Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The Su(var)3-9/eIF2γ locus of Drosophila melanogaster and the advantages and disadvantages of this compound… Harrington, Michael John 2001

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

Item Metadata


831-ubc_2001-611043.pdf [ 14.09MB ]
JSON: 831-1.0090483.json
JSON-LD: 831-1.0090483-ld.json
RDF/XML (Pretty): 831-1.0090483-rdf.xml
RDF/JSON: 831-1.0090483-rdf.json
Turtle: 831-1.0090483-turtle.txt
N-Triples: 831-1.0090483-rdf-ntriples.txt
Original Record: 831-1.0090483-source.json
Full Text

Full Text

T H E SU(VAR)3-9/eIF2Y L O C U S O F DROSOPHTLA M E L A N O G A S T E R A N D T H E A D V A N T A G E S A N D D I S A D V A N T A G E S O F THIS C O M P O U N D G E N E A R R A N G E M E N T by M I C H A E L J O H N H A R R I N G T O N B.Sc., The University of British Columbia, 1990 A THESIS SUBMITTED IN PARTIAL FULFILLMENT O F T H E REQUIREMENTS FOR T H E DEGREE O F D O C T O R OF PHILOSOPHY i n 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 February 2001 © Michael John Harrington, 2001 In p r e s e n t i n g this thesis in partial fu l f i lment of the r e q u i r e m e n t s fo r an a d v a n c e d d e g r e e at the Univers i ty of Brit ish C o l u m b i a , I agree that t h e Library shall m a k e it f reely available fo r re fe rence and study. I further agree that p e r m i s s i o n fo r ex tens ive c o p y i n g of this thesis fo r scholar ly p u r p o s e s may b e g ran ted by the h e a d of m y d e p a r t m e n t o r by his o r her representat ives . It is u n d e r s t o o d that c o p y i n g o r p u b l i c a t i o n of this thesis for f inancial ga in shall no t b e a l l o w e d w i t h o u t m y w r i t t e n p e r m i s s i o n . D e p a r t m e n t T h e Un ivers i ty of Brit ish C o l u m b i a V a n c o u v e r , C a n a d a Date F ^ 8 2 2 2 0 0 1 D E - 6 (2788) 11 A B S T R A C T The Su(var)3-9/eIF2ycompound locus of Drosophila melanogaster is comprised of two genes. Alternative splicing produces two transcripts which encode very different proteins. Su(var)3-9 is a putative chromatin protein while eIF2y is a translation initiation factor. A genetic dissection of this locus has allowed two hypotheses to be tested. The genes in compound loci have a complex arrangement and share regulatory and transcribed regions. The first hypothesis proposed that these loci are therefore more vulnerable to mutations than independent genes. A characteristic of many compound loci is an ancestral or functional relationship between the pair of genes. This formed the basis of the second hypothesis, whether there is a relationship between the function of the Su(var)3-9 and eIF2y proteins. In summary, this thesis investigated one possible benefit (coordinated regulation of functionally related genes) and one disadvantage (increased susceptibility to mutations) of compound loci. Using Southern blot, D N A sequencing, Northern blot, and phenotypic analyses, mutant alleles were characterized and could be divided into three categories. Those that affected both genes equivalently (322,P25,P17,P17rvl2, and P17rv9), those that disrupted only the Su(var)3-9 gene (311,318, and 330), and one that primarily affected the Su(var)3-9 gene (336). These alleles and expression of a transformed eIF2y gene allowed mutant phenotypes to be associated with each gene. Four independent alleles were due to insertions of transposable elements into the region of the locus common to both genes. Each mutation disrupted two genes and hence two cellular processes. This reveals the sensitivity of this compound locus to mutations and thus supports the first hypothesis. Mutations in the Su(var)3-9 gene cause a dominant suppressor of PEV phenotype and interact with the Histone gene cluster (HIS-C). Mutations in the eIF2ygene are recessive lethal and possess several mutant phenotypes associated with Minute loci. Most Minute loci encode ribosomal proteins, mutations in this translation initiation factor share their dosage sensitive phenotypes. Because the functions of each gene, as determined by this genetic analysis, were independent, there was no evidence supporting the second hypothesis which examined the relationship between the two genes. iv T A B L E O F C O N T E N T S Abstract ii List of Tables ix List of Figures • xi Acknowledgments xii Chapter 1: Introduction 1 1.1 The Su(var)3-9/eIF2y\ocus of Drosophila tnelanogaster 1 1.2 Gene clusters and polycistronic loci in metazoans 4 1.3 The eIF2y gene 12 1.4 The Su(var)3-9 gene 16 1.5 Summary • - 21 Chapter 2: Materials and methods 24 2.1 Classical genetics 24 2.2 Genomic D N A isolation 24 2.3 Polymerase chain reaction (PCR) . - 24 2.4 D N A sequencing . 25 2.5 Computer analysis 25 2.6 Southern blots 26 2.7 Northern blots. 28 2.8 Germline transformation 28 2.9 General molecular biology techniques 29 Chapter 3: Characterization of D N A alterations in mutant alleles of the Su(var)3-9/eIF2ylocus. . . 30 3.1 The 17-series of alleles 30 3.1.1 The P element induced alleles 2A5,17A3,17D4, and 17F2 are recessive lethal or semilethal 32 3.1.2 The Su(var)3-9/eIF2y\ocus is disrupted by P element insertions in each of these alleles 32 3.1.3 P element inserts cause the Su(var) and the lethality phenotypes of the 2A5,17D4, and 17F2 alleles 36 3.1.4 Nature of the 2A5,17D4,17D4rvl2, and 17D4rv9 a l l e l e s . . . . . 45 3.2 The EMS-series of alleles 50 3.2.1 Recombination mapping confirms that 311, 318,319, and 330 are alleles of the Sufi;ar)3-9/eIF2ylocus 51 3.2.2 Sequence of the Su(var)3-9/eIF2/locus in each mutant allele .54 3.3 The tobo-series of alleles 56 3.3.1 306,310,322,328, and 340 are lethal alleles of Su(var)3-9/ eIF2y while 336 is a non-lethal allele 57 3.3.2 Southern blots reveal two lesions in the 5' end of the locus. . 59 3.3.3 The 336 allele has a hobo element inserted into the first intron. 64 3.3.4 The 322 allele has an insert of 8.3 kb of unknown D N A that may be a retrotransposon 65 3.4 Screen for new alleles with P element mutagenesis 70 3.5 Summary 73 Chapter 4: Transcript and phenotypic analysis of the mutant alleles of the Su(var)3-9/eIF2y\ocus .79 4.1 Northern blot analysis of the Su(var)3-9/eIF2/mutant alleles 79 4.2 Phenotypic analysis of the Su(var)3-9/eIF2/mutant alleles. 88 4.3 Summary 95 Chapter 5: Assignment of mutant phenotypes to the genes of the Su(var)3-9/eJF2y locus using a transformed eiF2ygene 100 5.1 Germline transformation with the pP{hs-eIF2y.H} construct 101 5.2 Characterization of the P{hs-eIF2y.H} transformant lines 101 5.3 Successful rescue of the dominant bristle phenotype 104 5.4 Successful rescue of the recessive lethality phenotype. 110 5.5 The eIF2 y transgene has no discernible influence on PEV 116 5.6 Summary 117 Chapter 6: A potentially new class of alleles of the Su(var)3-9/eIF2y locus 121 6.1 Genetic interactions with certain SuCz;flr)3-9/eIF2yalleles 123 6.2 Recombination mapping the E(var) mutations 127 6.3 Southern blot analysis of the E(var) mutations 129 6.4 Can the P{hs-eIF2y.H} transgene rescue the E(var) mutations? 134 6.5 Northern blot analysis of the E(var) mutations 136 6.6 Summary. 139 Chapter 7: The function of the Su(var)3-9 protein . . 143 7.1 Genetic interactions between Su(var) mutations and the Histone gene cluster 144 7.2 Histone transcript levels may be altered by certain Su(var)3-9 alleles. . 147 7.3 Summary 150 Chapter 8: Discussion 152 8.1 The Su(var)3-9 gene of Drosophila 152 8.1.1 The role of the Su(var)3-9 protein at the HIS-C locus 152 8.1.2 The dominant Su(var) phenotype of the Su(var)3-9 g e n e . . . . 158 8.2 The first hypothesis and the nature of each mutation 163 8.2.1 The 322 allele (8.3 kb insert of unknown D N A / 5 ' end of locus) 163 8.2.2 The P25 allele (2.8 kb P element insertion/5' UTR) 164 8.2.3 The P17, P17rvl2, and P17rv9 alleles (1.8,1.3, and 0.4 kb P element insertions, respectively/first intron) 166 8.2.4 The 336 allele (hobo element insert/first intron) 168 8.2.5 The 311,318, and 330 alleles (G521D, S616L, D536N) . 171 8.2.6 The vulnerability of compound genes to mutations 173 8.3 The eIF2ygene of Drosophila 174 8.3.1 The Drosophila eIF2ygene is an essential locus 174 8.3.2 Mutations in the Drosophila eIF2ygene have Minute phenotypes 175 8.4 The second hypothesis and the origin of the Su(var)3-9 / eIF2y\ocus. . 180 8.5 Summary 187 viii Chapter 9: Conclusions 192 Chapter 10: Future research 194 References 199 LIST OF T A B L E S Table 1: D N A fragments used as probes for Southern and Northern blots . . . 27 Table 2: Complementation tests using the 17-series 33 Table 3: Somatic reversion of the 17-series alleles 38 Table 4: Germline reversion of the Su(var) phenotype of the 17-series alleles 39 Table 5: Germline reversion of the lethality phenotype of the 17D4 and 17F2 alleles .41 Table 6: Recombination mapping the EMS-series alleles relative to the P25 allele 53 Table 7: Complementation tests with the hobo-series alleles 58 Table 8: Screen for new alleles using P element mutagenesis 72 Table 9: Summary of the mutant alleles 76 Table 10: Quantified amounts of the Su(var)3-9 and eIF2y transcripts in the mutant alleles 85 Table 11: The bristle lengths of Su(tw)3-9/eIF2ymutants 93 Table 12: P(hs-eIF2y.H} transformed lines 102 Table 13: Rescue of the Minute bristle phenotype by P{hs-eIF2y.H I l l Table 14: Rescue of recessive lethality by single copies of P{hs-eIF2y.H}... 113 Table 15: Rescue of recessive lethality by double copies of P{hs-eIF2y.H} 115 Table 16: Genetic interactions with the E(var) mutants 124 Table 17: Rescue of the E(var) mutations using P{hs-eIF2y.H} 135 X Table 18: Genetic interactions involving the HIS-C locus 146 Table 19: Quantified amounts of the Hisl and His4 transcripts in the Su(var)3-9/eIF2ymutai\ts. . 151 xi L I S T O F F I G U R E S Figure 1: The Su(var)3-9/eIF2ycomplex locus of Drosophila melanogaster .2 Figure 2: Southern blot of the 17-series alleles 34 Figure 3: Southern blot of the 17D4 revertant alleles 43 Figure 4: PCR analysis of the 17-series alleles 47 Figure 5: Southern blot of the /zobo-series alleles using the E probe 60 Figure 6: Southern blot of the /zobo-series alleles using the A+B and hobo probes . 6 2 Figure 7: Southern blot of the 322 allele 66 Figure 8: The location of the mutant alleles in the Su(var)3-9/eIF2y locus 74 Figure 9: Northern blots of the Su(var)3-9 and eIF2/transcripts in the mutant alleles 81 Figure 10: Southern blot of the P{hs-eIF2y.H} transformed lines 105 Figure 11: Rescue crosses using theP{hs-eIF2y.H} transgenes. 108 Figure 12: Southern blot of the E(z;ar) mutations using the A+B probe. . . 130 Figure 13: Southern blot of the E(var) mutations using the E probe 132 Figure 14: Northern blot of the Su(var)3-9 and eIF2y transcripts in the E(var) mutant flies 137 Figure 15: Northern blot of the Hisl and His4 transcripts in the mutant alleles 148 Xl l A C K N O W L E D G M E N T S First off I would like to thank the people who donated the mutant strains used in this thesis, John Locke, Randy Mottus, and Gunter Reuter. Without your flies, this thesis would have had nothing between the introduction and the discussion sections. Thanks to the members of my lab past and present, especially Bob Argiropolous, Erick James, Peter Knight, Dean Mulyk, and Jim Whalen. May your labs always be filled with loud music and crude put-downs. Special thanks to the undergraduate students who contributed to this project, Judy Tsai and Patrick Whalen. Thanks to my supervisory committee, Hugh Brock, Don Moerman, Ann Rose, and George Spiegelman for guiding this project to completion after these many years. I would also like to thank the staff at the E M lab and the Nucleic Acid-Protein Service Unit for doing in hours what would have taken me weeks. But mostly I want to extend a grateful thanks to my family and friends for smiling and nodding when I explained to them what it was I did with my flies and for understanding when I had to disappear for a few years to get the job done. 1 C H A P T E R 1: I N T R O D U C T I O N 1.1 The Su(var)3-9/eIF2y\oc\is of Drosophila tnelanogaster Most metazoan genes are found in a discrete region of the chromosome and are expressed using their own promoters. Gene clusters and polycistronic loci are rare exceptions in which two genes are transcribed from a common promoter (Blumenthal 1998). This nonrandom arrangement of genes represents a seemingly non-ideal and needlessly complex gene organization. Why juxtapose genes when there is enough D N A to have each transcript encoded by a discrete region on the chromosome? If the genes are homologous then the arrangement is likely the result of a duplication of an ancestral gene. If the genes are not related the gene cluster or polycistronic locus is probably the result of a chromosomal rearrangement or an insertion that has brought the two genes together. These new arrangements may have been detrimental for one or both genes and may have necessitated changes in their structure and regulation. Conversely, these arrangements may have proved beneficial in that they allow the co-regulation of the two genes and may have allowed duplicated genes to diverge in structure and function. A n example of a compound locus is Su(var)3-91eIF2yo( Drosophila melanogaster (Reuter et al. 1986; Tschiersch et al. 1994). This locus is located on the third chromosome at 3-56.4 on the genetic map and cytologically at 88D10-E1. Alternative splicing produces two transcripts that encode very different proteins (see Figure 1). The larger 2.4 kb transcript is made of exons A , B, and C and 2 Figure 1: The Su(var)3-9/eIF2ycomplex locus of Drosophila melanogaster The figure shows the genetic organization of this locus based on Tschiersch et al. (1994). Exons are shown as boxes and are labeled A through F. Open reading frames are shown as shaded boxes. The two alternately spliced transcripts are shown below. The Su(var)3-9 transcript is comprised of exons A , B and C while the eIF2/transcript is made of exons A , B, D , E , and F. The sites of the restriction enzymes BamHI (B), EcoRI (R), Pvull (P), and Xhol (X) are shown relative to the transcription start site being zero. Note that not all Drosophila strains have the Pvull site at position -200. At the top of the figure are shown the sites of the PCR primers (5R1 and 3R1) and the regions used as probes in this thesis. The latter are designated A , A+B 3 Probes and Primers A A + B 5R1 3R1 Locus A B 1 kb D E Transcripts Su(var)3-9 eIF2y Map 1 kb X o (M CN -// 13 1 P 1 R (P) 1 'L_ B R 1 1 . P P P B B V l R 1 P 1 B 1 X 1 o o it-o o — • 1 o o O O o o o o o o o o T—1 o o T—i O fN T - H T T < cn m co ro O N 00 CN O N rH <N CN N 00 rn \D in rH rr> r H 1 1 fN m ^t1 ^ m 35 00 00 4 encodes Su(var)3-9, a putative chromatin protein currently being studied in our lab. The smaller 2.0 kb transcript is comprised of exons A , B, D, E , and F and encodes the eIF2Y translation initiation factor. For simplicity, the Su(var)3-9 / eIF2y complex will be referred to as one locus comprised of two genes, Su(var)3-9 and eIF2y. This thesis describes the genetic dissection of the SH(iw)3-9/eIF2ylocus. A series of mutations were obtained or generated and then tested to determine how each disrupted the structure of the locus, the abundance of each transcript, and the phenotype of the flies. This work was used to address two hypotheses related to compound loci as well as to investigate the function of the Su(var)3-9 and the eIF2y genes in Drosophila. These questions are outlined below in an introduction to the three themes of this thesis: gene cluster arrangements and the roles of the eIF2y and Su(var)3-9 proteins. 