UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

An analysis of polycomb group protein interactions Kyba, Michael Stephen 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_1998-345734.pdf [ 7.4MB ]
Metadata
JSON: 831-1.0088980.json
JSON-LD: 831-1.0088980-ld.json
RDF/XML (Pretty): 831-1.0088980-rdf.xml
RDF/JSON: 831-1.0088980-rdf.json
Turtle: 831-1.0088980-turtle.txt
N-Triples: 831-1.0088980-rdf-ntriples.txt
Original Record: 831-1.0088980-source.json
Full Text
831-1.0088980-fulltext.txt
Citation
831-1.0088980.ris

Full Text

A N ANALYSIS OF POLYCOMB GROUP PROTEIN INTERACTIONS by MICHAEL STEPHEN K Y B A B.Sc, The University of Alberta, 1991 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF T H E REQUIREMENTS FOR T H E D E G R E E OF DOCTOR OF PHILOSOPHY in T H E F A C U L T Y OF G R A D U A T E STUDIES (Department of Zoology) We accept this thesis as conforming to the required standard  T H E UNIVERSITY OF BRITISH C O L U M B I A April 1998  © Michael Stephen Kyba, 1998  In presenting this thesis degree  at the  in partial fulfilment  of the  requirements  University of British Columbia, I agree that the  for an advanced  Library shall make it  freely available for reference and study, i further agree that permission for extensive copying of this thesis for scholarly purposes department  or  by  his  or  her  representatives.  may be granted It  is  by the head of my  understood  that  copying  or  publication of this thesis for financial gain shall not be allowed without my written permission.  Department The University of British Columbia Vancouver, Canada  DE-6 (2/88)  Abstract  The Polycomb Group (PcG) of proteins are global regulators of transcription. PcG mutants display posterior homeotic transformations, the result of ectopic expression of homeotic selector genes of the Bithorax and Antennapedia Complex, demonstrating that the PcG is required for the repression of target genes outside of their normal spatial boundaries of expression. Coimmunoprecipitation, cofractionation, and colocation on larval salivary gland chromosomes suggest that PcG proteins act through large multimeric complexes formed at their target sites. This thesis is a characterization of the protein interactions that underlie multimeric complex formation. Using the yeast two-hybrid system and an in vitro co-affinity precipitation assay, I demonstrate direct interactions between Polycomb (Pc) and Posterior Sex Combs (Psc), and between Psc and polyhomeotic (ph). I also show that Psc, ph, and Asx have self-interacting domains, and perform a detailed analysis of the selfinteracting domain of ph. For the most part, these interacting domains are highly conserved between the Drosophila proteins and their mammalian counterparts. Because Asx shows no direct interact interactions with Pc, Psc, or ph, I screen Asx for interacting proteins within a two-hybrid library and within a two-hybrid panel of other chromatin proteins. Several interactors are identified, including the Drosophila homologue of cyclin G, and z40, a previously unknown protein which interacts strongly with Pc. In addition, an interaction is demonstrated between the respective carboxyl termini of Asx and trithorax (trx), a protein required for activation of homeotic selector genes. I show that Psc can repress transcription in Saccharomyces cerevisiae, and show that this repression does not require interactions with a variety of yeast proteins required for repression of various loci in the S. cerevisiae genome. These data enlarge our understanding of the structure of PcG complexes, and suggest that PcG proteins interact with one another promiscuously, enabling them, in theory, to form a large number of different complexes each tailored to a particular chromosomal neighbourhood.  ii  Table of Contents  Abstract  ii  List of Figures  vii  Acknowledgments  ix  Introduction  1  The Polycomb Group  2  Cis Elements Recognized by the PcG  4  The Trithorax Group  5  The Mechanism of PcG Silencing or Repression  6  Other Silencing and Repression Systems  8  Chapter I: Interacting Domains of Asx, Pc, Psc, and ph.  12  Sequence Motifs Present in Asx, ph, Psc, and Pc.  13  Co-immunoprecipitation of Pc, ph, and Psc  14  Two-Hybrid Interactions  19  In-Vitro Interactions  27  Ternary vs. Binary Complexes  31  Isolated Domain Interactions can be Modulated by External Sequences  32  Interactions of the Vertebrate Homologues of Psc, Pc, and ph  35  The Role of Multiple Interacting Domains in PcG Complexes  36  iii  Table of Contents, cont.  Chapter H : The S A M Domain.  38  The S A M Domain  38  The ph S A M Domain Mediates Self-Association In Vitro  39  Homologues of the ph S A M Domain  42  Homotypic and Heterotypic Self-Association of ph S A M Homologues  43  Mutations in Conserved Residues Have Different Effects on Binding to Different SAMs  49  SAM: A Self Association Motif  49  Key Structural and Functional Residues of the S A M Domain  52  The Potential for Promiscuous Oligomerization  53  Chapter HI: Asx Interactions  55  A Conserved Domain in the C-terminus of Asx  55  The Asx Two Hybrid Library Screen  58  Interactions with Other PcG Proteins  60  Evaluation of AsxC Interacting Proteins  60  Possible Roles for Bowel and Cabeza  62  The Complete Sequences of z34 and z40  63  z34 Contains a Cyclin Box  63  An Asx-Cyclin G interaction?  73  The Asx-z40 Interaction  76  Asx Interacts with the trx SET Domain  76  The Transcriptional Consequence of an Asx-trx Interaction  82  iv  Table of Contents, cont.  Chapter TV: PcG Functional Interactions in Yeast  85  Pc, Psc, and ph Can Repress Transcription Directly  85  Psc is a Transcriptional Repressor In Yeast  88  Telomeric Effects of the PcG and trxG in Yeast  94  Implications for the Mechanism of PcG Action  98  Chapter V: Materials and Methods  99  Subcloning  99  Mutagenesis  102  Primers  103  Sequencing  105  GST-fusion protein expression and purification  106  In vitro coprecipitations  106  Binding to Ni-NTA agarose  107  Transformation and Culturing of Yeast Strains  108  Yeast Strains  108  Two-Hybrid Interaction Assays  109  [3-galactosidase assays  110  Co-immunoprecipitation from Kc nuclear extracts  110  SL2 Transfections  111  C A T assay  111  Telomeric Variegation Assays  112  v  Table of Contents, cont.  Chapter VI: Conclusion  113  The Assembly of PcG Complexes  113  Dimerization  116  Silencing or Repression?  119  The Large Membership of the PcG  120  Nomenclature  124  Bibliography  127  Appendix A: Interactor  Preliminary Sequences  142  Appendix B: Interactor  Sequence Comparisons  151  vi  List of Figures  Chapter I Figure 1.1  Sequence motifs of the four PcG proteins studied.  15  Figure 1.2  Pc, ph, and Psc proteins coimmunoprecipitate.  17  Figure 1.3  Two-hybrid interaction assay results for ph, Psc, Pc, and carboxyl deletions.  21  Figure 1.4  Asx interactions.  23  Figure 1.5  Two-hybrid interaction assay results for conserved sequence constructs.  Figure 1.6  25  In vitro binding of reticulocyte lysate-generated S-labeled 35  Psc constructs to bacterially produced GST-fusions. 29 Figure 1.7  Domains involved in interactions between Pc, ph, and Psc.  33  Figure 2.1  Self-binding activity of the carboxyl terminus of ph.  40  Figure 2.2  Alignment of relatives of the ph S A M domain.  44  Figure 2.3  Homotypic and heterotypic binding of various  Chapter II  S A M domains. Figure 2.4  46  Mutations in pSAM and their effects on binding.  50  Figure 3.1.  hAsx sequence.  56  Figure 3.2  Sequence of the z34 cDNA.  64  Figure 3.3  Sequence of the z40 cDNA.  67  Figure 3.4  z34 Sequence Alignments.  69  Figure 3.5  Asx-trx interactions.  79  Chapter III  vn  Figures, cont.  Chapter IV Figure 4.1  Transcriptional repression by PcG proteins in transiently transfected SL2 cells.  86  Figure 4.2  Transcriptional repression by Lex-Psc.  Figure 4.3  Transcriptional repression by Lex-Psc in various mutant backgrounds.  Figure 4.4  89  91  Telomeric silencing effects of trx sequences.  96  Figure 6.1  A model for regulated PcG complex assembly.  114  Figure 6.2  Self-association of PcG proteins may facilitate inter-  Conclusion  homologue interactions Figure 6.3  Different loci require different PcG proteins for silencing  viii  117 121  Acknowledgments  Thanks are due to many people:  Dr. Hugh Brock, for allowing me great freedom throughout his supervision of this work.  Members of the Brock Lab, especially Dr. Jacob Hodgson, for much good advice, many stimulating discussions and encouragement.  Wilfred Lim for help with the two-hybrid library screen.  Michael O'Grady for his two-hybrid panel of chromatin proteins.  Erica Golemis, Fred Winston, Jasper Rine, Ira Herskowitz, Carrie Brachmann, Randall Mann, and David Stillman for providing yeast strains, Paul Adler, Rick Jones, Yasushi Matsui, Renato Paro, Ivan Sadowski, Rob Saint, Jeff Simon, Kazunori Shimada, and Don Sinclair for providing cDNAs, and Paul Adler and Jacob Hodgson for providing antibodies. i  The National Science and Engineering Research Council of Canada for funding part of this work.  Mum and Dad for putting food on the table when my pockets were empty, and for their great moral and spiritual support.  ix  Introduction  The central dogma of molecular biology postulates that information flows from DNA to RNA to protein. The relative rate at which different segments of information (genes) are transcribed through the first step in this progression is a critical element, sometimes the only necessary element, in defining a given biological state. Likewise, rate changes are linked to biological state changes: ontogeny, differentiation, response to external stimuli, and neoplasia, to name some of the more important ones. The understanding of transcriptional regulation, then, is of crucial importance to the understanding of biological phenomena.  If one considers the rate of transcription to be governed by a balance of positive and negative factors, then to date our understanding of the positive factors far outstrips our understanding of the negative factors. This is partly due to the relative diversity of positive factors, partly due to the existence of lower thresholds that can be crossed by the removal of positive factors by mutation, but a relative paucity of upper thresholds that can be crossed by the removal of negative factors, and partly due to the logical requirement for a functioning in vitro system of transcription to exist (the functioning of which presumes the contribution of certain positive factors) before one can study negative transcriptional factors in vitro. This biased apprehension has traditionally contributed to a view of transcription that overvalues the role of positive factors, and ignores or undervalues the role of negative factors. However, recent advances have begun to bring the role of negative factors into focus. This thesis addresses the less-charted negative side of transcription through a molecular study of one group of negative transcriptional regulators, the Polycomb group (PcG).  1  The Polycomb Group Restricting the spatial and temporal expression of the transcription factors that control development is a central part of the mechanism by which pattern and structure come into being in a developing organism. In Drosophila, cellular identity along the anteriorposterior (AP) axis is specified by the homeotic selector genes of the Antennapedia and Bithorax Complexes (AntC, and BxC, respectively) [1; 2]. Particular patterns of homeotic gene expression are thought to define particular positions along the AP axis, and changes in these expression levels in a given axial segment cause corresponding changes of fate for that segment. The domains of expression of the homeotic selector genes are set up by gap and pair-rule genes which act early in development [3]. The proteins of the PcG maintain the fidelity of these domains of expression through later divisions when the early regulators are no longer present [4].  The PcG was originally identified in Drosophila as a group of genes whose mutation causes multiple homeotic transformations similar to gain-of-function mutations of the homeotic selector genes. This similarity is the result of derepression of the homeotic genes outside of their proper spatial boundaries [1; 5-8]. Because derepression in PcG mutants characteristically occurs between 5 and 6 hours of development, before which time selector gene expression is normal in both degree and pattern, the PcG is thought to be required for maintenance of the repressed state, but not for initiation of the repression [8-10].  There are 14 genetically characterized members of the PcG: Polycomb (Pc) [1], extra sex combs (esc) [11], Polycomblike (Pel) [12], Enhancer of Polycomb (E(Pc)) [13], super sex combs (sxc) [14], polyhomeotic (ph) [15], Sex combs on midleg (Scm) [16] Sex combs extra (See) [17], Posterior sex combs (Psc) [IS], Additional sex combs (Asx) [19], Suppressor two ofzeste (Su(z)2) [20], Enhancer of zeste (E(z)) [10], pleiohomeotic (pho) [21], multi sex combs (mxc) [22], and cramped (crm) [23]. Each member of this group is  2  required for repression of one or more of the homeotic selector genes outside of their proper domains of expression (although in the case of Su(z)2 and E(Pc) this requirement is minimal except in a genetically sensitized background, see below). The other hallmark of PcG mutants is that they display dominant enhancement [16; 20; 24; 25], and in some cases antipodal suppression [24; 25 ; 26; 27] of each other's phenotypes, implying that the function of the group as a whole is sensitive to the dosages of its members [16; 28]. However such interactions are not seen for every mutant combination, nor for every phenotype, and in many cases interactions are allele-specific [25]. With respect to this thesis, some details of these genetic interactions are worth pointing out. When all possible PcG (with the exception of ph) double heterozygotes were generated, Asx, Pc, Pel, Psc, See and Sem enhanced each other's adult homeotic phenotypes in every pairwise combination (with the exception of Psc/Asx and Psc/Sce) [25]. Double heterozygous combinations of ph , 503  a null, with Pc , Pc , or Su(z)2 were lethal, with Psc strongly 4  16  1  1  semilethal, while with other PcG genes, non lethal [24]. Combinations hemizygous for ph ®9, a strong hypomorph, and heterozygous for Pc , E(z) , Psc , Sem, and See, were 4  1  1  lethal, while other PcG genes did not interact lethally with ph  1  409  [24]. Thus ph, Pc, and  Psc form a strongly interacting leash within a larger set of mutually interacting elements, which includes Pel, Sem, See, and Asx. Other PcG genes tend to interact more sporadically with members of this set and with each other.  A model that explains this genetic synergism postulates that the PcG acts as a multimeric protein complex with phenotypic enhancement being the result of increased perturbation of the complex with an increased number of mutant members [28]. All PcG proteins that have been tested immunohistochemically are present at specific sites along polytene chromosomes. Pc, Pel, and ph bind the same approximately 100 sites [29; 30]. E(z) shares most of these sites but binds to unique sites of its own [31], as do Psc and Su(z)2 [32; 33]. Colocation on polytene chromosomes, co-immunoprecipitation (co-IP) of ph and  3  Pc [29], as well as the presence of approximately 10 unidentified proteins in the Pc immunoprecipitate [29] support the idea that large, multimeric PcG protein complexes reside at various sites in the chromatin of Drosophila. However the fact that many loci stain for some members but not others, as well as the fact that different PcG mutants display different levels of selector gene derepression [8] suggests that if a complex exists it must be heterogeneous.  Mammalian Hox gene expression boundaries appear to be maintained through a mechanism similar to that seen in Drosophila, via mammalian homologues of PcG proteins. Targeted gene replacements of the Psc homologues Bmi-1 [34] and Mel-18 [35], the ph homologue rae28 [36], and the Pc homologue M33 [37] cause posterior transformations of the axial skeleton, due to the anterior shift of several Hox gene expression boundaries. Overexpression of Mel-18 in transgenic mice confers the opposite phenotype [38]. The surprising result that an M33 transgene partially rescues a Pc mutation in Drosophila demonstrates that there has been remarkable conservation of the mechanism of PcG function between flies and mammals [39].  Cis Elements Recognized by the P c G D N A sequences that restrict or reduce the expression of reporter genes in a PcG-dependent manner (PREs, for PcG Response Elements) have been found in the regulatory regions of the homeotics [40-44] as well as upstream of ph [45] and engrailed (en) [46]. Although D N A crosslinking experiments suggest that PcG proteins interact with D N A a few kb upstream and downstream in addition to the PRE [47], the vast distances between regulatory elements of the homeotics evoke an image of PcG complexes binding discretely to a few sites along the BxC and AntC. It was noticed that expression of the white gene, an eye colour marker used to score for P-element-mediated transformation, was suppressed when PRE-containing transgenes were made homozygous, and that this suppression was  4  dependent on the PcG [43; 45; 46] Such pairing-sensitive repression suggests that the PREs on homologous chromosomes interact with each other via PcG protein interactions.  By juxtaposing PREs with other cis elements, it has been shown that a PRE can block nearby binding of other DNA-binding proteins, and that this blockage is competitive, being overcome by higher expression levels of the protein being blocked [48]. Other experiments have shown that insulating sequences (gypsy and ses) interposed between the PRE and a reporter gene can block silencing [49]. Gypsy elements surrounding a PRE-reporter construct prevent silencing at some insertion sites, provided that the transgene is heterozygous, showing that silencing involves interactions with other sequences, perhaps other PREs in the vicinity, on the homologous chromosome, or even on non-homologous chromosomes [49]. The phenomenon of homing, whereby a transgene carrying a PRE inserts preferentially very near to an endogenous PRE [45], is also suggestive of promiscuous, interlocus PRE interactions.  The Trithorax Group Mutations in a group of 12 other genes (collectively named the trithorax Group, trxG ) suppress the homeotic transformations of the PcG [50]. The trxG is also required for proper expression patterns of the homeotics, and in many ways behaves similarly, although in the opposite direction, to the PcG. trxG mutations cause transformations similar to loss of function mutations of the homeotics [51], the result of reduced expression of the homeotics, with different homeotics showing different sensitivities to a given mutant [52]. Double heterozygotes for some pairs trxG mutants show enhanced homeotic phenotypes relative to single heterozygotes [53].  The trxG protein brahma (brm) contains six blocks of sequence similar to those found in the yeast general transcriptional activator, SWI2, a DNA-dependent ATPase [54]. SWI2 is  5  part of a multiprotein complex (the SWI/SNF complex) that enables transcription by relieving chromatin-mediated repression [55; 56]. Another trxG protein, snrl, is the homologue of SNF5 (also a member of the SWI/SNF complex) and is present with brm in a megadalton complex in Drosophila nuclear extracts [57]. Trithoraxlike (Trl) encodes G A G A factor [58], a protein required for transcriptional activation of many genes, including hsp70 [59]. In vitro, G A G A factor can relieve repression by histone HI [60], and can, in the presence of the ATP-dependent nucleosome remodeling factor (NURF), alter the chromatin structure of the hsp70 promoter [61].  From these examples, it would appear that the function of the trxG is to enable transcription by altering the chromatin structure of a targeted locus. That is likely the case for the examples given above, however the mechanism of action of other trxG proteins is less clear. Interestingly, several trxG proteins share sequence motifs with PcG proteins. E(z), ashl and trx contain SET domains [62]. Pel, ashl and ash2, contain PHD fingers [62; 63]. This sequence homology may mean that certain PcG and trxG proteins compete for the same factors, a scenario that would explain the suppression of PcG haplo insufficient phenotypes by trxG mutations.  The Mechanism of PcG Silencing or Repression Based on sequence similarity between Pc and HP1, a heterochromatin protein [64], Renato Paro has suggested that the PcG may organize the genes that they regulate into compact higher order heterochromatin-like structures that are inaccessible to RNA polymerase II (pol II) and transcriptional activators [65; 66]. Thanks to the trxG findings mentioned above, and a possible link to suppressors of position effect (mentioned below), heterochromatin has figured heavily into current thinking of PcG function. It is worth pointing out that the chromatin state of G A G A factor-, or SWI/SNF-regulated genes is not heterochromatic, even in the absence of transcription. Biochemical evidence of  6  heterochromatic changes induced by PcG proteins has yet to be presented, worse there is an absence of such change at the BxC in Pc mutants [67].  Vincenzo Pirrotta has suggested a model based on cooperative binding, whereby weak individual binding sites are distributed at strategic points along a transcriptional domain, and a PRE, which consists of multiple binding sites clustered together, serves as a nucleation center, looping in the single sites from adjacent DNA only when a complex of sufficient size has assembled at the PRE [68; 69]. The looping would prevent enhancers from mteracting with the promoter of the gene that they regulate. Determination of DNAbinding activity of PcG proteins or interacting factors, and the fine scale mapping of the sites they recognize along the BxC will be necessary to judge the validity of this model.  Other models have been proposed. One PcG protein, esc, has been suggested to interact with and inactivate the basal transcription complex [70], although as yet there is no biochemical evidence for this suggestion. Subnuclear localization to a silent compartment has been suggested [67]. The evidence for this is shaky at present: Pc, ph, and Psc are present at multiple nuclear foci in interphase cells [71; 72]. When overexpressed in transfected cells, the mammalian Psc homologue Bmi-1 is visible at the nuclear periphery [72], however other studies of endogenous Bmi-1 and other mammalian PcG proteins report a speckled distribution [73]. If a silent PcG compartment exists, it is most likely distributed through the nucleus, thus rendering questionable the term "compartment".  Whether a single model can explain the biochemical function of the PcG has not really been questioned. However, early opinion held that at least two genes, Pc and esc, must have independent functions. This was based on the fact that the phenotype of individuals with a complete absence of esc (maternal") was made more severe by reducing the dose of Pc [74], and by the fact that esc was required earlier than Pc [5]. Notwithstanding the genetic  7  interaction studies cited above, this result precludes a single model based on the formation of a complex with a threshold concentration, unless one assumes that there is redundancy built into the complex (i.e. that some members can substitute for one another).  Other Silencing and Repression Systems Drosophila has two other classical systems that monitor transcriptional silencing or repression: Position Effect Variegation (PEV) and transvection. PEV is the stochastic, clonally inherited silencing of a gene that has been brought into proximity to heterochromatin by chromosomal rearrangement. This proximity is thought to convert the transcriptional domain into heterochromatin in some cells and their descendants. Mutations of chromatin modifiers or assembly factors are thought to enhance or suppress the frequency of this incorporation [75; 76]. Tti is an enhancer of variegation (E(var)) [58], while E(Pc) and Asx are suppressors of variegation (Su(var)s) [77]. Extra copies of the human homologue of E(z) are E(var)s [78]. Other PcG genes have not convincingly been shown to affect PEV. These data would lend strength to the idea that at least some PcG proteins are heterochromatin factors, except for the inability to determine whether the modifier effect is direct, or the result of derepression of a true modifier.  Transvection is the pairing-sensitive genetic interaction of homologous loci [79]. It is not a system of repression per se, however transvection-dependent repression of the white gene has been noted in a zeste mutant background. Modifiers of this repression are thought to 1  encode euchromatic chromatin proteins that influence inter homologue interactions. Transvection has been shown to occur at the BxC [80], and interestingly, several PcG genes are modifiers of the zeste-white interaction: Psc, Su(z)2, E(z), and Scm, [81-83]. It is tempting to speculate that the same PcG modifiers of transvection could be involved in crosstalk at the BxC, perhaps between PREs.  8  In the yeast Saccharomyces cerevisiae, two classes of loci, telomeres and the silent mating type loci, are kept transcriptionally inert through a mechanism involving a compact chromatin state. This type of permanent transcriptional inhibition is referred to as silencing, in contradistinction to regulatable transcriptional inhibition, or repression. Silencing at both telomeres and the silent mating type loci requires the action of Sir2, 3, and 4 (Sir, Silent Information Regulator) [84; 85], histones, and assorted other factors [86]. Sir3 and 4 are recruited by sequence-specific factors, and interact with the N-terminal tails of histones H3 and H4 [87], setting up a compact structural configuration that spreads some distance along the chromosome. The stochastic silencing of genes near telomeres is strongly reminiscent of PEV.  The conditional repression of many inducible genes in S. cerevisiae depends on the action of Ssn6 and Tupl. A pentamer containing four molecules of Tupl and one molecule of Ssn6 [88] is recruited by a variety of sequence-specific DNA-binding factors including the cc2 repressor, required for repression of a-specific genes [89], and M i g l , required for catabolite repression [90]. The mechanism by which Ssn6-Tupl represses transcription is unclear, but may involve the positioning of nucleosomes. Deletion of SSN6 or TUP1 disrupts the positioned nucleosomes of the oc2 operator, independently of whether transcription is occurring at the locus [91]. Tupl interacts directly with the N-terminal tails of histones H3 and H4, and these interactions are strongest when the histones are underacetylated [92]. On the other hand, Tupl/Ssn6 repression can be seen in an in vitro system lacking nucleosomes , suggesting an interaction with the basal transcriptional machinery itself [93].  The study of a large number of factors identified through different repression assay systems in yeast has converged on the carboxyl terminal repeat domain of pol II (CTD) as being a target of inducible transcriptional repression. Many members of the protein  9  complex associated with the C T D (named the mediator [94; 95], for mediator of transcriptional activation) can be mutated to confer repression-insensitivity to pol II. The inducible genes in which these repression-defective mutants were originally identified include SUC2 (ssn mutants), HO (sin mutants), Ty insertions (spt mutants), and CYC7 (rox mutants) (reviewed in [96]). Repression in some of these systems also requires Ssn6/Tupl.  Understanding the relationship of the PcG to the well defined transcriptional repression and silencing systems of yeast will be helpful in piecing together the mechanism by which the PcG mediates transcriptional repression. Likewise, understanding the nature of the involvement of the PcG in PEV and transvection ought to shed light on its action at the BxC and AntC. Protein-protein interactions figure centrally in each of the other systems of transcriptional repression or silencing described above. Studying the protein interactions within the PcG then, is relevant not only to the structure of the PcG complex(es) but also to their function. This is especially true given the behaviour of the cis elements with which the PcG proteins interact, namely their PcG-dependent interactions with PREs on homologous chromosomes, and with other cis elements on the same or other chromosomes, and the influence that these interactions have on repression.  A clear understanding of how the PcG proteins carry out their function will ultimately require knowing not only the composition and structure of protein complexes that they form, but how these complexes are assembled, and the factors (DNA, chromatin, transcriptional activators, basal transcription factors, replication factors, or others) with which different members of the PcG interact.  My approach has been two-pronged: the first objective was to uncover and characterize interactions between known and cloned PcG proteins using the yeast two-hybrid system  1 0  and in vitro co-affinity precipitation. The second objective was to discover other interacting factors through a random screen of a two-hybrid library, a rational screen of a panel of chromatin proteins, and by assaying PcG proteins for function in a heterologous environment, and studying the requirement for other factors known play a role in silencing or repression in that heterologous environment.  11  Chapter I: Interacting Domains of Asx, Pc, Psc, and ph.  