of Antennapedia and bithorax gene expression by the Polycomb locus in Drosophila. Cell 44, 739-748. Wedeen, C., Harding, K., and Levine, M . (1986b). Spatial regulation of Antennapedia and bithorax gene expression by the Polycomb locus in Drosophila. Cell 44, 739-748. Yamamoto, Y., Girard, F., Bello, B., Affolter, M. , and Gehring, W. (1997). The cramped gene of Drosophila is a member of the Polycomb-group, and interacts with mus209, the gene encoding Proliferating Cel l Nuclear Antigen. Development 124, 3385-3394. Zhang, P., and Spradling, A. C. (1993). Efficient and dispersed local P element transposition from Drosophila females. Genetics 133, 361-373. Zink, B., Engstrom, Y., Gehring, W., and Paro, R. (1991). Direct interaction of the Polycomb protein with Antennapedia regulatory sequences in polytene chromosomes of Drosophila melanogaster. EMBO J. 10, 153-162. Zink, D . , and Paro, R. (1995). Drosophila Polycomb-group regulated chromatin inhibits the accessibility of a ?rans-activator to its target D N A . EMBO J. 14,5660-5671. 124 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0088433/manifest

Comment

Related Items