1.2 Gene clusters and polycistronic loci in metazoans Blumenthal (1998) has recently reviewed the subject of gene clusters and polycistronic loci. There are a variety of different examples in metazoans of genes that are more complex than the "one gene, one functional R N A or protein" norm. They will be collectively referred to as compound loci in this thesis. A similar term, complex locus, is defined by King and Stansfield (1997) as closely linked but non-overlapping groups of functionally related genes. Compound loci are those rare examples whereby a single locus is responsible for producing two or more distinct transcripts. They are more than just genes in 5 which the content of an m R N A transcript can be varied by using alternative transcriptional start sites, alternative splicing in the body of the transcript, or alternative polyadenylation sites. There are several types of compound loci. How they function in the organism depends on how the particular arrangement arose and how the arrangement has impacted the regulation of each gene separately and together. Compound loci may involve compromises or benefits for one or both of the genes involved. The following examples, from Drosophila where possible, are discussed in terms of these points. Duplications of a single ancestral gene have allowed the formation of many gene complexes where homologous genes are adjacent to each other. These duplications have allowed the subsequent diversification and specialization of the repeated genes. There are, however, few examples of duplicated genes in which the transcription of both is dependent on a single promoter. A n example of a compound locus of this type is the Alcohol dehydrogenase (Adh) locus in Drosophila melanogaster (Brogna and Ashburner 1997). Both an abundant monocistronic Adh transcript and a rare dicistronic transcript are produced. The dicistronic transcript contains an additional open reading frame of a downstream gene, Adh-related. The generation of the longer Adhr transcript involves bypassing the usual 3' end processing of the Adh transcript. This was determined as flies mutant for a subunit of the 3' cleavage stimulation factor (CstF) had an increase in the amount of Adhr transcripts (Brogna and Ashburner 1997). 6 The Adhr transcript encodes the A D H R protein, which is 37% identical to A D H (Brogna and Ashburner 1997). Though A D H R is detectable in adult flies by immunostaining, its function is unknown. It is unlikely to be an alcohol dehydrogenase as mutations mapping to the Adh gene can eliminate all alcohol dehydrogenase activity (Jeffs et al. 1994). Furthermore, A D H R carries an amino acid substitution that has been shown to render A D H non-functional. Despite this, A D H R is likely to be more than just an artifact of this gene arrangement as it has been present in flies for more than 40 million years (Albalat and Gonzalez-Duarte 1993). Dicistronic transcripts do not necessarily imply that the genes involved are homologous. About 25% of the genes in Caenorhabditis elegans are transcribed as polycistronic R N A from a single promoter (reviewed by Blumenthal and Spieth 1996). These transcripts are processed into monocistronic mRNAs before they are translated, unlike in bacteria operons. A few genes in these C. elegans operons encode proteins that function in the same processes and are thus candidates for co-regulation. This situation allows for transcriptional regulation as both transcripts will be expressed in the same tissues in the same amounts. A different mechanism that allows the duplicated genes in a compound locus to both be functional is with alternative splicing. One such example is the unc-60 locus of C. elegans (McKim et al. 1994). Two transcripts, uncSOA and unc-60B, share a promoter and the first exon while the remainder of the unc-60A region is in the first intron of unc-60B. The two transcripts encode actin binding proteins that are 38% identical with each other. Although the proteins are 7 homologous, they have slightly different roles in regulating actin filament dynamics (Ono and Benian 1998). The two classes of compound loci discussed above involved duplicated genes that use dicistronic transcripts or alternative splicing to allow both genes to function. A third, more unusual class of compound loci are those that, because they include non-homologous genes, have a less clear origin. The cholinergic locus of Drosophila encodes two transcripts; one nested in the first intron of the other (Kitamoto et al. 1998). This gene arrangement was first found in C. elegans (Alfonso et al. 1994) and is also present in mammals (reviewed in Eiden 1998). The proteins encoded are not homologous but, intriguingly, they are functionally related. V A C h T is a vesicle transporter and C h A T is a biosynthesis enzyme, both of the neurotransmitter acetylcholine. In all three species the locus is arranged common exon(s), VACht specific exons and then C h A T specific exons. Despite this general similarity, there is a variation in the position of exons and introns. For example, in nematodes and mammals the common 5' exon(s) are non-coding, yet in Drosophila the 5' common exon contains the beginning of the C h A T open reading frame. Most mutations in the C. elegans cholinergic locus map to either the unc-17 gene, which encodes V A C h T , or the cha-1 gene, which encodes C h A T (Rand 1989). Three unusual alleles failed to complement with alleles of either gene. One of these was mapped by recombination to lie between the two genes (Rand 1989). It was subsequently shown to be a small deletion that removes the 3' end of the last unc-17 specific exon and part of the intron that precedes the first cha-1 8 specific exon (Alfonso et al. 1994). How this mutation affects the downstream cha-1 gene is not known, especially considering that a slightly smaller deletion in the same region only affects the unc-17 gene. The other two mutations that failed to complement alleles of both genes were recombination mapped to be in the 5' regulatory region where they presumably interfere with the transcription of both genes (Rand 1989). Compound loci are rare in Drosophila. Aside from the Adh/Adhr and cholinergic loci mentioned above there are three other studied examples. The stoned gene is thought to produce a dicistronic transcript that encodes the dissimilar S T N A and STNB proteins (Andrews et al. 1996). These proteins are involved in the recycling of synaptic vesicle membranes (Fergestad et al. 1999). This gene arrangement may allow synthesis of specific amounts of the two proteins (Andrews et al. 1996), or perhaps it may even facilitate the formation of S T N A / S T N B heterodimers (as speculated by Blumenthal 1998). Another example is the DUb80/IP259 locus (Mottus et al. 1997). These genes share a common promoter and a small non-coding exon. The coding region of the IP259 gene is found in a single exon located in the first intron of the DUb80 gene. DUb80 is an ubiquitin/ribosomal protein fusion protein. The function of IP259 is not yet known although it has homologues in mammals, C. elegans, and Saccharomyces cerevisiae (Mottus et al. 1997) A third recent example are the meiotic recombination genes, mei-217 and mei-218, which are expressed as a dicistronic transcript (Liu et al. 2000). The proteins they encode are not similar despite their being functionally related. The two openreading frames, which are in different frames, overlap by about six codons. 9 The Su(var)3-9/eIF2ycompound locus of Drosoiphila has similarities with some of these examples. As the two genes are not homologous, this locus is not the result of a duplication of an ancestral gene. Alternative splicing allows both genes to be expressed. The locus is different from other examples of compound loci in that there is an overlap in the two open reading frames. Exons A and B, which are common to both transcripts encode 80 amino acids shared by both conceptual proteins (see Figure 1 and Tschiersch et al. 1994). When this project began, genetic analysis suggested that mutations in the Su(var)3-9/eIF2y\ocus were of two classes that had opposite phenotypes. As only the identity of the Su(var)3-9 gene had been established, it was hypothesized that the locus produced two proteins with opposing roles. The mutations had different phenotypes because they were lesions in one gene or the other. This would have implied that two antagonist proteins that were produced by the same locus were involved in the same cellular process. This raised the intriguing possibility that the reason these two genes were together in one locus was so that the relative abundance of their proteins could be regulated at the level of alternative splicing. This would have been a fascinating explanation as to why this compound locus possessed the gene arrangement it did. However appealing this model was, it was nevertheless ruled out when the identity of the eIF2 y gene was determined in yeast and mammals (as described in the next section). Because the genes that comprise a compound locus are overlapping the locus is more complicated than would be the case if the genes were independent. 10 Because of this complexity these loci may be relatively more sensitive to mutations for a variety of reasons. This statement forms the basis for the first hypothesis: Are compound loci more vulnerable to spontaneous and generated mutations? Five reasons why this may be true are as follows. (1) It is possible that a single mutation could affect both genes and potentially compromise two separate processes. These mutations would have more severe consequences than if they only affected a single gene. How much each gene is affected by these mutations depends on the presence of mechanisms that can compensate for the mutation and also on the relative sensitivity of each cellular process to mutations. (2) The two genes in the Su(var)3-9/eIF2y locus and other compound loci overlap at the 5' end. Because transposable elements preferentially insert into the 5' end of genes they would be expected to compromise both genes if they inserted into a compound locus. (3) A mutation in just one of the genes may still have a detrimental affect on the other. For example, a defect in a splice site may prevent the production of one transcript but may cause an increase in the other. (4) The genes in compound loci may also be more vulnerable to mutations because they have a non-optimal structure or regulation. For example, the first two exons of the Su(var)3-9/eIF2y locus encode 80 amino acids common to both 11 proteins. Even if these residues are not necessary for one of the proteins, a nonsense mutation in this region would still prevent its translation. (5) The regulation of compound loci may be more precisely controlled than would be typical for these genes if they were not overlapping. Because of this, what might otherwise have been inconsequential mutations may still produce a range of different mutant phenotypes due to the two genes being affected in unexpected ways. As discussed in the next sections, the Su(var)3-9 and eIF2y genes are involved in seemingly unrelated cellular processes. While there is no obvious reason why these two genes are superimposed it cannot be concluded that it was just chance that brought these genes together and allowed this gene arrangement to persist. Many examples of compound loci comprised of dissimilar genes still show evidence of a functional relationship between the two proteins. This was the case for the cholinergic, stoned, and mei-217/ mei-218 loci as discussed above. This was tested as the second hypothesis: Are the two genes in the Su(var)3-9/eIF2ylocus functionally related? If there were an overlap in the function or regulation of these two genes it would help explain how compound gene arrangements form, what allows them to be maintained during evolution, and how they can become optimized. When the Su(var)3-9 / eIF2jlocxLS was created there must have been positive and negative selective pressure acting on it. The second hypothesis illustrates an example of positive selection. If the genes are related in structure 12 or function this new arrangement may be beneficial. For example it may allow the coordinated expression of these genes. O n the other hand the first hypothesis proposes a negative consequence of compound loci, the increased vulnerability to mutational damage that these types of gene arrangements subject the organism to. To determine the function of each gene the mutant phenotypes of alleles of this locus were investigated. Mutations at this locus are expected to fall into two classes, as was found for the unc-17/cha-1 locus discussed above. Those that affect a single gene are useful to establish the consequences of loss of each gene. The mutant phenotypes associated with each gene will help reveal the function of the two proteins, as is required to test the second hypothesis. The other class of mutations, those that affect both transcripts, will be used to test the first hypothesis. This genetic dissection will also allow an examination of the function of these two genes in Drosophila. The next sections discuss what is known about the eIF2yand Su(var)3-9 proteins and what questions can be addressed by this mutational analysis. 1.3 TheeIF27gene The 2.0 kb transcript encodes the putative Drosophila homologue of the eukaryotic translation initiation factor two gamma (eIF2y; Erickson et al. 1997). This was not known at the time the locus was first sequenced by Tschiersch et al. (1994). eIF2y and the non-identical proteins eIF2oc and eIF2p form the eukaryotic translation initiation factor two (eIF2) complex. eIF2 is responsible for the 13 recruitment of the initiator methionyl-tRNA (Met-tRNAi) to the 40S ribosomal subunit and ensuring that it recognizes the A U G start codon (reviewed by Trachsel 1996; Kimball 1999; Kozak 1999). The first step in translation initiation is the formation of a ternary complex between eIF2, GTP, and Met-tRNAi. This ternary complex attaches to the 40S ribosomal subunit creating the 43S preinitiation complex. The 43S complex attaches to an m R N A at its 5' m7G cap. The preinitiation complex then scans the transcript until there is an anticodon to codon match. At this point the GTP attached to eIF2 is hydrolyzed, eIF2-GDP and the other initiation factors disassociate, the 60S ribosomal subunit can then bind, and translation elongation commences. eIF2-GDP must be recycled by the guanine exchange factor, eIF2B, before it can join a new Met-tRNAi. eIF2 is regulated by the phosphorylation of the eIF2a subunit at its Ser50 residue (Ser51 in Drosophila; reviewed by De Haro et al. 1996). When eIF2a is phosphorylated, eJF2 binds irreversibly to eIF2B and sequesters it. As eJF2B is in limiting amounts, translation initiation is repressed by the lack of eIF2-GTP. eIF2a is phosphorylated in S. cerevisiae cells as a response to amino acid starvation and in mammalian cells in response to viral infection. Although protein translation is down regulated during heat shock treatment (stress response) in Drosophila, there is only a slight increase in the phosphorylated form of eIF2a (Duncan et al. 1995). A Drosophila eIF2a kinase has been cloned, though the tissues and situations in which it down regulates translation initiation remain unknown (Santoyo et al. 1997; Olsen et al. 1998). 14 The eIF2y protein encoded by the Drosophila Su(var)3-91eIF2y locus is approximately 75% identical to its human, S. cerevisiae, and Schizosaccharomyces pombe homologues and very similar to a bacterial / mitochondrial translation elongation factor, EF-Tu (Erickson et al. 1997). The open reading frame encodes a complete and presumably functional eIF2y protein. The protein has the same three highly conserved domains as its eukaryotic (eIF2y) and prokaryotic (EF-Tu) homologues (Keeling et al. 1998). Domain I is a guanosine triphosphate (GTP) binding site with the characteristic amino acid sequence: GxxxxGK(S/T)...DxxG...NKxD...