As a first step towards understanding the structure of PcG complexes, I tested four PcG proteins, Asx, ph, Psc, and Pc, for interactions with each other, and identified domains essential for these interactions. At the initiation of this work, there was a complete lack of knowledge about PcG intra-complex physical interactions. The current knowledge of these interactions in Drosophila is still very sparse. Temperature shift experiments with a temperature-sensitive allele of E(z) suggested that in vivo binding of ph, Psc, and Su(z)2 to most but not all of their sites on salivary gland chromosomes was dependent on E(z) protein [32]. This may mean that E(z) plays a role in targeting or fixing certain PcG complexes to their sites. An E(z)-esc interaction has been detected by Jeff Simon using the two-hybrid system (personal communication). Somewhat more work has been done with vertebrate PcG homologues. A possible ph-Psc interaction is suggested by recent twohybrid experiments with the mouse homologues of these proteins, Mphl and Bmi-1 respectively [73]. The C-terminal 292 amino acids of Mphl interacted with Bmi-1, and a 220 amino acid putative helical domain of Bmi-1 interacted with Mphl. However these two domains were not tested against each other, so it is not known whether they interact with each other or with other parts of their respective proteins. The human homologues of ph, HPH1 and HPH2, were recently cloned using Xenopus Bmi-1 as bait in the two hybrid system [97]. This interaction was delimited to a 295 amino acid C-terminal fragment of HPH2. The conserved amino-terminal 188 amino acids of the Xenopus homologue of Psc, XBmi-1, has been shown to bind to the Xenopus homologue of Pc, XPc, and Xpc has been shown to bind to itself [98]. The mouse homologues of Psc and Su(z)2, Bmi-1 and Mel-18 have been shown to coimmunoprecipitate with the mouse homologue of ph, Mph, and the mouse homologue of Pc, M33 [73], (N. Hashimoto, H.W. Brock, M . Nomura, M . Kyba, J. Hodgson, Y . Fujita, Y . Takihara, K. Shimada, and T. Higashinakagawa, submitted).  12  Sequence Motifs Present in Asx, ph, Psc, and Pc From the point of view of domain analysis, the four proteins studied have several interesting features, ph is a tandemly duplicated gene with the proximal and distal transcription units coding for two nearly identical proteins of 167 and 149 kDa. The proximal ph product has 193 amino-terminal amino acids that are absent from distal ph, and in addition makes use of internal initiation to give an alternate product shorter by 244 amino acids [99]. A notable feature of this unique proximal domain is the presence of a PxxPxxPxxP motif (aa 156-165) with proline spacing the same as that of the polyproline type-II helix recognized by the SH3 domain [100]. ph also has many glutamine repeats and a serine/threonine rich region. Near the carboxyl terminus are two blocks of sequence (aa 1297-1388 and aa 1511-1576) that are shared with the mammalian ph homologues [73; 97; 101]. The first sequence, named HI consists of 28 highly conserved amino acids followed by an unusual C4 zinc finger with intercysteine spacing CX2C...CX3C. The second sequence has been variously referred to as H2 [101] or SEP [73] in the mouse homologue, SPM in the PcG protein Scm [102] as well as in the human ph homologues, HPH1 and HPH2 [97] and S A M in a variety of yeast signal transduction proteins [103]. I have shown that this domain can mediate homotypic and heterotypic self-association between ph and Scm proteins in vitro (Chapter 2). In view of this result, I refer to the domain in general as a Self Association Motif, and keep the acronym S A M , but refer to the specific subset of SAMs with greatest similarity to ph and Scm as SPM. The only internal region of sequence dissimilarity between proximal and distal ph are the 52 amino acids immediately preceding the SPM domain. This work (with the exception of chapter 2) has exclusively used the proximal isoform of ph.  Psc is a 170 kDa protein with several stretches of repeated amino acids. Strong similarity to amino acids 261-467 of Psc has been found in the Drosophila PcG protein Su(z)2 and the mammalian homologues Bmi-1 [104; 105], and Mel-18 [106]. This block of conserved  13  sequence includes a potential C2HC3 ring finger at the amino end and a putative helix-turnhelix (HTH)  motif at the carboxyl. Interestingly, another Drosophila homologue has been  cloned which consists of these two conserved sequences and nothing else [107].  Pc is a 44kDa protein with two histidine repeats and two proline rich regions, the first of which partly overlaps with interspersed glutamine repeats. Amino acids 26-62 of Pc are conserved with HP1, a Drosophila heterochromatin protein, and the mammalian protein M33,  and have been named the 'chromobox'. [64; 108] In addition, Pc and M33 share a  short sequence near their respective carboxyl termini.  Asx is a 182 kDa protein. Like ph, it has extensive glutamine repeats. It also has a run of 20 alanines near the amino terminus. At the extreme carboxyl terminus is a cluster of cysteines with spacing C - X - C - X 7 - C - X 2 - C - X 3 - C - H - X 2 - C - X 6 - C - X 2 - C , which could contain a zinc finger. There are two domains that have a high degree of sequence conservation with mammalian ESTs (expressed sequence tags): the putative zinc finger just mentioned (and described in more detail in Chapter 3), and a stretch of sequence from aa 201-318. This latter sequence similarity was not known at the time that these experiments were initiated.  Co-immunoprecipitation of Pc, ph, and Psc Pc and ph have previously been shown to colocalize on polytene chromosomes and to immunoprecipitate with each other as well as at least 10 unidentified proteins [29]. Given the high level of overlap between the polytene chromosome binding sites of these two proteins with Psc [32; 33] as well as the coimmunoprecipitation of the mammalian homologues of all three proteins [73] (N. Hashimoto, H.W. Brock, M . Nomura, M . Kyba, J. Hodgson, Y . Fujita, Y . Takihara, K. Shimada, and T. Higashinakagawa, submitted), it seemed likely that Psc would complex with ph and Pc in vivo. I therefore performed an immunoprecipitation of a nuclear extract with an antibody to Pc. As shown in Figure 1.2, ph and Psc proteins are both present in the immunoprecipitate of the Pc  14  Sequence Features  PC  chromo | | |Hlfq[|p| SAM  P  h  Q  Q  Q  T/S  ring HTH  Psc Asx  Q  Q  Scale: |-  H lOOOaa  Zinc or Ring Finger  Mammalian Homology S  I  Amino A c i d Repeats  15  Figure 1.1 Sequence motifs of the four PcG proteins studied. Regions of sequence conservation with mammalian homologues are marked black. Zinc-finger and ring-fingers are striped, and regions containing a predominance of a particular amino acid are shaded and labeled with the one letter designation for that amino acid. SAM: self-association motif, chromo: chromobox, ring: ring finger, HTH: helix-turn-helix  16  212— 156114— 97— 66 55  Figure 1.2 Pc, ph, and Psc proteins coimmunoprecipitate. The nuclear extract immunoprecipitate of a Pc antibody and its cognate pre immune serum were electrophoresed in two lanes each, and electrophoretically transferred to a nitrocellulose filter. The filter was then cut into three pieces, each probed with a different antibody. The reconstructed filter is shown where (A) is the part probed with ph antibody, (B) is the part probed with Psc antibody, and (C) is the part probed with Pc antibody. All three proteins are present in the Pc IP, but not in the preimmune IP. The large band at 55 kDa is the IgG heavy chain of the immunoprecipitating antibody, which reacts with the secondary antibodies.  18  antibody, but not present in the immunoprecipitate of the preimmune serum. Asx could not be tested due to the lack of an antibody. Very recently, the two reciprocal immunoprecipitations have been performed by Strutt and Paro [109], who show that the ph immunoprecipitate contains Psc, and the Psc immunoprecipitate contains ph, completing the circle of interactions.  Two-Hybrid  Interactions  To identify potential direct contacts between the ph, Pc, and Psc proteins, I generated DNA-binding and activator fusions to all three proteins and carboxyl deletion derivatives and tested them for interaction in the yeast two hybrid system [110]. All possible pairwise combinations were tested. Shown in Figure 1.3 are the most informative pairs. All pairs not shown were negative. Three interacting combinations were detected: Psc-Pc (Figure 1.3a), ph-ph (Figure 1.3b), and ph-Psc (Figure 1.3d). There were no self-interactions seen for either Pc or Psc (Figure 1.3c). The deletion derivatives locate a ph-ph interacting region in the amino terminal 522 amino acids, although in Chapter 2,1 demonstrate that ph also has a carboxyl terminal self-interacting domain. The ph-Psc interacting domains were mapped to between amino acids 523 and 1418 of ph, and amino acids 205 to 696 of Psc. This interaction occurred with deleted versions of each protein, but not with full length. The Psc-Pc interaction also mapped to amino acids 205 to 696 of Psc, and was similarly not observed with full length Psc.  Three Asx constructs were generated and tested for two-hybrid interactions. An amino construct, AsxA (aa 1-335), a construct from the glutamine-rich central portion of Asx, AsxQ (aa611-1138), and a carboxyl construct, AsxC (aa 1139-1668). All three constructs were tested against the entire previous panel, in both DNA-binding and activator fusion combinations. No interactions were seen between any Asx construct and any ph, Psc, or  19  Pc construct. When the Asx constructs were tested against themselves, the only interaction detected was between with the AsxC construct, which interacted with itself (Figure 1.4).  To better define the ph-Psc-Pc interactions, I generated a set of smaller constructs. Because all of the interactions mapped to areas that contained sequence similarity to mammalian homologues, these sequences alone were tested against each other and against the previous panel of constructs (Figure 1.5).  The smallest fragment of ph to interact with Psc was the HI domain, amino acids 12971418. The minimal Psc element required for the same interaction was the H T H fragment, amino acids 336-473. The minimal domains interacted with each other and are therefore sufficient. DNA-binding fusions to both phHD (amino acids 1297-1576) and the subfragment HI activated transcription alone as assayed by their ability to promote growth on leucine deficient medium in the absence of any other plasmid. It was therefore impossible to test these domains reciprocally. However by using the phAN construct (amino acids 1-1417), which contains the HI domain and does not activate transcription alone, I could demonstrate reciprocity for the H T H domain of Psc. An interesting modulating effect was noted with the SPM domain of ph (amino acids 1511-1576): when the SPM domain was present in a construct, the interaction with Psc was weaker, or as in Figure 1.4d absent.  The domain of Pc required for the interaction with Psc was shown to reside in the 320 amino acids C-terminal to the chromobox (Figure 1.5c, referred to as AchrPc). Surprisingly, the chromobox was not required for this interaction, nor did it or AchrPc show interactions with any Pc construct or with any ph construct from the panel (not shown). The Psc domain required for the Pc interaction was also located within the region of amino acid conservation. Minimally, the H T H domain showed interaction with Pc as  20  I  •  •  11  1  o  I  I  I  I  I  I  I  1  Zi  •  c  IE  E  Z  l/l  Z  </!  HI HI  i  1818  I  I  +  I  C  •  '7  I •a  S5  IS  ••e  9-  B  B  1118  g  5  c  0 '7  .s S  '•5  c  <  7  z D  2/  Figure 1.3 Two-hybrid interaction assay results for ph, Psc, Pc, and carboxyl deletions. D N A binding fusions represent protein fusions to LexA, a bacterial DNA-binding protein, and activator fusions represent protein fusions to B42, a short acidic transactivation sequence. Constructs were expressed in the yeast strain EGY48, which has the LEU2 gene downstream from LexA binding sites. All pairwise combinations were tested. Combinations not shown were negative. Strong positives (1mm colonies after four days of growth on selective medium) are indicated by a large plus, weak positives (<lmm colonies after four days) by a small plus, (a) Psc interacts with Pc, however full length Psc must be deleted for this interaction to be seen, (b) ph interacts with itself through a domain or domains in the smallest amino-terminal construct, (c) Self-interactions were not seen with either Pc or Psc. (d) Psc interacts with ph, and this interaction requires carboxyl deletions of both proteins to be detected. Shadings are as described for Figure 1.1  22  DNA-binding fusions  Activator fusions  AsxA  AsxA I M l AsxQ AsxC  01 AsxQ  All other PcG  Ql M  1 Q 1  AsxA I |A||  o  1 Q |  AsxQ  1  0  1  AsxC  1  0  1  All other PcG  M  Q M  M  AsxC  AsxA I M l AsxQ AsxC All other PcG  All other PcG  AsxA 1 M | AsxQ AsxC  23  Figure 1.4 Asx interactions. No Asx construct interacted with any ph, Psc, or Pc construct. A selfinteraction was seen, however, with AsxC. Shadings are as described for Figure 1.1  24  3.  3.  -ii  z  z  0 z >  I  4  I  3 >  E = fe  C  a. o  I  s.  I  +  mm mm mm wm wm  i  3 3.  s  +  I I X  1  E T 3  • ill + + 25  mm  3.  X H X  r  3.  ma  +  X H X  B  ra ra  -  mm--  •3  lI  S  B  3  "3  5•  A  Figure 1.5 Two-hybrid interaction assay results for conserved sequence constructs, (a) Psc-ph interacting constructs. The interaction is delimited to the HI domain of ph and the H T H containing region of Psc. It is stronger in the absence of the S A M domain of ph. (b) PscPsc interacting constructs. This interaction was only seen with the isolated domains and was dependent on the ring finger, (c) Psc-Pc interacting constructs. The interaction appears dependent on sequences carboxyl to the chromobox of Pc and the HTH-containing region of Psc, although an interaction is seen with the ring finger in one pair. Shadings are as described for Figure 1.1  26  both a DNA-binding fusion, and as an activator fusion. The ring finger of Psc showed weak interaction with the activator-fusion of AchrPc. This may mean that although Pc makes contacts primarily with the H T H domain, it also makes weaker contacts with the ring finger domain.  When expressed in the absence of surrounding sequence, the ring finger of Psc dimerized (Figure 1.5b). This was surprising as dimerization of Psc had not been observed with any larger construct. A weak interaction between the ring finger construct and the H T H domain was seen in one orientation but not the other. This interaction may occur simply because these domains fit together naturally in the tertiary structure of the protein, or it may be part of a true Psc dimerization domain.  Caution should be used in relating the strength of interactions seen in the two-hybrid system with presumed affinities of individual proteins for one another. Two-hybrid analysis done with interactors of known affinities has shown that while interaction strength generally correlates with in vitro affinity, the response curve is not linear, and in many cases shows a threshold below which no response is seen [111].  In-Vitro  Interactions  The two-hybrid interaction assay takes place within the yeast nucleus. Because PcG proteins are transcriptional repressors, this environment is likely very close to their natural environment. However for the same reason it may also contain confounding influences. Any of these interactions could be mediated by an endogenous yeast nuclear protein with enough similarity to the Drosophila protein that actually functions as the mediator, hence the observed interaction may not be direct. Likewise, there may exist yeast proteins capable of interacting with the Drosophila fusion proteins which would occlude or prevent their interaction with each other. I therefore sought to test the identified interactions in  27  vitro. Interacting proteins and domains were subcloned into pGEX4T-l for bacterial GST (glutathione-S-transferase)-fusion protein expression, and pET28a for 17 transcription and in vitro translation in a rabbit reticulocyte lysate. The T7 constructs were translated in the presence of S-labeled methionine and incubated with GST-fusion protein immobilized on 35  glutathione agarose. An interaction between the S-labeled protein and the GST-fusion 35  protein results in the co-precipitation of both on the affinity resin. Bound protein was then washed extensively, eluted with reduced glutathione, run on SDS-PAGE and autoradiographed.  The construct PscAB (aa 1-696) originally implicated in the two-hybrid interaction was shown to interact specifically with both the minimal HI domain of ph and with phHD (amino acids 1297-1576), the larger construct which contains the HI domain and the SPM domain. It also bound the chromobox-deleted Pc fusion. However it did not bind a ph construct that does not contain HI, nor did it bind any Psc construct, nor GST alone (Figure 1.6a). These data corroborate the two-hybrid data. The construct PscHD (amino acids 250-473), which contains only the conserved sequences of Psc (the ring finger followed by the HTH-containing region) showed similar behaviour, although a new interaction with itself was detected (Figure 1.6b). When the homology region was broken into the ring finger and the H T H domain, the H T H domain interacted with HI while weaker interactions were seen between H T H and Pc as well as H T H and ring-finger containing constructs (Figure 1.6c). The ring finger did not interact with HI, but did show an interaction with itself, and a weaker interaction with Pc and H T H (Figure 1.6e). Full length Psc interacted with both HI and chromobox-deleted Pc (Figure 1.6f) recapitulating the behaviour of PscAB.  In the translation of HTH-containing constructs of Psc, I observed smaller labeled fragments derived most likely from weak internal initiation or possibly from breakdown of  28  A  / #  PscAB*  //»>  B  PscHD*  <T * <f #  D  HTH*  ^  F  ring*  •4*  Zf  <z i* h  PscHD* <f  ^  <t  Psc*  c?V ^ /  •$  9  Figure 1.6 In vitro binding of reticulocyte lysate-generated S-labeled Psc constructs to bacterially produced GST-fusions. GST-fusions were expressed in bacteria, bound to glutathione agarose beads, and blocked with BSA. S-labeled Psc constructs were transcribed and translated in a reticulocyte lysate, and added to the blocked beads. Beads were washed, then eluted with reduced glutathione. The labeled construct used in each experiment is denoted by an asterisk, (a) Labeled PscAB (aa 1-696) binds to GST fusions to regions of ph which contain HI (aa 1297-1418) and to a GST-fusion of Pc deleted for the chromobox. It does not bind phAS (aa 1-522 of ph) or other Psc constructs, (b) Labeled PscHD (aa 250-473) bind the same GST-fusions as well as GST-fusions containing aa 250-473 of Psc. (c) The Psc HTH-containing region (aa 336-473) binds Hl-containing ph constructs strongly, and Pc and Psc constructs weakly, (d) PscHD (aa 250-473) binds as strongly to the ring finger alone (aa 250-335) as it does to the ring finger plus the H T H region, and only weakly to H T H . (e) The ring finger of Psc binds to itself, and also more weakly to Pc and HTH. (f) Labeled full length Psc binds GST fused to phHl (aa 12971418) and to the GST-fusion of Pc deleted for the chromobox, but not to PscHD (aa 250473). 35  35  30  the full length products (Figure 1.6a, b, c, d). These bound to HI-containing constructs but not to Pc. I interpret this as evidence that ph and Pc bind to different regions of PscHD. Furthermore, while HI-containing GST-fusions strongly bound both PscHD and the H T H domain, Pc strongly bound only the complete PscHD, and bound both H T H and the ring finger more weakly. This is further evidence that Pc makes use of a different interaction surface than that used by ph, and that this interaction surface is likely made up of elements from both the ring finger and HTH. The self-interaction of PscHD required the ring finger (Figure 1.6d), and was not seen with the larger construct PscAB.  The amount of sample loaded in each experiment was such that a bound band of equal intensity to the input band represents approximately 10% of input labeled protein remaining bound through multiple wash steps of increasing stringency, and eluting with reduced  1  glutathione. By comparing the relative intensity of bound band to input band between experiments, the most stable association under these conditions is seen between the labeled H T H fragment and ph constructs containing the HI domain. This level of bound to input protein is similar to that seen in experiments with the SPM domain interactions of ph and Scm [112].  Ternary vs. Binary Complexes Independently, the co-IP and domain analysis are consistent with both a ternary complex or multiple binary complexes, however a ternary complex seems more likely considering the data together. The co-IP demonstrates the existence of protein complexes containing Pc-ph and Pc-Psc, while the domain analysis only gives evidence for direct interactions between Pc-Psc and ph-Psc. A ternary complex with Psc as the bridge explains both sets of data. Alternatively a direct Pc-ph interaction may have eluded these assays, or may be mediated by another unidentified protein in the nuclear extract.  31  Isolated Domain Interactions can be Modulated by External Sequences In the domain analysis, some interactions were affected by parts of the proteins not implicated in binding. In the case of the ph-Psc interaction, the presence of the ph SPM domain weakened the interaction in most two-hybrid combinations although not in the in vitro assay. Since the SPM domain has the potential for heterologous self-association, and since yeast proteins with this domain exist [103] the modulation might be an artifact of ph interacting with endogenous yeast proteins. In Drosophila there are at least two nuclear proteins that contain the SPM domain: Scm [102] and l(3)mbt [113]. Whether the Scm-ph interaction affects the ph-Psc interaction is an open question. The two-hybrid interactions were also attenuated by full length Psc. This may be due to the ability of full length Psc to repress transcription in yeast (Chapter 4): Consistent with this, the full length protein does interact with the expected domains of ph and Pc in the in vitro assay.  The greatest inconsistency between the two-hybrid results and the in vitro results was seen with the Psc-Psc interaction. In the two-hybrid system, self interactions were only seen with the isolated ring-finger domain. In vitro, self interactions were seen with the ring finger and with the complete conserved region which includes the ring finger, but not with larger constructs. The most likely reason for the discrepancy is the fact that these assays employ proteins produced from three different sources: yeast cells, bacterial cells, and a reticulocyte lysate. A protein expressed in a heterologous system will not necessarily have the same folding and covalent modifications as its native cognate. A given domain may be prevented by its expression context from attaining the fold or covalent modifications required for interaction. The fact that large parts of Psc from outside of the homology domain prevent the self-interaction may mean that the interaction is spurious, an artifact of the isolation of individual domains, or that dimerization is cryptic, and normally modulated by other parts of Psc, with dimerization only happening under certain conditions such as binding to D N A or binding other PcG proteins.  32  Interacting domains HI  Q  HQ  S/T Psc  chromo  Pc  i  mm  t  Psc  Psc  W s  R HTH  Psc  ph  Pc  55  S  A  M  Figure 1.7 Domains involved in interactions between Pc, ph, and Psc. The Psc-interacting domain of ph spans aa 1297-1418. The Psc-interacting domain of Pc is within aa 70-390. The homology domain of Psc (aa 250-473) binds to Pc, while the H T H subregion (aa 336-473) binds to ph. Shadings are as described for Figure 1.1.  34  Interactions of the Vertebrate Homologues of Psc, Pc, and ph My results are similar in general, but differ in detail from those reported for the various mammalian homologues of ph and Psc. Although the isolated Mph HI domain and Bmi-1 H T H domain were not tested with each other, the presence of both HI and the SPM domain of Mph was required for the interaction with Bmi-1 [73], leading the authors to speculate that Mph dimerization was a prerequisite for Bmi-1 binding. I do not see such a requirement for ph binding to Psc. The issue is complicated by the fact that besides Psc, there are two other ring-HTH containing proteins in the fly, Su(z)2 [104; 105], and L(3)Ah [107], and at least one other in the mouse, Mel-18 [106]. The mammalian complex members may truly behave differently from their fly cognates, or perhaps Mel-18, and not Bmi-1 is the functional homologue of Psc.  In this work, the Psc-Pc interaction was seen with both the ring-finger and the H T H domain of Psc. Alkema etal.[73] do not see a two-hybrid interaction between the mouse homologues, Bmi-1 and M33. However Hashimoto etal. ( N . Hashimoto, H.W. Brock, M . Nomura, M . Kyba, J. Hodgson, Y . Fujita, Y . Takihara, K. Shimada, and T. Higashinakagawa, submitted) have reported such an interaction with an in vitro binding assay similar to that used in this work, and in one orientation in the two-hybrid system, and show that the H T H domain-containing region is required. The Xenopus homologues, XPsc and XPc, have been shown to interact with each other, however this interaction was shown not to require the H T H domain of XPsc [98], requiring instead the 188 upstream amino acids which contain the ring finger. While these differences may reflect true differences between fly, frog and mouse, given the sequence conservation of these domains, it is more likely that the differences arise from differences in the assays, specifically in the sizes and imprecise overlap of the constructs used. Since I have seen interactions with both the ring-finger and the HTH-containing region in both two-hybrid and in vitro assays, I speculate that Pc primarily contacts the HTH-containing region but  35  also contacts thering-fingerdomain weakly. Alternatively, Pc may contact the region in between thering-fingerand the H T H domain proper, and some level of binding to each half is seen even when this region is divided. Pc and XPc differ also in their observed self affinity: Reijnen et al. [98] reported that full length XPc was able to interact with both its amino terminus and its carboxyl terminus, whereas I see no Pc-Pc self-interaction.  It has been shown that full length Mel-18 has the ability to bind DNA whereas a deleted version of Mel-18 lacking the ring finger does not [106]. It is possible that Psc also has this ability, and it would be interesting to know whether the binding of ph and Pc, so close to and perhaps directly on the putative DNA-binding domain would influence the putative DNA-binding properties of Psc.  The Role of Multiple Interacting Domains in PcG Complexes Using a formaldehyde crosslinking assay, Strutt and Paro [109] have recently shown that the composition of PcG complexes is not the same at all target loci. The partially but not completely overlapping patterns of PcG protein binding to polytene chromosomes also suggest PcG complexes that are heterogeneous in composition, being different at different target sites. The interaction domains that I have described may facilitate this heterogeneity. Psc has a domain with the ability to bind either ph or Pc, or perhaps both, while ph has two very distinct domains with the ability to bind Psc on the one hand, and ph or Scm (Chapter 2) on the other. These interaction domains make possible multiple protein contacts, not all of which necessarily occur at every site. By allowing different complexes to form at different sites, more complex regulation of target genes is permitted.  All of the conserved sequences of ph and Psc have now been shown to function as proteinbinding domains. This raises the question of what purpose the nonconserved sequence, which forms the vast majority of these proteins, serves. A putative complex involving only  36  a single copy of each of ph, Sem, Psc, and Pc would be on the order of 0.5 MDa, although the interacting amino acid sequences would account for less than 80 kDa. One possibility is that the nonconserved sequence has a direct transcriptional repression function that is conserved in the absence of sequence conservation. An alternative is that transcriptional repression is an indirect result of the bulk of the protein complex, which either excludes transcriptional activators from the vicinity of their binding sites, or prevents their interaction with the basal transcription machinery. If this were the case, the PcG proteins could be described as very large molecules with small domains that can interact with each other promiscuously, allowing bulky heterogeneous complexes to form at their various sites of action.  37  Chapter II: The S A M domain  To explore interaction-space outwards from the initially defined ph-Psc-Pc and Asx interactions, I tested other PcG proteins against the two-hybrid panel. LexA-fusions of esc and Pel provided by Jeffery Simon and Rob Saint, respectively, did not interact with any members of the panel. An activator fusion of E(z) showed a weak interaction with Lex-esc (an interaction also noted by Jeffery Simon, personal communication), but not with any other member of the panel. Finally, in a collaboration with Jeffery Simon, an interaction between Sem and ph was discovered. The minimal domains required for this interaction were the carboxyl terminal SPM domains of ph and Sem (aa 1511-1576 of ph, and 797877 of Sem).  The S A M Domain The SPM domain is unusual among conserved domains of PcG proteins in that there are distinct paralogous sequences not only in other chromatin proteins but also in cytoplasmic proteins. This domain was in fact originally identified in comparisons of cytoplasmic proteins, and named the S A M domain [103]. The name S A M is an acronym for Sterile Alpha Motif, reflecting the putative alpha helical structure that is strongly predicted for this sequence [103]. Because the proteins Byr2p [114; 115] and C33B4.3 [116], have S A M domains at their extreme N (amino acids 1-66) and C (amino acids 1045-1110) termini respectively, the boundaries of this domain are clearly defined. Database searches have identified S A M domains in over 60 other proteins [117] that share no obvious common function.  