(C/S)Axxx (Kjeldgaard et al. 1996). The eIF2y proteins differ from the EF-Tu proteins in two respects (Erickson et al. 1997). They have a non-conserved amino terminus -about 30 amino acids long in D. melanogaster. They also contain a 35 to 37 amino acid insert in the GTP binding domain that, though not conserved, is rich in hydrophilic, proline, and cysteine residues. The role of eIF2y in translation initiation has been well characterized in S. cerevisiae using specific mutations in its elFly gene, GCD11. eIF2y is responsible for correct recognition of the A U G start codon (Dorris et al. 1995), for binding GTP (Naranda et al. 1995; Erickson and Hannig 1996), for binding the Met-tRNAj (Gaspar et al. 1994; Erickson and Hannig 1996), and it is necessary for the hydrolysis of GTP to GDP (Huang et al. 1997). The only uncertainty is the function, if any, of the non-conserved amino terminus. S. cerevisiae gcdll knockout mutants can be (partially) rescued by both the S. pombe protein which has a different and much shorter amino terminus and the GcdllD3-77 truncated protein which lacks this amino terminus altogether (Erickson et al. 1997). 15 However, a missense mutation in this region, gcdll-D6N, has a gcdll mutant phenotype. Genes encoding eIF2Yhave been isolated and characterized from S. cerevisiae (Hannig et al. 1993), mammals (Gasper et al. 1994; Ehrmann et al. 1998), and S. pombe (Ericksoh et al. 1997). Al l of these genes were biochemically demonstrated to encode the eIF2y subunit. Since the 2.0 kb transcript produced by the Su(var)3-9/eIF2/locus contains a complete openreading frame which encodes a protein that is about 75% identical to the others it seems reasonable to conclude that it does indeed encode the fly eIF2y protein. There has been some work done on the other proteins present in the Drosophila eIF2 complex. The Drosophila genes encoding eIF2a and eIF2p have been cloned (Qu and Cavener 1994; Ye and Cavener 1994). Q u et al. (1997) sought to artificially increase or decrease the global amount of protein translation during development by using transgenes that expressed different mutant forms of eIF2a. These mutant eIF2a proteins either mimicked phosphorylated eIF2a (constitutively off) or were unable to be phosphorylated (constitutively on). Flies expressing the first protein were slow to develop and small in size while those flies that expressed the second protein were quick to develop and larger than wild type flies. While many of the Drosophila translation initiation factors have been cloned, only eIF2a (see above) and eIF4A have been analyzed genetically. eIF4A is one of the factors that bind the 5' end of m R N A (reviewed by Merrick and 16 Hershey 1996). As discussed later in this thesis, mutant alleles at the Drosophila eIF4A locus have a variety of mutant phenotypes (Dorn et al. 1993b). However, it is not clear which of these phenotypes were due to loss of the eIF4A gene and which were due to other nearby genes that were also affected. Aside from these examples, the eIF2/mutant alleles described in this thesis represent one of the first opportunities for mutational analysis of translational initiation in Drosophila. The role of eIF2yin Drosophila will be investigated through characterization of these mutant alleles and generation of transgenic strains that carry ectopic eLF2ygenes. These experiments will be interpreted to determine which tissues and developmental stages are most sensitive to alterations in the amount of eIF2y transcription. 1.4 The Su(var)3-9 gene In contrast to the biochemically defined role of erF2y, the function of the other protein product by this locus, Su(var)3-9, is not known. A s its name implies, it was originally identified by mutant alleles that were dominant suppressors of position effect variegation (PEV). The 3-9 refers to this being the ninth such gene on the third chromosome to be identified (Reuter et al. 1986). PEV is a Drosophila phenomenon that was first described seventy years ago (Muller 1930). It occurs when euchromatic genes are moved by a chromosome rearrangement to a new position adjacent to centromeric heterochromatin. In their new location they are silenced in some cells yet remain transcriptionally competent in others (reviewed by Weiler and Wakimoto 1995). Although the 17 silencing of euchromatic genes at inappropriate locations on the chromosome has been described in many eukaryotes, only in Drosophila can it be caused by the juxtaposition of euchromatic and heterochromatic regions by chromosome rearrangements ("classic PEV"). There are two current models for P E V in Drosophila. Both models suggest that the inactivation of the euchromatic gene is a consequence of it taking on a characteristic of heterochromatin. The "spreading model" proposes that in some cells there is an inappropriate spread of heterochromatic D N A packaging into the euchromatic region (see, for example, Locke et al. 1988; Grigliatti 1991). This type of chromatin packaging is not conducive to transcription and the gene is silenced. Although conceptually simple, the spreading model is rapidly losing ground to the "pairing-looping model" (Henikoff 1996; Sabl and Henikoff 1996; Talbert and Henikoff 2000). This model states that tandem repeats such as those present in heterochromatin pair in somatic tissues. These paired structures are sequestered (by looping of the chromosome) into transcriptionally silent heterochromatin compartments in the nucleus. Euchromatic genes variegate when they are too close to these repeats so that in some cells they are dragged into the wrong part of the nucleus. In Drosophila the standard variegating rearrangement is Inversion (1) white mottled four (tum 4; Reuter et al. 1983). It is a paracentric inversion of the X-chromosome that has placed the white* gene to within 25 kb of the centromeric heterochromatin (Tartof et al. 1994). The White protein is necessary for the cell autonomous red pigmentation of the Drosophila eye. The eyes of wmi flies are a 18 mosaic of red (white+ gene transcriptionally competent) and white (white* gene silenced) ommatidia (facets). The extent of this variegation is sensitive to any quantitative or qualitative changes in the nucleus's chromatin. Because of this, PEV has been used successfully as a reporter system to identify genes involved in chromatin structure in Drosophila (reviewed in Wustmann et al. 1989; Grigliatti 1991; Wallrath 1998). Mutations that suppress the variegation (inactivation) are termed Su(var)'s while those that enhance variegation are called E(var)'s. The eyes of wm4; Su(var) flies are 100% pigmented (uniformly red) while those of wm4;E(var) flies lack pigment (uniformly white). Two of the best studied Su(var) genes are Su(var)205 and Su(var)3-7. They encode the polytene chromocentre associated proteins Heterochromatin Protein 1 (HP1; James et al. 1989) and Su(var)3-7 (Cleard et al. 1997), respectively. The chromocentre is the densely packaged central region of the larval polytene chromosomes. It is comprised of the pericentric heterochromatin from all the chromosome arms. The presence of these two proteins at the chromocentre is interpreted to mean that they are structural components of heterochromatin. Mutations in their genes cause the Su(var) phenotype because either there is a reduced spread of heterochromatin packaging (spreading model) or there is less pairing and sequestering of tandem repeats (pairing-looping model). The Su(var)3-9/eIF2ylocus was initially identified by twenty mutant alleles, all of which had a dominant Su(var) phenotype (Reuter et al. 1986). Initial 19 analysis associated this phenotype with loss of the larger 2.4 kb transcript based on three lines of evidence (Tschiersch et al. 1994). First, twelve of these alleles had some disruption in exon C (see Figure 1), six of which would result in truncated proteins. Second, the 2.4 kb transcript, but not the 2.0 kb transcript, was reduced in Su(var)3-914 mutant embryos and absent in Su(var)3-906 embryos. And third, an 11.5 kb fragment of genomic D N A spanning the region was transformed into flies. Eleven independent inserts were able to rescue the Su(var) phenotype demonstrating that it had to be either the 2.4 or the 2.0 kb transcript that was responsible. The 2.4 kb Su(var)3-9 transcript encodes a 635 amino acid conceptual protein with two key domains. In the middle of the protein is the 40 amino acid "chromatin modifier organizer" (chromo) domain first found in the Drosophila proteins Polycomb (PC) and HP1 (Paro and Hogness 1991). Experiments in which the chromo domains of PC and FTP1 were interchanged have shown the chromo domain to be necessary for the normal localization of HP1 to heterochromatin and PC to its euchromatic sites (Messmer et al. 1992; Powers and Eissenberg 1993; Platero et al. 1995). At the carboxyl-terminus of Su(var)3-9 is the 130 amino acid "SET" domain, found in four Drosophila proteins; Su(var)3-9, Enhancer of Zeste (E(Z)), Trithorax (TRX), and Absent Small or Homeotic 1 (ASHl)(Tschiersch et al 1994; Jenuwein et al. 1998). Each of these Drosophila proteins has been implicated in global or regional chromatin structure; responsible for the packaging of heterochromatin (HP1), negative regulation of the homeotic genes (PC, E(Z)), and positive regulation of the homeotic genes (TRX, ASH1) (Pirrotta 1998). 20 Genes homologous to Su(var)3-9 have been isolated from humans (SUV39H1) and mice (Suv39hl; Aagaard et al. 1999). Antibodies raised against the mouse protein were used to show that Suv39hl has a dynamic chromatin distribution. In mouse interphase nuclei there are cytologically visible blocks of heterochromatin which, due to their high A / T content, preferentially stain with the dye 4'-6'-diamidino-2-phenylidole (DAPI). These heterochromatic foci are also the site of the M31 protein (a mammalian HP1 homologue; Wreggett et al. 1994). The Suv39hl protein was found to co-localize to these heterochromatic foci as well (Aagaard et al. 1999). In metaphase chromosomes Suv39hl had a more restricted distribution as it was only present at the centromeres and not the flanking heterochromatin. The S. pombe Su(var)3-9 homologue, cryptic loci regulator 4 (clr4), was identified as a gene necessary for transcriptional silencing of the mating loci, mat! and mat3 (Ekwall and Ruusala 1994; Thon et al. 1994). It was subsequently shown to be necessary for transcriptional silencing at the centromere and for proper chromosome segregation (Ekwall et al. 1996; Ivanova et al. 1998). Like its homologues, Clr4p has a chromo domain at its amino terminus and a SET domain at its carboxyl terminus (Ivanova et al. 1998). The four homologues characterized so far have the two defining protein domains (chromo and SET) as well as a cysteine-rich region adjacent to the SET domain (Aagaard et al. 1999). The fly and mammalian proteins have a forth region in common, the "Su(var)3-9 specific" region upstream of the chromo domain. The only structural differences between the homologues are at the 21 amino end (Aagaard et al. 1999). The Clr4 protein begins at the chromo domain while there is an additional 215 or 40 amino terminal residues in the fly and mammalian proteins respectively (Aagaard et al. 1999). Three lines of evidence suggest that the Drosophila Su(var)3-9 protein functions as a chromatin protein. First, its mammalian and yeast homologues are components of chromatin. Second, Su(var)3-9 has domains shared with other Drosophila proteins present in chromatin. Third, mutations in the gene suppress PEV. Several, but not all, Su(var) loci encode chromatin proteins (for exceptions see Wallrath 1998). This thesis will address three questions concerning the Su(var)3-9 gene. First, how has its presence in this compound gene arrangement affected its functioning? Second, are only those mutations that map to exon C able to cause the Su(var) phenotype? For example, the 80 amino acids that it shares with eIF2y may be dispensable for its own functioning. Mutations in this region may change the amino acid sequence of the Su(var)3-9 protein but not have any phenotypic consequences. And third, what is the role of the Su(var)3-9 protein in chromatin structure? Understanding the function of the Su(var)3-9 protein is necessary to test the second hypothesis as well. 1.5 Summary Compound loci are somewhat of an enigma. O n one hand they can allow the coordinated regulation of functionally similar or related genes, a seemingly 22 beneficial gene arrangement. Yet, despite the abundance of operons in bacteria, most metazoans studied (with the obvious exception of C. elegans) possess only a few such examples of multiple gene arrangements. It was proposed that one reason for their scarcity is that the genes in compound loci are more vulnerable to mutations than would be the case if these genes were independent on the chromosome. These ideas formed the basis for the two hypotheses tested in this thesis. The first hypothesis proposed that compound loci are relatively more vulnerable to spontaneous and induced mutations than are independent genes. Several reasons why this may be true are listed in Section 1.2. The second hypothesis reflects a characteristic of many compound loci, that of a homologous or functional relationship between the pair of genes. If this was true for the Su(var)3-9 / eIF2ylocus it may explain why this arrangement came about and has persisted. Thus the second hypothesis asked if there is a relationship between the Su(var)3-9 and e/F2y genes beyond their presence in this compound locus. To test these hypotheses required a genetic dissection of the locus. The tools used to achieve this were mutant alleles in which one or both genes were damaged as well as ectopic copies of the eIF2ygene under the control of an induceable promoter. These tools were also used to investigate the cellular processes of protein translation and chromatin structure which are dependent on the eIF2y and the Su(var)3-9 genes, respectively. 23 The data obtained in the course of these experiments is described in Chapters 3 through 7. Chapter 3 describes the generation of a variety of mutations in this locus. The first hypothesis predicts that several of these alleles would compromise both genes. To test this, the site and nature of each mutation in the locus is described in this chapter. The effect each allele has on the transcription pattern from the locus and the adult morphology is the subject of Chapter 4. This analysis shows which gene or genes are damaged by each mutation. The phenotypic analysis presented in Chapter 4 was also necessary to determine the function of each gene, as required to test the second hypothesis. A n overlap in the phenotypes of the two genes may reflect an overlap in their function. However, there were not enough mutant alleles obtained to conclusively establish the relationship between the mutant phenotypes and each gene. This was rectified with experiments that used an transformed eIF2ygene under the control of an induceable promoter. These experiments are described in Chapter 5. Chapter 6 presents a class of mutations that interact with the locus in an unusual way and may represent an unique category of new alleles. They were investigated because they may represent a novel way in which this compound loci can be compromised (pertinent to the first hypothesis) or they may reveal an overlap in the function of the two genes (pertinent to the second hypothesis). Finally, Chapter 7 examines the function of the Su(var)3-9 protein to determine if there are any similarities with the functioning of eIF2y, as predicted by the second hypothesis. 24 C H A P T E R 2: M A T E R I A L S A N D M E T H O D S 2.1 Classical genetics Flies were raised on standard yeast cornmeal media in vials (38mL) or bottles (250mL). Crosses were done at 25°C unless otherwise noted. Mutant strains are described at FlyBase <>. The posterior scutellar bristles were measured using an ocular micrometer once the flies had been immobilizing in a drop of Gary's Magic Mountant (1 part Canadian Balsam extract, 1 part methylsalicylate) and crushed between a coverslip and a slide. 