A possible function of the S A M domain is to associate with other S A M domains, either homotypically whereby two identical S A M domain-containing proteins associate, or heterotypically whereby two different S A M domain-containing proteins associate through  38  the interaction of their S A M domains. Two S A M domain-containing proteins required for mating in S. pombe, Ste4p and Byr2p, have been shown to interact with each other [118]. The interaction occurs through regions of both proteins that contain S A M domains: amino acids 1-160 of Ste4p and amino acids 1-392 of Byr2p. The S A M domain of Byr2p is essential for this interaction, as a single base substitution in the Byr2p S A M domain abolishes Ste4p-binding activity [118]. A subset of the ETS family of transcription factors including ETS-1, ERG-2, and T E L from vertebrates, and Pointed-P2 and Yan from Drosophila have a S A M domain near their amino termini, referred to previously as the B domain [119] or the pointed domain [120]. The S A M domain of T E L when fused to either the PDGF-b receptor t(5;12) or the AML1 gene t(12;21) induces oncogenic transformation [121; 122] and oligomerization through the S A M domain is essential for the constitutive activation of TEL-PDGF-b's protein kinase activity and mitogenic properties [123]. The presence of this domain in two PcG proteins, as well as in a wide variety of other proteins, prompted a more detailed functional analysis. I was interested to know what sequence features of this domain governed its association behaviour, and if this behaviour had implications for the complexes formed by the PcG.  The ph S A M Domain Mediates Self-Association In Vitro I generated a variety of GST-fusions to carboxyl sequences of ph (Figure 2.1a) and tested each for the ability to bind the in vitro translated S-labeled polypeptides phHD and pSAM 35  from the carboxyl terminus of ph. The GST-fusions were immobilized on glutathione agarose, mixed with labeled polypeptide, washed extensively in 500mM NaCl and eluted with reduced glutathione. Labeled bound polypeptide was detected by gel autoradiography. Both labeled polypeptides tested, phHD (amino acids 1297-1576) which contains the HI domain and the S A M domain, and the smaller pSAM (amino acids 15111576) containing only the S A M domain, showed similar behaviour (Figure 2.1b and c). They bound only to GST-fusions with an intact S A M domain. phHD showed aberrantly  39  A  ph fusions  •  129/  I30B  phHD  GST  H1  H1  GST  H1  I SI I  SAM  H2  SAM  H2AC  S4  H24L>4A  S M  SAM  SAM  B  phHD' 4^  pSAM*  ln/6  Figure 2.1 Self-binding activity of the carboxyl terminus of ph. (a) GST-fusion constructs, (b) Binding of S-labeled phHD to the GST fusions, (c) Binding of S-labeled pSAM to the GST fusions. S-labeled constructs were transcribed and translated in a rabbit reticulocyte lysate, and binding and elution was done as described above (Figure 1.6). The labeled construct used in each experiment is denoted by an asterisk. 35  35  35  41  slow migration due to the high content of proline in the sequence just upstream of the S A M domain. On binding to and elution from the H2 construct, phHD showed a slight and reproducible shift in migration which I attribute to a conformational change in the proline rich region (possibly a peptide bond isomerization) that is resistant to denaturation in the SDS loading buffer.  The first lane in all gels contains a fraction of the total labeled polypeptide representing one tenth of the amount added to each individual binding reaction. A bound band equal in intensity to the input band therefore represents 10% of the total labeled polypeptide binding, remaining bound through the wash steps, and eluting with reduced glutathione. By comparing the input lanes with the bound lanes (Figure 2.1b and c), the notable difference between the larger phHD construct and the minimal pSAM construct is that binding of the larger construct under these conditions is much stronger than binding of the minimal S A M domain. Nevertheless the fact that the minimal S A M domain binds to itself demonstrates that this domain alone is sufficient for the self-association.  Homologues of the ph SAM Domain The S A M domain of ph has several close relatives from within Drosophila and from other species. Figure 2.2 shows an alignment of S A M domains, where amino acid identities are boxed and conservative substitutions relative to ph residues, as determined by the Kyte Doolittle matrix and visual inspection, are shaded. Characteristic features of the domain sequence are at the amino-terminal end, where it initiates with a conserved tryptophan followed 5 residues later by a valine, in the middle where a perfectly conserved glycine is followed closely by a hydrophobic block with a strong preference for the sequence A L L L L , and towards the carboxyl-terminal end where there is an almost perfectly conserved glycine. Features in particular that separate this group from other proteins  42  falling under the broad consensus suggested by Schultz et al. [117] are the conservation of both glycines previously mentioned (neither of which are part of the consensus of Schultz et al), and a strong preference for the conserved tryptophan to be followed by the sequence [ST]-X-[DE]-[DE].  Homotypic and Heterotypic Self-Association of ph SAM Homologues I selected several members of this group of S A M domains to determine whether the selfassociation function was conserved with the sequence. Proximal and distal ph are semiredundant genes [124]. Therefore I was interested to determine whether the binding properties of their S A M domains differed, particularly when the domain was present in the context of its upstream sequence, the only region of interstitial sequence dissimilarity between proximal and distal ph. The S A M domain of distal ph is nearly identical to that of proximal, with only three conservative substitutions (Figure 2.2). RAE-28 was chosen as it is the recognized mammalian orthologue of ph [73; 101] and has 40/64 amino acid identities. Scm represented a paralogous S A M domain from within the same species [102], with 25/64 identities. BEB1 represented the most related S A M domain from S. cerevisiae [125; 126] and shares 16/64 amino acids with ph over the S A M domain. The proximal and distal ph S A M domains were tested as isolated domains as well as in the context of their unique upstream sequences. I use the nomenclature bSAM, rSAM, sSAM, pSAM, and dSAM to designate the minimal S A M domains of BEB1 (amino acids 263-329), RAE28 (amino acids 945-1012), Scm (amino acids 797-877), proximal ph (amino acids 15111576), and distal ph (amino acids 1338-1403) respectively. Scm2, pH2, and dH2 represent sequences from Scm (amino acids 767-877), proximal ph, (amino acids 14271576) and distal ph (amino acids 1249-1404) respectively, which include the S A M domain and upstream sequence.  43  Scrn(d)  D - F 1 Q Q E 1 D G D - F .1 Q Q E 1 D G D. L F R K H E 1 D G  L(3)rnbb(d)  V  389108(rn)  T A F  ph_prox(d) ph_disb(d)  HIBBU50(h)  1 F - E Q D 1 D G Q E Q E 1DG E - F R S Q E 1 D G E - F_ R A Q E 1 D G  TEL(h)  1 D S N T F EM'N G  RAE28(rn)  TEL(rn)  1 E S N K F E M N G  ETS(d)  L D " L I  YAN(d)  D L F Q M N G D L F iQ. M N G  C-ETS-2(h)  V N L G R F G M.N. G  BOII(y)  DAGkina9s;ri)  K - F K E HQ VS G R - F Q K H K 1 S G 1 - F T RH D 1 R G  mMG11(m)  D  BEB1(y)  36 ph_dist(d) 36 Scrn(d) 37 L(3)mbb(d) 36 3891 08(m) 3 5 RAE28(tn) 36 HIBBU50(h) 33 TEL(h) 37 TEL(m) 37 ETS(d) 37 YAN(d) 37 C-ETS-2(h) 3 7 36 BOI1(y) BEB1(y) 36 DAGkirase(ri) 3 5 m M G 11 (rn) 3 7  ph_prox(d)  1 F R D N K 1 A G  Q A L L L L K E,.K.H L V N A M G M K L P, A L L L L K E . N . H L V N A M G M K L  P A  K I  A K V E S 1 K E V  P A  K I  K A L L L L N S E M | M M K Y|MJ3  K.L.  P A  K I  A K V E S 1 K E V N L V N K V N G R  E.L.  T Y L"K  K' V  P A! V ! K L  S.i L 1 L N L R 1 A  L L L|M Q R T D V E H Q A L L L L K  T G L S  L  P A  K I Y E H H  S A M N  L  P A  S A M N I  L  PA  K I C A K 1 N V L K E T K J Y X R 1 S M L K  K A L L L L T  Y R - S P  S  D V  Y E L L  K A L L L L T  Y R - S P  S  D V  Y E L L  L T  H R - C P  A  D Y  H N V L  L L T  H R - C P  D V  Rj D  A E L A P D V E L E I N S F E L D I N S F Dl Q Y T K V  D R  D L G V S 3 L  E R K  u g  R A L  G RK  K  Q A L L L L  K A L  KA  L  QjM L K  c  -  K  D H  c N L G  1 L L E L E  K 1 L L E L E  5 E L L H L E S F .L P F I D  Q Q Q Q  1 K V  R F Q  KM  I  70  65 &g  69 69 69 70  I F K E 1 R N 1 K J L C G  70 70 71 70 66  - H 1 L K Q R - M L 1 1 E S  A  T R F E I F K E 1 E. K 1 K E A H M K R  34  36  H 1 L K Q R  -  - M L 1 1 E S H N.V L D I L.NV E H L E - 0 M 1 K E N  I  35 35 36 35 34 35 32 36 36 36 36 36 35 35  1 K E L S R 3  E C 1 Q G L "S Q 3  70  70 69 71  Figure 2.2 Alignment of relatives of the ph S A M domain. Amino acid identities are boxed. Conservative changes from the S A M domain of ph are shown shaded. The organism from which the protein sequence is taken is denoted by a letter in parenthesis: d - Drosophila, m Mouse, h - Human, y - yeast (S. cerevisiae). The sequence shown begins 5 amino acids upstream of the S A M domain proper, of which the conserved tryptophan is residue number  45  pSAM*  .  B  dSAM*  rSAM*  3  sSAM*  bSAM*  F  sSAM*  —  raw  Ni NTA- bound  Figure 2.3 Homotypic and heterotypic binding of various S A M domains, (a) S-labeled polyhomeotic proximal S A M domain tested for binding to GST and 7 G S T - S A M domain fusions. The lane marked 'input' contains a fraction of the S-labeled polypeptide before binding, (b) S-labeled polyhomeotic distal S A M domain, (c) S-labeled RAE-28 S A M domain, (d) S-labeled Scm S A M domain, (e) S-labeled BEB1 S A M domain, (f) S-labeled Scm S A M domain tested for binding to Ni-NTA agarose. The arrowhead indicates the shorter product lacking the 6xHis tag. 35  35  35  35  35  35  35  47  The labeled polypeptides pSAM and dSAM showed similar behaviour. They both showed weaker binding to ph S A M domains than to those of sSAM or rSAM. Interestingly, both pSAM and dSAM showed stronger binding to dSAM than to pSAM. Neither bound bSAM or GST alone (Figure 2.3a, b). rSAM and sSAM behaved similarly to each other, but differently from pSAM and dSAM. Both bound pSAM and dSAM more strongly than they bound themselves and each other, and their respective homotypic interactions were stronger than those of pSAM or dSAM (Figure 2.3c, d). Neither bound bSAM or GST alone. bSAM was distinctive in that it bound only to itself, but not to any of the other SAMs, nor to GST alone (Figure 2.3e). These data confirm the hypothesis that the S A M domain is a self-association motif. Each S A M domain tested has the ability to bind to itself in vitro, albeit weakly in some cases. Furthermore with the exception of the yeast S A M domain which is most divergent in sequence, each is also capable of binding other S A M domains. However these heterotypic interactions occur with different affinities, and do not occur in any combination with bSAM. Because these domains all share the amino acids of the consensus sequence yet behave differently, the specificity of association must be derived from the nonconserved amino acids.  A protein doublet was seen in the in vitro translation products of some constructs, particularly with sSAM. The in vitro translated products were transcribed from the vector pET28a, which in addition to providing an initiation methionine, also adds a 6xHis tag followed by a thrombin cleavage site to all proteins translated. Only the upper band of the doublet (the full length product) bound to Ni-NTA agarose (Figure 2.3f). The lower band therefore lacks the 6xHis tag, and is consistent in size with the proteolytic product resulting from cleavage at the thrombin site, most likely by minor contaminating proteases in the reticulocyte lysate.  48  Mutations in Conserved Residues Have Different Effects on Binding to Different SAMs The evidence presented above shows that the S A M domain self-associates and that specificity arises from the non-conserved amino acids. I wished to investigate the role of the conserved amino acids in the S A M domain. Therefore I created 5 mutations in residues that were conserved within the group of proteins tested and within the broader group of S A M domains (Figure 2.4a). The mutations were made in the pSAM domain and expressed as fusions to GST. When tested for binding to labeled pSAM, all 5 mutations abolished binding (Figure 2.4b). However when the mutants were tested with the larger phHD construct, weak binding was seen with the mutant I62D (Figure 2.4c). The strongest binders to wild type pSAM were rSAM and sSAM (Figure 2.3). They showed some binding to I62D, and in addition showed binding to L33A and L41A (Figure 2.4d,e). No construct bound the mutants W l A or G50A. By comparing the intensities of mutantbound bands to wild-type-bound bands, it is apparent that no mutant S A M bound as strongly as wild type. An allelic series can be constructed where (W1A, G50A) > (L33A, L41A) > I4D with severity being measured by the number of constructs whose binding is abolished, or by the levels of binding compared to wild type pSAM.  SAM: A Self Association Motif The carboxyl terminal region of ph can self-associate in vitro, and this self-association is a function of the S A M domain. I have demonstrated in vitro self-association of four other S A M domains from different proteins derived from fly, mouse, and yeast. In accordance with these observations, I propose to keep the name suggested for this sequence by Ponting, namely S A M [103], but to redesignate the acronym: Self Association Motif.  Although homotypic self-association was seen for all S A M domains tested, their relative affinities for self varied, with the S A M domains of RAE28 and Sem showing high affinity  49  A W  1  I  L33A  A  L41A  I  G50A  I  I  I62D  I  WSVDDVSNFI RELPGCQDYV DDFIQQEIDG QALLLLKEKH LVNAMGMKLG PALKIVAKVE SIKE A A A DK*  A  B  pSAM*  C  phHD"  t D  rSAM*  * sSAM*  E  •  So  *  Figure 2.4. Mutations in pSAM and their effects on binding, (a) The amino acid sequence of the wild type pSAM domain and substitution mutations. The mutation names are given above their respective positions in the wild type ph S A M domain and the substituted amino acids are given below, (b) S-labeled polyhomeotic proximal S A M domain tested for binding to GST, GST-pSAM, and mutants of pSAM. (c) S-labeled phHD binding to the mutant panel, (d) S-labeled RAE-28 S A M domain binding to the mutant panel, (e) S-labeled Scm S A M domain binding to the mutant panel. 35  35  35  35  51  and proximal and distal ph, and BEB1 S A M domains showing low affinity. In addition, various heterotypic interactions were noted between all S A M domains with the exception of that of BEB1. These interactions varied in strength, with affinities between RAE28-ph SAMs and Scm-ph SAMs being even stronger than the strongest homotypic interactions. These observations demonstrate that while association in general may be a function of the conserved residues of the domain, the specificity of the interaction is determined by nonconserved residues. Thus, particular sets of nonconserved residues generate particularly good interaction surfaces, such as sSAM-pSAM, while others such as bSAM-pSAM do not. The lack of any heterotypic interaction with the S A M domain of BEB1 is perhaps not surprising, given that it is the most divergent of the group. It is notable that bSAM is the only member tested in which the four leucine run is broken by a charged amino acid. Incompatibility between bSAM and the other SAMs might be expected if this region of the domain formed part of the binding surface.  The GST fusions, pH2 and dH2, which contain the S A M domains of proximal and distal ph preceded by their respective upstream unique sequences behaved similarly. Since the upstream unique sequences contain the most significant interstitial sequence divergence between proximal and distal ph, it has been suggested that the functional differences between the two isoforms arise from this region [99]. The similar binding behaviour of pH2 and dH2 suggests that the functional differences between proximal and distal ph do not arise as a consequence of SAM-mediated interactions.  Key Structural and Functional Residues of the SAM Domain Without knowing the structure of the domain it is difficult to speculate which particular amino acids comprise the interaction interface, however the mutational analysis provides some hints. Two mutations abolished interaction between pSAM and any other construct: W1A and G50A. It is most likely that the conserved tryptophan and glycine are key  52  elements of the domain fold, and altering them prevents the domain from folding properly. However three mutations showed weaker binding to some constructs, and the absence of binding to others. The fact that binding was seen at all demands that the domain fold itself must be intact. Assuming this, these three mutations, L33A, L41A, and I62D, must alter key residues of the binding interface, either directly by side chain substitution, or indirectly by inducing a change in neighbouring residues that are themselves part of the binding interface.  The Potential for Promiscuous Oligomerization The S A M domain has been found in proteins of very diverse function. In this study, pSAM, dSAM, sSAM, and rSAM come from chromatin proteins involved in transcriptional repression, while bSAM is from a cytoskeletal protein that interacts with BEM1 and is required for proper cell polarization [125; 126]. The S A M domain-containing proteins from S. pombe, Ste4p and Byr2p are involved in mating pheromone signal transduction. One feature that unifies this group of proteins is that they are all involved in multimeric protein complexes, ph is in a complex with at least 10 other proteins [29]. Beblp interacts with Bemlp, and has several protein interaction domains besides the S A M domain: an SH3 recognition motif, an SH3 domain of its own, and a PH domain [125; 126]. Ste4p and Byr2p from S. pombe and their homologues, Ste50p and Stel lp from S. cerevisiae form a complex web of interactions with other components of the M A P kinase module, including Ste5p, Stellp, Ste7p, and Fus3p [118; 127-130].  From these examples, it would appear that the S A M domain is a protein interaction domain that is particularly well suited to joining members of multiprotein complexes. In the case of the PcG, the S A M domain could facilitate heterodimerization of ph and Sem, or homodimerization of either. In addition it is possible that the S A M domain could mediate an interaction between PcG complexes containing ph or Sem with S A M domain-containing  53  transcription factors such as dETS, Yan or Pointed. As yet no interaction between a PcG protein and a sequence-specific DNA-binding transcription factor has been demonstrated, and it will be interesting to know whether these proteins associate with ph or Scm.  The different affinities noted for particular S A M domains may have biological relevance. In the case of proximal and distal ph and Scm, which are present in the same organism, heterotypic interactions appear to be more stable than homotypic. This would suggest that PcG complexes might prefer ph-Scm heterodimers to Scm-Scm or ph-ph homodimers. The promiscuous nature of S A M domain binding suggests that the complexes containing ph and Scm could be highly heterogeneous. Whether PcG protein complexes take advantage this potential heterogeneity for regulatory purposes whereby the exchange of one S A M domain-containing protein for another would enable or disable the silencing activity of the complex, or functional complexity whereby the complex would employ certain S A M domain-containing proteins for certain tasks such as recognizing a specific target locus, remains to be determined.  54  Chapter III: Asx Interactions  In the two-hybrid matrix, none of the three Asx constructs interacted with any other PcG construct. This raised the possibility that Asx might function independently from ph, Pc, and Psc at the molecular level, perhaps being a member of a different protein complex, or that Asx might function at a step in the regulatory hierarchy above or below the ph-Pc-Psc complex. In an attempt to clarify the issue, I undertook a broader search for Asxinteracting proteins by screening Asx against a two-hybrid library of embryonic cDNAs, and by screening Asx against a panel of chromatin proteins generated by Michael O'Grady.  A Conserved Domain in the C-terminus of Asx From the three LexA-fusions to different domains of Asx, I chose Lex-AsxC as the bait for the library screen after discovering that it contained a short domain of high similarity to a previously unknown mammalian gene: I performed an exhaustive B L A S T search of dbEST using the entire Asx protein sequence, and identified two human ESTs whose conceptually translated sequence matched that of the putative Asx zinc finger (Figure 3.1b). No other region of non-redundant Asx sequence gave any significant matches to any other ESTs at the time, or to any other known proteins. Both cDNAs were from the I M A G E consortium: cDNA 42515 (Genbank Accession T16795) and cDNA 840471 (Genbank Accession AA485878). I obtained cDNA 42515 from the IMAGE consortium, and sequenced it. The 1.5 kb insert encoded an open reading frame that was open through the 5' end and ended in two stop codons very near the 3' end (Figure 3.1a). Apart from the putative zinc finger, there was no similarity to Asx. Given that the interactions discovered so far between Psc, ph and Pc had been mediated by conserved domains, the C-terminus of Asx, which contained the sequence conservation, seemed the best candidate to use as bait in a two-hybrid screen.  55  3. hAsx sequence: GCTTGGAACGANGCAAGGNCAGATG A  W  N  X  A  R  X  GNATNGTTGGTCCTCAGAGATGGGT  D  G  X  V  G  P  Q  R  W  GTCTCGAGTATGTGCGGTCCGCCAA  V  S  R  V  C  A  V  R  Q  >  100 AAGATCCCAGATTCCCTACTGCTGG K  I  P  D  S  L  L  TCAGTACTGAGTACCAGCCAAGAGC  L  V  S  T  E  Y  Q  P  R  CGTGTGCCTGTCCATGCCTGGGTCC  A  V  C  L  S  M  P  G  S  >  200 TCAGTGGAGGCCACTAACCCACTTG S  V  E  A  T  N  P  TGATGCAGTTGCTGCAGGGTAGCTT  L  V  M  Q  L  L  Q  G  S  GCCCCTAGAGAAGGTTCTTCCACCA  L  P  L  E  K  V  L  P  P  >  300 ACCCACGATGACAGCATGTCAGAAT T  H  D  D  S  M  S  CCCCACAAGTACCACTGACCAAAGA  E  S  GGATCTTTACATGGTCTTGGAAAAA G S  L  H  G  L  G  K  .  P  Q  V  P  L  T  K  D  Q  ACAGTGGCATGGTTGATGGAAGCAG N  S  G  M  V  D  G  CCAGAGCCATGGCTCGCTACGCATG  S  S  H  G  S  L  R  M  >  CCCCAGTTCTATAAGGGCTTTGAAG  S  P  S  S  I  R  A  L  K  >  400 GAGCCTCTTCTGCCAGATAGCTGTG E  P  L  L  P  D  S  C  AAACAGGCACTGGTCTTGCCAGGAT E  T  G  T  G  L  A  R  I  TGAGGCCACCCAGGCTCCTGGAGCA E  A  T  Q  A  P  G  A>  500 CCCCAAAAGAATTGCAAGGCAGTCC P  Q  K  N  C  K  A  V  CAAGTTTTGACTCCCTCCATCCAGT P  S  F  D  S  L  H  P  V  GACAAATCCCATTACATCCTCTAGG T  N  P  I  T  S  S  R  >  600 AAACTGGAAGAAATGGATTCCAAAG K  L  E  E  M  D  S  K  E  TCACAGGACAGTAATTCAAATGCTG S  Q  D  S  N  S  N  AGCAGTTCTCTTCCTTTAGTTGTGA  A  Q  F  S  S  F  S  C  E  CTCCAGGAAAGAGCCCAGGAGATCT A  P  G  K  S  P  G  D  L  AGATCAGAAGGAAGTCCGTGCTATG D  Q  K  E  V  R  A  M  >  TACTACCTCGAGAACACCTCGTTTC T  T  S  R  T  P  R  F  >  700 T C A T C T C C A A A T G T G A T C T C C T T T G GTCCAGAGCAGACAGGTCGGGCCCT S  S  P  N  V  I  S  F  G  P  E  Q  T  G  R  A  L  GGGTGATCAGAGTAATGTTACAGGC G  D  Q  S  N  V  T  G  >  800 CAAGGGAAGAAGCTTTTTGGCTCTG Q  G  K  K  L  F  G  S  GGAATGTGGCTGCATCCCTTCAGCG G  N  V  A  A  S  L  Q  R  CTCCAGACCTGCGGACCCGATGCCT S  R  P  A  D  P  M  P  >  900 CTCCCTGGTGAGATCCCTCCAGTTT L  P  G  E  I  P  P  V  CAGACTCCAAGGGAAGACTGGGCTC Q  T  P  R  E  D  W  A  TTCCCAGTGGGAAGTTGGGACCAAG F  P  S  G  K  L  G  P  S  CAAAGCCACATGCCTTTGTTGGCAG P  K  P  H  A  F  V  5 6  G  S  CACAAACTCCATGTCTGGTGGGGTA T  N  S  M  S  G  G  V  >  CGTCAAGAATGAGAAGACTTTTGTG V  K  N  E  K  T  F  V  >  hAsx sequence, continued: 1000  GGGGGTCCTCTTAAGGCAAATACCG AGAACAGGAAAGCTACTGGGCATAG TCCCCTGGAACTGGTGGGTCACTTG G G P L K A N T E N R K A T G H S P L E L V G H L > 1100  GAAGGGATGCCCTTTGTCATGGACT TGCCCTTCTGGAAATTACCCCGAGA GCCAGGGAAGGGGCTCAGTGAGCCT E G M P F V M D L P F W K L P R E P G K G L S E P > 1200  CTGGAGCCTTCTTCTCTCCCCTCCC AACTCAGCATCAAGCAGGCATTTTA TGGGAAGCTTTCTAAACTCCAACTG L E P S S L P S Q L S I K Q A F Y G K L S K L Q L >  AGTTCCACCAGCTTTAATTATTCCT CTAGCTCTCCCACCTTTCCCAAAGG CCTTGCTGGAAGTGTGGTGCAGCTG S S T S F N Y S S S S P T F P K G L A G S V V Q L > 1300  AGCCACAAAGCAAACTTTGGTGCGA GCCACAGTGCATCACTTTCCTTGCA AATGTTCACTGACAGCAGCACGGTG S H K A N F G A S H S A S L S L Q M F T D S S T V > 1400  GAAAGCATCTCGCTCCAGTGTGCGT GCAGCCTGAAAGCCATGATCATGTG CCAAGGCTGCGGTGCGTTCTGTCAC E S I S L Q C A C S L K A M I M C 0 G C G A F C H> 1500  GATGACTGTATTGGACCCTCAAAGC TCTGTGTATTGTGCCTTGTGGTGAG ATAATAAATTATGGCCATGGGAAAC D D C I G P S K L C V L C L V V R * * ATTGT  t) Asx-hAsx homology: Asx: hAsx:  1 6 3 4 CACSLNAMVICQQCGAFCHDDCIGAAKLCVAC-VIR* 16 69 CACSL AM++C+ CGAFCHDDCIG +KLCV+C V + R * 1 5 2 2 CACSLKAMIMCQGCGAFCHDDCIGPSKLCVLCLWR* 1 5 5 8  Figure 3.1. hAsx sequence, (a) hAsx sequence derived from cDNA 42515 of the I M A G E consortium, Genbank Accession: T16795. The D N A sequence is given above, and the conceptual translation below. The region of seqence conservation with Asx is underlined, (b) Sequence similarity between hAsx and Asx proteins at the carboxyl terminus of both proteins.  57  The Asx Two Hybrid Library Screen I screened the activator fusion library RFLY1 [131] which was derived from poly(A ) +  R N A of 0-12 hour Drosophila embryos. The library contained 4.2xl0 independent 6  transformants (Russell Finley, personal communication). From approximately 10  5  cDNAs, 62 colonies that grew on selective medium in the presence of galactose as a sugar source were picked. These were named zl-z62. Of these, 20 grew when tested on dextrose and were discarded (transcription of the activator fusion is shut off by dextrose, therefore true interactions should only result in growth on galactose). Plasmids were rescued from the remaining 42, and retransformed into the bait strain, as well as a strain containing LexA alone as bait. Of these, 20 were specific to AsxC. Insert size and preliminary sequence from the 5' end classified these clones into 11 groups (Table 3.1). Clones were named after the number of the first member picked from the group.  Of the 11 interacting cDNAs identified, z28, z38, and z46 corresponded to previously identified Drosophila  genes,&owe/, klett, and Cabeza , respectively, and z7, z60, and z3  were obvious Drosophila homologues of genes cloned from other organisms, ribosomal protein L34, thioredoxin, and spermidine synthase, respectively. Other interactors showed smaller scale sequence similarities to known proteins: z34 contained glutamine repeats, shared by other members of the PcG and by many other transcription factors, and a run of seven alanines, similar to that seen in Asx itself, z2 showed similarity to thrombospondin through a cysteine-rich region, and z l 1 showed similarity to phosphatidylinositol (4,5)bisphosphate 5-phosphatase. Structural features of note are the presence of zinc fingers in at least two interactors, Cabeza and Bowel, and a cluster of cysteine residues in zl. z2 also contained a leucine-rich region strongly predicted to be ahelical, although not conforming to the leucine zipper motif. Preliminary sequence of z l and z40 showed no significant similarities to any known proteins. The preliminary  58  Clone  Insert Size  Sequence • Features  PcG  interactions  AsxC  AsxCl  Strong Interactors 1,6,13,25,29,57  3.3kb  -  ++  ++  11,27,41,50  0.9kb  -  ++  +  56,60  0.6kb  thioredoxin  -  ++  -  1,21  0.6kb  Aides ribosomal prot. L34  -  ++  na  3 4  2.2kb  polyQ, poly A , cyclin box  PscAB*  ++  ++  3 4 A B  1.3kb  polyQ, polyA  na  ++  ++  2  l.lkb  L-helix, cysteine cluster  -  +  +  3  0.75kb  spermidine synthase  -  +  Weak Interactors  2 8  2.0kb  Bowel (multiple Zn-fingers)  -  +  -  3 8  2.5kb  Klett  -  +  na  4 0  l.lkb  Pc, AchrPc  +  -  4 6  0.6kb  na  +  *  Cabeza (single Zn-finger)  Table 3.1 Grouping of the Asx interactors. The underlined clone number is used to designate the entire group. The fourth column designates interactions with any members of the PcG LexA fusion panel. The rightmost column designates interaction with the construct AsxCl, an amino derivative of AsxC, representing amino acids 1139-1420. ++ strong interaction, + weak interaction, - no interaction, na not assayed, * some very weak growth. z34AB is a deletion construct retaining the amino 306 amino acids of z34.  59  sequence for each interactor not shown in this chapter is given in Appendix A , and sequence similarities between interactors and known proteins are shown in Appendix B.  To determine which interactions, if any, might require the conserved sequence at the extreme carboxyl terminus, AsxC was divided roughly in half to create the constructs AsxCl (aa 1139-1420) and AsxC2 (aa 1412-1669). Unfortunately Lex-AsxC2 activated transcription alone as measured by its ability to promote growth on leucine deficient medium in the absence of any other plasmid, and was therefore not used in this assay. However AsxCl, which lacks the conserved sequence, was tested against each of the interactors. z l , 2, 11, and 34 retained an interaction with AsxCl, while the interaction was abolished for z28,40, and 60 (Table 3.1, column 5). A much reduced interaction occurred for z3 and z46.  Interactions with Other PcG Proteins If Asx were a member of the same complex(es) containing ph, Pc, and Psc, then Asxinteracting proteins might also interact with any of these three. I tested z l , 2, 3, 11, 28, 34,40, and 60 for interaction with each member of my panel of full length PcG constructs and carboxyl deletions. One strong interaction appeared between z40 and Pc as well as a weak interaction between z34 and Psc (Table 3.1, column 4).  