2.2 Genomic D N A isolation D N A was isolated using a protocol developed in the G . Rubin lab. Twenty flies were starved for an hour to minimize isolation of D N A from ingested yeast. They were then homogenized in 500uL H B buffer (7M urea, 2% SDS, 50mM Tris pH7.5, lOmM E D T A , and 0.35M NaCl). The D N A was extracted twice with phenol/ chloroform/ iso-amyl alcohol (25:24:1) and once with chloroform/ iso-amyl alcohol (24:1) before ethanol precipitation. The D N A was resuspended in lOOuL T E buffer with lOug of RNaseA. 2.3 Polymerase chain reaction (PCR) Oligonucleotides used as primers in P C R and sequencing reactions were made by the Nucleic Acid-Protein Service Unit (NAPS) at U B C . Most PCR 25 reactions used the touchdown program: 15 cycles of 94°C for 1 minute; 68°C less 1°C per cycle for 1 minute; 72°C for 2 minutes; and then 15 cycles of 94°C, 1 minute; 54°C, 1 minute; 72°C / 2 minutes. Annealing temperatures, extension times, and the number of cycles were adjusted as appropriate. P C R products were purified for sequencing or use as probes by gel purification followed by QIAquick gel extraction (Qiagen Inc). 2.4 D N A sequencing Sequencing reactions used the AmpliTaq D N A polymerase FS and Big Dye terminators supplied by the NAPS unit. The reactions were done in a P T C -100 Programmable Thermal Controller PCR machine (MJ Research Inc) and were ethanol precipitated as recommended by the NAPS unit. 2.5 Computer analysis The molecular mass of D N A and R N A bands was calculated by the distance that they had migrated using "Gel" version 1.01 (Jean-Michel Lacroix, 1993). The optical density of bands was calculated using "NIH Image" version 1.61. D N A sequences were analyzed using "MacVector" version 5.0.2 (Oxford Molecular Group, 1996) and "AssemblyLIGN" version 1.0.7 (Eastman Kodak Company, 1993). BLAST searches (Altschul et al. 1997) of the GenBank protein and D N A database were conducted at the NCBI website of the National Institute of Health <>. BLAST searches of the Drosophila sequences 26 exclusively were done at the Berkeley Drosophila Genome Project website <>. 2.6 Southern blots Each digest used 3-5 ug of D N A and 25U of the appropriate Boehringer Mannheim (Roche Diagnostics) restriction enzyme in a 50uL volume. The digested D N A was purified with a single phenol/ chloroform/ iso-amyl alcohol (25:24:1) extraction and an ethanol precipitation. The samples were resuspended in 18uL of T E buffer. The samples were separated by electrophoreses in an 0.8% T A E agarose gel in a BioRad gel apparatus at 20V for 18 to 20 hours. The gel was treated and a capillary transfer was set up as recommended for the Hybond-N nylon membrane (Amersham Pharmacia Biotech). Radioactively labeled probes were made with the Boehringer Mannheim random prime labeling kit using Redivue l a 3 2 P ] d A T P (Amersham Pharmacia Biotech). Table 1 describes the nature and source of the various probes used. After the probes were made, non-incorporated nucleotides were removed using S-300 H R microspin columns (Amersham Pharmacia Biotech). Hybridization was done as recommended for the Hybond-N nylon membrane (Amersham Pharmacia Biotech) using a SSPE/Denhardt's based solution at 60°C. Stringency washes were lx SSC, 0.1% SDS at 60°C for 15 minutes, twice, and then 0.5x SSC, 0.1% SDS at 60°C for 15 minutes, twice. Autoradiography was done using Kodak X-omat AR, Kodak BioMax M R , or LabScientific Full Blue film. 27 Table 1: D N A fragments used as probes for Southern and Northern blots probe description size A exon A 326 bp A+B exons A & B 1029 C exon C 1418 E exon E 870 hobo hobo element 820,870 1530 P P element 2907 RP49 Ribosomal protein 49 650 H I Hisl -H2B His2b 372 H4 His4 312 origin1 3 PCR product (P25 x 5R1/3PSEQ) note: contains 230 bp of the locus and 96 bp of P element D N A PCR product (OR c x 5R1/3R1) PCR product (OR x 39KYLE/3SET) PCR product (OR x 4105f/4974r) pRG2.6X plasmid (contains a 50:50 mixture of EcoRI and HmdJJJ. fragments) PCR product (pP7c25.1 x Pin/Pin) pH4.0 plasmid (EcoRI/Hindni fragment) a variety of restriction fragments from the plasmid pBAC-HIS-C PCR product (pBAC-HIS-C x DMH2B-F/DMH2B-R) P C R product (pBAC,HIS-C x D M H 4 - F / D M H 4 - R ) a b c = The first four probes are specific to the Su(var)3-9/eIF2/locus, see Figure 1. = The origin of PCR products is shown as template x primer/primer. = OR is the wild type fly strain Oregon-R. 28 2.7 Northern blots Total R N A was isolated from 50 young adult female flies using the TRIzol reagent protocol (Gibco BRL). m R N A was purified using the Oligotex m R N A midi protocol (Qiagen Inc.)- R N A was separated by electrophoreses according to the method of Farrell (1993). R N A molecular weight markers II were used according to the manufacturer's protocol (Boehringer Mannheim). Capillary transfer and hybridization were done as recommended for the Hybond-N nylon membrane (Amersham Pharmacia Biotech). The remainder of the procedure is as described for Southern blotting except the stringency washes were lx SSC, 0.1% SDS at 60°C for 15 minutes, three times. 2.8 Germline transformation The transformation vector used was pP{CaSpeR-hs} (Thummel and Pirrotta 1991). This construct uses a white* reporter gene and an inducible heat shock promoter to transcribe the gene of interest. The eIF2 y c D N A is contained in the GM03594 clone generated as part of the Drosophila genome project and was purchased from Genome Systems. The identity of this c D N A was confirmed by sequencing. The D N A sequence resembles that published except for a single base pair change in both untranslated regions and five silent substitutions in the following codons: Glyl47, Alal53, Ile242, Leu381, and Glu438. The complete c D N A was excised from GM03594 and ligated into pP{CaSpeR-hs} to make pP{w[+mC] elFltfhs.PJ.H^s.elFly.H}. The integrity of this construct was confirmed by restriction enzyme digestion, PCR, and finally D N A sequencing. 29 The transformation construct pP{hs-eIF2y.H}was purified by cesium chloride density centrifugation (Sambrook et al. 1989). The pP-Turbo plasmid, the source of the P Transposase protein, was purified using a Qiagen plasmid midiprep. Germline transformation of flies was done using standard procedures (Rubin and Spradling 1982; reviewed by Ashburner 1989). Egg lay chambers were set up using ~300 young adult white isogenic flies and sucrose plates (5% sucrose, 1.5% agar) with a bit of yeast paste (dry yeast resuspended in Five-Al ive™) . Eggs were collected every 30 minutes. Embryos were dechorionated in 50% bleach, rinsed well and lined up on a piece of double sided tape attached to a coverslip. Once lined up the embryos were dried with a blow dryer for 75 seconds and then overlaid with halocarbon oil. The embryos were injected with a mix containing 165 ng /uL transformation construct, 325 n g / u L pP-Turbo plasmid, O.lx PBS buffer, 5% glycerol, and 1.5% green food colouring. The needle used was a capillary tube pulled on a Brown-Flaming micropipette puller (model P-80, Sutter Instrument Co.) and the injection apparatus was a Eppendorf Microinjector (model 5242). Surviving embryos were collected and grown at 17°C. 2.9 General molecular biology techniques Al l general molecular biology techniques and solutions are described in Sambrook et al. (1989). 30 C H A P T E R 3: C H A R A C T E R I Z A T I O N O F D N A A L T E R A T I O N S IN M U T A N T A L L E L E S O F T H E SU(VAR)3-9/eIF2Y L O C U S This chapter describes the isolation and characterization of several mutant alleles of the Su(var)3-9 / eIF2y\ocus. The goal of this chapter is to describe the mutations that will form the basis of the genetic dissection of this locus. The mutations will be in the Su(var)3-9 or e/F2y specific regions or in the region common to both. It is predicted that many of the alleles will affect both genes no matter the site of their specific lesions. As proposed in the first hypothesis, compound loci are more vulnerable to spontaneous and induced mutations. While there are several alleles that are specific to the Su(var)3-9 gene (Reuter et al. 1986), there are no characterized alleles that affect the eIF2y gene. Alleles that affect both genes and just the eIF2y gene will be used to test the second hypothesis, to determine if there is any overlap in the function or regulation of the two genes. In addition, mutations in the eIF2y gene will allow a genetic dissection of translation initiation in Drosophila. The effects that these mutant alleles have on the abundance of each transcript and the resulting mutant phenotypes will be addressed in Chapter 4. 3.1 The 17-series of alleles Previous work on this locus done in the Reuter lab demonstrated that lesions in the Su(var)3-9 open reading frame were responsible for the suppressor of position effect variegation phenotype shown by the existing alleles (Tschiersch et al. 1994). While most alleles of this locus have just this Su(var) phenotype, our 31 lab isolated four alleles that had additional phenotypes. Most strikingly, they were recessive lethal, unlike Reuter's alleles which were homozygous viable. The simplest explanation was that mutations with these additional phenotypes affected both genes at this locus, Su(var)3-9 and elFly. Thus the working hypothesis was that the Su(var) phenotype was due to the absence of one transcript while the recessive lethality and the other phenotypes were due to the absence of the other. This project originated with four alleles of Su(var)3-9/elFlythat are collectively known as the 17-series. For simplicity, mutant alleles of the Su(var)3-9/eIF2y\ocus will be referred to as just allele rather than their proper though somewhat misleading designation of Su(var)3-9allele. These mutations, 2A5,17A3,17D4, and 27F2, were generated in a mutagenesis screen for new Su(var) mutations performed in our lab by Randy Mottus (R. Mottus, personal communication). The screen involved mobilizing the non-autonomous P transposable elements from the Birml chromosome (Engels et al. 1987) and examining w m 4 offspring for red eyes. P elements are a class of Drosophila transposable elements that move via a D N A intermediate (reviewed by Capy et al. 1998). The P elements on the Birml chromosome all have internal deletions of various extents and cannot encode functional P Transposase proteins. These P elements were mobilized using the P Transposase source, P{ry+ A2-3j(99T3) (P{A2-3); Robertson et al. 1988). P{A2-3} is a modified P element that has been stabily inserted on the third chromosome. 32 3.1.1 The P element induced alleles IAS, 11 A3,17D4, and 17F2 are recessive lethal or semilethal Complementation tests confirmed that these four Su(var) mutations were alleles of the same gene (Table 2). (Note that TM3,Sb is a third chromosome balancer chromosome that carries the dominant mutation Stubble - bristles are very short and thick.) Al l allelic combinations were lethal except 17D4/17F2 and 17F2/17F2. The relative viability of the survivors was calculated using the formula shown. In both genotypes there were more female survivors than male. Not only were there fewer 17D4/17F2 flies, the survivors were physically defective; they had frail wings, tergite defects (tergites are the dorsal cuticle plates on the abdomen), the bristles on their thorax were very short while those on the abdomen were mostly absent, and they were sterile. There were many dead pupae in these crosses as well. The 17F2/17F2 flies were healthier in appearance though still sterile. The nature of these mutant phenotypes will be returned to in the next chapter. As the same crosses done at 22°C and 17°C yielded similar results (data not shown) these data show that all four alleles have disruptions of the same essential gene or genes. 3.1.2 The Sw(i;ar)3-9/eIF2ylocus is disrupted by P element insertions in each of these alleles Preliminary work had shown that some of these alleles had a P element inserted into the first intron (R. Mottus, personal communication). To confirm this, Southern blots were performed using D N A from all the 17-series strains digested with Apal, BamHl, Bglll, or EcoRI. Figure 2 shows one of these Southern blots prepared using the restriction enzyme EcoRI. The P probe, which is specific Table 2: Complementation tests using the 17-series (A) Genetic crosses performed wm4 17-series wm4 17-series Q — . cr QO — > - -f-Y TM3,Sb wm4 TM3,Sb (B) Offspring obtained3: female parent 2A5 male parent 2A5 17A3 17D4 17F2 0/0/62 0/0/67 0/0/95 0/0/95 17A3 0/0/142 0/0/123 0/0/118 0/0/162 17D4 0/0/124 0/0/114 0/0/116 2/14/102 17F 2 0/0/146 0/0/190 1/7/155 14/29/143 (C) Relative viablity of the exceptional genotypes: 17D4/17F2 males 0.05b females 0.33 17F2/17F2 males 0.39 females 0.81 a = The numbers of offspring from each cross are listed as: exceptional males / exceptional females / TM3,Sb offspring. b = A sample calculation of relative viability: observed male offspring 2+1 / expected male offspring (102+155) 34 Figure 2: Southern blot of the 17-series alleles This Southern was prepared with genomic D N A digested with EcoRI. The membrane was hybridized with probe P (P element), stripped, hybridized with probe A+B (5' end of the locus), stripped, and finally hybridized with probe C (middle of the locus). The figure shows the resulting films at their actual sizes. The genotypes are indicated above each lane. The sizes of relevant bands are indicated in kilobases. \ 35" probe P •*"-- CO cN •2 <C D & <3 r>. t> r>. CN r H r H r H w e l l s 4.1 2.7 2.3 probe A+B w e l l s ^^ ^^  4.1 2.7 2.2 probe C »»,• ^ ^ w e l l s ^ ^ ^ ^ ^ ^ *r | 36 to P element D N A , was used to determine the number of P element inserts present in each strain. The pattern of bands shown in this Southern blot and others were interpreted as follows. The two bands in the 2A5 and 17A3 lanes result from a single P element that contains an internal EcoKL restriction site. There are two P element inserts in the 17D4 strain, neither of which contains an EcoRI site, and eight or more in the 17F2 strain. To determine which region of the locus was disrupted in the mutants, the blot was hybridized with the A+B and the C probes which, as shown in Figure 1, are complementary to the 5' end and middle of the locus respectively. As the mutant strains used were heterozygous, all the lanes contain bands corresponding to the wild type restriction fragments from the wild type chromosomes. These are the 2.2 kb bands detected by the A+B probe and the 5.4 kb bands recognized by the C probe. Each mutant strain shows a larger aberrant band detected by the A+B probe. In the 2A5 and 17A3 lanes there is a 2.7 kb band while 17D4 and 27F2 have a 4.1 kb band. These bands are equivalent in size to one of the bands recognized by the P probe. These results show that all four alleles have disruptions at the 5' end of the locus which are associated with P element D N A , a conclusion confirmed by the next series of experiments. 3.1.3 P element inserts cause the Su(var) and the lethality phenotypes of the 2A5,17D4, and 17F2 alleles If these P element inserts are responsible for the dominant Su(var) phenotype of the 17-series alleles, one would predict that precise excision of the P element during development would revert this phenotype. Table 3 shows the 37 crosses used to produce flies that were wm4; 17-series allele/ P{A2-3}. Any-somatic excision of a P element that was causing the Su(var) phenotype should result in a Su(var)+ sector in the eye. These patches of 10% pigmentation in an otherwise red eye were observed in most 17D4/P{A2-3} and 17F2/P{A2-3} flies. Only a few 2A5/P{A2-3} females showed revertant sectors, and these sectors were less obvious. Unlike the others, the 17A3/P{A2-3} flies never showed sectors. The simplest explanation for these results is that the Su(var) phenotypes of the 2A5,17D4, and 27F2 alleles are due to insertions of mobile P elements. The above results imply that the P element insertions detected in the 5' end of the Su(var)3-9 / eIF2ylocus are the cause of the Su(var) phenotype of the 2A5,17D4, and 17F2 alleles. But what of the lethality phenotype? Recessive lethality is not associated with any previously described mutant allele of this locus (Reuter et al. 1986; Tschiersch et al. 1994). To prove that a single P element insertion is responsible for both the Su(var) and the lethality phenotypes of these alleles requires the generation of germline revertants. Revertant offspring isolated on the basis of having lost one mutant phenotype should also have lost the other phenotype. Table 4 shows the crossing schemes used to revert the Su(var) phenotype of all four alleles. Offspring that had 10% eye pigmentation were counted as revertants as this is the typical level of pigmentation shown by xvm4 flies. The 2A5 and 17D4 alleles reverted at a frequency of about 20%; that is, 20% of the flies that could have been revertants were revertants. 17F2 reverted at a lower frequency (8%) while no revertants were obtained of the 17A3 allele. These 38 Table 3: Somatic reversion of the 17-series alleles (A) Genetic crosses performed wm4 TM3,Sb,P{A2-3} ^, wm4 17-series _ ; c? ® —; $ Y Ly | wm4 TM3,Sb 17-series TM3,Sb,P{A2-3} do their eyes have Su(var)+ sectors? (B) Offspring obtained strain offspring showing Su(var)+ sectors male female 2A5 0 / 2 3 5 / 2 4 17A3 0 / 2 8 0 / 1 8 17D4 14/18 23 /23 17F 2 7 / 9 18 /22 39 Table 4: Ge rml ine reversion of the Su(var) phenotype of the 17-series alleles (A) Genetic crosses performed wm4 TM3,Sb,P{A2-3) _ wm4 17-series . Y L y T w m 4 TMS/Sb w m 4 17-series ^ wm4 Ly Y ' TM3,Sb,P{A2-3} w m 4 ' TM3,Sb 17-series TM3,Sb or Lyare these pheno typ ica l ly Su(var)+? (B) Offspr ing obtained strain offspring: r evers ion sex Su(var)+ Su(var) frequency 2A5 males 72 255 22% females 27 126 18% 17A3 males 0 358 0% females 0 343 0% 17D4 males 30 108 22% 17F 2 males 21 255 8% (C) Revertant strains isolated homozygous viable: 2A5rv2,2A5rv3,2A5rv7,2A5rv8, a n d 17F2rv4 homozygous lethal yet v iable w i t h 2A5: 2A5rv4,17D4rv4,17D4rv5, 17D4rv9,17D4rvlO, 17D4rvll, a n d 17F2rv8 40 crosses show that the alleles that were previously demonstrated to revert in somatic tissue were also the ones that reverted in the germline. Twelve independent Su(var)+ revertants were established as stocks (all were from different vial crosses). Eleven of these were either homozygous viable or were viable and fertile when crossed with 2A5. This implies that the lethality phenotype mapping to the locus has also been reverted in these eleven strains; some must carry second site mutations that prevent the recovery of homozygous flies. The exception was the partial revertant 17D4rv9. Its level of pigmentation in a wm4 background was intermediate, ranging from 30 to 50% in males and 50 to 90% in females (as estimated visually). 17D4rv9/2A5 flies had a slightly reduced viability (0.89) and lower than normal female fertility. Table 5 shows the crosses used to revert the lethality phenotype of the 17D4 and 17F2 alleles. As shown in Table 2, flies that are genotypically 2A5/17D4 or 2A5/17F2 are inviable. Only if the lethality phenotype were reverted (as indicated by the asterisk) in the germline of the male parents would there be any non-TM3,Sb offspring. While no 17F2*/2A5 male offspring were recovered, twenty one 17D4*/2A5 males were obtained. These males were tested to see if they had also reverted for the Su(var) phenotype. In an outcross, sixteen of the lethality"1" revertants were found to be also Su(var)+ revertants (four of the males produced no offspring and therefore could not be tested). The sole exception was the 17D4rvl2 line. 2A5/17D4rvl2 flies had moderate viability (0.70) and the females had very low fertility. The 17D4rvl2 flies had a moderate 41 Table 5: G e r m l i n e reversion of the lethality phenotype of the 17D4 and 17F2 alleles (A) Genet ic crosses performed wm4 TM3,Sb,P{A2-3} _ w™4 17D4 or 17F2 _ ; & ® — > °_ Y Ly ^ wm4 TM3,Sb wm4 17D4 or 17F2 „ w™4 2A5 cT « — / ^ w / Y TM3,Sb,P{A2-3} T w™4 TM3,Sb wm4 17D4* or 17F2* -— ; cT Y 2A5 do these survive? (B) Offspr ing obtained strain male offspring: exceptional o ther 17D4 21 17D4*/2A5 255 17F2 0 17F2*/2A5 110 (C) Revertant strain isolated 17D4rvl2 (homozygous lethal, semi-v iable w i t h 2A5) 42 Su(var) phenotype, males had 20-80% pigmentation, and females had 40-100% pigmentation. These results support the proposition that single P element insertions in the 2A5,17D4, and 17F2 lines are the cause of both the Su(var) and the recessive lethality phenotypes. Of the twelve Su(var)+ revertants, eleven had also reverted the lethality phenotype. Sixteen of the seventeen lethality"1" revertants of 17D4 had also lost the Su(var) phenotype. The two exceptions, 17D4rv9 and 17D4rvl2 behave genetically as partial revertants which is consistent with their D N A lesions (as revealed next). The 17F2 allele may have reverted at a lower frequency because this strain contains many more P elements that were also mobilized in these crosses (as seen in the Southern blot in Figure 2). Why the 17A3 allele did not revert either somatically or in the germline is not understood, especially in view of the results that follow. Southern blots were used to confirm that excision of the P elements was responsible for the reversion of the mutant phenotypes described above. Figure 3 shows a Southern blot prepared with D N A from the 17D4 strain and four 27D4 revertant lines. The 2.4 kb band detected with the A+B probe (Figure 3, left) is the wild type EcoRI restriction fragment (see Figure 1). It is present in all lanes as the strains used were heterozygous. As was seen before (Figure 2) there was a larger band in the 17D4 lane of 4.2 kb. This band was absent in the complete revertant lines, 17D4rvlO and 17D4rvll. The partial revertant lines 17D4rvl2 and 17D4rv9 had smaller aberrant bands of 3.7 and 2.8 kb, respectively. 43 Figure 3: Southern blot of the 17D4 revertant alleles For this Southern blot, genomic D N A was isolated from the heterozygous strains shown and digested with EcoRL The membrane was hybridized with P probe (P element), stripped, and hybridized with the A+B probe (5' end of the locus, see Figure 1). The figure shows the resulting films at their actual size. There are three lanes containing D N A from the 17D4rvl2/+ strain in the gel. The size of relevant bands are shown in kilobases beside large arrow heads. The location of the molecular weight marker bands (lambda Hmdlll) are indicated with small arrow heads, these correspond to 23.1,9.4, 6.6,4.4,2.3, and 2.0 kb. 44 probe A+B + + + + + + CN CN CN ~~ O i—i r H r H r H O r—l t-H + > > > > > > J _ I !-i S H U> S H *rv ^ 1 TT probe P + + + CN CN CN + > u > + O N > + + O r-H > 5-< > M Tf* ^ 45 The aberrant bands detected by the A+B probe in the 17D4,17D4rvl2 and 17D4rv9 lanes were also recognized by the P probe (Figure 3, right). This indicates that in each case the disruption of the locus is associated with P element D N A . (Note that even the complete revertant lines still contain P element insertions in their genomes.) These results show that complete reversion of the Su(var) and lethality phenotypes shown by the 17D4rvlO and 17D4rvll alleles is correlated with excision of the P element from the Su(var)3-9/eIF2/locus. A second Southern blot (not shown) confirmed the complete excision of the P elements in the revertant lines 2A5rv2,2A5rv3,17D4rv4,17D4rv5, and 17F2rv4. 3.1.4 Nature of the 2A5,17D4,17D4rvl2, and 17D4rv9 alleles The Southern blot in Figure 3 also reveals the sizes of the inserts in the 17D4,17D4rvl2 and 17D4rv9 alleles. Each of the restriction fragments detected by both the A+B and the P probes must be the sum of the 2.4 kb genomic region plus the P element insert. Thus, the P elements in these alleles must be approximately 1.8,1.3 and 0.4 kb long, respectively. A similar analysis confirmed that the 2A5 insert is a 2.8 kb long P element. The full length P element is 2907 base pairs long (O'Hare and Rubin 1983). If these insert sizes are correct, then polymerase chain reaction (PCR) amplification of the 5' end of the locus should produce products that are larger than normal. Using the 5R1 and 3R1 primers the wild type product would be 1.0 kb (see Figure 1). With the 2A5/+, 17D4/+, and 17D4rvl2/+ heterozygous lines there was an additional larger product, of a size consistent with the predicted size of the P element insert. A Southern blot on these PCR products confirmed that 46 they did in fact contain P element sequences (data not shown). Thus the P element inserts present in these mutant alleles are of the sizes predicted based on the Southern blot analysis and are found in the region delimited by this pair of primers. The orientation and location of the inserts in the 17-series alleles were determined by PCR. The P element primers, 5PSEQ and 3PSEQ, prime out of the P element from its 5' and 3' ends respectively (see Figure 4). By using one of these primers and either the 5R1 or 3R1 primer, the orientation of the P element relative to the Su(var)3-9/eIF2ylocus was determined by which primer pairs gave products. Figure 4 shows the successful PCR reactions. The Oregon-R wild type strain functioned as a negative control because, as it lacks P elements, there should never be any PCR products obtained when it was used as a source of template D N A . When D N A from the 2A5 and 17A3 strains was amplified, the 5R1/3PSEQ and 3R1/5PSEQ primer pairs produced PCR products. In these alleles the P element insert must be in a 3' to 5' orientation; that is, the P element has inserted into the locus such that the direction of its own transcription would be opposite that of the locus itself. When the 17D4 and 17F2 strains were used as a source of template D N A , the 3R1 /3PSEQ and 5R1 /5PSEQ primer pairs successfully produced amplified products. Therefore, these alleles have P element inserts that are in a 5'to 3'orientation. Because the location of each primer site is known, the sizes of the PCR products revealed the location of the P element inserts relative to the 5R1 and 3R1 sites (Figure 4). Surprisingly, it seems that the inserts in 2A5 and 17A3 are at 47 Figure 4: PCR analysis of the 17-series alleles P C R reactions were performed using genomic D N A from wild type (Oregon-R) or mutant flies. Each reaction used one genomic primer (5R1 or 3R1) and one P element primer (5PSEQ or 3PSEQ). 5ul of each reaction (or lOul in the case of 5R1/3PSEQ) was electrophoresed in a 1.5% agarose gel which was subsequently stained with ethidium bromide. At the top of the figure are shown two gels with the fly strain and primer pair used for each reaction. The gels themselves contain molecular weight markers (Gibco BRL) in the left most lane, the size of the PCR products is indicated in base pairs. At the bottom of the figure are shown the locations of the primers at the Su(var)3-9/eIF2y\ocus (left) and the P element (right). The lengths of the PCR amplified regions are indicated as is the length of the full sized P element. ( 43 3R1 5R1 5PSEQ 3PSEQ 3R1 5R1 3PSEQ 5PSEQ 822 318 5R1 3R1 5PSEQ 3PSEQ A B C P element 1029 bp 96 bp 60 bp 49 the same location as their amplified products appear identical in size. For both of these alleles, the P element must be about 35 bp from the 5R1 primer site and 970 bp from the 3R1 primer site. This location is in the 5' untranslated region (UTR) of exon A . Because 17D4 and 17F2 produced PCR products that appear to be the same size, these alleles also seem to possess P element inserts at the same position. The P element insert is approximately 220 bp from the 5R1 primer site and 760 bp from the 3R1 primer site, a location in the first intron. The exact locations of the 2A5 and 27D4 inserts were determined by sequencing these PCR products. 27D4 has a P element inserted into the beginning of the first intron. It has created a duplication of the genomic sequence 5 ' C G C G T C G T on either side of the P element sequences. These are the 69th to 76th basepairs of the 569 bp long first intron. The 2A5 insert caused a duplication of the sequence 5 ' T G C C G A G C that is bp 19 to 26 of the 90 bp 5' UTR. P elements usually cause eight base pair target duplications when they insert (O'Hare and Rubin 1983). Neither of these target duplications resembles the consensus sequence 5 ' G N C C A G A C in either orientation but they do share its G C richness (Ashburner 1989). Since the lesions in 2A5 and 17D4 are in the common region shared by the two genes, the phenotypes they have are likely due disruptions of both the Su(var)3-9 gene (suppression of PEV) and the eIF2y gene (recessive lethality). The most plausible explanation for 17D4 and 17F2 having the same insert is that they are the result of the same mutational event. This would imply that during the original P element mutagenesis screen, a mutagenized fly with this mutation 50 in its germline sired both the 17D4 and 17F2 lines. The only difference between 17D4 and 17F2 would have been the number of additional P element insertions elsewhere on the chromosome (one in the case of 17D4 and eight or more in the case of 17F2). A similar explanation is also likely for 2A5 and 17A3. There are, however, differences between each pair. For example, the 17F2 allele was slightly more viable than 17D4; the 2A5 allele could revert while the 17A3 allele did not. Because the explanation for these differences may be trivial or may never be found it was decided to continue this project using just the 2A5 and 17D4 alleles. To reflect this shift and to bring their names more in line with standard Drosophila nomenclature, the 2A5 allele was renamed P25 while the 17D4 mutation became P17. The location of these inserts is shown in Figure 8 at the end of this chapter. As listed in Table 9 (also at the end of this chapter), the revertant lines that were analyzed further are the P25rv2 and P17rvl0 complete revertants and the P17rvl2 and P17rv9 partial revertants. 3.2 The EMS-series of alleles There are many ethylmethanesulfonate (EMS) induced third chromosome Su(var) mutations generated by our lab (Sinclair et al. 1983). EMS is a chemical mutagen commonly used to generate small deficiencies and basepair changes in Drosophila (reviewed by Grigliatti 1986). Seventeen of these mutations were crudely mapped by recombination to the vicinity of the Su(var)3-9 locus at 3-56.4. With two exceptions (see below), none of these potential alleles interacted with the P25 allele; transheterozygous flies were viable, healthy and fertile. 51 As most alleles of the Su(var)3-9 locus are homozygous viable, allelism must be determined by recombination mapping the Su(var) phenotype of a putative allele relative to the Su(var) phenotype of a known allele of Su(var)3-9. This cross is depicted in Table 6. While all the offspring with parental genotypes will have the Su(var) phenotype (100% eye pigmentation), one half of the recombinant offspring will have the Su(var)+ phenotype (10% eye pigmentation). If both Su(var) mutations were in the same gene, the number of recombinant offspring would be exceedingly low. Patrick Whalen, an undergraduate student in our lab, used this method to identify six new alleles of Su(var)3-9, namely 311, 318,319,327,329, and 330. 3.2.1 Recombination mapping confirms that 321,318,319, and 330 are alleles of the Su(var)3-9/eIF2ylocus As part of this project, four of the EMS-series of alleles were characterized. 319 and 330 were chosen as they were typical of most mutations at this locus as they had no phenotypes other than dominant suppression of PEV. 311 and 318 were chosen because although they too were homozygous viable Su(var) mutations, initial results suggested that they may affect the splicing of the transcripts produced by the Su(var)3-9 / eIF2ylocus. Flies that were 311/P25 and 318/P25 had unusual pleiotrophic phenotypes suggesting that 311 and 318 affected more than just the Su(var)3-9 gene. However, further genetic analysis suggested that these interactions with the P25 allele were most likely due to second site mutations on the 318 and 311 chromosomes (data not shown). 