Evaluation of AsxC Interacting Proteins Both the preliminary sequence and the interaction results from the PcG panel were helpful in deciding which interactors to pursue further. Ribosomal proteins (z7) are well-known false positives in two-hybrid screens (Erica Golemis, personal communication) and thioredoxin (z60) and spermidine synthase (z3) were excluded because the link between these enzymatic activities and PcG function is tenuous. From the set of previously cloned genes, klett had been isolated by Gunter Reuter and Michael O'Grady (personal  60  communication). The clone that I isolated is in fact the very clone that they had isolated as a Lex-Su(var)3-9 interactor. There are as yet, no klett mutants or information on klett expression. While its separate isolation in a Su(var)3-9 screen could be indicative of a physical link between Asx and Su(var)3-9, the fact that it interacted with many constructs from the O'Grady panel (Michael O'Grady, personal communication) suggested that it might just as easily be a sticky protein, showing spurious interactions with many different Lex-A fusions.  On the other hand, bowel and Cabeza were more promising. Bowel is a multiple zincfinger transcription factor, a member of a paralogous gene family including odd-skipped and sob [132]. bowel homozygotes show cephalopharyngeal and hindgut defects [133] consistent with the restricted anterior and posterior expression pattern of bowel mRNA in developing embryos. Cabeza is a glycine rich RNA-binding protein made up of five domains: An amino-terminal glycine-rich domain followed by an RRM (RNA-Recognition Motif) domain, a central glycine-rich domain, a C2C2 zinc finger, and a carboxyl-terminal glycine rich domain [134]. The z46 clone, which encodes amino acids 246 through to the stop codon of Cabeza, lacks the RRM and amino-terminal glycine-rich domain. Cabeza is expressed ubiquitously early, is later enriched in the embryonic brain and CNS, but is absent from larval brain. It is also present in eye-antennal, wing, and leg imaginal discs, and is enriched in adult head vs. body [134]. It is located in the nucleus of all cell types examined (larval fat body and imaginal disc).  Asx protein is ubiquitous in the early embryo, and concentrated in the neurectoderm and CNS in late embryos [135]. Asx mRNA levels drop markedly in larvae and increase again in pupae and adults. Asx mutants die as larvae with mild posterior transformations of the cuticle of abdominal and thoracic body segments. They show striking cephalopharyngeal defects including the complete failure of head involution [19]. In addition, Asx mutations  61  cause derepression of a Ubx-lacZ reporter gene in the CNS of parasegments 2-5 [8; 135] indicative of a requirement for Asx in transcriptional repression in the CNS of these parasegments. The expression pattern of Asx as well as the domain of effect of Asx mutations coincide well with those of both bowel and Cabeza making these two proteins attractive prospects for in vivo binding partners of Asx. At the time of this writing, studies by others are underway to ascertain the role of Asx in Bowel and Cabeza function.  Possible Roles for Bowel and Cabeza Since Bowel is likely to be a sequence-specific DNA-binding protein, it may play a role in recruiting Asx to target genes in Bowel-expressing tissues. It remains to be seen what genes Bowel regulates, and whether it is a positive or negative transcriptional regulator.  If Asx interacts with Cabeza in vivo, one would assume that RNA is a component of Asx complexes, or that Asx complexes have the ability to bind to RNA. Renato Paro has claimed (unpublished) that Polycomb fractionates differently in the presence vs. the absence of RNase, suggesting that RNA is a component of Polycomb-containing complexes. It may be that Cabeza is the RNA-binding protein responsible. One could also imagine a regulatory function for RNA. One of the perennial questions about PcG protein function is why some targets are silenced in one set of cells but not in others, when the PcG proteins responsible for maintenance of the silenced state are present in both cell types. The corollary is also puzzling: how is it that some loci are active in salivary glands, while others are silenced, when PcG proteins are found at both? One solution to this problem is to postulate that the PcG proteins are present at all target sites, but are either active or inactive, depending on the transcriptional state of the locus at the time of complex assembly. Negative regulatory proteins such as Hunchback, bound to D N A at the time of PcG programming, could provide input favouring the silencing state, while positive factors could provide antipodal input. mRNA itself, transcribed from the locus in question, could  62  be such a positive factor, signaling into the complex through an RNA-binding protein such as Cabeza.  The Complete Sequences of z34 and z40 The interactions noted for z34 and z40 with the PcG panel, as well as the presence of glutamine and alanine repeats in z34 prompted the complete sequencing of these two interacting clones. I subcloned the inserts into pBluescript, and using a combination of forward/reverse primers and 5' and 3' restriction site deletions, sequenced them in their entirety (Figures 3.2 and 3.3). Both contained open reading frames that extended through the 5' end of the sequence, and eventually terminated before reaching the 3'end. Although neither gene has previously been described in Drosophila or any other organism, both genes are represented by multiple Drosophila ESTs. By analyzing 5' EST sequences, the 5' ends of both open reading frames could be determined. EST sequence 5' to the z34 clone contained stop codons in all three reading frames. The first methionine codon of the z34 cDNA is in my interacting clone, therefore the z34 clone contains a complete open reading frame (Figure 3.2). z40 on the other hand, was found to be missing 89 nucleotides of open reading frame at the 5' end. Figure 3.3 shows the melded sequence of my z40 clone and the Drosophila EST 1032424 (Genbank accession: AA391083).  z34 Contains a Cyclin Box I rescreened the protein sequence databases using the full length sequences, and was still unable to find any significant match for z40. z34 however, now gave high scoring matches to cyclins, with the highest being to a human G-type cyclin. The sequence similarities were strongest at the cyclin box, which encodes the C D K (cyclin dependent kinase)-binding domain of the cyclin. Figures 3.4a and 3.4b show the cyclin box sequence of z34 aligned with the same for all of the Drosophila cyclins (a) and human and mouse cyclin G (b). Cyclin G has not yet been described in Drosophila. As is clear from inspection of the  63  z34 c D N A  sequence  GAATTCCGGACGAGGCGATTTTTTG  GAAATAAGAAGCAAGAAAAGCAGAT  ACTGATCAAAACGCAGAGGCATCCG  100 *  *  *  GGTACTAGGCCAGCCCTCACAATGT M  CTGTCCCTGTACGCTACTCCTCTGC S V P V R Y S S A  TGCCGCCGAATACGCCGCCGAAGTT A A E Y A A E V>  200 *  *  *  GATTGTGAGTTGGAGAGCACTCTGC D C E L E S T L  AACAGCAGCAACAGTTGCACTTGCA Q Q Q Q Q L H L Q  ACAGCAATACGAGCAATACCAGCAC Q Q Y E Q Y Q H >  *  *  *  TACCAGTACCAACGGGAGCAGGATA Y Q Y Q R E Q D  TCGCCTACTATTGCCAGTTGCAGGC I A Y Y C Q L Q A  GGCGCGGCAGCAGGAGCAGTTGATG A R Q Q E Q L M >  CAGCAGAGGACATCGATGTCGTCGT  CGGTTATGCCAGGCCTAGCCTTGCC  CCAGGATCACCAGGACCACCCAGCC  300  Q  Q  R  T  S  M  S  S  S  V  M  P  G  L  A  L  P  Q  D  H  Q  D  H  P  A  >  400 *  *  *  GCCCTTTTGAACGGACCCCACAACA A L L N G P H N  ACAACATCGGACTCGCCATGGACGC N N I G L A M D A  CCACAGCATCAACGCCATTCTGGTC H S I N A I L V >  500 *  *  *  GACGACGAGCAGGCCTCGACTTCGG  CCCAGGCTGCCGCCGCTGCTGCCGC  ATCCGCGGGTGGATCTGCTGGTGCG  D  D  E  Q  A  S  T  S  A  Q  A  A  A  A  A  A  A  S  A  G  G  S  A  G  A> 600  *  *  *  GGATCGGGATCGGGATTGGGTGGTG G S G S G L G G  CTATCCGTGGGGGCAAGCTGGGCAA A I R G G K L G N  CGCGATTAACCGCAATGCAGAGATG A I N R N A E M>  CCAACTGATTGGATGAGGATTGCGG  ACGAGGGCCGGTATGGGACACCGGG  TGCTGCTGGCTTGGAATATCAGAAG  P  T  D  W  M  R  I  A  D  E  G  R  Y  G  T  P  G  A  A  G  L  E  Y  Q  K>  700 * TACGAACAACAACAACAACTGGAGG Y E Q Q Q Q L E  * ATCTGGCGGAGTCCGAGGCAGGAGC D L A E S E A G A  * GTACGGTGGAGCCAGCAACAACAAC Y G G A S N N N >  800 *  *  *  GGCGAATCGTCGTCGTCCTTGAAAA G E S S S S L K  AGCTAGAGGATCAGCTGCACGCCCT K L E D Q L H A L  CACCTCGGACGAGTTGTACGAAACC T S D E L Y E T >  900  *  *  *  CTCAAGGAGTACGACGTCCTGCAGG  ACAAGTTCCACACGGTGCTGCTGTT  GCCCAAGGAATCAAGGCGTGAGGTT  L  K  E  Y  D  V  L  Q Bam  D  K  F  H  T  V  L  L  L  P  K  E  S  R  R  E  V  >  *  *  *  ACTGCCGGAGGACGAGATGGATCCG  CCTACGTGCTGCGCTGCCTGAAGAT  GTGGTACGAGCTGCCCTCCGACGTC  T  A  G  G  R  D  G  S  A  Y  V  64  L  R  C  L  K  M  W  Y  E  L  P  S  D  V  >  z34 cDNA sequence, cont. 1000 CTGTTCTCGGCCATGAGCCTGGTGG L  F  S  A  M  S  L  V  ACCGCTTCCTGGATCGCATGGCCGT D  R  F  L  D  R  M  A  CAAGCCGAAGCACATGGCCTGCATG  V  K  P  K  H  M  A  C  M  >  1100 * AGCGTGGCCTCGTTCCACTTGGCCA S V A S F H L A  * TCAAGCAGCTGGACTTGAAACCCAT I K Q L D L K P I  * TCCCGCCGAGGATCTGGTTACAATA P A E D L V T I > 1200  *  *  *  TCTCAGTGTGGTTGTACCGCTGGTG S Q C G C T A G  ATCTGGAACGCATGGCCGGCGTGAT D L E R M A G V I  TGCCAACAAGCTGGGCGTCCAGATG A N K L G V Q M>  GGACATGCACCGATCACTTCTGTGA  GCTACCTGCGCATCTACTACGCCCT  CTTCCGCAACTTGGCGAAGGAGATC  G  H  A  P  I  T  S  V  S  Y  L  R  I  Y  Y  A  L  F  R  N  L  A  K  E  I  >  1300 * GGCGGCGACTTTTTCAAGTTCTACC G G D F F K F Y  * AGCAGCTCATCAAGCTGGAGGAACT Q Q L I K L E E L  * GGAGAACCGCCTGGAGATCCTGATG E N R L E I L M >  1400 * TGCGACGTGAAGACCACGGTGATCA C D V K T T V I  * CGCCCTCGACGCTGGCGCTGGTGCT T P S T L A L V L  * CATCTGCCTGCACCTGGACTTCCAC I C L H L D F H> 1500  *  *  *  ATCAAGGAGTCGTACACCCGCGGCA I K E S Y T R G  GTCCGGAGCTGAACACTCTTCAATT S P E L N T L Q L  ACATTCTCTCCTGCAGCAGTACATG H S L L Q Q Y M >  AGGATTCCTGATCGCGTTTTCACCT  GCGGCTTCAGCATCGTTTCGGGTAT  TCTGTCCCATTACAACGGGCAGAAC  R  I  P  D  R  V  F  T  C  G  F  S  I  V  S  G  I  L  S  H  Y  N  G  Q  N  >  1600 * AAGGCGCCCTACAAGCAGCGGCTTG K A P Y K Q R L  * TCTGGAAGCTGTCCAGTCGCACGCT V W K L S S R T L  * GCGCGTCTTGCGCCCGATCAACCGC R V L R P I N R>  1700 *  *  *  TTCTCCTCCGACCTGCCCACCATTG AGGAAGGCATCCCCAACGCCCTCGA  CGATGGCCTGCGTTCTAGAACCGAG  F  S  S  D  L  P  T  I  E PEST  E  G  I  P  N  A  L  D  D  G  L  R  S  R  T  E>  -2.1 1800  *  *  *  AGTATTAGCTCCGAGGAGGAGGAGG S I S S E E E E  ACTGGCCCACTTCACCCATAATTCC D W P T S P I I P  AATTTTCGAACAATGTTAGTATCCA I F E Q C *  GCGACAGCCAGCAGCATTAGAGCAG  CAGGGTCAACAGCTGGGCAAAACTG  GAGCAGCGGACAAAAGGACCATCAA  1900  * CCAGACGTAATCGCGGGACACACCT  * GACACCTCACACCTCCTGCTTTCCT  65  * GGCACTAGGCATATCCCATAGCAGC  z34 cDNA sequence, cont. 2000 *  * AGCAGGAGGGTGGAAATCTCCGCAG  AGGATGTGTGTCCGTGAGAGCGTTC  * CGCGTGCGAGTGTAGTAGTGCTGGT 2100  *  *  *  TGAGGTAGAGTGTGTATATTTTGCT  AAAGTGCATCATAATTCTTTTGGCA  TACACACAAATGTGTATTTGGTAGC  *  *  *  GCGCTGGTTCTAACATTTAAGAAAC  TATTGAGATACGGAAATGTGGACGC  CAAGGTGACGCAGCCACGCCCACCC  2200 *  *  *  CTTTTGAAAAGTTACCCGAGAGTAG  CGCCCCCACATTTTTTCGATTTATT  TGCGGCAGACAAAAAAATACGTGAA  AAAAAAACAAAAGCAAAAAAAAAAA  AAAAAAAAAA  Figure 3.2 Sequence of the z34 cDNA. The D N A sequence is given above, and the conceptual translation below. The internal BamHI site used to generate the AB construct is shown as a horizontal bar. The PEST sequence is underlined and the PESTfind algorithm score given below.  66  z40 cDNA sequence AAAAAGAACACAAAAGTTATAATTC  TGTCATTTAATTGTCTGCAAAATAG  GACTTTCCAGCATCATTGCATGTAG  100 *  *  *  TTAAATCCAGATTTCTGAAAAACCT  TATTTGGCAGCGATATGTTTAGCCA  AAAAATAGGAATTTGCGGATAAATT  200 *  *  *  ACCCCCCGGAGTGACAAAGATCTGT  ATTGTCATACAGACGCCTGCAGAAA  AAAGGGTTTTACAAGAAGAAAATGC  *  *  *  GCTATTTTGCAAAGGAATGATTTCT  AGAATTCAACGTCAAATATAATGGA M D  TAACATAGTCTACGACTTTGCCAAG N I V Y D F A K>  300  *  *  *  ATCACGTTTCAAGCAAAAGACAACA  GGTCACCCACTACTAACTCGAATCT  GTCGTGGCAACTAAATCAGATGGCT  I  T  F  Q  A  K  D  N  R  S  P  T  T  N  S  N  L  S  W  Q  L  N  O  M  A>  400 *  *  *  TTGTCGGACATGGAGGAGATGCAGG L S D M E E M O  ACACATCCGAGCCCATAGCTCCACC D T S E P I A P P  CGAATCCGATGACAATGTCAGCAGT E S D D N V S S >  PEST  +13  . 4 500  *  *  *  GAATCGCAAGACTCCGACGATGTGG E S O D S D D V  ACTCGCAATTGAGTCGCTGCGAGGA D S O L S R C E D  CAACGATGACGACAGCGATTGCATC N D D D S D C I >  *  *  *  AGTGGATCCTCCAGACGCAGTTCCA S G S S R R S S  CTTCTGGAGCTCGAGCGGGCGTGGC T S G A R A G V A  TCGTCGCAGAATGCCCGCCAGGGTG R R R M P A R V>  600  *  *  *  TCCAAGGACAACTTTAACCGGATCT  GGAGCGCCATCATGAAACCCATCAA  AAAGAAGCAACGCAAAGAGCTGAAC  S  K  D  N  F  N  .  R  I  W  S  A  I  M  K  P  I  K  K  K  Q  R  K  E  L  N  >  700 *  *  *  ACAAATGCCCAAACCCTTAAAAGCA T N A Q T L K S  TCGAAAGGATCCACACCAGCAGGCG I E R I H T S R R  CATGAAAAAGTTCACGCCCACCAAT M K K F T P T N>  800 * CTGGAGACAATCTTCGAGGAACCCA L E T I F E E P  * GCGATGAGAATGCCGCCGATGCAGA S D E N A A D A E PEST  * GGACGACAGCGAGGAGTGCTCCATC D D S E E C S I >  +11.7 900  *  *  *  AGCAGCCAAGTGAAAGTAGTTAAGG S S 0 V K V V K  TGTGGGGTCGCAAACTCCGCCGGGC V W G R K L R R A  AATATCCTTCAGCGATGGCCTGAAC I S F S D G L N>  6  7  z40 cDNA sequence, cont.  AAGAACAAAATCCTGTCGAAGAGAC K  N  K  I  L  S  K  R  GCCGCCAGAAGGTGAAGAAGACCTT R  R  Q  K  V  K  K  T  F  TGGCAAGCGTTTCGCACTCAAGAAA G  K  R  F  A  L  K  K>  1000 * ATCTCCATGACCGAGTTCCACGATC I S M T E F H D  * GTCTGAATAAGAGCTTCGACAGTGC R L N K S F D S A 1100  * GGGCGGGAGGATCGGCGGAGGCCGT G  R  E  D  R  R  R  P  * CATGCTGGAGGGGGATGATGCCAGA M L E G D D A R>  *  CAACATTCCCCAAGACATCCATGAC S  T  F  P  K  T  S  M  T  * CATGGAGGACATACAGCTGCCGACA M  E  D  I  Q  L  P  T  >  1200 *  *  *  ATGAGCAGCCAGCACCAGTTCTTCA M S S Q H Q F F  TGCAACCGGCGGGCTTTGAGTAGAG M Q P A G F E *  AGACTGAATGATCCATCAAATACGC  CCCACATTGATTTGCATTGCATTAA  AACTAGGTAAATAGTGCCCAAAAAT  AAATGTACTGATTACCAATTATGAA  1300  *  *  *  GTTAGGATTAACGTTCGTTTGGTTA  ACTTCTCACCTTAGTCTTAAGCCCC  ATAAAAGTTATAAATGAGTGTAAAT  1400 * AGCATGTAGAAGAAAAGAAAATAAG  * AGCTATACCTAGAGCTAAACTTATC  * CAGCCATAGAATACGATTCTGTGCT  TAGCCATTAAGATAATAAATAAAA  Figure 3.3 Sequence of the z40 cDNA. Note: nt 1-360 are from the Drosophila EST 1032424, Genbank accession: AA391083 nt 360-1449 are from the two hybrid interacting clone, z40. The D N A sequence is given above, and the conceptual translation below. The caret denotes the first in frame amino acid of the z40 construct from the interacting clone. PEST sequences are underlined with the PESTfind algorithm score given below.  68  a  M WY E P SD V D T S I L I DW L V E V S E E Y K T V S H K M R A V L I D M I N E V H L Q F H A A T E Y 0 K V F I F F A N V j Q V L O E Q L K[Lj R Q O V V I T P PM R K I V A E M M M E V C A E E N C Q E L Q P RM R A I L L D H L I E V C E V Y K | L ] H R T  7.34  V T A  A  I  It c; u K z34  A U  •» -ll -it  C I)  •»  K z34  A It c I)  K  G G [ R | DG S A Y V L R C L K  SH N  M A M A  -  V K D  -  M R  V K  P K H M A C M S V A S F H  V V R S K  L A  L Q L V O T A A M Y  t  K R T Y L Q L ; V O V T A L F  L  K N  V  R K T Q  I D P L L  N  S  K  S  A  H K V Q K T H  CG C D S Y DTY  L A P T C  0 I L A A  L  C  L O  I G  I K Q L I A A K Y E  k k  I A T  I L L A  S  A C L L L A  S  - L K P  I P A E D L V T  -  P  Y E  E -  - L F P  V  E -  - F G V I  E  S K L RJE -  A NKL L K I A R Q I R O M F KA Q S A I K T K F S Y A Y A O R TN Y T D N S I Y K D D L I K W L SRD - - G A C [ T ] E R D I L N HLEJK I L L LQ QAA -  y,34 mG2 NG1G2  D E -  I T C L F V A A K V  OV  K A Q V L R M  L F S[A]M S L V D R F LD R L Y s v FLYJL D R F L S 0 F 0 L A VA I I ^RLYJ Q V I A TJA T V F K]R F Y A R M | D R F LIS S V L L A L N FY L AV D p RIYILIH V  I  E  A  0  -  1 Y P  E V O E  P A  I  S  F V F L  T  I G D F V F  S N S R L  -  P S C R A L  -  I Y 0 P K 1 O E  I  I  C  V  V  V  T  S V D L L  E  E -  V L I H L L  s v s Y[L]R I Y Q MG HA A Y V F I N T A S F DL C S N C NL S I I L EC E G W D L S S VlTlP L D FlLlE L L M DWD I s  •  P SDV L SGT E T F FDT E T F  L  S[A1M  S  L A V N L A V N L  F A Y  VjD R F L D R F L D R F  L D R L A L  S  L  O 7.34  [MIAJV K P K H [ M A  mG2 hGlG2 41  M K wM.^wffi'* M K violP k H L  ™ mG2 811 hC.lC.2 8 i  7.34  7.34  F R N L A K E  734 238 mC2 2221 hGlG2 2 2 3 I z34 mG2  I G G D F F V F C H T  LOENLP  734 mG2 hGIG2 152 734 im I11G2 182 h(ilG2 183  o  FSITJL  P P  S V L A L  0 Y MR KHIL K H  K  IIP. L \ R ) V  'S; D.pr E YlN G | R ] D  R T | L  V S; V S:  A L  S V L A L  R R  RJV  TA )N T A RIO  R  L  - L K P I F A ] E  A  0  R MJA G y j T | A NfkJL II|.S E K RM E S P L MIR M E [W L E k  mG2 1191 hG[G2 u « |  -  E E EG D J'vHs > E E ER N  ®;i .A^g|p :  L D  1 E 0  A  E  0  G 0 O M G  v v  D  L V  p  PRTH D|V PJLIAIT D L  SO SO SO  T | I  I R  1  I R  I  K  H A P Ipflg V S Y j L ] R j j Y Y A L j L Y[HJA A L N F L L Y Y K|A T T A FQ F L r  r  r  i  r  L H Y E L E|A T T  ma: W K  K F Y Q Q L[JJK|L S[FR]K  T  S I  I L  EjE L ~ E 1 N R I U I E  S'LIETK L E A  L[E_RJR N S[JJN  F[EjR  I L MICIDIVI K T T V  Q L K A C  L E A O L K A C  C R VJV F S K C_RJI I F S K  L~T1c[LlH L P F H I K E S Y T R G S P E L N T [ L | Q L H S|L L Ol E L N L E V E L L E I L L V K T I A O rojA L E o v E L P T I E I G I E c JZQ\ F T ) C G  F S  jjy~slo  E L Y LEFIWIQ  E  L  I [ L | S H [ Y 1 N O P N K A J p ]Y KIOIRIL R  S K C L  V SK C  L  R P I N R F S S P.L P T H S S Y YS V P §L P T K H S Y Y R I T HL P T  I S S.E E - E E DW P T S. p I I p I F S C O E E S L S S S£ P s D Q E C T F  E YS E YS ILEJFG  r  PR NK i  K  E  N A L D D G L R  I P EGJG c  F D G S E S E D  1 p  E(MV[P1  R T M L V  VWIK  K L V WI N V [ Q K Llklw I  D L  S SD S  0 0  H  F F D F Q V A Q T L C F P P  R  T  E S  G E D  M  £  —  |  s  0= 1  O  j  Figure 3.4 z34 Sequence Alignments, (a) Comparison of z34 cyclin box sequence to Drosophila cyclins A - E . Boxes surround sequence identities within the group. Shading represents sequence similarity to z34. (b) Comparison of z34 cyclin box and following sequence to G cyclins from mammals; h, human; m, mouse. Numbering begins at the first residue of the cyclin box for each sequence given. Arrowheads mark the conserved alanines at putative interhelical crossing points. Asterisks mark two charged residues critical for cyclin A - C D K contact; z34 has an alanine substitution for the second of these. Circles identify the positions of residues that make the up hydrophobic pocket of cyclin A responsible for binding the central portion of the PSTAIRE helix. The vertical bar separates the cyclin-box domain (on the amino side) from the C-terminus of the protein, (c) Phylogenetic tree of cyclin box sequences generated by the cluster algorithm. z34 was compared to Drosophila cyclins A , B, C, D, and E , as well as mammalian cyclins A and G.  71  alignments (Figure 3.4a and 3.4b) or the phylogenetic tree generated using cyclin-box sequences (Figure 3.4c) z34 is more closely related to mammalian cyclin G than to any known Drosophila cyclin. z34 can be divided into three sequence domains. The amino domain contains the poly-glutamine and poly-alanine stretches, the central domain is the cyclin box, and the carboxyl domain contains sequence that is conserved with cyclin G but to a lesser extent than the cyclin box. Mouse cyclin G2, the mammalian G cyclin most similar to z34, is 35% identical and 73% similar in sequence over the cyclin box, and 24% identical and 55% similar over the sequence following the cyclin box. A BamHI site lies at the junction between the amino domain and the cyclin box, and was used to generate a deletion construct, z34AB, that contained only the amino domain. This construct interacted strongly with AsxC and AsxCl (Table 3.1). Asx therefore interacts with sequences amino to the cyclin box.  The cyclin box folds into a 5 helix bundle with short inter-helical distances [136]. In particular, the tight packing between the second and third helices requires alanine residues at their crossing points. These are conserved in z34 (arrowheads, Figure 3.4-b). The cyclin A-CDK2 interface consists of many interactions, with the focal point being the PSTAIRE helix of CDK2 [136]. Key among these are four hydrogen bonds involving two conserved cyclin residues, a lysine and a glutamate (asterisks, Figure 3.4b). The lysine donates two hydrogen bonds to the PSTAIRE helix, the glutamate accepts one hydrogen bond, and there is an additional hydrogen bond between the lysine and glutamate themselves. Whereas these two residues are conserved in human and mouse cyclin GI and G2 [137], z34 has an alanine substitution for the glutamate. In addition, the alanine is followed by glycine, suggesting that the pattern of interactions for a z34-CDK binding interface would be significantly different in this region from that of cyclin A-CDK2. This could mean that the z34-CDK interaction is weak, perhaps requiring a cofactor, or that the  72  C D K bound by z34 or z34 itself has compensatory substitutions that make up for the lack of hydrogen bonds involving this glutamate. It may also mean that z34 does not bind a C D K , however given the striking conservation of other residues throughout the cyclin box of z34, this is unlikely. Identified with circles in Figure 3.4b are the positions of residues that make the up hydrophobic pocket of cyclin A which binds the central portion of the PSTAIRE helix. The hydrophobic character is conserved in all but one of the five residues. While this nonconserved residue is a glutamate in z34, in the mouse and human homologues it is a leucine and a serine, respectively, indicating that even among the human cyclin Gs, this residue does not necessarily need to be hydrophobic.  An Asx-Cyclin G interaction? Cyclin G was originally cloned accidentally in a low stringency screen for src family kinases from rat. Its sequence was found to be most similar to the A cyclins, however it was shown to lack a PEST sequence, destruction box, or any other sequence implicated in protein turnover, and to have relatively constant mRNA levels through the cell cycle in a rat cell line [138]. Mouse cyclin G was discovered in a screen for genes activated by p53 in a leukemic cell line [139]. Cyclin G mRNA levels increased in response to an increase in p53 expression levels or y-irradiation (which induces p53) prompting the suggestion that cyclin G may regulate apoptosis. Two types of cyclin G were subsequently shown to exist in mammals, G I and G2 [137]. Both are tissue specific: GI is expressed at high levels in skeletal muscle, ovary and kidney, while G2 is expressed in cerebellum, thymus, spleen, kidney and prostate. In contrast to cyclin GI, cyclin G2 mRNA levels oscillate with the cell cycle, showing maximum expression in late S phase. The C-terminus of cyclin G2 contains a PEST sequence. z34 has a weak PEST sequence at its C-terminus, as determined by the PESTfind algorithm [140]. Cyclin G immunoprecipitates with the kinases CDK5 and G A K [141]. It also immunoprecipitates with the regulatory subunit of protein phosphatase 2A (PP2A-B'a), whereas other cyclins do not [142]. B'oc is  73  predominantly nuclear and is thought to play a role in the translocation/localization of PP2A in the nucleus [143] making it likely that cyclin G is also nuclear (no data on the subcellular location of cyclin G is yet published.) Whether cyclin G is concurrently or alternately associated with a kinase and a phosphatase, and whether these associations are generally applicable or specific to particular cell types, is unknown.  Two questions arise: "Is z34 Drosophila cyclin G?" and "What is the significance of an Asx-cyclin (cyclin G) interaction?" The answer to the former is probably yes. Until we better understand the function and properties of mammalian cyclin G, and are able to determine the same for z34, one cannot be sure. While z34 is more similar to the G type cyclins than to any other known proteins, there may exist a protein in Drosophila that is more similar to mammalian cyclin G. As Figure 3.4c shows, z34 and human cyclin G2 are significantly more divergent than are Drosophila and human cyclin A. In addition, the alanine substitution at a critical PSTAIRE-interacting residue raises the possibility that z34 will not show the same CDK-binding activity as mammalian cyclin G. On the other hand, I have done B L A S T searches for mouse cyclin G2-related sequences on the Drosophila  EST  database, and z34 ESTs (three as of February 10, 1998) are the most similar sequences in the database. Given that the cyclin boxes of cyclins A - E are all represented multiple times over in the EST database, it seems unlikely that there exists a more similar cyclin box sequence expressed in Drosophila.  The significance of an Asx-cyclin G interaction will remain somewhat unclear until the properties of cyclin G are better worked out. Assuming that Asx interacts with some type of cyclin, it would then recruit a C D K (or perhaps PP2A) to its chromosomal sites of action. At least three potential targets of such a kinase (or phosphatase) can be imagined. The first potential target is his tone HI. Phosphorylation of histone HI is a property of MPF, being accomplished by CDC2, and is in fact the classic test for C D K activity [144].  74  In the non-mitotic phases of the cell cycle, HI phosphorylation is thought to be associated with opening up of chromatin for transcriptional competence [145]. An HI kinase activity could then antagonize the predicted function of Asx, which is transcriptional repression. The second potential target is the R N A polymerase holoenzyme or initiation complex. There are in fact two CDKs already known to be associated with R N A polymerase: CDK7 and cyclin H are part of TFIJH [146; 147] and SRB10 and SRB11 are a kinase-cyclin panin the R N A polymerase U holoenzyme itself, required to phosphorylate the carboxylterminal domain (CTD) of pol II [148]. A cyclin G - C D K pair brought to a site of transcription could enhance or potentiate transcription by helping the pol II-resident CDKs, while a cyclin G-PP2A pair could suppress or prevent transcription by antagonizing the same. Finally, the PcG proteins themselves are likely targets for a C D K brought in by Asx. Many PcG proteins, including Asx itself, have serine, threonine, or serine/threonine repeats. Although it has not been shown, regulation of PcG protein activity through phosphorylation of these repeats is an attractive possibility. The PcG proteins would be the most likely targets of a C D K for another reason, however. PcG proteins have a definite requirement to be sensitive to the phase of the cell cycle, and thus could take advantage of a cell-cycle phase dependent kinase activity. Whereas the transcriptional repression of a PcG target is maintained through many cell divisions, rendering the polymerase or chromatin at that target no different from those at any other locus in their requirement for input from the cell cycle, the PcG protein complexes must duplicate themselves with each cell division in order to ensure this maintenance. Phosphorylation events associated with S phase would be very useful in triggering a change in a given complex that would ensure its replication following cell division. Indeed, the behaviour of PcG proteins is reported to be cell cycledependent. Some stay at their chromosomal sites through nuclear division, while others are shunted out of the nucleus, only to return after cell division is complete [72]. Phosphorylation and dephosphorylation could be initiating these translocations. Those that stay behind must in some way allow polymerase to pass through their sites, and then  75  reform after replication, spreading to both daughter strands. Such spreading, either along a D N A molecule, or from one to another, would be disastrous if not restricted to the appropriate phase of the cell cycle.  The Asx-z40 Interaction Although the sequence of z40 does not give any clues to its function, it is nevertheless the most salient of the interactors by virtue of the very strong interaction that it shows with Pc. This interaction addresses the question posed at the beginning of this chapter namely, "Can evidence be found supporting a physical connection between the complex containing ph, Psc, and Pc, and Asx?" The answer is yes, with z40 as the putative bridge. Evidence has since come out that Asx and Pc do colocalize. On polytene chromosomes, 64 of the 90 Asx binding sites reliably detected correspond to previously determined Pc binding sites, including the AntC and BxC loci [135]. Whether Asx or z40 are intimately associated with ph and Pc, or are peripheral remains to be determined. Given that the z40-Asx two-hybrid interaction is much weaker than the z40-Pc interaction, it may be that z40 is an integral component of complexes containing Pc, while Asx is more peripheral. Evidence has recendy been obtained for z40 being required for maintenance of homeotic gene expression boundaries. Homozygotes for a deficiency that uncovers the z40 locus at 65A show extensive ectopic expression of the homeotic gene Scr (Tom Milne, personal communication). If this phenotype can be narrowed down to the z40 locus itself, z40 would become the newest bonafidemember of the PcG, being required for homeotic gene regulation and showing an interaction with Pc.  Asx Interacts with the trx SET Domain Several PcG genes have been shown to be modifiers of position effect variegation. Asx is an E(var), E(Pc) is a Su(var) [77], and a transgene expressing the human homologue of  76  E(z) is an E(var) [78]. This pointed to Su(var)s and E(Var)s as potential candidate binding partners of Asx.  I therefore collaborated with Michael O'Grady to test his unpublished  two-hybrid panel of Su(var)s and E(var)s for interaction with the three Asx constructs as both DNA-binding fusions and as activator fusions. The O'Grady panel consisted of LexA fusions to Su(var)3-9 [149], and klett (discovered in a two-hybrid screen using Su(var)39) and activator fusions to klett, PP1 (the product of the Su(var)3-6 locus) [150], and trxC (the C-terminal 553 amino acids of trx which has a domain of high similarity to Su(var)3-9, but behaves as an E(var), Sarb Ner, personal communication). Two interactions were seen, both with Lex-AsxC. This fragment interacted weakly with the activator fusions to klett (which had independentiy been isolated as z38 in my library screen) and with the activator fusion to trxC (Figure 3.5a).  The trx interaction is very significant. In PcG/trxG double heterozygotes, homeotic transformations are suppressed, producing a wild type fly [51; 151]. The basis of this antagonistic behaviour is not understood at the molecular level. The Asx-trx interaction is the first evidence of a direct protein-protein interaction between the PcG and the trxG.  The trxC construct contains a sequence motif known as the SET domain. This domain is shared in common with another trxG gene, ash-1 [62], the PcG gene, E(z) [31], and Su(var)3-9 [149] from Drosophila. It is also present at the C terminus of ALL-1, the human homologue of trx [152-154]. A clue to the function of the SET domain comes from the yeast gene, SET1 [155] Strains lacking this SET domain-containing gene are defective in telomeric silencing. This defect is corrected by expressing a mini-gene that consists of the SET domain alone. It is also corrected by expressing hE(z), the human homologue of E(z) [78]. Another clue comes from ALL-1. Chromosomal translocations involving A L L 1 are seen in 80% of cases of infantile acute lymphoblastic leukemia [156] and are involved in many other leukemias [153]. These rearrangements replace the C terminus of the  77  protein, including the SET domain, with in-frame sequence from a variety of other genes (12 to date) [157]. Many of these other genes have been shown to be transcriptional activators [158; 159]. In fact a recombinant gene composed of the amino terminus of A L L 1 and a minimal transcriptional activation domain is sufficient for cellular transformation in vitro [160]. The final clue comes from the trxC construct itself. As a LexA fusion, this construct strongly activates transcription from LexA reporter genes (Figure 3.5e). These observations point to the SET domain being involved in regulating both chromatin and transcription, although its presence in both enhancers and suppressors of position effect, and activators and repressors of transcription makes it difficult to speculate what its mode of action is likely to be.  To further refine the interaction between Asx and trx, I generated three smaller constructs: AsxCl (aa 1139-1420), AsxC2 (aa 1412-1669), and trxSET (the SET domain alone, aa 3608-3759). Figure 3.5a shows the Asx-trx interaction with the constructs Lex-AsxC and Lex-AsxCl assayed by measuring the frequency of leucine prototrophs (see below). Both interact strongly with trxC, and with trxSET, although in the case of Lex-AsxC, the latter interaction is significantly weaker than the former. In both cases, transcriptional activation is dependent on galactose (compare the upper bar with the lower bar in the graph for each experiment) meaning that they require expression of the activator fusion. These data point to an interaction between the SET domain and AsxCl.  Unfortunately, Lex-AsxC2 and Lex-trxSET both activated transcription alone, although not as strongly as trxC, and so could not easily be used to assay for interaction. In an attempt to circumvent the problem of transcriptional activation, I developed a limiting dilution assay to determine the frequency of cells that were converted to leucine auxotrophy. By comparing the number of colony forming units in a given drop spotted on leucine deficient galactose and dextrose plates with the number spotted on leucine supplemented plates, I  78  Figure 3.5 Asx-trx interactions, (a) Two-hybrid interactions between non-activating LexA-Asx fusion constructs and activator fusions to trx constructs. The graph shows the frequency of prototrophs on galactose (upper bar) and dextrose (lower bar). Dextrose shuts off transcription of the activator fusion. A D neg is the activator fusion plasmid with no insert, (b) Interactions bewteen LexA fusions to AsxC2, which activates alone, and activator fusions to trx constructs, (c) Interactions between Lex-trxSET and Asx activator fusions, (d) In vitro co-affinity precipitation with glutathione agarose of reticulocyte lysate in vitro translated Asx constructs and bacterially produced GST fusions to trx constructs, (e) f$gal assays to monitor transcription in the presence to two different LexA fusions.  80  could measure the relative effect on transcription caused by the expression of the activation domain fusion. With LexA-fusions that activate transcription alone, the self-activation may be unaffected by the binding of the second fusion, in which case activation levels will go up because now there are two activation domains at the operator, or the self-activation may be blocked by the binding of the second fusion, in which case the activation levels may go down, depending on the relative contributions of the first activation domain (which is now blocked) and the second, which is brought in by the interaction. However one would not expect activation levels to drop to zero, because even if the activation activity of the LexAfusion is completely blocked, the second fusion still carries its own activation domain.  Figure 3.5b shows the results of interactions with the Lex-AsxC2 construct. Activation caused by this fusion is actually reduced 18 fold by coexpression of the trxC activation fusion, and somewhat reduced (two fold) by coexpression of the SET domain activation fusion alone. This suggests that the AsxC2 can also interact with trxC, and that this interaction reduces the ability of AsxC2 to activate transcription. Whether the residual transcription in the presence of coexpressed trxC is due to residual activity of the LexAsxC2 construct, or due to the activation domain of the trxC activation fusion, or a combination of both, is impossible to tell. In any case these data suggest that AsxC2 can also interact with trxC, although the SET domain itself may not be sufficient for this interaction (the 50% reduction is not statistically significant).  Figure 3.5c shows the converse experiment, using activation domain fusions of Asx constructs and a LexA fusion of trxSET. In this conformation, trxSET-induced transcription is strongly enhanced with AsxC, as expected (upper panel) and also with AsxC2 (panel 3). The interaction with AsxC2 suggests that the statistically insignificant interaction from 5b, above, is the result of a real interaction, in other words that the AsxC2 domain can in fact interact with the SET domain. Unexpectedly, no enhancement or  81  suppression was seen between trxSET and AsxCl (panel 2). However a lack of interaction in this assay is not strong proof of the absence of a physical interaction, since there are still colonies that come up on leucine deficient medium. In other words, there is transcription at the reporter locus. Depending on the interaction between the two transcriptional activators at the locus, transcription with both present where one may be blocking the other need not necessarily be at a different level than transcription with only one present.  To ascertain whether these interactions could be due to direct protein-protein contact, I performed an in vitro GST-fusion protein binding assay. In vitro translated AsxC binds to GST-trxC coupled glutathione agarose, but not to glutathione agarose coupled to GST alone (Figure 3.5d, panel 1). In vitro translated AsxCl binds to both GST-trxC and GSTtrxSET but not to GST alone (Figure 3.5d, panel 2) confirming that the minimally defined interaction, that between AsxCl and trxSET can occur directly.  The Transcriptional Consequence of an Asx-trx Interaction The genetics of Asx suggest that the interaction demonstrated above may have transcriptional consequences. One Asx allele, Asx , shows anterior homeotic pi  transformations typical of the trxG in addition to posterior transformations typical of the PcG [19]. Recent work has now shown that this allele when homozygous, in addition to enhancing the posterior transformations of Pc/+ flies, also enhances the anterior transformations of trx/+ flies. In addition, several Asx/trx double heterozygotes show enhanced anterior transformations (Tom Milne, personal communication). This is in opposition to the usual behaviour of PcG/trxG double heterozygous allelic combinations, which is cosuppression. One could think of Asx as behaving as a member of the PcG with respect to interactions with other PcG genes, but as a member of the trxG with respect to interactions with trx. While the data from Figure3.5b and c are a result of transcriptional interactions, because one of the constructs is fused to an exogenous activating sequence,  82  they do not isolate the influence of Asx and trx sequences per se on one another. I therefore generated a second set of LexA fusions using a vector with a different selectable marker, allowing the coexpression of two LexA-fusions in the same cell. Since Lex-trxC activates transcription, coexpressing Lex-Asx constructs gives a measure of the transcriptional interaction between Asx and trx, together at one locus. Figure 3.5e shows that the AsxCl and AsxC2 constructs have different effects on trxC-induced transcription. Lex-AsxCl is capable of enhancing trxC-mediated transcription to levels approximately twice those seen with coexpression with LexA alone or Lex-AsxC, while Lex-AsxC2 abolishes trxC-mediated transcription completely. Since there are 5 LexA sites at the operator, the most likely interpretation of these data is that both proteins are present together at the operator, and that Asx is blocking or enhancing trx through a physical association. This is backed up by the evidence from Figure 3.5b that AsxC2 can reduce trx-induced transcription even when fused to an activation domain instead of a DNAbinding domain. However, an alternative explanation would be that the LexA fusions compete for the operator, and that Lex-AsxC2 always wins. Such a competition bias could be the result of trx sequences occluding LexA, somewhat weakening its D N A binding ability, or it could be a result of a physical interaction between trx and AsxC2.  Although the system is heterologous, these data are compelling evidence of a functional interaction between Asx and trx in transcription. Considered together with the genetics of Asx, they suggest a model whereby Asx is a component of the system that integrates the repression signal of the PcG with the activation signal of the trxG. Depending on what inputs are received from other upstream proteins, Asx can either allow or suppress transcription mediated by trx at the various homeotic loci, this allowance or suppression being dependent on different domains of Asx. Furthermore, they suggest a mechanism for the oncogenic transformation caused by ALL-1-transcriptional activator fusions. If the SET domain-containing C-terminus of ALL-1 is a transcriptional activation domain that is  83  subject to regulation by hAsx, then replacement of that domain with a different transcriptional activation domain that is not subject to regulation by Asx would generate a protein that activates transcription inappropriately at various loci, including loci that govern cell division.  84  Chapter IV: PcG Functional Interactions in Yeast  The PcG had been shown to be essential for transcriptional repression of the homeotic genes of the BxC and AntC by genetic means, but at the outset of this work there was no evidence that this repression was direct. It had been shown that ph and Pc were present on salivary gland chromosomes, but since they were present near genes that were repressed (the BxC) as well as near genes that were expressed (ph and Pc) it was not known whether their presence alone near a target gene was sufficient for transcriptional repression. Early on in this work, I demonstrated that Pc, Psc, and ph could repress transcription directiy, simply by being targeted to a locus of transcription (see below). Having created a panel of PcG proteins for expression in yeast, I sought to use this panel to develop a functional assay system in yeast, with the long term goal of identifying non-PcG factors from yeast that are required for transcriptional repression by PcG proteins. The ability of these proteins to repress transcription in a heterologous system would mean either that they require no cofactors to repress, or that required cofactors are sufficientiy conserved between flies and mammals to allow their functional interaction with the Drosophila PcG proteins. If such cofactors could be identified, the mechanism of silencing by the PcG could be characterized as similar to one or another of the well defined silencing or repression systems of yeast.  Pc, Psc, and ph Can Repress Transcription Directly I set up a producer/reporter two plasmid system for use in the Drosophila tissue culture cell line SL2 [161]. The reporter plasmid consisted of the bacterial chloramphenicol acetyl transferase (CAT) gene under the control of the HSP70 promoter and upstream heat shock response elements, with 5 Gal4 binding sites 500 bp upstream (Figure 4.1a). Gal4 alone, Gal4-PcG fusions, and PcG nonfusions were expressed constitutively by the actin 5C promoter. SL2 cells were transiently transfected with both reporter and producer plasmids  85  Figure 4.1 Transcriptional repression by PcG proteins in transiently transfected SL2 cells, (a) Schematic diagram of the reporter construct, (b) C A T activity in heat-shocked SL2 cells cotransfected with the reporter plasmid and producer plasmids that express only full length cDNAs of ph, Pc, and Psc. Vector: producer plasmid with no insert, (c) C A T activity in heat-shocked SL2 cells cotransfected with the reporter plasmid and producer plasmids that express Gal4 (pG), or Gal4-PcG fusions.  87  together, subjected to heat shock, and assayed for transcription of the CAT gene. As Figure 4.1b and c show, ph, Pc, and Psc are all able to repress heat-shock activated transcription when targeted upstream of the HSP promoter via Gal4 fusion, but not when expressed as nonfusions. Bunker and Kingston have since done similar experiments in mammalian cells [162] and shown that LexA fusions to Pc, Psc, and Su(z)2 can repress transcription induced by a variety of activators.  Psc is a Transcriptional Repressor In Yeast Having yeast expression constructs for LexA-PcG fusions in hand, I tested them for transcriptional repression in yeast. The reporter plasmid in this case, JK1621 [163], has the lacZ gene under the control of the constitutive CYC1 promoter, with 5 LexA sites upstream. As shown in Figure 4.2a, Lex-Psc is able to repress transcription from the yeast CYC1 promoter, while Lex-ph, -Pc, -E(z), -esc, and -Pel are not. Lex-Psc is also able to repress transcription from an integrated CYC1 reporter with upstream LexA sites (Figure 4.2b).  If Psc represses transcription in yeast in a manner homologous to its repression in the fly, then this demonstration of function in yeast could be an important step forward in understanding the mechanism of PcG transcriptional repression, the reason being that transcriptional repression and silencing are well studied and better understood in yeast. In an attempt to identify other factors necessary for repression by Psc, I repeated the two plasmid repression assay in a variety of mutant strains defective for components of known silencing or repression systems.  The products of the SIR genes are required for transcriptional silencing of the silent mating type loci [84] and telomeres [85]. When assayed in sir' strains, Lex-Psc behaved as it did in wild type strains: transcription in the presence of Lex-Psc was always reduced relative to  88  Figure 4.2 Transcriptional repression in yeast by Lex-Psc. (a) LexA and various full length PcGLexA fusions monitored for their effect on transcription of a plasmid bearing a constitutive promoter with upstream LexA binding sites, (b) Transcription from a chromosomally integrated reporter gene driven by the CYC1 promoter with upstream LexA binding sites in the presence LexA or Lex-Psc.  90  Figure 4.3 Transcriptional repression by Lex-Psc in various mutant backgrounds. The left bar in each case is the level of reporter gene expression in the presence of LexA; the right bar is the level in the presence of Lex-Psc. (a) sir mutants and isogenic wild type, (b) hst (Homologous to Sir Two) mutants, including sir2 and isogenic wild type, (c) Histone H2A,H2B deficiency (which reduces the dose of H2A and H2B by half) and isogenic wild type, (d) Histone H4 (One copy of hhf'has been deleted, the other has been replaced by the silencing-defective mutant H4A4-29). (e) sin mutants and isogenic wild type, (f) srb mutants and isogenic wild type.  92  LexA alone (Figure 4.3a). However in the case of sir2, repression was not as strong as in the other sir mutants, or in the isogenic wild type strain.  SIR2 has four known homologues in S. cerevisiae, three of which are involved in telomeric silencing (HST1,3,4), and one of which (HST1) when overexpressed can complement sir2 [164]. Because of the potential mitigation of Psc repression seen in sir2, I repeated the assay in hst mutant strains (Figure 4.3b). Individually, and in the multiple mutant, sirl; hstl; hst2; hst3, none of the hst mutant genotypes had significantly different levels of repression, relative to an isogenic wild type genotype. Intriguingly, repression in the sir2 single mutant strain was again somewhat less severe than that seen in any other strain. The significance of this weak reduction is questionable however, given that it was not seen in the sir2; hstl; hst2; hst3, quadruple mutant.  Intact histones are required for silencing at the silent mating type loci and at telomeres, however neither a dose reduction of H2a/H2b nor a silencing-abolishing mutation of the amino tail of H4 prevented repression by Lex-Psc (Figure 4.3c,d). It would appear from these data that the repression due to Psc is not being mediated by the factors that are responsible for the maintenance of silent loci in S. cerevisiae. Since silenced loci in yeast are permanently devoid of transcription, it may not be surprising that Psc, which regulates loci that may be on or off, does not make use of silencing factors. Such silenced loci are more akin to transcriptionally inert heterochromatin than to the homeotic loci that are regulated by the PcG.  There are many factors that have been shown necessary for repression of regulated loci in S. cerevisiae. These fall into two broad classes: sequence-specific factors that are activated conditionally, and general factors that are recruited by activated sequence-specific factors. It is the latter class that is capable of shedding light on the mechanism of PcG silencing.  93  The products of the SIN genes are required for repression of the HO locus and other genes [165]. I tested Lex-Psc repression in sinl and sin3 mutants, and found that neither of these mutants relieved repression by Psc (Figure 4.3e). Finally I tested two components of the pol II holoenzyme, SRB10 and SRB11, that mediate transcriptional repression [166; 167]. These also had no effect on repression by Psc (Figure 4.3f).  Before ruling out the involvement of yeast cofactors, several other general repressors would need to be tested, however until such a cofactor is found, it remains a possibility that Psc is repressing transcription without the help of yeast cofactors. In order to prevent transcription, Psc would need to do more than merely bind upstream of a transcribed gene. Perhaps Psc is able to assemble into a higher order homomultimer that occludes the binding of activators to the CYC1 promoter.  Telomeric Effects of the PcG and trxG in Yeast Yeast telomeres are transcriptionally silent, but are derepressed by mutations in histones, SIR genes, the telomeric repeat-binding protein RAP1, and select other factors. SIR3 enhances telomeric silencing when tethered to a telomere via LexA in a strain with telomeric LexA binding sites [168]. Telomeric silencing in this strain can be assayed by monitoring variegation of expression of two sub-telomeric reporter genes: ADE2, which when not transcribed gives rise to red colonies, and URA3, which when expressed causes toxic sensitivity to the purine analogue, 5-fluoroorotic acid (FOA). Using this same strain, I assayed telomeric silencing in the presence of my panel of LexA fusions. Despite the lack of a demonstrated interaction for Psc, if any protein from the panel interacted with silencing factors, such an interaction could enhance or disrupt (perhaps in a dominant negative manner) the function of these factors at telomeres.  94  The PcG proteins tested were ph, Pc, Psc, E(z), AsxA, A s x C l , and AsxC2. I also tested the trx constructs, trxC and trxSET. In the case of the PcG, none of the LexA fusions enhanced or disrupted telomeric silencing. The trx constructs, however, did have a measurable effect. Compared to LexA alone, Lex-trxSET enhanced telomeric silencing moderately, expanding the extent of red sectoring and increasing the resistance of the strain to FOA. The larger construct, trxC reduced telomeric silencing in both assays (Figure 4.4). The enhancement of silencing with targeted trxSET is significant in light of the fact that mutants in the yeast SET1 gene cause derepression of telomeres and that this derepression is corrected by expression of the minimal SET domain of SET1 [155]. It would appear from the behaviour of Lex-trxSET that the SET domain is capable of interacting with and enhancing the activity of silencing factors. The opposite behaviour of the larger construct, trxC, which contains the SET domain, demonstrates that silencing enhancement by the SET domain can be overcome by other sequences. It is possible that in this case, the other sequences are interacting erroneously, generating a dominant negative disruption of silencing. However, at the chromosomally integrated LEU2 reporter gene, trxC acts as a strong activator (chapter 3). The blocking of silencing by trxC may then be due to its transcriptional activation potential overcoming the repressive environment of the telomere. Whatever the reason for trxC activating transcription and blocking silencing in yeast, in the fly trx is a bona fide activator, and as such, it is puzzling that it would contain a domain associated with silencing. It may be that the transcriptional activation activity of the C-terminus of trx is directed against nucleosomes or other higher order silencing complexes, and that the SET domain, with its ability to interact with silencing factors is the Trojan horse that brings trx to the silencing complexes that are slated for disruption.  E(z) has a SET domain, but had no effect on telomeric silencing. Since the minimal SET domain of E(z) was not tested, one can not directiy compare the behaviour of trx with E(z). However if the SET domain is a protein module that recognizes silencing factors, E(z) may  95  a  LexA Lex-TrxSET Lex-TrxC  b  LexA Lex-TrxSET  # # •  #a  <s*  • # * < •  Lex-TrxC  96  **•  J  1 1  Figure 4.4 Telomeric silencing effects of trx sequences, (a) A lOx limiting dilution series of the telomeric reporter strain bearing various LexA fusions plated out on medium lacking FOA. Variegated repression of the telomere manifests as red colony sectoring, (b) The identical lOx hmiting dilution series plated out on medium containing FOA. Colony growth is a measure of telomeric repression.  97  be using it to interact with these factors at the homeotic loci, helping to lock in the silenced state of this chromatin domain. In such a model of SET domain function, the genetic result that E(z) mutants enhance ash-1 and show trxG phenotypes in some tissues at certain times [169] would mean that E(z) can also unlock a silenced domain using the SET domain as a key.  Implications for the Mechanism of PcG Action The silencing assay and the repression assay appear to be giving two different results. In the silencing assay, one protein, trx, has the ability to interact with yeast silencing factors. In the other assay, Psc, which represses, does so independently of silencing factors. One interpretation would be that the mechanism of activation by the trxG is fundamentally different and unrelated to the mechanism of silencing by the PcG. To be rigorous though, the results tell us only that the mechanism of trx function in yeast is distinct from the mechanism of Psc function in yeast. If the paradigm is valid, these results actually invite the hypothesis that PcG/trxG repression/activation makes use of at least two distinct mechanisms, one involving histones, higher order chromatin structures, silencing factors and trx, the other which is mediated by Psc, not involving these. One group of PcG proteins may be involved in silencing, while another group may be involved in repression. The homeotic genes of the BxC and AntC do occupy a curious regulatory niche, requiring permanent silencing in certain tissues in order for determination to hold, yet not requiring ubiquitous silencing as do the silent mating type loci, which are never to be transcribed under any circumstances. In the years before the proliferation of PcG genes and the mass action/multimeric complex model of Locke et al. [28], the genetic interactions within the PcG were taken as evidence for multiple independent pathways to repression of the homeotic loci [74]. It may be time to revisit this interpretation.  98  Chapter V : Materials and Methods Subcloning I used linker PCR to generate appropriate restriction enzyme sites at the 5' and 3' ends of subcloned fragments. PCR products were digested, ligated into pBluescript, and sequenced to confirm the absence of PCR-induced mutations. They were then subcloned into pET28a (Novagen), pGEX-4T-l (Pharmacia), pEG202 [110], pJG4-5 [170], and p B T M l 16 [171] as EcoRJ/XhoI fragments. In the case of p B T M l 16, the Xhol site was ligated into a Sail site, destroying both. pBTM-ph, Pc, and Psc were exceptions, using instead EcoRI/BamHI digestion; see below. The standard PCR reaction contained the following: ljLLg template, 0.5|iL lOmg/mL acetylated BSA, 5|iL lOx buffer (NEB), 0.7(iL 25mM dNTPs,lLiL lOOmM MgSCU, 1|J.M each primer, lixL (2 units) Vent polymerase (NEB), and H2O to make 50LIL final volume, overlayed with 50|iL mineral oil. The temperature cycles were: 5 minutes at 95°C, 2x(l minute at 4°C, 1 minute at 72°C, 1 minute at 95°C), 7x(l minute at 45°C, 1 minute at 72°C, 1 minute at 95°C). I used a low number of cycles with a large amount of template (and an error-correcting polymerase) in order to minimize the chance of PCR-induced mutagenisis.  The template cDNAs used for these constructs were proximal ph: c4-l 1 [172], distal ph: c4-7 [99], RAE28: RAE-28[101] provided by Kazunori Shimada, BEB1: GSTboi2P[126] provided by Yasushi Matsui, Pc: Pc-12c [64] provided by Renato Paro, Psc: PscUIA [104] provided by Paul Adler, E(z) e32 [173] provided by Richard Jones, Asx: Asxfl provided by Don Sinclair.  Pc:  Primers Pc5 and Pc3 were used to generate the full length Pc EcoRI-ATG/BamHI  fragment. This fragment was subcloned into EcoRI/BamHI digested p B T M l 16 to create pBTM-Pc. AchrPc was created using the primers Pc208f and Pc3. The minimal chromobox-containing fragment was generated with the primers chr5 and chr3.  99  Psc:  I used the primers Psc5 and Psc3 to create the EcoRI-ATG/BamHI fragment PscAB  which contains amino acids 1-696. Full length Psc was created in all subsequent constructs by ligating the BamHI/BamHI fragment from the Psc cDNA into the BamHI site of PscAB. Psc constructs designated AN were deleted for the 3' sequence following the NotI site (corresponding to amino acid 1460) by NotI digestion and religation, which liberated a Nod fragment, and those designated AS were deleted for the 3' sequence following the Sail site (corresponding to amino acid 205) in the same way. I created PscHD with the primers Psc748f and Pscl 149r, ring with the primers Psc748f and Pscl005r, and H T H with the primers Pscl006f and Pscl 149r.  ph:  An EcoRI site was generated direcdy upstream of the first A T G of ph by PCR with  the primers ph5 and ph255r. This EcoRI/XhoI fragment replaced the 5' Eco/Xho fragment of c4-ll (full length proximal ph cDNA.) ph contains a BamHI site 3 codons before the stop codon. ph was subcloned as an EcoRI/BamHI fragment, ph constructs designated A N retain amino acids 1-1418, and are deleted for the 3' sequence following the Ncol site by Ncol digestion and religation, which liberated an Ncol fragment. Those designated AS retain amino acids 1-522, and are deleted for the 3' sequence following the first Sail site in the same way. I created phHD using the primers phD5 and phD3. I created HI by digesting phHD with Ncol, which liberates 3' sequence (corresponding to amino acid 1418 ff.) as an Ncol fragment, and recircularizing the plasmid. The construct H2 was created by cutting pETphHD with EcoRI and Ncol to remove the intervening sequence, filling in with Klenow, and religating, which regenerated the EcoRI site in the correct reading frame, as any fool can see. H2AC was created by PCR with the primers H2f and M D L . dH2 was created by PCR with the primers H2f and H2r, but with the distal cDNA, c4-7. pSAM was created with the primers SAMf and H2r; dSAM was created using the same primers, but with the distal cDNA, c4-7.  100  Sem:  The constructs Scm2 which encodes a GST fusion to amino acids 767-877 of Sem  and sS A M which encodes a pET fusion to amino acids 797-877 of Sem were provided by Jeffrey Simon.  Pel:  The construct Lex-Pel which encodes a fusion of full length Pel to LexA, was  provided by Rob Saint.  BEB1: The EcoRI/XhoI BEB1 S A M fragment was created with primers BEBf and BEBr.  RAE28: The EcoRJTXhoI RAE28 S A M fragment was created with primers RAEf and RAEr.  E(z):  I subcloned the 2.5kb Bglll/NotI fragment containing the entire E(z) orf into  BamHI/Notl cut pBluescript. The pBluescript polylinker contains an EcoRI site upstream of the BamHI site, which was in the correct reading frame after the BamHI/Bglll ligation, so E(z) was subcloned from this construct as an EcoRI/NotI fragment.  Asx:  I created AsxA with the primers Asx5 and Asx 1005r, AsxQ with the primers  Asxl831f and Asx3414r and AsxC with the primers Asx3415f and Asx3.  Gal4 Fusion Plasmids: The first GAL4 fusion construct, pGPc was generated by subcloning the Eco RI-Bam HI Polycomb PCR product into Eco RI-Bam HI digested p M l [174]. This generates a mammalian Gal4-Pc expression construct. Gal4-Pc was then subcloned as a Bgl II-Bam HI fragment into Bam HI digested phsNeoAct(Bam) [175]. This final construct, pGPc, has a single Eco RI site between Gal4 and Pc, and a single Bam HI site, at the 3' end of Pc. Upstream of Gal4 there is no restriction site as Bam HI has ligated into Bgl U. The constructs pGph and pGPsc were created by removing the Pc  101  sequence with Eco RI and Bam HI, and substituting ph or Psc DNA. The construct pG was created by digesting pGPc with Eco RI and Bam HI, blunting with Klenow and religating to remove the fragment containing Pc. The control non-fusions were created by cutting pHSNeoAct(Eco) [175] with Eco RI, and inserting the Eco RI cDNA of Pc, ph, or Psc.  