5 2 These four alleles were remapped relative to a known allele of the Su(var)3-9/eIF2y\oc[xs, P25 (Table 6). As controls, P25 was mapped relative to another allele of Su(var)3-9 called 336 (it will be discussed in Section 3.3), and to a nearby Su(var) mutation, Su(var)3-85A4 (R. Mottus, personal communication; Reuter et al. 1986). Female offspring were examined and any with less than 70% pigmentation were retested by outcrossing them and examining their progeny for Su(var) individuals. The only verified Su(var)+ recombinant female offspring were those recovered in the Su(var)3-85A4 cross. As only half of the recombinant offspring were detectable, the recombination frequency was twice the number of recombinants observed divided by the total number of offspring scored. The calculated distance between the 5A4 and the P25 mutations of 2.8 c M is in close agreement with the previously determined positions of Su(var)3-8 and Su(var)3-9 at 3-53.5 and 3-56.4 respectively (Reuter et al. 1986). As no recombinant offspring were recovered from the other crosses, a maximum distance was calculated as if the very next offspring observed had been a recombinant. The maximum distances between each of the EMS-series alleles tested and P25 are too small for any of them to be mutations in a closely flanking gene (calculation of distances between intragenic mutations is confounded by gene conversion events). As predicted, there were no recombination events that separated the 336 mutation from the P25 mutation, since both are lesions in the Sw(z;ar)3-9/eIF2y locus. As the results with 5A4 and 336 confirm the validity of this mapping protocol, the Su(var) mutations 311,318, 319, and 330 are all lesions in the Su(var)3-9 gene. 53 Table 6: Recombination mapping the EMS-series alleles relative to the P25 allele (A) Genetic crosses performed wm4 + + •* ~ ; C f ® Y - + + | Su(var) + + + (parental) + + 7i  w m 4 Su(var) + d1 ($ — ; m4 + Su(var) Su(var) + + (parental) w" Su(var) Su(var) + + (recombinant) + + + + (recombinant) (T3) Results obtained Allele Su(var)+ female Su(var) female offspring offspring 3 Distance between Su(var) mutations 5A4 336 311 318 319 330 13 0 0 0 0 0 918 1062 516 + 2252 548 + 1796 819 + 2036 978 +1680 2.8 c M b <0.19 c M c <0.072 c M <0.085 c M <0.070 c M <0.075 c M a = Where there are two numbers the first is the data obtained by P. Whalen. b = The recombination frequency was calculated as: RF= (2 x Su(var)+ recombinants) / total female offspring, c = The maximum recombination frequency was calculated as if the next female offspring had been a Su(var)+ recombinant: RF= (2x1 recombinant) / (total + 1). 54 3.2.2 Sequence of the Su(var)3-9/eIF2ylocus in each mutant allele To determine the precise lesion associated with the 311,318,319 and 330 alleles, the Su(var)3-9/e/F2ylocus was sequenced in each mutant strain. The sequence obtained spanned the length of the locus from the transcription start site to the polyadenylation sequence at the end of exon F (see Figure 1). Due to the selection of PCR primers used, there were gaps of only partially readable sequence in the middle of the second intron and the middle of the 3' untranslated region in exon C. The sequence of all four alleles had consistent differences with the published c D N A sequences (Tschiersch et al. 1994) and an unpublished genomic D N A sequence (courtesy of Gunter Reuter). There were single base pair changes in untranslated regions, the middle of introns, as well as those affecting the coding regions. Of the latter, three cause silent substitutions, one in exon C (Asn614) and two in exon E (Ile242, Leu381). More importantly, there were two changes affecting the amino acid sequence of the Su(var)3-9 protein found in all four mutant alleles. The first, Glu509 to Gly ( G A A to G G A ) is likely an error in the published sequence as this residue is a glycine in the three Su(var)3-9 homologues (human, mouse, and S. pombe) and 12 other SET domain proteins (Aagaard et al. 1999; Jenuwein et al. 1998). The second, Arg627 to Ala (CGT to GCT) is a non-conserved residue but it is an alanine in four other SET domain proteins (Jenuwein et al. 1998). This region of the Su(var)3-9 gene was sequenced from the Oregon-R and Canton-S wild type strains to confirm these two discrepancies with the published 55 Su(var)3-9 protein sequence. As both these wild type strains, and the recently compiled complete Drosophila genome sequence (Adams et al. 2000), possessed these differences, they likely represent corrections of errors in the published sequence. The genome project's sequence (Adams et al. 2000) does show a third amino acid discrepancy in the Su(var)3-9 protein (Thr275 to He) but the sequencing results from five different fly strains shows this residue to be a threonine as was initially published (Tschiersch et al. 1994). Initial phenotypic analysis suggested that while 329 and 330 would be lesions affecting just the Su(var)3-9 gene, the 311 and 318 alleles may affect splicing of one or both transcripts. However, no D N A sequence changes affecting splice junctions were found. Instead the sequence analysis identified three single base pair changes that all cause missense mutations in exon C affecting the Su(var)3-9 protein. The sequences from 321,318, and 330 showed that each of these strains had a unique mutation. The 329 sequence appears to be a mixture of sequences from the 330 and 318 strains. This conclusion was drawn from the raw sequence data by comparing the relative heights of the peaks for the wild type and mutant bases. This implies that the 329 fly stock is actually a mixture of two similar mutant strains, 318 and 330; it is not considered further. The three mutations identified were each confirmed by sequencing four independent PCR products. In each case the change was a G C to A T transition which is the predominant change caused by the mutagen used, E M S (Krieg 1963). The 322 strain contains a glycine (polar residue) to aspartate (negatively charged residue) substitution (G521D) in the amino portion of the SET domain. 5 6 Gly521 is conserved in the mammalian Su(var)3-9 homologues and five other SET domain proteins - although this residue is an aspartate in Clr4p, the S. pombe homologue (Aagaard et al. 1999; Jenuwein et al. 1998). The 318 strain has a serine (polar residue) to leucine (hydrophobic residue) substitution (S616L) in a non-conserved region between the "core SET domain" and the cysteine-rich region at the extreme carboxyl-terminus (Aagaard et al. 1999; Jenuwein et al. 1998). Finally, the 330 strain has an aspartate (negatively charged residue) to asparagine (polar residue) change (D536N) in the middle of the SET domain. Asp536 is present in the mammalian and S. pombe homologues and seven other SET domain proteins - but this residue is an asparagine in four SET domain proteins (Aagaard et al. 1999; Jenuwein et al. 1998). As described in Figure 8 and Table 9 at the end of this chapter, all alleles were missense mutations in the Su(var)3-9 open reading frame. 3.3 The Jiobo-series of alleles John Locke's lab at the University of Alberta isolated six Su(var) mutations that were recombination mapped to 3-56.5, the same location as Su(var)3-9 (Locke et al. 1993). Although isolated in a P element screen for new Su(var) mutations they were not associated with new P element insertions. Since they all had new hobo transposable elements inserted at the cytological position of Su(var)3-9 gene (88D-E), they were possibly caused by insertions of this other element, hobo elements are similar to P elements as both move via a D N A intermediate and encode a single protein, their Transposase (Streek et al. 1986). The dysgenic cross that had been used to mobilize P elements was subsequently 57 found to be capable of mobilizing hobo elements as well (Locke et al. 1993). Five of the mutations; 306,310,322,328, and 340, were lethal both as homozygotes and in trans-heterozygous combinations, while the sixth, 336, was homozygous viable yet semilethal with the 340 mutation (Locke et al. 1993). These were therefore excellent candidates for alleles that disrupt the e7F2ygene and interfere with the normal functioning of the locus. 3.3.1 306,310,322,328, and 340 are lethal alleles of Su(var)3-9/eIF2y while 336 is a non-lethal allele As described above, the Su(var) phenotype of these alleles had been recombination mapped to the vicinity of the Su(var)3-9 / eIF2y\ocus. The first question that needed to be addressed is whether the recessive lethality phenotype was also due to disruptions of this locus. Table 7 depicts the complementation tests done using the hobo-series of mutations and the P25 and PI 7 alleles. The five mutations that were homozygous lethal were also lethal in combination with P25. Thus both the Su(var) phenotype and the recessive lethality shown by these five alleles has the same basis as the P25 allele. These five alleles, 306, 310,322,328, and 340, likely affect both the Su(var)3-9 gene (Su(var) phenotype) and the elF2y gene (recessive lethality phenotype). The sixth putative allele, 336, is homozygous viable yet was shown to be semilethal with 340 (Locke et al. 1993). During the genetic analysis depicted in Table 7, the 336 allele was viable with PI7 but semilethal in combination with either 322 or P25. These genotypes were less viable when the female parent in the complementation crosses was 336/336. The 336/322,336/P25 and 336/P17 58 Table 7: Complementation tests with the Ziobo-series alleles male female offspring: parent 3 parent 3 numbers*5 relative viablity 0 306 P25 0 / 0 / 156 310 P25 0 / 0 / 182 328 P25 0 / 0 / 142 340 P25 0 / 0 / 1 4 7 322 P25 0/0 / 209 322 P17 0 / 0 / 5 0 P25 322 0 / 0 / 181 P17 322 0 / 0 / 119 336 P25 29 /25 / 6 6 0.82 336 P17 21/1% / 34 1.18 P25 336 3 / 4 / 4 4 0.16 P17 336 16 / 25 / 25 1.60 322 336 3 / 1 /51 0.08 336 322 12 / 11 / 88 0.26 a = The mutant alleles in the parents were balanced over TM2,Ubx (306,310,322, and 340); balanced over TM3,Sb (P25 and PI7); or were homozygous (336). b = The offspring are listed as: exceptional males / exceptional females / TM3,Sb or TM2,Ubx offspring, c = As the 336 parent was homozygous, the expected number of exceptional offspring was equal to the number of TM2,Ubx or TM3,Sb offspring recovered. 59 flies were physically defective. They were small and had disrupted tergites on their abdomen and very short bristles. There were also many dead pupae in these crosses. The significance of these mutant phenotypes will be addressed in the next chapter. In conclusion, all six of the hobo-series mutations are new alleles of the Su(var)3-9 / eIF2y\ocus. 3.3.2 Southern blots reveal two lesions in the 5' end of the locus Southern blot analysis revealed the molecular basis of these new alleles. Figures 5 and 6 show a Southern blot of D N A from all six alleles digested with either BamHI or Pvull. To reduce possible confusion with restriction fragment length polymorphisms, the flies used to obtain D N A were from an outcross of the frofro-series strains to the wild type Canton-S strain. D N A from Canton-S flies was also included as a control. The blot was hybridized with the A+B and the E probes which detect disruptions at the 5' and 3' ends of the locus, respectively (see Figure 1). Figure 5 shows that when this blot was probed with E there were only the wild type length restriction fragments detected, a 3.8 kb BamHI band and 3.0 and 1.7 kb Pvull bands. As there were no aberrant bands, this demonstrated that there were no detectable abnormalities in 3' end of the locus in any of the mutant alleles. Figure 6 (top) shows disruptions affecting the 5' end of the locus in all the mutant alleles. As heterozygotes, all mutants show the wild type restriction fragments of 4.1 kb (and 3.9 kb) when digested with BamHI and 4.1 kb when digested with Pvull. The faint 3.9 kb in the BamHI digest is due to a 100 bp overlap between the A+B probe and the 3.9 kb BamHI restriction fragment from 60 Figure 5: Southern blot of the hobo-series alleles using the E probe Genomic D N A was isolated from flies which were +/+ (Canton-S), and hobo-series/+ (the offspring from a hobo-series/TM2,Ubx <S> Canton-S outcross). The D N A was digested with either BaraHI or Pvull. The resulting membrane was hybridized with the A+B probe (5' end of the locus, see Figure 1), stripped, hybridized with the hobo probe (hobo element), stripped, and finally hybridized with the E probe (3' end of the locus). The sizes of the bands were calculated from their distance from the origin relative to the size markers and are indicated in kilobases. The 322/+ BamHI lane was distorted by a bubble during the capillary transfer. BamHI P v u l l )be E + + + + vO o 7- O H + co co CN CO + 00 CN CO + + CO CO o co + + + + + + + + vo o CO CO CN CN CO CO CN CO vO CO CO o CO 1.7 • 62 Figure 6: Southern blot of the hobo-series alleles using the A+B and hobo probes The same blot as described in Figure 5 is shown here. Note that during the capillary transfer a bubble distorted the 322/+ BamHI lane. To allow comparison between the two films, the location of the molecular weight markers (lambda HmdIII) are shown with small arrowheads. These correspond to 23.1, 9.4, 6.6,4.4,2.3 and 2.0 kb. Bands which were detected with both the A+B and the hobo probe are indicated with asterix. 63 64 the middle of the locus (see Figure 1). Al l five lethal alleles had an identical pattern of disruptions as seen in the BamHL and Pvull lanes, while the 336 allele had a different pattern of aberrant bands. As these alleles may have been caused by insertions of hobo elements, the same blot was hybridized with hobo probe. As seen by the multiple bands in the film shown in Figure 6 (bottom), all six mutant strains contain several hobo elements. However, only in the 336 lanes did the hobo probe hybridize to some of the same bands as the A+B probe. These bands are indicated with asterisk. The next section examines the 336 allele while the following section investigates the other alleles. 3.3.3 The 336 allele has a hobo element inserted into the first intron The bands seen in the 336 lanes (Figure 6, top) are consistent with a 2.9 to 3.3 kb long insert in the 5' end of the locus in this allele. In the BamHl digest the aberrant 7.4 kb band is 3.3 kb longer than the wild type 4.1 kb restriction fragment. The Pvull aberrant bands of 4.7 kb and 2.3 kb are together 2.9 kb longer than the wild type 4.1 kb fragment. The insert would have one Pvull restriction site but no BamHl sites. As the hobo probe (Figure 6, bottom) also recognized the 7.4 kb BamHl band and the 4.7 kb Pvull band, these restriction fragments presumably contain hobo element sequences. As the full length hobo element is 3016 bp (Streck et al. 1986), the 2.9 to 3.3 kb insert in this allele is likely comprised of a full length or near full length hobo element. PCR analysis confirmed the presence of a hobo element in the 336 allele using the same strategy as was used to confirm the presence of P elements in the P25 and P17 alleles (Section 3.1.4). D N A from 336/336 flies was amplified using 65 hobo element primers in combination with the 5R1 and 3R1 primers (see Figure 1). The PCR products obtained confirmed that there is a hobo element inserted into the first intron in a 5' to 3' orientation. The size of the 5R1 / 3R1 product (calculated as 4017 bp long) indicates that this is either a full length (3016 bp) or nearly full length hobo element (Streek et al. 1986). These PCR products were sequenced and analyzed to determine that the hobo element insert is near the end of the first intron. The target site duplication, 5 ' A T A C A C A C , is the 437th to 444th basepairs of the 569 bp long intron. hobo elements create an eight base pair target duplication, usually at the consensus site 5' N T N N N N A C (Saville et al. 1999; Streek et al. 1986) as was the case for this insertion. The location of the hobo element in the 336 allele is shown in Figure 8 at the end of this chapter. 3.3.4 The 322 allele has an insert of 8.3 kb of unknown D N A that may be a retrotransposon The other five recessive lethal alleles have an apparently identical lesion in the 5' end of the locus (Figure 6, top). Unexpectedly, there was no indication that any of these alleles contained a hobo element insertion as neither the 13.1 kb BamHI fragment or the 2.8 and 7.8 kb PvuH. fragments were recognized by the hobo probe. It is not known whether the alleles were the result of the same type of mutation event that occurred repeatedly or the same mutation event, isolated repeatedly. In any case, it was decided to continue this project with the 322 allele. To determine the nature of the 322 lesion more definitively required a second Southern blot (Figure 7). D N A from Oregon-R and 322/+ flies (the offspring of an outcross with Oregon-R females) was digested with four 66 Figure 7: Southern blot of the 322 allele For this Southern blot, genomic D N A was isolated from wild type (Oregon-R) and 322/+ flies and digested with BamHI, EcoRI, Pvull, and Xhol. The membrane was hybridized with the A+B probe (5' end of the locus, see Figure 1), stripped, hybridized with the A probe (extreme 5' end of gene), stripped, and finally hybridized with the E probe (3' end of the gene, data not shown). The figure shows the eight adjacent lanes of the blot separated for clarity. The sizes of bands are shown in kilobases. 67 BamHI EcoRI Pvull Xhol + c o + probe A+B 12.5 + + CN CN ^ ~ CO + + + CN CN CO + 7.9 + + CN CN CO + 21.0 9.5 7.9 4.2 3.8 £ WW 5.1 27 Z m 2*4 JIB 2.9 2.4 probe A 12.5 7.9 21.0 7.9 4.2 2.7 • 2.4"- 2.4 68 restriction enzymes. The blot was hybridized with the A+B probe as before (Figure 7, top). Al l the +/ + and 322/ + lanes show bands corresponding to the wild type length restriction fragments. Note that the wild type Pvull restriction fragment on the Oregon-R chromosome is 2.4 kb rather than the 4.1 kb as seen with Canton-S in Figure 6. As indicated on Figure 1, the Pvull site immediately upstream of the gene is absent in some fly strains. The 322/+ lanes in Figure 7 (top) show one (BamHl) or two (EcoRI, Pvull, and Xhol) aberrant bands when hybridized with probe A+B. A n interpretation of these results is that there is an 8.3 to 8.4 kb insert in the 5' end of the locus. The explanation, described next, is also shown in diagram form at the bottom of Figure 8. The restriction enzymes either did not cut in this predicted insert (BamHl) or cut it once (Pvull or Xhol) or more than once (EcoRI). The insert must be in the middle of the region recognized by the A+B probe - the portion of the locus between the transcription start site and the beginning of the second intron. This is because the probe binds strongly to two restriction fragments when the insert is cleaved by EcoRI, PvuU, and Xhol. To account for the different sized restriction fragments seen in 322 it is best to explain each restriction digest in turn. The aberrant 12.5 kb BamHl restriction fragment is 8.3 kb longer than the wild type 4.2 kb fragment. The simplest explanation for this would be if there were 8.3 kb of new D N A between the BamHl sites in the locus (see Figure 8, bottom). The PvuU aberrant fragments, 7.9 and 2.9 kb are together 8.4 kb longer than the 2.4 kb wild type fragment, consistent with the insertion of an 8.4 kb piece of D N A that contains a 69 single Pvull site. As the 5.1 and 2.7 kb bands from the EcoRI digest are together only 5.4 larger the 2.4 kb wild type restriction fragments this suggests that EcoRI cuts the insert more than once. The Xhol aberrant fragments are together not as large as the wild type restriction fragment. The only way to account for this in the context of the proposed 8.3 to 8.4 kb insert is if there is a Xhol site closer to the 5' end of the locus on the 322 chromosome than the Oregon-R chromosome. If there is an Xhol site at the position indicated in Figure 8, then a single'X/zoI cut site in the insert could account for these results. If this explanation, shown in map form in Figure 8, is true, the reason there were two aberrant bands in the EcoRI, Pvull, and Xhol digest is that these enzymes cut inside the insert. During the Southern blot the A+B probe could therefore hybridize with the regions of the Su(var)3-9/eIF2y\ocus both upstream and downstream of the insert. This proposition was tested by reprobing the membrane with the A probe which recognizes only the first 230 bp of the locus. As predicted, this probe only hybridized to one of the two aberrant fragments (Figure 7, bottom). This has allowed a tentative restriction map of the insert to be determined. Figure 8 shows the insert in the first intron and the location of the restriction sites based on this position. Also shown are the sizes and orientation of the restriction fragments recognized by the A+B and A probes. These results are all consistent with the presence of an insertion of 8.3 to 8.4 kb of unknown D N A into the Su(var)3-9/eIF2y locus between the Pvull and BamHI restriction sites as indicated on Figure 8. The Southern blot shown in Figure 7 was hybridized with the E probe but there were no detectable 70 alterations at the 3' end of the gene (data not shown), further supporting this conclusion. Attempts to PCR amplify the region containing this putative insert were not successful. Although other interpretations of the data may be possible, a small insertion is the simplest way to explain this pattern of bands. The insert of unknown D N A in the 322 allele is likely a retrotransposon. These transposable elements differ from P and hobo elements in that they move via an R N A intermediate. Several Drosophila retrotransposons are 7.0 kb or larger (reviewed by Arkhipova et al. 1995). A n insertion of a large retrotransposon such as the 8.7 kb B104 element into the 5' end of the Su(var)3-9 / eIF2y\ocus may have occurred during the screen for Su(var) mutations (Locke et al. 1993). If this insertion occurred in the germline of a fly that was part of an early generation during the genetic crosses it may have been isolated repeatedly. This would explain why the 306, 310, 322, 328, and 340 alleles all contained identical disruptions despite being isolated from independent parents (J. Locke, personal communication). In summary, the presence of a retrotransposon in the 322 allele accounts for the restriction fragments seen on the Southern blots, the mechanism by which the mutation was created, and the reason why four other independently isolated alleles appear to share this lesion. 3.4 Screen for new alleles with P element mutagenesis A l l alleles of the Su(var)3-9 / eIF2y\ocus were generated as Su(var) mutations, therefore all alleles affect the Su(var)3-9 gene. Mutations exclusive to the eIF2 y gene would enable better genetic dissection of this gene both by 71 themselves and in combination with other mutant alleles. However, none of the P element induced mutations generated as part of the Drosophila genome project are allelic. Ten P element insertions that had been mapped to the vicinity of the Su(var)3-9/eIF2y\ocus (88D10-E1) were tested for allelism but all were viable and fertile in combination with P25 (complementation test data not shown). There are no known deficiencies that include this locus. Since deficiencies are by definition null mutations, they make good baselines when describing the phenotypes of hypomorphic alleles. Schmucker et al. (1997) did generate four X-ray induced deficiencies that included the 88E region but these strains had poor viability and soon perished (Ulrike Gaul, personal communication). P element mutagenesis was used in an attempt to generate new alleles. Judy Tsai, an undergraduate student in our lab, assisted with this project. Nearby P element inserts, obtained from the Fly Genome Project, were mobilized to generate either new insertions into the Su(var)3-9/eIF2ylocus or deletions that extended into the locus (a "local P hop" screen). Table 8 details the genetic cross scheme used to mobilize five different P element inserts. Males were generated (line 2) that had both the P{A2-3} Transposase source and one of the P{PZ} or P{lacW} inserts. From their offspring, males were recovered (line 3) that had a potentially mutagenized third chromosome (as indicated by the asterisk). These males were individually mated to P25/TM3,Sb females. Had any of these males carried a new recessive lethal allele, there would be only TM3,Sb offspring produced (line 4). Thus each cross of a male from line 3 represents a single chromosome scored. The putative allele could then be 72 Table 8: Screen for new alleles using P element mutagenesis (A) Genetic crosses performed (1) wm4 TM3,SbP{A2-3j , + P{PZ) n ; O ® ; — — — - ¥• Ly + TM3, Sb (2) + PfPZj ^ W 4 Ly Y TM3, Sb Pf A2-3) wm4 TM3, Sb (3) wm4 PiPZr Y TM3, Sb ? u;7"4 P25 w m 4 TM3, Sb (4) P{PZ}* TM3, Sb P25 TM3, Sb P{PZ}* P25 absent? (B) P element insert lines used Insert line Cytological location number of chromosomes scored P{PZ}ejfl462 g 8 D 5 6 i q 4 7 P{PZ}1(3)0355003550 88E8-9 631 P{lacW}l(3)sl782s1782 8 8 D 5 . 6 5 4 6 P{PZ}1(3)0071600716 8 8 E 1 _ 2 1 0 1 g P{PZ}1(3)0471304713 8 8 E 1 . 2 795 total= 4037 73 recovered from the surviving P{PZ}*/TM3, Sb flies and retested. About 4037 chromosomes were tested for lethality with P25. Unfortunately, none of the putative alleles isolated passed retesting. A modified form of the screen was undertaken but it too was unsuccessful despite testing a further 2178 chromosomes. A different type of P element mutagenesis was then attempted. The Su(var)3-914 allele has a P element inserted into the middle of exon C (Tschiersch et al. 1994). If this P element could be mobilized it might generate recessive lethal alleles if it excised imprecisely and removed at least a portion of the eIF2ygene. To this end, crosses were undertaken to generate flies that were Su(var)3-914/ P{A2-3}. However, this genotype was lethal. A Southern blot confirmed the presence of at least 16 P elements in the Su(var)3-914 strain (data not shown). With so many P elements being mobilized by the P{A2-3] Transposase there would have been lethal levels of genetic damage in the Su(var)3-914/P{A2-3} flies. Continuing this imprecise excision screen would have required removing these additional P elements using meiotic recombination. 3.5 Summary This chapter has described the generation and characterization of a variety of mutant alleles of the Su(var)3-9 / eIF2y locus. The location of these lesions is summarized in Figure 8 while their characteristics are listed in Table 9. While none of the mutations obtained mapped to the eIF2/specific regions 74 Figure 8: The location of the mutant alleles in the Su(var)3-9/eIF2ylocus The figure is as described for Figure 1 except the positions of the mutations are indicated. The insertions of P elements (P25, P17, P17rvl2, P17rv9), hobo elements (336), and unknown D N A (322) are shown to scale. The location of the three missense mutations in exon C are also indicated. The diagram at the bottom shows the approximate location of the insert in the 322 allele, the possible location of restriction sites in this insert, and the restriction fragments generated. 75 A B C D E F <-7 .9P2.9 - * + 2.7 R R5.1 > <-7.9X9.5 -> o o o o o o o o o o o o o o o o rfr1 O r - i O O H O S (N r - i r-< T * rO CO CO CO rO CN CO ON 0 0 ( N O r H CN (N Ir^  00 r H ID r i N ~. co r-< 1 1 r n C N l r O T f in ^ O t r ^ o o o o CN I 1 76 Table 9: Summary of the mutant alleles Al le le Description P25 2.8 kb P element inserted into the 5'UTR in 3' to 5' orientation P25rv2 complete revertant of P25 P17 1.8 kb P element inserted into the first intron in 5' to 3' orientation P17rvl2 partial revertant of PI 7 that retains 1.3 kb of the P element P17rv9 partial revertant of PI 7 that retains 0.4 kb of the P element P27rvl0 complete revertant of P27 322 G521D missense mutation in the SET domain of the Su(var)3-9 open reading frame 318 S616L missense mutation adjacent to the SET domain of the Su(var)3-9 open reading frame 330 D536N missense mutation in the SET domain of the Su(var)3-9 open reading frame 322 8.3 to 8.4 kb insert of unknown D N A into the 5' end of the locus, possibly a retrotransposable element 336 complete or nearly complete hobo element inserted into the first intron in 5' to 3' orientation 77 downstream of exon C, several of the mutations are in the region common to both genes. A l l of these were insertions of transposable elements (or unknown D N A ) in or between the two shared exons at the 5' end of the locus. Drosophila transposable elements preferentially insert at the 5' end of genes into promoter regions and the first exons and introns. This was the case with all four independent insertion alleles. P25 and PI 7 were induced P element insertions in the 5' UTR and first intron, respectively. The 322 allele was likely caused by a retrotransposon that inserted early in a screen for new Su(var) mutations as proposed in Section 3.3.4. If true, this mutation would be spontaneous in origin. The 336 allele, although not due to the insertion of a P element as first thought is due to another transposable element that was mobilized during the Su(var) screen (Locke et al. 1993). It can be considered induced or spontaneous as, though it was the consequence of a hobo dysgenic cross, it was not an insertion of the desired transposable element. The remaining alleles, 311,318, and 330, were missense mutations induced by the chemical mutagen EMS which preferentially causes the G C to A T transitions found in each. Returning to the hypotheses under test, three of the insertion alleles are candidates for disrupting both genes, consistent with the proposed vulnerability of compound loci to spontaneous and induced mutations. These mutations are like those predicted in Section 1.2 whereby a single mutation has disrupted two genes. These were anticipated to be a common occurrence for compound loci but a rare event for independent genes. While no mutations affecting the splice 78 sites directly were recovered, the 336 allele is a good candidate for affecting the alternative splicing pattern of this locus, as shall be investigated in the next chapter. The remainder of this thesis will use these mutations to address questions concerning how this locus functions and the roles of the eIF2/and Su(var)3-9 proteins in Drosophila. This relates to the second hypothesis, how the roles and the regulation of these two proteins are a consequence of their presence in a compound locus. The effect of these mutations on the transcription of these two genes and the resulting mutant phenotypes is the subject of the next chapter. 79 C H A P T E R 4: T R A N S C R I P T A N D PHENOTYPIC A N A L Y S I S O F T H E M U T A N T A L L E L E S O F T H E SU(VAR)3-9/eIF2y L O C U S This chapter seeks to connect the mutations described in the previous chapter to their consequences for the organisms that carry them. This is necessary for testing both hypotheses. Most of the mutations were found to be in the common 5' end of the locus and may therefore affect both genes. If this were the case, these alleles would demonstrate how a single mutation could have an atypically large effect due to the overlapping nature of compound loci. Thus it was necessary to confirm that these mutations do or do not affect both genes. This was done using two complementary approaches. Each mutation was tested for alterations in the Su(var)3-9 and eIF2ytranscript amounts in adult females and for changes in adult morphology in different genetic backgrounds. The latter phenotypic analysis was also necessary to test the second hypothesis. If there were to be an overlap in the mutant phenotypes that could be assigned to each gene, there may be an overlap in the function of their proteins. 