CAT Reporter Plasmid: The reporter plasmid pG5hspCAT was created by inserting a Hind Ul-Bam HI fragment containing the hsp70 promoter and upstream P element sequence from the expression plasmid pNHT4 [176] into Xba I-Bam HI digested pG5BCAT [177]. The Hind III and Xba I overhangs were made compatable by filling in using only the nucleotides C and T for pG5BCAT (which leaves a T-C-5' overhang) and the nucleotides A and G for pNHT4 (which leaves a 3 - A - G overhang).  Mutagenesis The mutants W l A and I62D were close enough to the 5' and 3' ends of the S A M sequence to be incorporated into their respective end primer-linkers (of the same names) and created by linker PCR. For all other mutants, a two step PCR protocol was used. Overlapping forward and reverse mutant primers were synthesized and used in separate reactions with the appropriate forward or reverse end primer linker. Vent polymerase (NEB) was used in the standard reaction through the following cycles: 5 min at 96°C; 2x(l min at 96°C; 1 min at 4°C; 1 min at 72°C); 7x(l min at 96°C; 1 min at 42°C; 1 min at 72°C). The products of the first step were purified by phenol/CHCl3 extraction, 2% agarose T A E gel electrophoresis, and Qiaex gel extraction (Qiagen) and because their mutant ends overlapped, were used together as the template for the second step. The ligation of the two overlapping fragments was achieved with a single PCR using the forward and reverse end primer linkers using the following cycles: 2x(l min at 96°C; 1 min at 4°C; 1 min at 72°C); 7x(l min at 96°C; 1 min at 42°C; 1 min at 72°C). For the mutant construct H24L>4A, the  102  end primers were H2f and H2r, and the mutant primers were M L A and HLP. For all S A M domain mutants the end primers were SAMf and H2r. L33A used mutant primers L>Af and L>Ar. L41A used mutant primers L2>Af and L2>Ar. G52A used mutant primers G2>Af and G2>Ar.  Primers Bold face denotes a restriction site  Pc5  5'-GGAGCGAATTCATGACTGGTCGAGGCAAGG-3'  Pc3  5-GGGGGGGATCCCGACATTGTTTGGGTC-3'  Pc208f  5 '-CCC A TA T G A A T T C G A C A T C T A C G A A C A A A C G AAC-3'  chr5  5'-CCCATATGAATTCGATCCAGTCGATCTAGTGTAC-3'  chr3  5'-GTGGGGATCCGATGAGGCGGCGATCCAGGAT-3'  Psc5  5'-GGAGCGAATTCATGATGACGCCAGAATCG-3'  Psc3  5'-AACGACTTGAGGAACTCCGAC-3'  Psc748f  5-CGCATATGGAATTCAGGCCACGCCCCGTCCTTCTA-3'  Psc 1149r  5 ' - C G C C G G A T C C C T G G G G C G A C T C A T A A A C ACG-3'  Pscl005r  5'-GCGGCTCGAGTCATTCCCGTTCGTAAAGGCCCGG-3'  Pscl006f  5'-CCGCGAATTCCTGATGCGCAAAAGGGCCTTC-3'  ph5  5-GCGAATTCATGGATCGTCGTGCAT-3'  ph255r  5'-GGCCGCTCGAGCTGCTTGCCACCC-3'  phD5  5'-CCACGAATTCCCCAAGGCGATGATTAAG-3'  phD3  5'-GTGGGGATCCTCCTTAATGGACTCCACCTT-3'  H2f  5-CCGCGAATTCATGGCTGAGGAGGAGAT-3'  H2r  5'-CCGCCTCGAGTCACTCCTTAATGGACTC-3'  103  Primers, cont.  MDL  5'-CCGCCTCGAGTCACGCTTGGCCGTCGATCTCCT-3'  MLA  5'-ATCGACGGCCAAGCGGCTGCGGCGGCCAAGGAGAAGCA TTTGGTG-3'  HLP  5'-CGCTTGGCCGTCGATCTCCT-3'  SAMf  5'-GGCGGAATTCAGCAGCTGGAGTGTGGAC-3'  W>A  5'-CCGCGAATTCAGCAGCGCGAGTGTGGACGATGTC-3'  I4>D  5 ' - C G C C C T C G A G T C A C T T A T C G G A C T C C A C C T T G G C - 3'  L>Af  5 '-GGCC A A G C G G C T C T G T T G C T C AA G G A G - 3'  L>Ar  5 '-G A G C A A C A G A G C C G C T T G G C C G T C G AT-3'  L2>Af  5'-CTCAAGGAGAAGCATGCGGTGAACGCTATGGGC-3'  L2>Ar  5 '-GCCC A T A G C G T T C A C C G C A T G C T T C T C C T T G AG-3'  G2>Af  5'-GGCATGAAGCTGGCTCCAGCTCTTAAAATT-3'  G2>Ar  5'-AATTTTAAGAGCTGGAGCCAGCTTCATGCC-3'  RAEf  5-CCGCGAATTCCCTAGCCAGTGGAGC-3'  PvAEr  5-CGCCCTCGAGGGTCACTTAGGT-3'  BEBf  5'-CCGCGAATTCGCAGAGTTTTGGTCACCCGAA-3'  BEBr  5'-GCGGCTCGAGTCACTCTTTGATTTTTTCTATTTC-3'  Asx5  5'-CCGGGAATTCATGAAAACCATTACGCCG-3'  Asxl005r  5'-CCGGCTCGAGCTCGCCCCAGAAGGGCTC-3'  Asxl831f  5'-CCGGGAATTCATGATTTCGTTTTCTCAG-3'  Asx3414r  5 ' - G G C C G G A T C C T G A G A T G A T A T T T A G T G A-3'  Asx3415r  5'-CCGGGAATTCATGACGCGTCCTGCCAAT-3'  Asx3  5'-CCGGGGATCCGTTATCCACCTCATCTA-3'  104  Primers, cont.  JGf  5'-CTGAGTGGAGATGCCTCC-3'  Sequencing All sequencing was done on an automated sequencer (NAPS unit) using fluorescent dye termination. Sequence from PCR products was obtained after subcloning into pBluescript, and priming with 17 and T3 standard primers.  Sequencing ofhAsx The insert from cDNA 42515 of the Image Consortium (Genbank Accession T16795) was subcloned into pBluescript in three pieces: a 750 bp Hind/Hind fragment, a 550 bp Xho/Xho fragment, and a 420 Hind/Hind fragment. These were sequenced with T7 and T3 primers. The sequence of the three fragments as well as the end sequence from the ESTs were melded into a single 1505 bp contig.  Preliminary sequencing of the Asx-interacting two-hybrid clones Preliminary forward sequence through the EcoRI restriction site was obtained using the primer JGf.  Complete sequencing of the z40 interacting clone The entire insert was subcloned into pBluescript as an EcoRI/XhoI fragment and end sequence was obtained with T7 and T3 primers. This was augmented by sequencing a BamHI deletion derivative (BamHI site at 365 bp 3' of the EcoRI site). These sequences were melded into a single 1108 bp contig.  105  Complete sequencing of the z34 interacting clone The entire insert was subcloned into pBluescript as an EcoRI/XhoI fragment and end sequence was obtained with T7 and T3 primers. Several 5' deletion derivatives were generated using the enzymes BamHI, BxtXI, Pstl, and Xbal, which cut at positions 919, 1184, 1490, and 1714 respectively. This sequence was augmented with sequence from a Drosophila EST, HL02032 5' (Genbank accession: AA567945) and melded into a single 2286 bp contig.  GST-fusion protein expression and purification pGEX-4T-l derivative plasmids were transformed into the bacterial strain AD202. Single colonies were grown to an ODgoo of 0.6 in 250mL LB at 37 °C and induced with the addition of 250uL of I M IPTG. Induction was carried out for 15 hours at 25°C. Cells were collected by centrifugation, resuspended in 15mL of 20mM Tris Cl/100mM #  NaCl/lmg/mL lysozyme, and left at room temperature for 1 hour. 5(lL of B-mercapto ethanol was added, and the resuspended cells were subjected to 6 cycles of freeze/thaw with N2(i). The extract was cleared by centrifugation for 40 minutes at 12500 rpm (SS34) at 4°C, and filtered through miracloth.  In vitro co-affinity precipitations 35  S-methionine labeled proteins were generated using the Promega TNT rabbit reticulocyte  lysate transcription/translation kit according to the manufacturer's instructions. Templates were uncut plasmid DNA. cDNAs with appropriate initiator methionine codons were transcribed by T7 or T3 polymerase from pBluescript constructs, and inserts lacking an initiator methionine were transcribed by T7 from pET28a (Novagen) constructs which provided the initiator methionine. GST-fusion protein bound to glutathione agarose beads were prepared by incubating an aliquot of raw bacterial extract with 50|iL of a 50% slurry of reduced glutathione agarose (Sigma) in lOOmM NaCl/20mM Tris«Cl ph 7.5 (TBS), in  106  l m L of TBS/1%NP40/0.5%PMSF saturated isopropanol (PMSF) for 30 minutes with gentle rocking at 4°C. The amount of bacterial extract was normalized to give l|lg of fusion protein in each experimental tube. The bound beads were washed twice in TBS/1%NP40, and once in TBS. They were then blocked in a solution of 5% skim milk in TBS for 30 minutes at 4°C. The S-labeled proteins from the in vitro translation 35  reactions were precleared with the addition of GST-bound glutathione agarose in TBS, followed by incubation at 4°C with gentle rocking for 30 minutes. For each 200|iL in vitro translation reaction, 100|iL bed volume of glutathione agarose coupled to lOLtg of GST in a volume of 500|iL was used in the preclearing step. 70iiL of precleared lysate and 5LLL of 10% NP40 (to 0.1% final) was added to the blocking mixture in each experimental tube and these were incubated for 30 minutes at 4°C. The bound beads were washed twice in TBS/0.5% NP40, twice in 500mM NaCl/20mM Tris'Cl pH 7.5, and once in TBS, followed by elution in 30(iL TBS/20mM reduced glutathione pH7.5. The eluate was analyzed by Tricine SDS P A G E [178] on a 10%/16% discontinuous gel for labeled minimal S A M domains and SDS P A G E on a 12% gel for the larger phHD construct. 1/3 of the eluate was loaded in each experimental lane, and 2.5 LtL of the pre-bound lysate was loaded in the control lane.  Binding to Ni-NTA agarose 100|iL of in vitro translated S-labeled sSAM was mixed with 50LIL of Ni-NTA agarose 35  (Qiagen) in lmL TBS and rocked for 30 minutes at 4°C. The beads were then washed 3x in TBS, and eluted with lOOmM imidazole/TBS ph 7.5. Equal volumes of pre-bound translation reaction and bound/eluted polypeptide were loaded and run on 10%/16% discontinuous Tricine SDS P A G E and visualized by autoradiography.  107  Transformation and Culturing of Yeast Strains Yeast were grown nonselectively on YPD or selected on C M dropout medium lacking uracil, tryptophan, histidine, or leucine. For transformations, 50mL of fresh yeast culture at an OD600 of 1.0 were collected by centrifugation for 5 minutes at 2000 rpm at room temperature on a tabletop centrifuge (Clay-Adams) and resuspended in 40mL of drJ^O. Cells were pelleted again and resuspended in 1.5mL of freshly prepared lOOmM LiOAc/TE. 200uL of cells were added to glass tubes containing lu.g of the plasmid to be transformed plus 200|Hg of denatured herring sperm D N A as carrier. 1.2 mL of PEG solution (8 parts sterile 50%PEG/1 part IM LiOAc/1 part lOx TE) was added and the tubes were set turning at 3 0 ° C for 30 min. A 15 minute heat shock at 42°C was applied and yeast were plated directly onto selective plates.  Yeast Strains  EGY48  M A T a ura3-52 his3 trpl leu2::lexA-LEU2  JRY4012  M A T a canl-100 his3-ll leu2-3,112 lys2A trpl-1 ura3-l gal+  JRY4622  isogenic to JRY4012 sirl A: :LEU2  JRY4588  isogenic to JRY4012 sir2A::LEU2  JRY4606  isogenic to JRY4012 sir3A::LEU2  JRY4581  isogenic to JRY4012 sir4A::LEU2  IH2534  M A T a ura3-52 leu2 his3 trpl ade2 lys2 gal-  IH2536  isogenic to IH2534 sin3A::TRPl  IH2542  isogenic to IH2534 sinlA::TRPl  RMY202I  M A T a ade2-101 his3A200 lys2-801 trplA901 ura3-52 hhtl,hhfl::LEU2 HHT2,HHFA1-17  DY1571  M A T a ade2 canl his3 leu2 trpl ura3 lexA:UAS y :lacZ:URA3  DY1609  M A T a ade2 can 1 his3 leu2 trp 1 ura3 U A S y dexA:lacZ:URA3  c  c  108  c l  cl  Yeast Strains, cont.  FY250  MATcc his3A200 leu2Al ura3-52 trplA63  FY604  M A T a his3A200 leu2Al ura3-52 trplA63 (hta2-htab2)ATRPl  YPH680  M A T a ura3-52 his3A200 leu2Al trplA63 lys2A202  YCB428  isogenic to YPH680 sir2A2  YCB532  isogenic to YPH680 hstlA3::TRPl  YCB494  isogenic to YPH680 hst2A2::TRPl  YCB424  isogenic to YPH680 hst3A3::TRPl  YCB644  isogenic to YPH680 hst4Al::TRPl  YCB483  M A T unknown ura3-52 his3A200 leu2Al trplA63 lys2A202 ade2A sir2A2::TRPl hstlA3::TRPl hst2A2::TRPl hst3 hst3A3::TRPl  These strains were the gifts of the following people: EGY48 Erica Golemis[110]; FY250, FY604 Fred Winston [179]; JRY4012, JRY4581, JRY4588, JRY4606, JRY4622 Jasper Rine [180]; JH2534, IH2536, IH2542 Ira Herskowitz; YPH680, YCB428, YCB532, YCB494, YCB424, YCB644, YCB483 Carrie Baker Brachmann [164]; RMY202I, Randall K. Mann [181]; D Y 1571, DY1609 David StiUman  Two-Hybrid Interaction Assays The strain EGY48 was transformed with derivatives of plasmids EG202 and JG4-5 [110] encoding the LexA- and activator-fusion proteins respectively. Three individual transformed colonies from each plate were streaked out on both dextrose and galactose plates containing complete minimal medium lacking uracil, histidine, and leucine. Growth was scored after 4 days: a strong interaction was deemed to have occurred if the colonies reached 1mm in diameter. Plates with slower-growing colonies were scored as weak interactions, and the absence of growth indicated no interaction. For the quantitative assays  109  used in Chapter 3, an overnight culture was grown in galactose medium supplemented with leucine but lacking histidine, tryptophan, and uracil (3"). Cells (0.5 ml) were pelleted and resuspended in H2O (150 uL). A limiting dilution series (lOx steps, 10 ml per sample) was spotted onto leucine supplemented 3~galactose, leucine deficient 3~galactose, and leucine deficient 3~dextrose plates. Colony number was counted in the spot with the most colonies that could reliably be scored, and extrapolated to the first sample. Frequency of prototrophs was taken to be (cfu from leucine deficient 3" galactose) divided by (cfu from leucine suplimented 3" galactose plates).  P-galactosidase assays Yeast were grown in triplicate overnight in 2mL of selective medium, pelleted, and resuspended in l m L Z-buffer (60mM Na2HP0 ,40mM N a H P 0 , lOmM KC1, ImM 4  2  4  M g S 0 , 50mM P-mercaptoethanol, pH 7). 100uX was diluted in 900|iL dF^O to a 4  disposable cuvette for absorbance readings at 600nM. To the remaining 900|iL, one drop (25uL) of 0.1%SDS and two drops (50|iL) of CHCI3 were added. Samples were vortexed 15 sec and left 15 min at 30°C to equilibrate. In 10 second intervals, 200uL of 4mg/mL ONPG was added to each tube, and mix by vortexing. Reactions were stopped in the same order in the same 10 sec intervals by adding 500(iL of lMNa2C03. The samples were pelleted for 5 minutes at max. on a benchtop centrifuge, and the supernatent saved for absorbance readings at 420nM. Units of (3-galactose were equal to: 1000(OD42o)/(t)(v)(OD60o)> where v is the volume (in mL) of yeast used (.9 in this protocol) and t is the time of the reaction in minutes. This protocol is modified from Breeden and Naysmith [182].  Co-immunoprecipitation from Kc nuclear extracts Nuclear extracts were prepared from 2L of Kc cells at a cell density of 2xl0 cells/mL 6  according to Heberlein et al. [183] and Parker & Topoi [184] Antibody to Pc was kindly  110  provided by Dr. Jacob Hodgson. 2|iL of pre-immune serum was added to 200 LtL of nuclear extract and incubated at 4°C with gentle rocking for 30 minutes. 80|iL of a 50% slurry of proteinA sepharose in H E M G (25mM HEPES-K+ pH 7.6, lOOmM KCI, 12.5mM M g C b , O.lmM E D T A , O.lmM E G T A , 15% glycerol, 1.5mM DTT) was added and the tube was rocked for a further 60 minutes. The beads were removed by centrifugation, and the cleared extract divided evenly between two tubes containing equal amounts of IgG, either 0.5^tL of pre immune serum or lfiL of affinity-purified anti-Pc antibody. The antibody was bound for 60 minutes, 20LLL of 50% proteinA beads were added and bound for 30 minutes. The bound beads were then washed 3x in H E M G and eluted with SDS P A G E loading buffer, run on an 8% gel, transferred to nitrocellulose and blocked in 3% BSA. The filter was then cut into high and low molecular weight pieces, and the bottom was probed with the same anti-Pc antibody, while the top was probed with anti-ph [172] and anti-Psc [33].  SL2 Transfections SL2 cells were seeded into 6x3 mL well plates and grown to 50% confluency in Schneider's Medium + 10% FCS. Transfections were carried out with Lipofectin from Gibco (Cat. No. 18292-011) according to the manufacturer's directions. Each experiment was performed in triplicate. Each replicate contained 0.11 Ltg of G5HSPCAT reporter, lttg of producer plasmid.  C A T assay 24 hours after transfection, cells were exposed to a lhr heat shock (37°C) and left to recover overnight at room temperature. The next day, cells were harvested, and C A T activity monitored using the Boehringer Mannheim CAT-ELISA kit (Cat. No. 1363 727) according to the manufacturer's directions.  Ill  Telomeric Variagation Assays Triplicate 2mL cultures were grown overnight. A limiting dilution series (lOx steps, 10 ml per sample, the same as for the two-hybrid quantitation, above) was spotted onto two sets of plates: one with 5-fluoroorotic acid (FOA) (lmg/mL), the other without. Colony number was counted in the spot with the most colonies that could reliably be scored, and extrapolated to the first sample. Frequency of FOA resistance was taken to be (cfu from the F O A plate) divided by (cfu from the control plates). To monitor ADE2 expression, plates with visible colonies were transferred to room temperature, and allowed to continue growing. Red colour appeared over a week.  112  Conclusion  The work described in this thesis has identified a network of protein-protein interactions within the PcG/trxG. Interprotein interactions, defined by domains that show interactions in one or more assays, can be described by a linear network: Scm-ph-Psc-Pc-z40-Asx-trx. E(z)-esc, and P e l do not as yet fit into this network. In addition to the heterotypic interactions, I have demonstrated homotypic interactions for Sem, ph, Psc, and Asx. A s more proteins are tested for interactions with the panel I have generated, the network may grow to include E(z)-esc and Pel, and w i l l probably change from a chain to a web of interactions.  The Assembly of PcG Complexes The network is not to be taken to describe the members of a particular complex and their interactions, rather it is a network of interaction possibilities. The constitution of a particular complex w i l l likely be governed by the availability of potential members, through the specificity of the various cis elements at the locus of complex assembly. The rather promiscuous interaction behaviour of the average P c G protein is consistent with the idea of higher order complexes brought about through multiple cooperative interactions. The constitution of a given complex w i l l likely further be subject to the effects of covalent modifications of monomers, allosteric interactions, steric hindrance, and cooperative assembly.  The following is a model for assembly of multimeric P c G complexes (Figure 6.1) based on some of the observations made in this thesis: P c G proteins are not synthesized in an assembly-ready state, and must be modified in some way to enable assembly. This is suggested by the absence of, or weakening of association when many of the interacting  113  Figure 6.1 A model for regulated PcG complex assembly, (a) Newly synthesized proteins are not assembly-competent. Cyt (cytoplasm); Nuc (nucleus). After nuclear import, proteins bind to their chromosomal target sites irrespective of transcriptional activity at the site, (b) Two possibilities exist: (left) the locus is transcriptionally active, in which case a transcriptional activator "A" directly or indirectly prevents assembly of the PcG complex, (right) The locus is transcriptionally silent, in which case a repressor "R" induces some modification (c) which renders the PcG proteins competent for complex assembly, (d) Once a complex has assembled, the inducing repressor is no longer needed. In addition, activators are now inhibited, either by occlusion of their cis elements as shown, or by being prevented from interacting with the basal transcriptional complex while bound to their cis elements.  115  domains that I have identified are expressed in the context of their full length proteins as opposed to in the absence of external sequences. Such suppression may be necessary to avoid the formation of complexes in the cytoplasm which would inhibit the nuclear import of the subunits. The modifications that allow assembly could happen before, after, or during (as a consequence of) recognition of DNA. Accepting for the moment Pirrotta's notion of a complex at the PRE inducing the recruitment of smaller, distant cis elements, complex assembly would happen after D N A recognition. The local presence of factors that are correlated with transcriptional activation (transcriptional activators, the holoenzyme, basal transcription factors, trxG proteins, or RNA) then block activation of assembly, while factors that are associated with repression (transient repressors such as hunchback protein) induce the activation of assembly. The activation event itself could be a covalent modification such as phosphorylation or dephosphorylation (given the presence of multiple serine/threonine rich domains in several PcG proteins) which would allow the factors present at the various cis elements to engage one another, catalysing the further inclusion of distant cis elements. This model allows PcG proteins to be present at a locus without necessarily repressing transcription there, as is seen for several loci in larval salivary gland nuclei. It also resolves the apparent inconsistency of the ubiquitous presence of PcG proteins with the limited domains of expression and repression of their target genes. Finally it makes two predictions. One, that some PcG proteins should be present in multiple isotypic forms, and two, that there should exist modifiers of PcG proteins, which if they are specific to PcG proteins, should themselves have PcG mutant phenotypes. The corollary of this latter prediction is that some PcG genes will encode enzymatic modifiers of PcG proteins. This has been suggested for E(Pc) [151].  Dimerization Crosstalk (transvection) between homologous chromosomes requires that a protein located at a particular site on one chromosome can bind to a protein located at the homologous site  116  Figure 6.2 Self-association of PcG proteins may facilitate inter-homologue interactions. (a) A locus bound by several proteins capable of heteromultimerizing, but incapable of homodimerizing. Protein interactions occur and are inimical to homologue synapsis. (b) A locus bound by proteins capable both of homodimerizing and of heteromultimerizing. Dimerization leads to homologue synapsis, which could enable transvection at the locus. Higher order complexes are also possible, with a lower degree of freedom than in case (a) due to obligatory synapsis.  118  on the homologous chromosome. If, as sequence homology makes likely, the same protein is bound at the same site on both homologues, then the ability of such a protein to dimerize will facilitate crosstalk, zeste, which mediates transvection, can bind D N A [185], and can self-aggregate [186]. I have shown that Asx, Psc, Sem, and ph have dimerization (or multimerization) modules. This means that the D N A target sites of these three proteins are potential homologue synapsis points. It invites the prediction that mutations that affect the dimerization of these proteins will also affect transvection at the BxC. Certain Psc and Son mutant alleles are already known to affect transvection at the white gene. Perhaps these genes affect zeste-white transvection because of a nearby PRE to which their products are bound, while the lack of effect on transvection at white by other PcG mutants means that this particular PRE is not a target site for these products.  PcG mutations cause a phenotype similar to gain of function mutations of the homeotics, so their requirement for transvection should be fairly easy to detect. Since the absence of transvection leads to loss of function at a locus showing pairing-dependent complementation, PcG mutations that suppress transvection should suppress this complementation, causing reduced expression, rather than their usual effect of causing increased expression of a given homeotic gene. An alternative test would be to examine transvection-suppressing rearrangements of the BxC for phenotypic enhancement of PcG heterozygotes. Finding an effect on transvection at the BxC would confirm the role of interhomologue interactions in PcG function at the BxC. It would mean that the Su(z) and E(z) phenotypes of PcG mutants and derepression of the homeotics are different manifestations of the same phenomenon.  Silencing or Repression? In chapter 4 it was suggested that some PcG/trxG factors might be involved in setting up a silenced chromatin conformation while others might be involved in conventional  119  transcriptional repression. In accordance with the model described earlier, two phases of PcG action could exist: the first competing directly with the transcriptional apparatus, the second locking in the silenced state. Silencing and repression could therefore both be occurring together at the homeotic loci. Given the large number of PcG proteins, it would not be surprising at all if the PcG restricted expression of target genes via multiple mechanisms. One should not rule out the possibility that different components of a multimeric complex, or even different domains within a given protein, are themselves functioning in different pathways to repression or silencing.  The Large Membership of the P c G Notwithstanding the possibility of multiple mechanisms of repression or silencing, why are so many different PcG proteins needed? Jurgens [16] and Landecker et al. [187] estimate that there are around 40 Pc-enhancing loci. Assuming that the untested PcG proteins behave similarly to the ones that I have tested, what can the sheer multitude of members of this group tell us about how the group functions? Multiple factors must be interacting to i) initiate silencing, ii) maintain silencing, or iii) reverse silencing. Beginning with the least likely of the three, if it is true that maintenance of silencing requires a large number of factors, the mechanism must involve more than simple occlusion of enhancer or promoter binding sites. Even so, if silencing can be achieved at some loci without the presence of certain PcG proteins, why is it that these are needed at other loci? Using multiple factors where only a few are necessary seems unparsimonious. It is most likely that the multiplicity of PcG members is related instead to initiation (and reversal, at loci where it occurs) of silencing. If one assumes that PcG silencing activity at a locus is dependent on the presence of other transcription factors, and that direct interactions with these factors either initiate silencing or prevent it, then the sheer multitude of transcription factors capable of acting on PcG-regulated loci would invoke a large number of PcG proteins. Every locus has its own particular constellation of transcription factors, therefore every locus capable of  120  Figure 6.3 Different loci require different PcG proteins for silencing. Transcription factors are shown in black. A potential functional interaction between a given transcription factor and a given PcG protein is designated by an indentation in the PcG protein of the same shape as the transcription factor. Loci (a) and (b) share one transcription factor in common but each have two which are unique. In this case, each transcription factor is specific to a different PcG protein, hence each share one PcG protein in common, but have two which are unique. The interacting transcription factor and PcG protein are shown adjacent to one another, suggestive of direct binding, but need not be, especially if the functional interaction were something other than physical binding. The relationship between which transcription factors bind to a locus and which PcG proteins do is governed by the functional consequences of potential interactions. Different loci have different patterns of upstream transcription factor binding, hence different loci have different requirements for PcG proteins, although in thefinalanalysis the activity accomplished by the PcG proteins may be the same from locus to locus (shown here as aggregation occurring at both loci).  122  being silenced by the PcG would have its own corresponding constellation of PcG proteins (Figure 6.3). One of the predictions of this explanation is that PcG proteins, in addition to interacting with each other, should also interact with other transcription factors. This seems to be the case for Asx, given its interactions with the zinc-finger transcription factor, Bowel and with trx.  It is clear that identifying interactions within the PcG is just the beginning of understanding how PcG-mediated repression works. It will be necessary to continue the search for protein interactions beyond the limits of the PcG, perhaps through such means as were used in chapter 3 for Asx, until a connection (direct or indirect) is made with the transcriptional apparatus itself.  123  Nomenclature  ALL-1  Acute Lymphoblastic Leukemia locus 1  AntC  Antennapedia Complex  AP  anterior-posterior  Asx  Additional sex combs  BxC  Bithorax Complex  BEB1  BEM1-binding protein 1  bmi-1  B-cell-specific Molony murine leukemia virus insertion site 1  brm  brahma  co-IP  co-immunoprecipitate  CAT  chloramphenicol acetyl transferase  CDK  cyclin dependent kinase  cfu  colony-forming units  crm  cramped  CTD  carboxyl terminal repeat domain of pol II  E(Pc)  Enhancer of Polycomb  E(var)  Enhancer of PEV  E(z)  Enhancer of zeste  en  engrailed  esc  extra sex combs  EST  expressed sequence tag  FOA  5-fluoroorotic acid  GST  glutathione-S-transferase  HEPES  N-[2-hydroxyethyl)piperazine-N'-(2-ethanesulfonic  hE(z)  human E(z)  hph  human ph  124  acid)]  Nomenclature, cont.  HST  homologous to SIR2  HTH  helix-turn-helix  IP  immunoprecipitate  mel-18  melanoma-specific cDNA 18  mph  mouse ph  mxc  multi sex combs  Ni-NTA  mckel-nitrilotriacetic acid  NURF  nucleosome remodeling factor  Pc  Polycomb  PcG  Polycomb Group  Pel  Polycomblike  PEG  polyethylene glycol  PEV  Position Effect Variegation  ph  polyhomeotic  pho  pleiohomeotic  PMSF  phenylmethylsulfonyl fluoride  pol II  R N A polymerase II  PP2A  protein phosphatase 2A  PRE  PcG Response Element  Psc  Posterior sex combs  RAE28  retinoic acid elevated cDNA 28  RRM  RNA Recognition Motif  SAM  Self-Association Motif (formerly Sterile Alpha Motif)  See  Sex combs extra  Sem  Sex combs on midleg  125  Nomenclature, cont.  Scr  Sex combs reduced  SET  Su(var)3-9-E(z)-Trx  SIN  SWI-independent  SIR  Silent Information Regulator  SL2  Schneider Line 2  SPM  Scm-ph-mbt  SRB  Suppressor of R N A polymerase B  Su(var)  Suppressor of PEV  Su(z)2  Suppressor two of zeste  sxc  super sex combs  TBS  lOOmM NaCl/20mM Tris'Cl ph 7.5  Trl  Trithoraxlike  trx  tri thorax  trxG  trithorax Group  Ubx  Ultrabithorax  126  Bibliography 1. Lewis, E.B., A gene complex controlling segmentation in Drosophila, Nature, 276, 565-570, 1978. 2. Kaufman, T . C , Lewis, R.A. and Wakimoto, B.T., Cytogenetic analysis of chromosome 3 in Drosophila melanogaster. the homeotic gene complex in polytene chromosome interval 84A-B., Genetics, 94, 115-133, 1980. 3. Harding, K. and Levine, M., Gap genes define the limits of Antennapedia andbithorax gene expression during early development in Drosophila., EMBO J., 7, 205214, 1988.  4.  Zhang, C.-C. and Bienz, M., Segmental determination in Drosophila  conferred by hunchback, a repressor of the homeotic gene Ultrabithorax, Proc. Natl. Acad. Sci. USA, 89, 7511-7515, 1992. 5. Struhl, G. and Brower, D., Early role of the esc gene product in the determination of segments in Drosophila., Cell, 31, 285-292, 1982. +  6.  McKeon, J. and Brock, H.W., Interactions of the Polycomb group of genes  with homeotic loci of Drosophila, Roux's Arch. Dev. Biol, 199, 387-396, 1991.  7.  Simon, J., Chiang, A. and Bender, W., Ten different Polycomb group  genes are required for spatial control of the abd-A and Abd-B homeotic products, Development, 114, 493-505, 1992.  8.  Soto, M.C., Chou, T.-B. and Bender, W., Comparison of germ-line  mosaics of genes in the Polycomb group of Drosophila melanogaster., Genetics, 140, 231243, 1995. 9. Struhl, G. and Akam, M.E., Altered distribution of Ultrabithorax transcripts in extra sex combs mutant embryos of Drosophila, EMBO J., 4, 3259-3264, 1985. 10. Jones, R.S. and Gelbart, W.M., Genetic analysis of the Enhancer of zeste locus and its role in gene regulation in Drosophila melanogaster, Genetics, 126, 185-199, 1990. 11. Struhl, G., A gene product required for correct initiation of segmental determination in Drosophila, Nature, 293, 36-41, 1981. 12. Duncan, I.M., Polycomblike: a gene that appears to be required for the normal expression of the bithorax and Antennapedia complexes of Drosophila melanogaster, Genetics, 102, 49-70, 1982. 13. Sato, T., Russell, M.A. and Denell, R.E., Homoeosis in Drosophila: a new enhancer of Polycomb and related homoeotic mutations, Genetics, 105, 357-370, 1983. 14. Ingham, P.W., A gene that regulates the bithorax complex differently in larval and adult cells of Drosophila, Cell, 37, 815-823, 1984.  127  15. Dura, J.-M., Brock, H.W. and Santamaria, P., Polyhomeotic: a gene in Drosophila melanogaster required for correct expression of segment identity, Mol. Gen. Genet., 198, 213-220, 1985. 16. Jurgens, G., A group of genes controlling the spatial expression of the bithorax complex in Drosophila, Nature, 316, 153-155, 1985. 17. Breen, T.R. and Duncan, I.M., Maternal expression of genes that regulate the bithorax complex of Drosophila melanogaster, Dev. Biol., 118, 442-456, 1986. 18. Adler, P.N., Martin, E.C., Charlton, J. and Jones, K., Phenotypic consequences and genetic interactions of a null mutation in the Drosophila Posterior Sex Combs gene, Dev. Genet., 12, 349-361, 1991.  19. Sinclair, D.A.R., Campbell, R.B., Nicholls, F., Slade, E. and Brock, H.W., Genetic analysis of the Additional sex combs locus of Drosophila melanogaster, Genetics, 130, 817-825, 1992. 20. Adler, P.N., Charlton, J. and Brunk, B., Genetic interactions of the Suppressor 2 ofzeste region genes, Dev. Genet, 10, 249-260, 1989.  21.  Girton, J.R. and Jeon, S.H., Novel embryonic and adult homeotic  phenotypes are produced by pleiohomeotic mutations in Drosophila, Dev. Biol, 161, 393407, 1994.  22. Docquier, F., Saget, O., Forquignon, F., Randsholt, N.B. and Santamaria, P., The multi sex combs gene of Drosophila melanogaster is required for proliferation of the germline, Roux's Arch. Dev. Biol., 205, 203-214, 1996.  23.  Yamamoto, Y., Girard, F., Bello, B., Affolter, M. and Gehring, W.,  The cramped gene of Drosophila is a member of the Polycomb-group, and interacts with mus209, the gene encoding Proliferating Cell Nuclear Antigen., Development, 124(17), 3385-3394, 1997.  24.  Cheng, N.N., Sinclair, D.A.R., Campbell, R.B. and Brock, H.W.,  Interactions of polyhomeotic with Polycomb group genes of Drosophila melanogaster., Genetics, 138, 1151-1162,1994.  25.  Campbell, R.B., Sinclair, D.A.R., Couling, M . and Brock, H.W.,  Genetic interactions and dosage effects of Polycomb group genes of Drosophila., Mol. Gen. Genet., 246, 291-300, 1995. 26. Kennison, J.A. and Russell, M.A., Dosage-dependent modifiers of homeotic mutations in Drosophila melanogaster., Genetics, 116, 75-86, 1987. 27. Phillips, M.D. and Shearn, A., Mutations in polycombeotic, & Drosophila Polycomb group gene, cause a wide range of maternal and zygotic phenotypes, Genetics, 125, 91-101, 1990. 28. Locke, J., Kotarski, M.A. and Tartof, K.D., Dosage-dependent modifiers of position-effect variegation in Drosophila and a mass action model that explains their effect, Genetics, 120, 181-198, 1988.  128  29. Franke, A., DeCamillis, M.A., Zink, D., Cheng, N., H.W. Brock and Paro, R., Polycomb and polyhomeotic are constituents of a multimeric protein complex in chromatin of Drosophila melanogaster, EMBO J., 11, 2941-2950, 1992.  30.  Lonie, A., D'Andrea, R., Paro, R. and Saint, R., Molecular  characterisation of the Polycomblike gene of Drosophila melanogaster, a trans-acting negative regulator of homeotic gene expression., Development, 120, 2629-2636, 1994. 31. Carrington, E.A. and Jones, R.S., The Drosophila Enhancer of zeste gene encodes a chromosomal protein: examination of wild-type and mutant protein distribution, Development, 122(12), 4073-4083, 1996. 32. Rastelli, L., Chan, C.S. and Pirrotta, V., 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, 1993. 33. Martin, E.C. and Adler, P.N., The Polycomb group gene Posterior sex combs encodes a chromosomal protein, Development, 111, 641-655, 1993.  34.  van der Lugt, N., Alkema, M., Berns, A. and Deschamps, J., The  Polycomb-group homologue Bmi-1 is a regulator of murine Hox expression, Mech. Dev., 58, 153-164, 1996.  35.  Akasaka, T., Kanno, M., Balling, R., Mieza, M.A., Taniguchi, M.  and Koseki, H., A role for mel-18, a Polycomb group-related vertebrate gene, during the anteroposterior specification of the axial skeleton, Development, 122, 1513-1522, 1996.  36. Takihara, Y., Tomotsune, D., Shirai, M., Katoh-Fukui, Y., Nishii, K., Motaleb, M., Nomura, M., Tsuchiya, R., Fujita, Y., Shibata, Y., Higashinakagawa, T. and Shimada, K., Targeted disruption of the mouse homologue of the Drosophial polyhomeotic gene leads to altered anteroposterior patterning and neural crest defects, Development, 124(9), 3673-3682, 1997.  37. Core, N., Bel, S., Gaunt, S.J., Aurrand-Lions, M., Pearce, J., Fisher, A. and Djabali, M., Altered cellular proliferation and mesoderm patterning in Polycomb-M33-deficient mice:, Development, 124, 721-729, 1997.  38. Alkema, M.J., Lugt, N.M.T.v.d., Bobeldijk, R . C , Berns, A. and Lohuizen, M.v., Transformation of axial skeleton due to overexpression of bmi-l in transgenic mice, Nature, 31 A, 724-727, 1995.  39.  Muller, J., Gaunt, S. and Lawrence, P.A., Function of the Polycomb  protein is conserved in mice and flies., Development, 121, 2847-2852, 1995. 40. Busturia, A. and Bienz, M., Silencers in Abdominal-B, a homeotic Drosophila gene, EMBO J., 12, 1415-1425, 1993. 41. Chan, C-S., Rastelli, L. and Pirrotta, V., A Polycomb response element in the Ubx gene that determines an epigenetically inherited state of repression., EMBO J., 13, 2553-2564, 1994.  129  42. Zink, B., Engstrom, Y., Gehring, W. and Paro, R., Direct interaction of the Polycomb protein with Antennapedia regulatory sequences in polytene chromosomes of Drosophila melanogaster., EMBO J., 10, 153-162, 1991. 43. Gindhart, J.G. and Kaufman, T.C., Identification of Polycomb and trithorax group responsive elements in the regulatory region of the Drosophila homeotic gene Sex combs reduced., Genetics, 139, 797-814, 1995.  44.  Simon, J., Chiang, A.C., Bender, W., Shimell, M.J. and O'connor,  M., Elements of the Drosophila bithorax complex that mediate repression by Polycomb group products., Dev. Biol, 158, 131-144, 1993.  45.  Fauvarque, M.-S. and Dura, J.-M., polyhomeotic regulatory sequences  induce developmental regulator-dependent variegation and targeted P-element insertions in Drosophila., Genes Dev., 7, 1508-1520, 1993. 46. Kassis, J.A., 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, 1994. 47. Strutt, H., Cavalli, G. and Paro, R., 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(12), 3621-3632, 1997.  48.  Zink, D. and Paro, R., Drosophila Polycomb-group regulated chromatin  inhibits the accessibility of a ?r<ms-activator to its target DNA., EMBO J., 14, 5660-5671, 49. Sigrist, C . and Pirrotta, V . , Chromatin insulator elements block the silencing of a target gene by the Drosophila polycomb response element (PRE) but allow trans interactions between PREs on different chromosomes., Genetics, 147(1), 209-221, 1997. 50. Kennison, J.A. and Tamkun, J.W., Dosage-dependent modifiers of Polycomb and Antennapedia mutations in Drosophila, Proc. Natl. Acad. Sci. USA, 85, 8136-8140, 1988. 51. Ingham, P.W., Differential expression of bithorax complex genes in the absence of the extra sex combs and trithorax genes, Nature, 306, 591-593, 1983. 52. Breen, T. and Harte, P., trithorax regulates multiple homeotic genes in the bithorax and Antennapedia complexes and exerts different tissue-specific, parasegmentspecific and promoter-specific effects on each, Development, 117, 119-134, 1993. 53. Shearn, A., The ash-1, ash-2 and trithorax genes of Drosophila melanogaster are functionally related, Genetics, 121, 517-525, 1989.  54. Tamkun, J.W., Deuring, R., Scott, M.P., Kissinger, M., Pattatucci, A.M., Kaufman, T.C. and Kennison, J.A., brahma: a regulator of Drosophila homeotic genes structurally related to the yeast transcriptional activator SNF2/SWI2., Cell, 68, 561-572, 1992. 55. Peterson, C L . and Herskowitz, I., Characterization of the yeast SWI1, SWI2, and SWI3 genes, which encode a global activator of transcription., Cell, 68, 573583, 1992.  130  56. Hirschhorn, J., Brown, S., Clark, C. and Winston, F., Evidence that SNF2/SWI2 and SNF5 activate transcription in yeast by altering chromatin structure., Genes Dev., 6(12(A)), 2288-2298, 1992.  57. Dingwall, A., Beek, S., McCallum, C , Tamkun, J., Kalpana, G., Goff, S. and Scott, M., The Drosophila snrl and brm proteins are related to yeast SWI/SNF proteins and are components of a large protein complex., Mol Biol Cell, 6(7), 777-791, 1995.  58. Farkas, G., Guasz, J., Galloni, M., Reuter, G., Gyurkovics, H . and Karch, F., The triihorax-like gene encodes the Drosophila G A G A factor., Nature, 371, 806-808, 1994.  59. Lee, H., Kraus, K., Wolfner, M. and JT, L., D N A sequence requirements for generating paused polymerase at the start of hsp70., Genes Dev, 6(2), 284-295, 1992.  60.  Croston, G., Kerrigan, L., Lira, L., Marshak, D. and Kadonaga, J.,  Sequence-specific antirepression of histone HI-mediated inhibition of basal R N A polymerase II transcription., Science, 251(4994), 643-649, 1991.  61.  Tsukiyama, T., Becker, P.B. and Wu, C , ATP-dependent nucleosome  disruption at a heat-shock promoter mediated by binding of G A G A transcription factor, Nature, 367, 525-532, 1994.  62.  Tripoulas, N., Lajeunesse, D., Gildea, J. and Shearn, A., The  Drosophila ashl gene product, which is localized at specific sites on polytene chromosomes, contains a SET domain and a PHD finger, Genetics, 143, 913-928, 1996. 63. Adamson, A. and Shearn, A., Molecular genetic analysis of Drosophila ash2, a member of the trithorax group required for imaginal disc pattern formation., Genetics, 144(2), 621-633, 1996. 64. Paro, R. and Hogness, D.S., The Polycomb protein shares a homologous domain with a heterochromatin-associated protein in Drosophila, Proc. Natl. Acad. Sci. USA, 88, 263-267, 1991. 65. Paro, R., Imprinting a determined state into the chromatin of Drosophila, Trends Genet., 6, 416-421, 1990. 66. Paro, R., Mechanisms of heritable gene repression during development of Drosophila., Curr. Op. Cell Biol, 5, 999-1005, 1993.  67.  Schlossherr, J., Eggert, H . , Paro, R., Cremer, S. and Jack, R.S.,  Gene inactivation in Drosophila mediated by the Polycomb gene product or by positioneffect variegation does not involve major changes in the accessibility of the chromatin fibre., Mol Gen. Genet., 243, 453-462, 1994. 68. Pirrotta, V. and Rastelli, L., White gene expression, repressive chromatin domains, and homeotic gene regulation in Drosophila., Bioessays, 16(8), 549-556, 1994. 69. Pirrotta, V., PcG complexes and chromatin silencing, Curr Opin Genet Dev, 7(2), 249-258, 1997.  131  70. Gutjahr, T., Frei, E., Spicer, C , Baumgartner, S., White, R.A.H. and Noll, M., The Polycomb-group gene, extra sex combs encodes a nuclear member of the WD-40 repeat family., EMBO J., 14, 4296-4306, 1995. 71. Messmer, S., Franke, A. and Paro, R., Analysis of the functional role of the Polycomb chromo domain in Drosophila melanogaster., Genes Dev., 6, 1241-1254, 1992.  72.  Buchenau, P., Hodgson, J., Strutt, H. and Arndt-Jovin, D., The  distribution of Polycomb-group proteins during cell division and development in Drosophila embryos: impact on models for silencing., J. Cell Biol, in press, 1998.  73. Alkema, M.J., Bronk, M., Verhoeven, E., Otte, A., Veer, L.J.v.t., Berns, A. and Lohuizen, M.v., Identification of Bmil-interacting proteins as constituents of a multimeric mammalian Polycomb complex, Genes and Development, 11, 226-240, 1997. 74. Struhl, G., Role of the esc gene product in ensuring the selective expression of segment-specific homeotic genes in Drosophila, J. Embryol. Exp. Morphol., 76, 297-331, 1983. +  75. Grigliatti, T., Position-effect variegation—an assay for nonhistone chromosomal proteins and chromatin assembly and modifying factors., Methods Cell Biol., 35, 587627, 1991. 76. Reuter, G. and Spierer, P., Position-effect variegation and chromatin proteins., Bioessays, 14, 605-612, 1992.  77. Sinclair, D.A.R., J., C.N., Antonchuk, J., Milne, T.A., Stankunas, K., Ruse, C., Grigliatti, T.A., Kassis, J.A. and Brock, H.W., Enhancer of Polycomb is a Suppressor of Position-Effect Variegation in Drosophila melanogaster, Genetics, 148, 211-220, 1998.  78. Laible, G., Wolf, A., Dorn, R., Reuter, G., Nislow, C., Lebersorger, A., Popkin, D., Pillus, L. and Jenuwein, T., Mammalian homologues of the Polycomb-group gene Enhancer of zeste mediate gene silencing in Drosophila heterochromatin and at S. cerevisiae telomeres., EMBO J, 16(11), 3219-3232, 1997. 79. Wu, C., Transvection, nuclear structure, and chromatin proteins., J Cell Biol, 120(3), 587-590, 1993. 80. Lewis, E., The theory and application of a new method of detecting chromosomal rearrangements in Drosophila melanogaster., Am. Nat., 89, 73-89, 1954. 81. Wu, C.-t. and Howe, M., A genetic analysis of the Suppressor 2 of zeste complex of Drosophila melanogaster, Genetics, 140,139-181, 1995. 82. Wu, C.-T., Jones, R.S., Lasko, P.F. and Gelbart, W.M., Homeosis and the interaction of zeste and white in Drosophila, Mol. Gen. Genet., 218, 559-564, 1989.  132  83. Kalisch, W.E. and Rasmuson, B., Changes of zeste phenotype induced by autosomal mutations in Drosophila melanogaster, Heriditas, 78, 97-104, 1974. 84. Rine, J. and Herskowitz, I., Four genes responsible for a position effect on expression from HML and HMR in Saccharomyces cerevisiae, Genetics, 116, 9-22, 1987.  85.  Aparicio, O., Billington, B. and Gottschling, D., Modifiers of position  effect are shared between telomeric and silent mating-type loci in S. cerevisiae., Cell, 66(6), 1279-1287, 1991.  86.  Laurenson, P. and Rine, J., Silencers, silencing and heritable transcriptional  states, Microbiol Rev, 56, 543-560, 1992.  87. Moretti, P., Freeman, K., Coodly, L. and Shore, D., Evidence that a complex of SIR proteins interacts with the silencer and telomere binding protein RAP1, Genes Dev, 8, 2257-2269, 1994. 88. Varanasi, U., Klis, M., Mikesell, P. and Trumbly, R., The Cyc8 (Ssn6)-Tupl corepressor complex is composed of one Cyc8 and four Tupl subunits., Mol Cell Biol, 16, 6707-6714, 1996. 89.  Smith, R., Redd, M. and Johnson, A., The tetratricopeptide repeats of  Ssn6 interact with the homeo domain of a2., Genes Dev, 8, 2857-2867, 1995. 90. Treitel, M. and Carlson, M., Repression by SSN6-TUP1 is directed by MIG1, a repressor/activator protein, Proc Natl Acad Sci, 92, 3132-3136, 1995. 91. Cooper, J., Roth, S. and Simpson, R., The global transcriptional regulators SSN6 and TUP1, play distinct roles in the establishment of a repressive chromatin structure., Genes Dev, 8, 1400-1410, 1994. 92. Edmondson, D., Smith, M. and Roth, S., Repression domain of the yeast global repressor Tupl interacts directly with histones H3 and H4, Genes Dev, 10, 12471259, 1996. 93.  Herschbach, B., Arnaud, M. and Johnson, A., Transcriptional repression  directed by the yeast oc2 protein in vitro, Nature, 370, 309-311, 1994. 94. Thompson, C , Koleske, A., Chao, D. and Young, R., A multisubunit complex associated with the RNA polymerase CTD and TATA-binding protein in yeast, Cell, 73, 1361-1375, 1993.  95.  Kim, Y.-J., Bjorklund, S., Li, Y., Sayre, M. and Kornberg, R., A  multiprotein mediator of transcriptional activation and its interaction with the C-terminal repeat domain of R N A polymerase II, Cell, 77, 599-608,1994. 96. Carlson, M., GENETICS OF TRANSCRIPTIONAL R E G U L A T I O N IN Y E A S T : Connections to the RNA Polymerase H C T D , Annu Rev Cell Dev Biol, 13, 1-23, 1997.  133  97. Gunster, M.J., Satijn, D.P.E., Hamer, K.M., Blaauwen, J.L.d., Bruijn, D.d., Alkema, M.J., Lohuizen, M.v., Driel, R.v. and Otte, A.P., Identification and Characterization of Interactions between the Vertebrate Polycomb-Group Protein BMI1 and Human Homologues of Polyhomeotic, Mol. Cell. Biol, 17(4), 23262335, 1997.  98.  Reijnen, M.J., Hamer, K.M., Blaauwen, J.L.d., Lambrechts, C ,  Schoneveld, I., Driel, R.v. and Otte, A.P., Polycomb and bmi-1 homologs are  expressed in overlapping patterns in Xenopus embryos and are able to interact with each other, Mech. Dev., 53, 35-46, 1995.  99.  Hodgson, J.W., Cheng, N.N., Sinclair, D.A.R., Kyba, M . ,  Randsholt, N.B. and Brock, H.W., The polyhomeotic locus of Drosophila  melanogaster is transcriptionally and post-transcriptionally regulated during embryogenesis, Mech Dev., 66(1-2), 69-81, 1997.  100. Yu, H., Chen, J.K., Feng, S., Dalgarno, D.C., Brauer, A.W. and Schreiber, S.L., Structural basis for the binding of proline-rich peptides to SH3 domains., Cell, 76(Cell 76 (5): 933-945), 933-945, 1994. 101. Nomura, M., Takihara, Y. and Shimada, K., Isolation and characterization of retinoic acid-inducible cDNA clones in F9 cells: one of the early inducible clones encodes a novel protein sharing several highly homologous regions with a Drosophila polyhomeotic protein., Differentiation, 57, 39-50, 1994.  102.  Bornemann, D., Miller, E. and Simon, J., The Drosophila Polycomb  group gene Sex comb on midleg (Scm) encodes a zinc finger protein with similarity to polyhomeotic protein, Development, 122, 1621-1630, 1996. 103. Ponting, CP., S A M : a novel motif in yeast sterile and Drosophila polyhomeotic proteins, Protein Sci, 4, 1928-1930, 1995. 104. Brunk, B.P., Martin, E . C and Adler, P.N., Drosophila genes Posterior sex combs and Suppressor two of zeste encode proteins with homology to the murine bmi1 oncogene, Nature, 353, 351-353, 1991.  105.  van Lohuizen, M., Frasch, M., Wientjens, E. and Berns, A . ,  Sequence similarity between the mammalian bmi-1 proto-oncogene and the Drosophila regulatory genes Psc and Su(z)2, Nature, 353, 353-355, 1991.  106. Tagawa, M., Sakamoto, T., Shigemoto, K., Matsubara, H., Tamura, Y., Ito, T., Nakamura, I., Okitsu, A . , Imai, K. and Taniguchi, M., Expression of novel DNA-binding protein with zinc finger structure in various tumor cells., JBiol Chem, 265(32), 20021-20026, 1990. 107. Irminger-Finger, I. and Nothiger, R., The Drosophila melanogaster gene lethel(3)Ah encodes a ring finger protein homologous to the oncoproteins MEL-18 and BMI-1, Gene, 163, 203-208, 1995.  108.  Pearce, J.H.H., Singh, P.B. and Gaunt, S.J., The mouse has a  Polycomb-like chromobox gene, Development, 114, 921-929, 1992.  134  109. Strutt, H. and Paro, R., The polycomb group protein complex of Drosophila melanogaster has different compositions at different target genes., Mol Cell Biol., 17(12), 6773-6783, 1997. 110. Golemis, E.A., Gyuris, J. and Brent, R., Two hybrid systems/interaction traps, in Current protocols in molecular biology, Ausubel, F.M., Brent, R., Kingston, R., Moore, D., Seidman, J., Smith, J.A. and Struhl, K., Eds., New York, John Wiley & Sons, 1993, 13.14.1-13.14.17. 111. Estojak, J., Brent, R. and Golemis, E.A., Correlation of two-hybrid affinity data with in vitro measurements, Mol Cell Biol, 15, 5820-5829, 1995.  112.  Peterson, A., Kyba, M., Borneman, D., Morgan, K., Brock, H. and  Simon, J., A Domain Shared by the Polycomb Group Proteins Sem and ph Mediates Heterotypic and Homotypic Interactions., Mol. Cell Biol., 17(11), 6683-6692, 1997.  113. Wismar, J., Loffler, T., Habtermichael, N., Vef, O., Geissen, M., Zirwes, R., Altmeyer, W., Sass, H. and Gateff, E., The Drosophila melanogaster tumor suppressor gene lethal(3)malignant brain tumor encodes a proline-rich protein with a novel zinc finger., Mech Dev., 53, 141-154, 1995. 114. Rhodes, N., Connell, L. and Errede, B., STE11 is a protein kinase required for cell-type-specific transcription and signal transduction in yeast., Genes Dev, 4, 1862-1874, 1990. 115. Wang, Y., Xu, H.P., Riggs, M., Rodgers, L. and Wigler, M., byrl, a Schizo saccharomyces pombe gene encoding a protein kinase capable of partial suppression of the rasl mutant phenotype., Mol. Cell. Biol, 11, 3554-3563, 1991.  116. Wilson, R., Ainscough, R., Anderson, K., Baynes, C , Berks, M., Bonfield, J., Burton, J., Connell, M., Copsey, T., Cooper, J., Coulson, A., Craxton, M., Dear, S., Du, Z., Durbin, R., Favello, A., Fulton, L., Gardner, A., Green, P., Hawkins, T., Hillier, L., Jier, M., Johnston, L., Jones, M., Kershaw, J., Kirsten, J., Laister, N., Latreille, P., Lightning, J., Lloyd, C , McMurray, A., Mortimore, B., O'Callaghan, M., Parsons, J., Percy, C , Rifken, L., Roopra, A., Saunders, D., Shownkeen, R., Smaldon, N., Smith, A., Sonnhammer, E., Staden, R., Sulston, J., Thierry-Mieg, J., Thomas, K., Vaudin, M., Vaughan, K., Waterston, R., Watson, A., Weinstock, L., Wilkinson-Sproat, J. and Wohldman, P., 2.2 Mb of contiguous nucleotide sequence from chromosome in of C. elegans, Nature, 368(6466), 32-38, 1994.  117.  Schultz, J., Ponting, C P . , Hofmann, K. and Bork, P., S A M as a  protein interaction domain involved in developmental regulation., Protein Sci, 6, 249-253, 1997.  118.  Barr, M.M., Tu, H., Aelst, L.v. and Wigler, M., Identification of Ste4  as a Potential Regulator of Byr2 in the Sexual Response Pathway of Schizosaccharomyces pombe, Mol Cell. Biol, 16(10), 5597-5603, 1996.  119.  Boulukos, K., Pognonec, P., Rabault, B., Begue, A. and Ghysdael,  J., Definition of an Etsl protein domain required for nuclear localization in cells and DNAbinding activity in vitro., Mol. Cell Biol, 9, 5718-5721, 1989.  135  120. Klambt, C , The Drosophila gene pointed encodes two ETS-like proteins which are involved in the development of the midline glial cells., Development, 117, 163-176, 1993.  121. Golub, T., Barker, G., Bohlander, S., Hiebert, S., Ward, D., BrayWard, P., Morgan, E., Raimondi, S., Rowley, J. and Gilliland, D., Fusion of the T E L gene on 12pl3 to the AML1 gene on 21q22 in acute lymphoblastic leukemia., Proc Natl Acad Sci U S A, 92(11), 4917-4921, 1995.  1 2 2 . Carroll, M., Tomasson, M., Barker, G., Golub, T. and Gilliland, D., The TEL/platelet-derived growth factor beta receptor (PDGF beta R) fusion in chronic myelomonocytic leukemia is a transforming protein that self-associates and activates PDGF beta R kinase-dependent signaling pathways., Proc Natl Acad Sci USA, 93(25), 1484514850, 1996.  123. Jousset, C , Carron, C , Boureux, A., Quang, C.T., Oury, C , Dusanter-Fourt, L, Charon, M., Levin, J., Bernard, O. and Ghysdael, J., A domain of T E L conserved in a subset of ETS proteins defines a specific oligomerization interface essential to the mitogenic properties of the TEL-PDGFR6 oncoprotein, EMBO J., 16(1), 69-82, 1997.  124. Dura, J.-M., Randsholt, N.B., Deatrick, J., Erk, L, Santamaria, P., Freeman, J.D., Freeman, S.J., Weddell, D. and Brock, H.W., A complex genetic locus, polyhomeotic, is required for segmental specification and epidermal development in D. melanogaster, Cell, 51, 829-839, 1987.  125.  Bender, L., Lo, H.S., Le, H., Kokojan, V., Peterson, V. and  Bender, A., Associations among PH and SH3 domain-containing proteins and Rho-type GTPases in Yeast, J. Cell Biol, 133, 879-894, 1996.  126.  Matsui, Y., Matsui, R., Akada, R. and Toh-e, A., Yeast src homology  region 3 domain-binding proteins involved in bud formation, J. Cell Biol, 133, 865-878, 1996. 127. Choi, K., Satterberg, B., Lyons, D.M. and Elion, E.A., Ste5 Tethers Multiple Protein Kinases in the M A P Kinase Cascade Required for Mating in S. cerevisiae, Cell, 78, 499-512, 1994. 128. Kranz, J.E., Satterberg, B. and Elion, E.A., The M A P kinase Fus3 associates with and phosphorylates the upstream signaling component Ste5, Genes Dev, 8, 313-327, 1994.  129.  Marcus, S., Polverino, A., Barr, M. and Wigler, M., Complexes  between STE5 and components of the pheromone-responsive mitogen-activated protein kinase module, Proc. Natl. Acad. Sci. USA, 91, 7762-7766, 1994. 130. Printen, J.A. and Sprague, G.F.J., Protein-Protein Interactions in the Yeast Pheromone Response Pathway: Ste5p Interacts With All Members of the M A P Kinase Cascade, Genetics, 138, 609-619, 1994.  136  131. Paroush, Z., Finley, R.L., Kidd, T., Wainwright, S.M., P. W. Ingham, R., Brent and Ish-Horowicz, D., Groucho is required for Drosophila neurogenesis, segmentation, and sex determination and interacts directly with hairy-related bHLH proteins, Cell, 79, 805-815, 1994. 132. Hart, M., Wang, L. and Coulter, D., Comparison of the Structure and Expression of odd-skipped and Two Related Genes That Encode a New Family of Zinc Finger Proteins in Drosophila, Genetics, 144, 171-182, 1996. 133. Wang, L. and Coulter, D., bowel, an odd-skipped homolog, functions in the terminal pathway during Drosophila embryogenesis., EMBO J., 15(12), 3182-3196, 1996. 134. Stolow, D. and Haynes, S., Cabeza, a Drosophila gene encoding a novel RNA binding protein, shares homology with EWS and TLS, two genes involved in human sarcoma formation., Nucleic Acids Res., 23(5), 835-843, 1995.  135. Sinclair, D., Milne, T., Hodgson, J., Shellard, J., Salinas, C , Kyba, M., Randazzo, F. and Brock, H., 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(7), 1207-1216, 1998.  136. Jeffrey, P., Russo, A., Polyak, K., Gibbs, E., Hurwitz, J . , Massague, J. and Pavletich, N., Mechanism of C D K activation revealed by the structure of a cyclinA-CDK2 complex, Nature, 376, 313, 1995.  137. Home, M., Goolsby, G., Donaldson, K., Tran, D., Neubauer, M. and Wahl, A., Cyclin G I and Cyclin G2 Comprise a New Family of Cyclins with Contrasting Tissue-specific and Cell Cycle-regulated Expression., J. Biol. Chem., 271(11), 6050-6061, 1996.  138. Tamura, K., Kanaoka, Y., Jinno, S., Nagata, A., Ogiso, Y., Shumizu, K., Hayakawa, T., Nojima, H. and Okayama, H., Cyclin G: a new mammalian cyclin with homology to fission yeast Cigl, Oncogene, 8, 2113-2118, 1993. 139. Okamoto, K. and Beach, D., Cyclin G is a transcriptional target of the p53 tumor suppressor protein, EMBO J., 13(19), 4816-22, 1994.  140.  Rechsteiner, M. and Rogers, S., PEST sequences and regulation by  proteolysis, Trends Biochem. Sci., 21(7), 267-271, 1996.  141.  Kanaoka, Y., Kimura, S., Okazaki, I., Ikeda, M. and Nojima, H.,  G A K : a cyclin G associated kinase contains a tensin/auxilin-like domain., FEBS Lett., 402, 73-80, 1997.  142. Okamoto, K., Kamibayashi, C , Serrano, M., Prives, C , Mumby, M. and Beach, D., p53-Dependent Association between Cyclin G and the B' Subunit of Protein Phosphatase 2A, Mol. Cell. Biol., 16(11), 6593-6602, 1996. 143. Tehrani, M., Mumby, M. and Kamibayashi, C , Identification of a novel protein phosphatase 2A regulatory subunit highly expressed in muscle., J. Biol. Chem, 111, 5164-5170, 1996.  137  144.  Arion, D., Meijer, L., Brizuela, L. and Beach, D., cdc2 is a component  of the M phase-specific histone HI kinase: evidence for identity with MPF., Cell, 55(2), 371-378, 1988.  145.  Lu, M., Mpoke, S., Dadd, C. and Allis, C , Phosphorylated and  dephosphorylated linker histone HI reside in distinct chromatin domains in Tetrahymena macronuclei., Mol Biol Cell, 6(8), 1077-1087, 1995.  146. Shiekhattar, R., Mermelstein, F., Fisher, R., Drapkin, R., Dynlacht, B. , Wessling, H., Morgan, D. and Reinberg, D., Cdk-activating kinase complex is a component of human transcription factor TFIIH., Nature, 374(6519), 283-287, 1995.  147.  Serizawa, H., Makela, T., Conaway, J., Conaway, R., Weinberg,  R. and Young, R., Association of Cdk-activating kinase subunits with transcription factor TFIIH., Nature, 374(6519), 280-282, 1995.  148. Liao, S., Zhang, J., Jeffery, D., Koleske, A., Thompson, C , Chao, D., Viljoen, M., van Vuuren, H. and Young, R., A kinase-cyclin pair in the R N A polymerase II holoenzyme., Nature, 374(6518), 193-196, 1995.  149.  Tschiersch, B., Hofmann, A., Krauss, V., Dorn, R., Korge, G. and  Reuter, G., The protein encoded by the Drosophila position-effect variegation suppressor gene Su(var)3-9 combines domains of antagonistic regulators of homeotic gene complexes., EMBO J., 13, 3822-3831, 1994. 150. Dombradi, V., Axton, J., Barker, H. and Cohen, P., Protein phosphatase 1 activity in Drosophila mutants with abnormalities in mitosis and chromosome condensation., FEBS Lett., 275(1-2), 39-43, 1990. 151. Kennison, J., The Polycomb and trithorax group proteins of Drosophila: transregulators of homeotic gene function., Annu Rev Genet., 29:, 289-303, 1995.  152.  Djabali, M., Selleri, L., Parry, P., Bower, M., Young, B. and  Evans, G., A trithorax-like gene is interrupted by chromosome llq23 translocations in acute leukaemias., Nat Genet, 2(2), 113-118, 1992.  153. Gu, Y., Nakamura, T., Adler, H., Prasad, R., Canaani, O., Cimino, G., Croce, C M . and Canaani, E., The t(4; 11) chromosome translocation of human acute leukemias fuses the ALL-1 gene, related to Drsophila trithorax, to the AF-4 gene, Cell, 71, 701-708, 1992. 154. Tkachuk, D.C, Kohler, S. and Cleary, M.L., Involvement of a homolog of Drosophila trithorax by llq23 chromosomal translocations in acute leukemias, Cell, 71, 691-700, 1992.  155.  Nislow, C , Ray, E. and Pillus, L., SET1, A Yeast Member of the  Trithorax Family, Functions in Transcriptional Silencing and Diverse Cellular Processes., Mol Biol Cell, 8(12), 2421-2436, 1997.  156. Cimino, G., Lo Coco, F., Biondi, A., Elia, L., Luciano, A., Croce, C , Masera, G., Mandelli, F. and Canaani, E., A L L - 1 gene at chromosome 1 lq23 is consistently altered in acute leukemia of early infancy., Blood, 82(2), 544-456, 1993.  138  157. Bernard, O. and Berger, R., Molecular basis of llq23 rearrangements in hematopoietic malignant proliferations., Genes Chromosomes Cancer, 13, 75-85, 1995.  158. Prasad, R., Yano, T., Sorio, C , Nakamura, T., Rallapalli, R., Gu, Y., Leshkowitz, D., Croce, C. and Canaani, E., Domains with transcriptional regulatory activity within the ALL1 and AF4 proteins involved in acute leukemia, Proc. Nat. Sci. USA, 91, 12160-12164, 1995. 159. Rubnitz, J., Morrissey, J., Savage, P. and Ceary, M., E N L , the gene fused with HRX in t(ll;19) leukemias, encodes a nuclear protein with transcriptional activation potential in lymphoid and myeloid cells, Blood, 84, 1747-1752, 1994.  160.  Slany, R., Lavau, C. and Cleary, M., The oncogenic capacity of HRX-  E N L requires the transcriptional transactivation activity of E N L and the D N A binding motifs of HRX., Mol Cell Biol, 18(1), 122-129, 1998. 161. Schneider, I., Cell lines derived from late embryonic stages of Drosophila melanogaster., J Embryol Exp Morphol, 27(2), 353-365, 1972. 162. Bunker, C A . and Kingston, R.E., Transcriptional repression by Drosophila and mammalian Polycomb group proteins in transfected mammalian cells., Mol. Cell. Biol., 14, 1721-1732, 1994.  163.  Keleher, C , Redd, M., Schultz, J., Carlson, M. and Johnson, A.,  Ssn6-Tupl is a general repressor of transcription in yeast., Cell, 68(4), 709-719, 1992.  164. Brachmann, C , Sherman, J., Devine, S., Cameron, E., Pillus, L. and Boeke, J., The SIR2 gene family, conserved from bacteria to humans, functions in silencing, cell cycle progression, and chromosome stability., Genes Dev, 9(23), 28882902, 1995.  165.  Sternberg, P., Stern, M., Clark, I. and Herskowitz, I., Activation of  the yeast H O gene by release from multiple negative controls., Cell, 48(4), 567-577, 1987. 166. Kuchin, S., Yeghiayan, P. and Carlson, M., Cyclin-dependent protein kinase and cyclin homologs SSN3 and SSN8 contribute to transcriptional control in yeast., Proc Natl Acad Sci USA, 92(9), 4006-4010, 1995. 167. Wahi, M. and Johnson, A., Identification of Genes Required for a2 Repression in Saccharomyces cerevisiae, Genetics, 140, 79-90, 1995.  168.  Lustig, A., Liu, C , Zhang, C and Hanish, J., Tethered Sir3p Nucleates  Silencing at Telomeres and Internal Loci in Saccharomyces cerevisiae, Moll Cell Biol, 16(5), 2483-2495, 1996.  169.  Lajeunesse, D. and Shearn, A., E(z): a polycomb group gene or a trithorax  group gene?, Development, 122, 2189-2197, 1996. 170. Gyuris, J., Golemis, E., Chertkov, H. and Brent, R., C d i l , a Human GI and S Phase Protein Phosphatase That Associates with Cdk2, Cell, 75, 791-803, 1993.  171.  Bartel, P. and Fields, S., Analyzing protein-protein interactions using two-  hybrid system., Methods Enzymol, 254, 241-263, 1995.  139  172.  DeCamillis, M.A., Cheng, N., Pierre, D. and Brock, H.W., The  polyhomeotic gene of Drosophila encodes a chromatin protein that shares polytene chromosome binding sites with Polycomb, Genes Dev., 6, 223-232, 1992. 173. Jones, R.S. and Gelbart, W.M., The Drosophila Polycomb-group gene Enhancer ofzeste shares a domain of sequence similarity with trithorax, Mol. Cell Biol, 13(10), 6357-66, 1993. 174. Sadowski, Bell, B., Broad, P. and Hollis, M., G A L 4 fusion vectors for expression in yeast or mammalian cells., Gene, 118(1), 137-141, 1992. 175. Thummel, C , Boulet, A. and Lipshitz, H., Vectors for Drosophila Pelement-mediated transformation andtissueculture transfection, Gene, 74, 445-456, 1988. 176. Schneuwly, S., Klemenz, R. and Gehring, W.J., Redesiging the body plan of Drosophila by ectopic expression of the homeotic gene Antennapedia, Nature, 325, 816-818, 1987. 177. Lin, Y., Carey, M., Ptashne, M. and Green, M., G A L 4 derivatives function alone and synergistically with mammalian activators in vitro., Cell, 54(5), 659-664, 1988. 178. Schagger, H. and Jagow, G.v., Tricine-sodium dodecyl sulfatepolyacrylamide gel electrophoresis for the separation of proteins in the range from 1 to 100 kDa., Anal. Biochem., 166, 368-379, 1987.  179.  Hirschhorn, J., Bortvin, A., Ricupero-Hovasse, S. and Winston, F.,  A new class of histone H2A mutations in Saccharomyces cerevisiae causes specific transcriptional defects in vivo., Mol. Cell. Biol., 15(4), 1999-2009, 1995. 180. Pillus, L. and Rine, J., Epigenetic Inheritance of Transcriptional States in S. cerevisiae, Cell, 59, 637-647, 1989. 181. Mann, R. and Grunstein, M., Histone H3 N-terminal mutations allow hyperactivation of the yeast GAL1 gene in vivo., EMBO J, 11(9), 3297-3306, 1992. 182. Breeden, L. and Nasmyth, K., Regulation of the yeast H O gene, Cold Spring Harb Symp Quant Biol, 50, 643-650, 1985. 183. Heberlein, U., England, B. and Tjian, R., Characterization of Drosophila transcription factors that activate the tandem promoters of the alcohol dehydrogenase gene., Cell, 41(3), 965-977, 1985.  184.  Parker, C. and Topol, J., A Drosophila R N A polymerase II transcription  factor contains a promoter-region-specific DNA-binding activity., Cell, 36(2), 357-369, 1984. 185. Benson, M. and Pirrotta, V., The product of the Drosophila zeste gene binds to specific D N A sequences in white and Ubx., EMBO J, 6(5), 1387-1392, 1987. 186. Bickel, S. and Pirrotta, V . , Self-association of the Drosophila zeste protein is responsible for transvection effects., EMBO J, 9(9), 2959-2967, 1990.  140  187. Landecker, H . L . , Sinclair, D . A . R . and Brock, H . W . , A screen for enhancers of Polycomb and Poly comblike in Drosophila melanogaster., Dev. Genet., 15, 425-434, 1994.  141  Appendix A:  Interactor Preliminary Sequence  zl *  *  GAATTCGGCACGAGGCGGTCCGACA AGGACTCGATGCCATGTTACGGCAG CTTAAGCCGTGCTCCGCCAGGCTGT TCCTGAGCTACGGTACAATGCCGTC E F G T R R S D K D S M P C Y G S  *  TGACTTCCAGATCACCACATCGGCG ACTGAAGGTCTAGTGGTGTAGCCGC D F Q I T T S A >  100  *  *  CAGTGTGACGAGCGAAAGCTGTATG CCCGCAAGGAGGACATTCTGCATGA GTCACACTGCTCGCTTTCGACATAC GGGCGTTCCTCCTGTAAGACGTACT Q C D E R K L Y A R K E D I L H E  *  AGTACTGAACATGCTGCCTCTGCTG TCATGACTTGTACGACGGAGACGAC V L N M L P L L >  200  *  *  *  AAGCCGGGCAATGAGGAGGCCAAGC  TTATCTACCTGACCCTCATACCAGT  TGCCGTCAAGGACACCATGCAGCAA  TTCGGCCCGTTACTCCTCCGGTTCG AATAGATGGACTGGGAGTATGGTCA  ACGGCAGTTCCTGTGGTACGTCGTT  K  P  G  N  E  E  A  K  L  I  Y  L  T  L  I  P V  A  V  K  D  T  M  Q  Q  >  300  *  *  *  ATTGTGCCCACGGAGTTGGTGCAGC AGATCTTCTCGTACCTACTCATCCA TCCAGCTATCACCAGCGAGGACAGA TAACACGGGTGCCTCAACCACGTCG TCTAGAAGAGCATGGATGAGTAGGT AGGTCGATAGTGGTCGCTCCTGTCT I V P T E L V Q Q I F S Y L L I H P A I T S E D R>  CGTTCGCTCAACATTTGGCTGCGTC ACTTGGAGGATCATATCCAAGCGGG GCAAGCGAGTTGTAAACCGACGCAG TGAACCTCCTAGTATAGGTTCGCCC R S L N I W L R H L E D H I Q A G  TTGTGGCGGGCCTGACAAATCGCAG AACACCGCCCGGACTGTTTAGCGTC C G G P D K S Q >  400  *  *  TTACTTCCTGCAGCCCTCGCCGCAA CTGGTCGCTGGTGGGTAGCTCAACA  *  GGCAGTGGTAGCTTGTTCCTCTTCC  AATGAAGGACGTCGGGAGCGGCGTT L L P A A L A A  CCGTCACCATCGAACAAGGAGAAGG A V V A C S S S >  GACCAGCGACCACCCATCGAGTTGT T G R W W V A Q Q 500  *  *  *  GGGGGACCANCTCTTCCGACAGGAT CCTGTTCCGTCGGGTGGCCTCATCC CCCCCTGGTNGAGAAGGCTGTCCTA GGACAAGGCAGCCCACCGGAGTAGG G G P X L P T G S C S V G W P H P  TCGGTTGGTGCCCCGCAAGCGGGAG AGCCAACCACGGGGCGTTCGCCCTC R L V P R K R E > 600  *  *  *  CCGTTCCTCAACGCACCAACGACTG GCAAACGGATCGCCCCGCCCAGGAA  GCAATTGGAAAACAAATTGGCCGGT  GGCAAGGAGTTGCGTGGTTGCTGAC  CGTTAACCTTTTGTTTAACCGGCCA  P  F  L  N  A  P  T  T  CGTTTGCCTAGCGGGGCGGGTCCTT G  K  R  GATTGG CTAACC D W>  142  I  A  P  P  R  K  Q  L  E  N  K  L  A  O  TTCCTTATGATGNCCCAGATATATG CCTCTCCCGACTTCGGCACGAGGCG GATCACCGAGCGTGAGAAGAACAAG AAGGAATACTACNGGGTCTATATAC GGAGAGGGCTGAAGCCGTGCTCCGC CTAGTGGCTCGCACTCTTCTTGTTC F L M M X Q I Y A S P D F G T R R I T E R E K N K>  100 * * * AAGCGCGATGTGCGCGGCTGGTATG AGCCAACGATCGCCCGGGAAGGAGT CGTGGATCACAGACACCAGGAGGTG TTCGCGCTACACGCGCCGACCATAC TCGGTTGCTAGCGGGCCCTTCCTCA GCACCTAGTGTCTGTGGTCCTCCAC K R D V R G W Y E P T I A R E G V V D H R H Q E V>  200 *  * * CCAACGGACGTGGAGCGCGGCGACA TTCCCGTTCTGAATGGCGATTGCGA AGACGCCCTCGCACGATCGCTCAGC GGTTGCCTGCACCTCGCGCCGCTGT AAGGGCAAGACTTACCGCTAACGCT TCTGCGGGAGCGTGCTAGCGAGTCG P T D V E R G D I P V L N G D C E D A L A R S L S >  300 * * * GATTTACTGGCTCTGGTGAAGCTGC TCCGCGAAGACGTCGCCCACCAGCG CCAGGAGATTGCCTACCTGCGTATG CTAAATGACCGAGACCACTTCGACG AGGCGCTTCTGCAGCGGGTGGTCGC GGTCCTCTAACGGATGGACGCATAC D L L A L V K L L R E D V A H Q R Q E I A Y L R M> CTCCTGGAGAACTGTGCCGGCTGCA AGAATCCCCTCACCACCGATAACCA GAGGACCTCTTGACACGGCCGACGT TCTTAGGGGAGTGGTGGCTATTGGT L L E N C A G C K N P L T T D N Q  ACTGCGCATCGAGCCCGACTGCCGT TGACGCGTAGCTCGGGCTGACGGCA L R I E P D C R >  400 * * * TCCGCCAATCCCTGTTATCCTGGAG TGGAGTGCTTGGACTCGGCGGNCGG TCCCCGATGTGGCACTTGTCCCCTT AGGCGGTTAGGGACAATAGGACCTC ACCTCACGAACCTGAGCCGCCNGCC AGGGGCTACACCGTGAACAGGGGAA S A N P C Y P G V E C L D S A X G P R C G T C P L > 500  * * * GGCTTCATTGGCGATGGNAAGAGCT GCAAGNCGGGNGTTACCTGCGCCCA TCACATGTGCTATCCAGGCGTCAAG CCGAAGTAACCGCTACCNTTCTCGA CGTTCNGCCCNCAATGGACGCGGGT AGTGTACACGATAGGTCCGCAGTTC G F I G D G K S C K X G V T C A H H M C Y P G V K >  600 * * * TGTCACGATACCGTGAATTGGAGCC AGTGTGATCCTGTCCAGCCGGCTAC GAGGGTGATGGACGACATGTCGGTA ACAGTGCTATGGCACTTAACCTCGG TCACACTAGGACAGGTCGGCCGATG CTCCCACTACCTGCTGTACAGCCAT C H D T V N W S Q C D P V Q P A T R V M D D M S V > CGTATCCTGGCTGGAACACCGTGCC CTCAGGACTCTTTGGTGCCTCGATT ATGTGAACATATCAGAGGGAGGCCA GCATAGGACCGACCTTGTGGCACGG GAGTCCTGAGAAACCACGGAGCTAA TACACTTGTATAGTCTCCCTCCGGT R I L A G T P C P Q D S L V P R L C E H I R G R P >  143  GAATTCGGCACGAGGCGCCTAATTC TGGACGGAATCATTCAGTGCACGGC CAGGGATGAGTTCTCGTACCAGGAG CTTAAGCCGTGCTCCGCGGATTAAG ACCTGCCTTAGTAAGTCACGTGCCG GTCCCTACTCAAGAGCATGGTCCTC E F G T R R L I L D G I I Q C T A R D E F S Y Q E >  100 * * * ATGATATCTTTCCTGCCGCTCTGCG CCCATCCCAATCCCAAAAAGGTCCT GATCGTGGGCGGTGGTGATGGCGGC TACTATAGAAAGGACGGCGAGACGC GGGTAGGGTTAGGGTTTTTCCAGGA CTAGCACCCGCCACCACTACCGCCG M I S F L P L C A H P N P K K V L I V G G G D G G>  200 * * * GTTGCTCGCGAGGTGGTAAAGCATC CACTGGTCGAGGAAGTGCATCAGGT GGAAATTGACGACCGTGTCGTCGAG CAACGAGCGCTCCACCATTTCGTAG GTGACCAGCTCCTTCACGTAGTCCA CCTTTAACTGCTGGCACAGCAGCTC V A R E V V K H P L V E E V H Q V E I D D R V V E>  300 * * * CTGTCCAAGCAATATCTCCCAGCGA TGGCCTGTGGTTTCGCCAACGAGAA GTTGAAGCTTACCATTGGCGATGGA GACAGGTTCGTTATAGAGGGTCGCT ACCGGACACCAAAGCGGTTGCTCTT CAACTTCGAATGGTAACCGCTACCT L S K Q Y L P A M A C G F A N E K L K L T I G D G> TTCGACTATATGAAGAAACACAAGA ACGAATTTGATGTCATCATCACCGA CAGCTCGGATCCCATTGGTCCGGCA AAGCTGATATACTTCTTTGTGTTCT TGCTTAAACTACAGTAGTAGTGGCT GTCGAGCCTAGGGTAACCAGGCCGT F D Y M K K H K N E F D V I I T D S S D P I G P A >  400 * * * GTGAGCCTGTTTCAGGAAAGCTACT ACGAGCTAATGAAACACGCGCTGAA GGATGACGGAATCGTGTGCTCCCAG CACTCGGACAAAGTCCTTTCGATGA TGCTCGATTACTTTGTGCGCGACTT CCTACTGCCTTAGCACACGAGGGTC V S L F Q E S Y Y E L M K H A L K D D G I V C S Q >  500 * * * GGCGGTAGCTTCTGGCTGGACCTGG ACTACATCAAGAAGACCATGTCCGG NTGCAAGGAGCACTTGGTTAGGTGG CCGCCATCGAAGACCGACCTGGACC TGATGTAGTTCTTCTGGTACAGGCC NACGTTCCTCGTGAACCAATCCACC G G S F W L D L D Y I K K T M S G C K E H L V R W >  CCTATGCCGTCACCTCCGTCCGTCC TATCCCTGGGCACATTG GGATACGGCAGTGGAGGCAGGCAGG ATAGGGACCCGTGTAAC P M P S P P S V L S L G T L >  144  GAATTCGGAACGAGGCGGAAAATGG TCCAACGTCTGACGCTCCGGAGACG CCTGTCCTACAACACACGCTCCAAC CTTAAGCCTTGCTCCGCCTTTTACC AGGTTGCAGACTGCGAGGCCTCTGC GGACAGGATGTTGTGTGCGAGGTTG E F G T R R K M V Q R L T L R R R L S Y N T R S N >  100 *  *  AAGCGGCGCATTGTTCGCACGCCCG GTGGTCGTCTGGTTTACCAGTATGT  *  GAAGAAGAACCCCACCGTGCCCCGT  TTCGCCGCGTAACAAGCGTGCGGGC CACCAGCAGACCAAATGGTCATACA K R R I V R T P G G R L V Y Q Y V  CTTCTTCTTGGGGTGGCACGGGGCA K K N P T V P R>  2 0 0  *  *  *  TGCGGNCAGTGCAAGGAGAAGTTGA  AGGGTATCACCCCCTCCCGCCCCAG  CGAGCGCCCCCGCATGTCCAAGCGC  ACGCCNGTCACGTTCCTCTTCAACT TCCCATAGTGGGGGAGGGCGGGGTC C G Q C K E K L K G I T P S R P S  GCTCGCGGGGGCGTACAGGTTCGCG E R P R M S K R > 3 0 0  *  *  *  CTGAAGACCGTGTCCAGGACCTACG GTGGAGTGCTGTGCCACAGCTGTCT  GCGCGAGCGTNTCGTGCGCGCCTTC  GACTTCTGGCACAGGTCCTGGATGC CACCTCACGACACGGTGTCGACAGA  CGCGCTCGCANAGCACGCGCGGAAG  L  K  T  V  S  R  T  Y  G  G  V  L  C  H  S  C  L  A  R  E  R  X  V  R  A  F>  *  CTCATCGAGGAGCAGAAGATCGTCA AGGCCCTGAAGAGCCAGCGNGAGGC GAGTAGCTCCTCGTCTTCTAGCAGT TCCGGGACTTCTCGGTCGCNCTCCG L I E E Q K I V K A L K S Q R E A  *  GCTCGTCAAGCCGGTGTAAGGCCCC CGAGCAGTTCGGCCACATTCCGGGG L V K P V *  4 0 0  *  *  *  AAGGNCAAGCCCGAGACCAAGAAGA  AGCCCGCTGCTGGAGCCAAGGGAAC  CAAGGGCGGTGNCGGTAAGGTCANC  TTCCNGTTCGGGCTCTGGTTCTTCT TCGGGCGACGACCTCGGTTCCCTTG  GTTCCCGCCACNGCCATTCCAGTNG  5 0 0  * ANGGGTGGGTGCTGGCGCCAAGGGA  *  *  GNCGCTGGNAAGAAGCCCGGNCAGA  AGCCAGCCGCTTGGAAAGCCAGGAA  TNCCCACCCACGACCGCGGTTCCCT CNGCGACCNTTCTTCGGGCCNGTCT  TCGGTCGGCGAACCTTTCGGTCCTT 6 0 0  *  *  *  GTNAACAGCCACGCAACGAGTCGGG NGTNTTGNTAANTAATTTTNAAATA  ATTGGGTTTTTTCCACTTGGAAAAA  CANTTGTCGGTGCGTTGCTCAGCCC NCANAACNATTNATTAAAANTTTAT  TAACCCAAAAAAGGTGAACCTTTTT  AAAAAAAAAAACTCGAG TTTTTTTTTTTGAGCTC  145  GAATTCGGCACGAGGCGGTTCCGCG AGGGCACCTCCGAGTACGACCTGAA GCGGCGGCCAGCCTGGACGGATCGG CTTAAGCCGTGCTCCGCCAAGGCGC TCCCGTGGAGGCTCATGCTGGACTT CGCCGCCGGTCGGACCTGCCTAGCC E F G T R R F R E G T S E Y D L K R R P A W T D R >  100 * * * ATAATGTACGCCGTGCAGCCACTGA ACCGGCAGCCCGGCATGCAGCTATC CATTGAGCAATGCTCGTATAAGTCC TATTACATGCGGCACGTCGGTGACT TGGCCGTCGGGCCGTACGTCGATAG GTAACTCGTTACGAGCATATTCAGG I M Y A V Q P L N R Q P G M Q L S I E Q C S Y K S >  200 * * * CATCCCCTGTACACCATCAGTGATC ACAAGCCGGTGACCAGTGACTTTAC CATCAAGCTCTACCCGAATGTACGG GTAGGGGACATGTGGTAGTCACTAG TGTTCGGCCACTGGTCACTGAAATG GTAGTTCGAGATGGGCTTACATGCC H P L Y T I S D H K P V T S D F T I K L Y P N V R>  300 * * * GCGCCCGGCGTGGTGTTCTCGCCTC TGTCGCTCTGGAAGATTGGGGACGA GAACACGGTGGAGTATCACAAGCAG CGCGGGCCGCACCACAAGAGCGGAG ACAGCGAGACCTTCTAACCCCTGCT CTTGTGCCACCTCATAGTGTTCGTC A P G V V F S P L S L W K I G D E N T V E Y H K Q > * * * GCAGAGTTCGACGAGGGGTCCAACG ACTGGATTGGNATCTTTCCGTCGGA GTACGCCAGTTTGGCGGATTACGTA CGTCTCAAGCTGCTCCCCAGGTTGC TGACCTAACCNTAGAAAGGCAGCCT CATGCGGTCAAACCGCCTAATGCAT A E F D E G S N D W I G I F P S E Y A S L A D Y V>  400 * * * GCCTACGAGTATGTCAATCAGGNTG AGTCGGCCTCATCCTCGGACTCCAA TCACCAACCGGGATCCGTTTGGAGA CGGATGCTCATACAGTTAGTCCNAC TCAGCCGGAGTAGGAGCCTGAGGTT AGTGGTTGGCCCTAGGCAAACCTCT A Y E Y V N Q X E S A S S S D S N H Q P G S V W R >  500 * * * CGGCCTCGCATCATCGAAGGGGGTC GGGCATCATACAGGAATCGCCATGC GACAGGTCGCCATCAGGAGGCTAAT GCCGGAGCGTAGTAGCTTCCCCCAG CCCGTAGTATGTCCTTAGCGGTACG CTGTCCAGCGGTAGTCCTCCGATTA R P R I I E G G R A S Y R N R H A T G R H Q E A N >  600 * * * GCCCAAGAGTTGGTGCGGCTAGATT TCGCCGACGATGTGGAACTGCGTCA CGGCGAGCAATACCTGTTGATATAT CGGGTTCTCAACCACGCCGATCTAA AGCGGCTGCTACACCTTGACGCAGT GCCGCTCGTTATGGACAACTATATA A Q E L V R L D F A D D V E L R H G E Q Y L L I Y> * * * TTCCGCAGCACCGGAGTCCGGGGCG TGACCAGTTTGGCCGGCGTCAGTGG TGTCTTTGTGGCGGAGAAGCGGCAC AAGGCGTCGTGGCCTCAGGCCCCGC ACTGGTCAAACCGGCCGCAGTCACC ACAGAAACACCGCCTCTTCGCCGTG F R S T G V R G V T S L A G V S G V F V A E K R H>  146  GAATTCGGCACGAGGCGGAAACCCT TCAAATGCACGGAATGCGGCAAGGG ATTTTGCCAATCGAGAACCTTGGCT CTTAAGCCGTGCTCCGCCTTTGGGA AGTTTACGTGCCTTACGCCGTTCCC TAAAACGGTTAGCTCTTGGAACCGA E F G T R R K P F K C T E C G K G F C Q S R T L A >  100 * * * GTCCACAAGATCCTGCACATGGAGG AATCACCCCACAAGTGCCCCGTCTG CAGTCGATCATTCAATCAGCGCTCC CAGGTGTTCTAGGACGTGTACCTCC TTAGTGGGGTGTTCACGGGGCAGAC GTCAGCTAGTAAGTTAGTCGCGAGG V H K I L H M E E S P H K C P V C S R S F N Q R S >  200 * * * AACCTGAAGACCCATCTGCTCACCC ACACGGATCACAAGCCCTACGAGTG CTCTTCATGCGGCAAAGTTTTCCGC TTGGACTTCTGGGTAGACGAGTGGG TGTGCCTAGTGTTCGGGATGCTCAC GAGAAGTACGCCGTTTCAAAAGGCG N L K T H L L T H T D H K P Y E C S S C G K V F R >  300 * * * CGTAACTGCGATCTACGACGCCATG NCTTGACCCATGCAGTGGGTGAGGT CAACTCCGGGGACTATGTGGATGTG GCATTGACGCTAGATGCTGCGGTAC NGAACTGGGTACGTCACCCACTCCA GTTGAGGCCCCTGATACACCTACAC R N C D L R R H X L T H A V G E V N S G D Y V D V > * * * GGCGAAGAGGATGAGGCCAGAAATT TANGTGGCGACGAGGAGGATTCGTT GCTGGAAGTGGACTCGCCCCGCCAG CCGCTTCTCCTACTCCGGTCTTTAA ATNCACCGCTGCTCCTCCTAAGCAA CGACCTTCACCTGAGCGGGGCGGTC G E E D E A R N L X G D E E D S L L E V D S P R Q >  400 * * * TCGCCAGTTCACAACTTGGGCGAGT CTGGTGGATCGGGTGAGAAATCTGA GTCCGAAAGAATGAGACTCAAGCGC AGCGGTCAAGTGTTGAACCCGCTCA GACCACCTAGCCCACTCTTTAGACT CAGGCTTTCTTACTCTGAGTTCGCG S P V H N L G E S G G S G E K S E S E R M R L K R >  500 *  * * AAGGCAGNCATCGATCATGAGGAAA GCGAAGAGGAGTTCGATGACTTCGA CGAGGAAGAGGAATGCAGGGATCTT TTCCGTCNGTAGCTAGTACTCCTTT CGCTTCTCCTCAAGCTACTGAAGCT GCTCCTTCTCCTTACGTCCCTAGAA K A X I D H E E S E E E F D D F D E E E E C R D L>  147  GAATTCGGCACGAGGCGGGCCTGCG AGAAGGCTTGGCGCGATTTTATTAT TGCAAAGATGACCCCCAAGCCGCCC CTTAAGCCGTGCTCCGCCCGGACGC TCTTCCGAACCGCGCTAAAATAATA ACGTTTCTACTGGGGGTTCGGCGGG E F G T R R A C E K A W R D F I I A K M T P K P P >  100 *  *  CGTATTCACCAGGTGGAGATGGGTT CGGAGCCAATGGATATCAACGAGGA  *  TGAGGCCGATGCACCGGATGATGAT  GCATAAGTGGTCCACCTCTACCCAA GCCTCGGTTACCTATAGTTGCTCCT ACTCCGGCTACGTGGCCTACTACTA R  I  H  -  Q  V  E  M  G  S  E  P  M  D  I  N  E  D  E  A  D  A  P  D  D  D>  200 *  *  *  CTGCCCATGTTGAATCTGGCCTCGT TTGCCATCTACAAGCTGTTCGCGGA GTGGGAACGGGAGGGCTATGTCGTG GACGGGTACAACTTAGACCGGAGCA AACGGTAGATGTTCGACAAGCGCCT CACCCTTGCCCTCCCGATACAGCAC L P M L N L A S F A I Y K L F A E W E R E G Y V V>  300 *  *  *  CCCGAGATGCACCCTTCGGCCAATG CTGCCCAACAGGCGGGAGGGGATGC  CGGAACTCCAGTTCCCCCCGTGCCG  GGGCTCTACGTGGGAAGCCGGTTAC GACGGGTTGTCCGCCCTCCCCTACG P E M H P S A N A A Q Q A G G D A  GCCTTGAGGTCAAGGGGGGCACGGC G T P V P P V P >  AAGGAGCCAAAGAAGCCGCCAGTGC GCACCGAGCTACCCTCTGGCTGGGA GACCATGCACCCGGCGACCATTCTT TTCCTCGGTTTCTTCGGCGGTCACG CGTGGCTCGATGGGAGACCGACCCT CTGGTACGTGGGCCGCTGGTAAGAA K E P K K P P V R T E L P S G W E T M H P A T I L>  400 *  *  *  TGNATTATGCGTCCGGGACTCAACT ACGTGGGACTACGGGTCATCTGGCG ACNTAATACGCAGGCCCTGAGTTGA TGCACCCTGATGCCCAGTAGACCGC X I M R P G L N Y V G L R V I W R  ACAAGANCAACGGCATGCAGCATCT TGTTCTNGTTGCCGTACGTCGTAGA Q X Q R H A A S >  500 GGGAATCATGGTGGACAACCAGGAG TCCACGCCAACGGGAGATCAAAGNA CCCTTAGTACCACCTGTTGGTCCTC AGGTGCGGTTGCCCTCTAGTTTCNT G N H G G Q P G V H A N G R S K X  148  ATGAGAAACGATGGAGGCATGCGCN ATCGCGGAGGAAGCGGTGGCGGTAA TGGAGGCGGTGGCGGCGGACGCTAC TACTCTTTGCTACCTCCGTACGCGN TAGCGCCTCCTTCGCCACCGCCATT ACCTCCGCCACCGCCGCCTGCGATG M R N D G G M R X R G G S G G G N G G G G G G R Y >  100 * GATCGCGGAGGAAGCGGTGGTGGTG  *  *  GCGGCGGCGGTGGCAACNTNCANCC  CCGTGATGGTGACTGGAAATGCAAC  CTAGCGCCTCCTTCGCCACCACCAC CGCCGCCGCCACCGTTGNANGTNGG D R G G S G G G G G G G G N X X P  GGCACTACCACTGACCTTTACGTTG R D G D W K C N >  200 *  *  *  AGCTGTAATAACACCAACTTCGCCT GGCGCAACGAATGCAATAGATGTNA  GACTCCCAAGGGCGACGACGAGGGC  TCGACATTATTGTGGTTGAAGCGGA CCGCGTTGCTTACGTTATCTACANT S C N N T N F A W R N E C N R C X  CTGAGGGTTCCCGCTGCTGCTCCCG T P K G D D E G >  300 *  *  *  TCTAGCGGAGGTGGTGGAAGCGGCG GCTACCGCGGCGGTGGTGGCGGAGG  AGGCTACGACCGAGGAAATGATCGT  AGATCGCCTCCACCACCTTCGCCGC CGATGGCGCCGCCACCACCGCCTCC TCCGATGCTGGCTCCTTTACTAGCA S S G G G G S G G Y R G G G G G G G Y D R G N D R > *  *  GGATCCGGCGGCGGTGGATATCACA ACAGAGATCGCGGTGGCAACTCGCA  *  GGGAGGCGAAGGCGGCGGCGGCGGT  CCTAGGCCGCCGCCACCTATAGTGT TGTCTCTAGCGCCACCGTTGAGCGT  CCCTCCGCTTCCGCCGCCGCCGCCA  G  S  G  G  G  G  Y  H  N  R  D  R  G  G  N  S  Q  G  G  E  G  G  G  G  G  >  400 *  *  *  GGTGGTGGCTACTCCCGCTTCNATG ACNACNATGGCNGAAGACGCCGTGG  CCCTTGAAGTGGTGGCGGCAATCCC  CCACCACCGATGAGGGCGAAGNTAC  GGGAACTTCACCACCGCCGTTAGGG  G  G  G  Y  S  R  F  X  TGNTGNTACCGNCTTCTGCGGCACC D  X  X  G  X  R  CGTGATTGTGGACCGATGAGAAACC ATGGAGGCNTGCGC GCACTAACACCTGGCTACTCTTTGG TACCTCCGNACGCG  149  R  R  G  P *  GAATTCGGNACGAGGCGGTCGGCTG CCGAGGAGTACCAGAAGTACATTAA CTTAAGCCNTGCTCCGCCAGCCGAC GGCTCCTCATGGTCTTCATGTAATT E F G T R R S A A E E Y Q K Y I N  TGCGGATAAGACGACCGTAGCTCTA ACGCCTATTCTGCTGGCATCGAGAT A D K T T V A L>  100 *  *  *  TTCGCCGCCGAATGGGCAGAGCAAT  GCGGTCAGGTGAAAGACGCGCTGGA  GGAGCTGGCCAAGATTACTGGCGAA  AAGCGGCGGCTTACCCGTCTCGTTA CGCCAGTCCACTTTCTGCGCGACCT F A A E W A E Q C G Q V K D A L E  CCTCGACCGGTTCTAATGACCGCTT E L A K I T G E>  200 *  *  *  AAACTGCAGTTCATCAGCCTAAACG CTGAACAATTTCCCGAGATTTCCAT  GAAACATCAGATCGAGGCCGTGCCC  TTTGACGTCAAGTAGTCGGATTTGC GACTTGTTAAAGGGCTCTAAAGGTA  CTTTGTAGTCTAGCTCCGGCACGGG  K  L  Q  F  I  S  L  N  A  E  Q  F  P  E  I  S  M  K  H  Q  I  E A V P >  300 *  *  *  ACAGTCATATTCTTCGCCAAGGGCT CCGCCGTTGACCGTGTCGATGGTGT  AGACATCGCCGCCATAAGCGCCAAA  TGTCAGTATAAGAAGCGGTTCCCGA GGCGGCAACTGGCACAGCTACCACA  TCTGTAGCGGCGGTATTCGCGGTTT  T  V  I  F  F  A  K  G  S  A  V  D  R  V  D  G  V  D  TCCAAAAAGTTGGCCGAAAACGCAA GCAGCGCGGCGGCAACAGGACAAAC AGGTTTTTCAACCGGCTTTTGCGTT CGTCGCGCCGCCGTTGTCCTGTTTG S K K L A E N A S S A A A T G Q T  I  A  A  I  S  A  K>  GTTGGAGGAACGCCTAAAGGCCCTA CAACCTCCTTGCGGATTTCCGGGAT L E E R L K A L >  400 *  *  ATCAATACAGCTCCGCTGATGATAT TCATGAAGGGCGACCGAAATGGACC  *  GCGTTGCGGATTCTCCAAGCAGCTC  TAGTTATGTCGAGGCGACTACTATA AGTACTTCCCGCTGGCTTTACCTGG I N T A P L M I F M K G D R N G P  CGCAACGCCTAAGAGGTTCGTCGAG R C G F S K Q L >  500 *  *  *  ATCGGCATTGTGAACGAAACCAACT TGCCGTACGAGACATTTGACATCCT  CGGCGACGAAGAAGTGCGTCAAGGC  TAGCCGTAACACTTGCTTTGGTTGA ACGGCATGCTCTGTAAACTGTAGGA GCCGCTGCTTCTTCACGCAGTTCCG I G I V N E T N L P Y E T F D I L G D E E V R Q O  600 *  *  *  CTGGTGAAAACTACTCCGACTGGCC ACATATCCCAGGTTTACGTCAAGGG  TGAACTTATCGGCGGCTCGATATTA  GACCACTTTTGATGAGGCTGACCGG TGTATAGGGTCCAAATGCAGTTCCC  ACTTGAATAGCCGCCGAGCTATAAT  L  V  K  T  T  P  T  G  H  I  S  150  Q  V  Y  V  K  G  E  L  I  G  G  S  I  L>  Appendix B: zl  Interactor Sequence Comparisons  no h i g h s c o r i n g  matches  z2  Smallest High Score  S e q u e n c e s p r o d u c i n g H i g h - s c o r i n g Segment P a i r s : sp|P49746|TSP3_HUMAN THROMBOSPONDIN 3 PRECURSOR  /pir||A57...  57  Sum Probability P(N) N 0.69  2  sp|P49746|TSP3_HUMAN THROMBOSPONDIN 3 PRECURSOR p i r | | A 5 7 1 2 1 t h r o m b o s p o n d i n 3 p r e c u r s o r - human g i | 8 8 6 2 9 9 t h r o m b o s p o n d i n 3 [Homo s a p i e n s ] L e n g t h = 95 6  (L38969)  S c o r e = 57 (26.5 b i t s ) , E x p e c t = 1.2, Sum P(2) = 0.69 I d e n t i t i e s = 7/17 ( 4 1 % ) , P o s i t i v e s = 13/17 (76%) Query: Sbjct:  123 DCRSANPCYPGVECLDS 13 9 +C A+PC+PG C+++ 319 ECAHADPCFPGSSCINT 335  S c o r e = 51 (23.7 b i t s ) , E x p e c t = 8.0, Sum P(2) = 1.0 I d e n t i t i e s = 7/13 ( 5 3 % ) , P o s i t i v e s = 11/13 (84%) Query: Sbjct:  12 6 SANPCYPGVECLD 13 8 S NPC+ GV+C++ 279 SPNPCFRGVDCME 291  S c o r e = 41 (19.0 b i t s ) , E x p e c t = 1.2, Sum P(2) = 0.69 I d e n t i t i e s = 7/24 ( 2 9 % ) , P o s i t i v e s = 14/24 (58%) Query: Sbjct:  85 REDVAHQRQEIAYLRMLLENCAGC 108 R+D+ Q +E++ +R + C C 246 RDDIRDQVKEMSLIRNTIMECQVC 269  z3 gi|309502  (L19311) s p e r m i d i n e s y n t h a s e [Mus m u s c u l u s ] gi|1061192 (Z67748) s p e r m i d i n e s y n t h a s e [Mus m u s c u l u s ] p r f | | 2 1 1 3 2 7 6 A s p e r m i d i n e s y n t h a s e [Mus m u s c u l u s ] L e n g t h = 3 02  S c o r e = 601 (278.9 b i t s ) , E x p e c t = 8.1e-80, P = 8.1e-80 I d e n t i t i e s = 111/163 ( 6 8 % ) , P o s i t i v e s = 137/163 (84%) Query: Sbjct: Query:  7 LILDGIIQCTARDEFSYQEMISFLPLCAHPNPKKVLIVGGGDGGVAREWKHPLVEEVHQ 66 L+LDG+IQCT RDEFSYQEMI+ LPLC+HPNP+KVLI+GGGDGGV REWKHP VE V Q 63 LVLDGVIQCTERDEFSYQEMIANLPLCSHPNPRKVLIIGGGDGGVLREWKHPSVESWQ 122  Sbjct:  67 VEIDDRWELSKQYLPAMACGFANEKLKLTIGDGFDYMKKHKNEFDVIITDSSDPIGPAV EID+ V+E+SK++LP MA GF++ KL L +GDGF++MK++++ FDVIITDSSDP+GPA 123 CEIDEDVIEVSKKFLPGMAVGFSSSKLTLHVGDGFEFMKQNQDAFDVIITDSSDPMGPAE  Query:  127  Sbjct:  183  SLFQESYYELMKHALKDDGIVCSQGGSFWLDLDYIKKTMSGCK 169 SLF+ESYY+LMK ALK+DGI+C QG WL LD IK+ CK SLFKESYYQLMKTALKEDGILCCQGECQWLHLDLIKEMRHFCK 22 5  151  126 182  z7 S e q u e n c e s p r o d u c i n g H i g h - s c o r i n g Segment P a i r s :  Score  sp|P45842|RL34_AEDAL 60S RIBOSOMAL  PROTEIN L34  (L31) / p i r . . .  sp|P45842|RL34_AEDAL  PROTEIN L34  (L31) p i r | | S 4 7 6 3 7  60S RIBOSOMAL  ribosomal protein ribosomal p r o t e i n L e n g t h = 13 0  L31 - f o r e s t day m o s q u i t o L31 [Aedes a l b o p i c t u s ]  396  gi|506631  P(N)  N  1.7e-50  1  (U03871)  S c o r e = 396 (182.7 b i t s ) , E x p e c t = 1.7e-50, P = 1.7e-50 I d e n t i t i e s = 76/99 ( 7 6 % ) , P o s i t i v e s = 84/99 (84%) Query:  2 5 MVQRLTLRRRLSYNTRSNKRRIVRTPGGRLVYQYVKKNPTVPRCGQCKEXLKGITPSXPS 84 MVQRLTLRRRLSYNT+ SNKRR+VRTPGGRLVY YVKK TVP+CGQCKE L GI PS PS 1 MVQRLTLRRRLSYNTKSNKRRWRTPGGRLVYLYVKKQRTVPKCGQCKEKLSGIKPSRPS 60  Sbjct: Query:  85 ERPRMSKRLXTVSRTXGGVLCHSXLRXRXVRASLIEEQR 123 ERPRM +RL TV+RT GGVLCH LR R +RA LI+EQ+ 61 ERPRMCRRLKTVTRTFGGVLCHRCLRERIIRAFLIDEQK 99  Sbjct:  zll High Score  S e q u e n c e s p r o d u c i n g H i g h - s c o r i n g Segment P a i r s : gi|1399101 gi|1399101  (U45973) p h o s p h a t i d y l i n o s i t o l  (4,5)b...  177  Smallest Sum Probability P(N) N 3.0e-22  (U45973) p h o s p h a t i d y l i n o s i t o l (4,5)bisphosphate 5 - p h o s p h a t a s e homolog; h a s s i m i l a r i t y t o m o t i f s c o n s e r v e d i n p h o s p h a t i d y l i n o s i t o l (4,5)bisphosphate 5-phosphatases, Swiss-Prot A c c e s s i o n Number Q01968, and t h e p r o d u c t o f GenBank A c c e s s i o n Numb.. L e n g t h = 32 9  S c o r e = 177 (81.6 b i t s ) , E x p e c t = 3.0e-22, Sum P(2) = 3.0e-22 I d e n t i t i e s = 34/93 ( 3 6 % ) , P o s i t i v e s = 49/93 (52%) Query: Sbjct: Query: Sbjct: Query: Sbjct:  2  16 KRRXAWTDRIMYAVQ 30 KR+ AWTDRI++ ++ 143 KRKPAWTDRILWRLK 157 3 6 PGMQLSIEQCSYKSHPLYTISDHKPVTSDFTIKLYPNVRAPGWFSPLSLWKIGDENTVE 95 P S+ Y SH Y ISDHKPV+ F ++L P V AP +V P LW + ++ V 170 PASHFSLSLRGYSSHMTYGISDHKPVSGTFDLELKPLVSAPLIVLMPEDLWTVENDMMVS 229 96 YHKQAEFDEGSNDWIGIFPSEYASLADYVAYEY 128 Y ++F DWIG++ + DYV+Y + 230 YSSTSDFPSSPWDWIGLYKVGLRDVNDYVSYAW 262  S c o r e = 50 (23.0 b i t s ) , E x p e c t = 3.0e-22, Sum P(2) = 3.0e-22 I d e n t i t i e s = 8/15 ( 5 3 % ) , P o s i t i v e s = 13/15 (86%)  152  z28  Sequences p r o d u c i n g H i g h - s c o r i n g gi|1388166 gi|1480194 sp|P23803|ODD_DROME gb|AA141582|AA141582  High Score  Segment P a i r s :  (U58282) Bowel [ D r o s o p h i l a m e l a n o g a s . . . (U62004) Sob p r o t e i n [ D r o s o p h i l a m e l . . . ODD-SKIPPED PROTEIN /gi|296793 (X574... CK02065.3prime D r o s o p h i l a  Embryo... +2  Smallest Sum Probability P(N) N  894 435 263  1.4e-121 1.2e-55 1.9e-29  421  8.3e-72  S c o r e = 894 (411.7 b i t s ) , E x p e c t = 1.4e-121, Sum P(2) = 1.4e-121 I d e n t i t i e s = 164/174 ( 9 4 % ) , P o s i t i v e s = 168/174 (96%) Query: Sbjct: Query: Sbjct: Query: Sbjct:  4 TRRKPFKCTECGKGFCQSRTLAVHKILHMEESPHKCPVCSRSFNQRSNLKTHLLTHTDHK 63 ++ KPFKCTECGKGFCQSRTLAVHKILHMEESPHKCPVCSRSFNQRSNLKTHLLTHTDHK 289 SKEKPFKCTECGKGFCQSRTLAVHKILHMEESPHKCPVCSRSFNQRSNLKTHLLTHTDHK 348 64 PYECSSCGKVFRRNCDLRRHXLTHAVGEVNSGDYVDVGEEDEARNLXGDEEDSLLEVDSP 123 PYECSSCGKVFRRNCDLRRH LTHAVGEVNSGDYVDVGEEDEARNL GDEEDSLLEVDSP 349 PYECSSCGKVFRRNCDLRRHALTHAVGEWSGDYVDVGEEDEARNLSGDEEDSLLEVDSP 408 124 RQSPVHNLGESGGSGEKSESERMRLKRKAXIDHEESEEEFDDFDEEEECRDLAK 177 RQSPVHNLGESGGSGEKSESERMRLKRKA IDHEESEEEFDDFDEEEE +DL + 409 RQSPVHNLGESGGSGEKSESERMRLKRKAAIDHEESEEEFDDFDEEEELQDLPR 462  gi|1480194  (U62004) Sob p r o t e i n L e n g t h = 577  S c o r e = 435 (200.3 b i t s ) , I d e n t i t i e s = 75/84 ( 8 9 % ) , Query: Sbjct: Query: Sbjct:  [Drosophila melanogaster]  E x p e c t = 1.2e-55, Sum P(2) = 1.2e-55 P o s i t i v e s = 79/84 (94%)  4 TRRKPFKCTECGKGFCQSRTLAVHKILHMEESPHKCPVCSRSFNQRSNLKTHLLTHTDHK 63 ++ KPFKC ECGKGFCQSRTLAVHKILHMEESPHKCPVC+RSFNQRSNLKTHLLTHTD K 445 SKEKPFKCAECGKGFCQSRTLAVHKILHMEESPHKCPVCNRSFNQRSNLKTHLLTHTDIK 504 64 PYECSSCGKVFRRNCDLRRHXLTH 87 PY C+SCGKVFRRNCDLRRH LTH 505 PYNCASCGKVFRRNCDLRRHSLTH 528  sp|P23803|ODD_DROME ODD-SKIPPED PROTEIN g i | 2 9 6 7 9 3 product [Drosophila melanogaster] L e n g t h = 3 92 S c o r e = 263 (121.1 b i t s ) , I d e n t i t i e s = 42/63 ( 6 6 % ) , Query: Sbjct: Query: Sbjct:  (X57480) o d d gene  E x p e c t = 1.9e-29, P = 1.9e-29 P o s i t i v e s = 54/63 (85%)  4 TRRKPFKCTECGKGFCQSRTLAVHKILHMEESPHKCPVCSRSFNQRSNLKTHLLTHTDHK 63 ++ KPFKC++CGKGFCQSRTLAVHK+ H+EE PHKCP+C RSFNQR+NLK+HL +H++ 271 SKDKPFKCSDCGKGFCQSRTLAVHKVTHLEEGPHKCPICQRSFNQRANLKSHLQSHSEQS 33 0 64 PYE 6 6 E 331 TKE 333  153  2 2 1  z40  no h i g h  scoring  matches  z46 Smallest Sum Probability  High Sequences  producing High-scoring  pir||S54729 pir||S54729  RNA-binding  Segment P a i r s : p r o t e i n cabeza  Score - fruit  f...  763  RNA-binding p r o t e i n cabeza - f r u i t f l y ( D r o s o p h i l a m e l a n o g a s t e r ) gi|532788 (U13178) RNA b i n d i n g p r o t e i n m e l a n o g a s t e r ] gi|567106 (L37083) RNA b i n d i n g p r o t e i n melanogaster] L e n g t h = 4 04  P(N)  N  2.2e-102  1  [Drosophila [Drosophila  Score = 763 (353.5 b i t s ) , Expect = 2.2e-102, P = 2.2e-102 I d e n t i t i e s == 129/129 (100%), P o s i t i v e s = 129/129 (100%) Query: Sbjct: Query: Sbjct: Query: Sbjct:  1 NVQPRDGDWKCNSCNNTNFAWRNECNRCKTPKGDDEGSSGGGGGGGYGGGGGGGGYDRGN 60 NVQPRDGDWKCNSCNNTNFAWRNECNRCKTPKGDDEGSSGGGGGGGYGGGGGGGGYDRGN 276 NVQ PRDGDWKCNSCNNTNFAWRNECNRCKT PKGDDEGS SGGGGGGGYGGGGGGGGYDRGN 335 61 DRGSGGGGYHNRDRGGNSQGGGGGGGGGGGYSRFNDNNGGGRGGRGGGGGNRRDGGPMRN 120 DRGSGGGGYHNRDRGGNSQGGGGGGGGGGGYSRFNDNNGGGRGGRGGGGGNRRDGGPMRN 336 DRGSGGGGYHNRDRGGNSQGGGGGGGGGGGYSRFNDNNGGGRGGRGGGGGNRRDGGPMRN 395 121 DGGMRSRPY 12 9 DGGMRSRPY 396 DGGMRSRPY 404  154  z60 Smallest Sum  High Sequences producing High-scoring pir||S51247  pir||S51247  Segment P a i r s :  thioredoxin homolog YDR098c - yeast  thioredoxin cerevisiae) cerevisiae]  Score ...  homolog YDR098c - y e a s t (Saccharomyces gi|633632 (Z47746) p r o b a b l e t h i o r e d o x i n  Probability P(N)  201 4.1e-27  [Saccharomyces  Length = 285 Score = 53 (24.1 b i t s ) , Expect = 4.1e-27, Sum P(3) = 4.1e-27 I d e n t i t i e s = 10/27 (37%), P o s i t i v e s = 13/27 (48%) Query: Sbjct:  19 DKTTVALFAAEWAEQCGQVKDALEELA 45 DK V F WAE C +K E ++ 57 DKLIVLYFHTSWAEPCKALKQVFEAIS 83  Score = 77 (35.0 b i t s ) , Expect = 4.1e-27, Sum P(3) = 4.1e-27 I d e n t i t i e s = 18/72 (25%), P o s i t i v e s = 36/72 (50%)  Query: Sbjct: Query: Sbjct:  48 TGEKLQFISLNAEQFPEISMKHQIEAVPTVIFFAKGSAVDRVDGVDIAAISAKSKKLAEN 107 + + F+S + +A++ EIS +1 AVP I KG+ + + G D + + + 87 SNSNVSFLSIDADENSEISELFEISAVPYFIIIHKGTILKELSGADPKEYVSLLEDCKNS 146 108 ASSAAATGQTLE 119 +S ++ T+E 147 VNSGSSQTHTME 158  Score = 201 (91.4 b i t s ) , Expect = 4.1e-27, Sum P(3) = 4.1e-27 I d e n t i t i e s = 39/89 (43%), P o s i t i v e s = 57/89 (64%) Query: Sbjct: Query: Sbjct:  114 TGQTLEERLKALINTAPLMIFMKGDRNGPRCGFSKQLIGIVNETNLPYETFDILGDEEVR 173 T + + RL L+N AP+M+FMKG + P+CGFS+QL+GI+ E + + FDIL DE VR 181 TEEQINARLTKLVNAAPVMLFMKGSPSEPKCGFSRQLVGILREHQVRFGFFDILRDESVR 240 174 QGLVKTTPTGHISQVYVKGELIGGSILLR 202 Q L K + Q+Y+ GE GG +++ 241 QNLKKFSEWPTFPQLYINGEFQGGLDIIK 269  155  N  

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.831.1-0088980/manifest

Comment

Related Items