4.1 Northern blot analysis of the Su(var)3-9feIF2y mutant alleles The alleles described in Table 9 were predicted to have different effects on the functioning of the Su(var)3-9/eIF2ylocus. The missense mutations, 311,318, and 330 have their impact by changing the amino acid sequence of the Su(var)3-9 protein. The other alleles, being insertions of P elements (P25, P17, P17rvl2, and P17rv9), a hobo element (336), or unknown D N A (322) may hinder the initiation of transcription or they may interfere with the transcript processing. These 80 mutant strains are therefore likely to possess altered amounts of each transcript or perhaps even aberrant transcripts. Certain mutant strains may also affect the ratio of the two transcripts that are present. This section describes a series of Northern blots used to establish how each of the key alleles changes the transcription pattern at the Su(var)3-9/eIF2y\ocus. Figure 9 shows the results of three Northern blots done with m R N A isolated from adult females. For the homozygous lethal lines, heterozygous flies were used that were generated by an outcross with Oregon-R wild type flies. The membranes were simultaneously probed three times to detect the 2.4 kb Su(var)3-9 transcript (probe C , see Figure 1), the 2.0 kb eIF2ytranscript (probe E), and the 0.6 kb Ribosomal Protein 49 transcript, a commonly used loading control (RP49 probe). For each hybridization a single film is shown, the image has been cropped to indicate which lanes were from different Northern blots. The difference in optimal exposure times, seven days in the case of the Su(var)3-9 transcript and two days with eIF2y, suggested that there was relatively more eIF2ythan Su(var)3-9 m R N A present in the adult females (results also obtained by Tschiersch et al. 1994). This occurred even though the probe for the e/F27transcript was made from a smaller PCR product (870 bp) than that used to make the Su(var)3-9 specific probe (1418 bp). However, differences in the specific activity and guanosine/cytidine content of each probe cannot be ruled out. This apparent difference in transcript abundance may be attributed to an overall bias favoring production of the eIF2 y transcript, the stability of this 81 Figure 9: Northern blots of the Su(var)3-9 and eIF2y transcripts in the mutant alleles. m R N A was isolated from mutant, revertant, and wild type adult females reared under identical conditions. The m R N A was separated by electrophoreses in three gels and then transferred to membranes. The three membranes were hybridized together with the probe C (Su(var)3-9), then probe E (eIF2y), and finally probe R P 4 9 (PR49). After each hybridization, film was exposed for an optimal period of time (seven days, two days, and four hours, respectively). The blots were never stripped because this causes spurious bands to appear when the membrane is reprobed (personal observation). The figure shows a single film (twice actual size) from each probe that has been cropped to indicate the lanes in each Northern blot. 82. M l 8-9 I Ml o W MJ , o +/+ C T E(var)3-5[01]/ + P17/ + P17rvl2/ + P17rvl0/ + t 1 P25/ + P25rv2/P25rv2 322/ + Dp(3;3)E8/ + + / + 336/336 311/311 318/318 330/330 83 m R N A , or differences in the temporal and spatial expression patterns of the two genes. Reuter's lab detected a third transcript of 2.4 kb that hybridized with an eIF2yspecific probe (Tschiersch et al. 1994). It was less abundant than the 2.0 kb eIF2/transcript. They speculated that this minor transcript was a variant of the eIF2 y transcript. If it does exist it may encode an isoform of the eIF2y protein, perhaps with a different amino terminus, the only variable part of the protein (Erickson et al. 1997). A n even more intriguing result would be if this transcript were non-functional. If a defective transcript was being produced at a detectable amount this would imply that the Su(z;ar)3-9/eIF2ylocus was not optimal. This would have interesting implications for the regulation of compound loci such as this one. No such transcript was detected in adult females from either wild type strain even at long exposure times (data not shown). As this is a negative result, the existence of a third minor transcript produced by this locus cannot be ruled out. However, as the third transcript is the same size as the normal Su(var)3-9 transcript its appearance in the Northern blot shown in Tschiersch et al. (1994) could be just an experimental error. Incomplete stripping of the blot between hybridizations or contamination of the probe with sequences that would recognize the 2.4 kb Su(var)3-9 transcript could account for this band. There is no other evidence that supports the presence of this third minor transcript. A screen of a c D N A library for Su(var)3-9 specific cDNAs (Sarb Ner, 84 personal communication) did not identify any that differed from the published sequence. A l l seven of the eJF2ycDNAs available from the Fly Genome Project resemble the published eIF2 y c D N A sequence. As they have only been sequenced from their 5' ends, their entire sequence is not known. Five of these cDNAs contain sequence from exons A , B, D, and E while the other two contain sequence from exons A , B, and D (see Figure 1). None of the Northern blots, including those shown in Figure 9, contained any aberrant transcripts in any of the mutant lines. This was a possibility if one of the insertion alleles was still transcriptionally competent. To use the PI 7 allele as an example, some of the 1.8 kb P element present in the first intron may have been included in either the Su(var)3-9 or elFly transcripts. Although none were observed, aberrant transcripts may be produced in one or more of the mutant strains but in quantities too small to be detected. Furthermore to be detectable, an abnormal transcript would have had to have been of a different size than the normal transcripts, polyadenylated, and contained enough of either exon C or E to hybridize with the probes used. As predicted, several of the mutant alleles have noticeable changes in the steady state levels of the Su(var)3-9 and eIF2/transcripts relative to the RP49 loading control. The bands were quantified and the results shown in Table 10. To quantitate the bands, the developed X-ray films were scanned into a computer using video image captures and then the optical density of each band (less the background) was determined using N I H Image software. Table 10 shows the optical density of the Su(var)3-9 and eIF2/bands expressed as a 85 Table 10: Quantified amounts of the Su(var)3-9 and eIF2y transcripts in the mutant alleles genotype Su(var)3-9 eIF2y +/+ 0.598 (1.08) 0.719 (0.95) E(var)501/+ 1.109 (2.00) 0.638 (0.84) P17/+ 0.477 (0.86) 0.537 (0.71) P17rvl2/+ 0.843 (1.52) 0.721 (0.95) P17rvl0/+ 0.786 (1.41) 1.000 (1.32) P25/+ 0.493 (0.89) 0.511 (0.68) P25rv2/P25rv2 0.657 (1.18) 0.939 (1.24) 322/+ 0.372 (0.67) 0.656 (0.87) Dp(3;3)E8 0.742 (1.34) 1.019 (1.35) +/+ 0.513 (0.92) 0.794 (1.05) 336/336 0.299 (0.54) 0.654 (0.86) 311/311 0.715 (1.29) 0.701 (0.93) 318/318 0.755 (1.36) 0.691 (0.91) 330/330 0.635 (1.14) 0.691 (0.91) The data presented is the adjusted amount of each transcript - the ratio OD$u(var)3_g/ODRp4g and ODeIF2y/ODRP4g calculated from each lane using N I H Image 1.61. To allow comparison between lanes, the values shown in parentheses are these ratios relative to the average wild type ratio for the Su(var)3-9 (0.556) and eIF27(0.756) transcripts. 86 fraction of the optical density of the RP49 band from each lane (referred to as the "adjusted amounts"). This standardization controlled for unequal loading between lanes. Also shown are these values relative to the average adjusted amount from the two wild type lanes. The Dp(3;3)E8 and E(var)3-501 strains will be discussed later in Section 6.5. The adjusted amounts of the Su(var)3-9 transcript were 0.513 and 0.598 in the wild type strains. The results for the mutant lines are listed relative to the average of these wild type values. The adjusted amounts of the Su(var)3-9 transcript were unexpectedly higher in the complete revertant lines (P17rvl0 and P25rv2) and in the EMS-series of alleles (311,318, and 330). As expected, there was a decrease in the mutant strains 336 (hobo element insert) and 322 (8.3 kb insert). Although P25 and P17 showed less Su(var)3-9 than their complete revertant alleles, P25rv2 and P17rvl0 respectively, there was only a small decrease relative to wild type. The adjusted amounts of the eIF2/transcript showed the same pattern. There was less eIF27in P17 than P17rvl2 and P17rvl0; less eIF2ym P25 than P25rv2. There was also a slight decrease in 336 and 332. It is notable that the 336/336 flies appear to have a greater decrease in the Su(var)3-9 transcript (0.54) than the eIF27transcript (0.86). Once again, however, the levels in the wild type lanes were less than the complete revertants. While these results are consistent with the complete revertants possessing an increased amount of eIF2 y in comparison with the levels present in the mutants, it is harder to make the 87 reciprocal argument, that the mutants caused a reduction in the amount of the eIF2 7 transcript in the first place. Whether these discrepancies were due to the known limitations of this method of analysis or represent genuine fluctuations in the abundance of the two transcripts is not clear. As most of the alleles tested are homozygous lethal, the effects of these mutations on the amounts of the transcripts has to be assessed in heterozygous flies. Because these flies were capable of producing at least 50% of the normal amounts of each transcript from their wild type alleles, the transcript analysis cannot be absolute. In addition, because the m R N A was obtained from whole adults, any effects the mutant lesions may have on the abundance of either transcript in specific tissues and during specific developmental periods cannot be determined. Perhaps the ratio of Su(var)3-9 to eIF2 7transcript production is sensitive to any change at the locus, or perhaps sensitive to even different genetic backgrounds. This second point could explain why the EMS-alleles appeared to have more Su(var)3-9 transcript yet less eIF2y relative to the wild type strains. As no aberrant transcripts were detected it seems most likely that the insertion alleles cause a decrease in both transcripts. This reduction in both transcripts may be due to a decrease or loss of transcription or transcript processing efficiency. Alternatively, the transcripts may be defective and quickly eliminated by the nonsense mediated m R N A decay pathway (reviewed by Culbertson 1999). Because Northern blots measure steady state amounts of R N A a fourth possibility, that the mutations cause decreased stability of the 88 transcripts cannot be ruled out. Only a single mutation, 336, appears to upset the ratio of the two transcripts. As presented in Table 10,336/336 females had a greater decrease in the Su(var)3-9 transcript than the eIF2y transcript. 4.2 Phenotypic analysis of the Su(i;ar)3-9/eIF2ymutant alleles The preceding section has allowed the mutant alleles to be divided into two groups. Three alleles affect the Su(var)3-9/eIF2y\ocus by changing the amino acid sequence of the Su(var)3-9 protein (311,318, and 330), while the remainder likely decrease the amounts of both transcripts (P25, PI7, P17rvl2, 322, and 336). Now that the nature of these mutations is known, their effects on fly morphology can be determined. The phenotypes of the mutant flies reveal the consequences of a decrease or loss of Su(var)3-9 and eIF2y protein activity. These results are necessary for understanding the function of each of these proteins and for determining if these roles overlap. The latter point is important because any similarity in function between these proteins suggests that there may be coordinated regulation and perhaps a selective advantage in this gene arrangement (as stated in the second hypothesis). The mutant alleles of the Sw(z?ar)3-9/eIF2ylocus are all dominant suppressors of the variegation of the white* gene in the wm4: rearrangement. This is to be expected as this is the criterion by which they were recovered from different mutant screens. Alleles specific to the eIF2ygene would not be expected to have this phenotype but, as Section 3.4 described, the screen for these alleles was unsuccessful. (This topic is returned to in Chapter 6 when a new class of 89 mutations is introduced.) With the exception of the partial revertants P17rvl2 and P17rv9, all alleles completely suppress wm4 variegation; their eyes are dark red and almost indistinguishable from those of wild type flies. Most Su(var) mutations are general in that they suppress all variegating genes (Sass and Henikoff 1998 and references therein). Unexpectedly when this was tested with P25,322, and other alleles with the chromosomal rearrangements Inversion(2R)brownVDe2 {bwVDe2)and Bar^oneVariegator ( B S C V ) it was found that they did not have the same strength as mutations in other Su(var) genes. Flies that are bwVDe2/+ ; +/ + have brown eyes as the brown gene, responsible for the deposition of the bright red pigments in the eye, variegates. bwVDe2 flies that also carried a mutation in the gene encoding HP1, that is bwVDe2/Su(var)20501 ; +/+ flies, had wild type-like dark red eyes as the variegating brown gene had been completely suppressed. Male and female flies that were bwVDe2/+ ; P25/+ had an intermediate ruddy brown phenotype, as did flies carrying the 318,322, and 336 alleles. bwVDe2/+; P25rv2/+ flies had brown eyes as expected. The results with B s c v were comparable. The variegating gene in this rearrangement is the Bar gene, responsible for the proper shape of the eye. The allele on this chromosome is a dominant mutant allele that causes the eyes to be narrower. Unlike the variegation of wild type genes, variegation in this case allows for a more wild type phenotype due to the inactivation of the mutant gene. Su(var) mutations allow more expression from this mutant Bar allele and 90 hence a narrower eye. Su(var)3-9/balancer flies were crossed to B s c v flies and the eyes of the offspring that were Bscv/+ ; Su(var)/+ were compared to those that were Bscv/+ ; balancer/+. The width of the right eye of the male offspring was measured with an ocular micrometer. These values are shown below together with the standard deviation and the sample size (N). cross Bscv/+; Su(var)3-9/+ Bscv/+; balancer/+ 322 177 ± 47 um (N=12) 245 ± 65 um (N=12) P25 259 ± 44 um (N=9) 310 ± 13 um (N=9) P25rv2 291 ± 28 um (N=10) 303 ± 15 um (N=8) The offspring with the 322 and P25 alleles had narrower eyes than their siblings, consistent with these alleles being general suppressors of PEV. However, in both crosses there was an overlap in the eye phenotypes of the Su(var) and the balancer flies. These results imply that none of these alleles of Su(var)3-9 have the same magnitude of effect on PEV as do alleles of Su(var)205 and other Su(var) genes. A genetic analysis of these mutant alleles can be used to reveal information about the other protein encoded by this locus, eIF2y. The role of the eIF2y protein at the cellular level is well understood from work done in S. cerevisiae and human cell cultures, as discussed in Chapter 1. However, the consequences of decreasing its activity in a multicellular organism have not yet been studied. Mutations that disrupt the Drosophila eIF2ygene and appear to decrease the steady state amount of its transcript were described above. What 91 can the mutant phenotypes they cause reveal about the tissues and developmental stages that are particularly sensitive to decreases in eIF2y activity? The complementation tests described in Tables 2 and 7 showed that the P25, P17, and 322 alleles were lethal in all genotypic combinations. It is consistent with the essential role of the eIF2y protein that mutations in its gene would be recessive lethal. Because three independent alleles were all recessive lethal, it is unlikely that there is a second, functionally redundant eIF2 ygene elsewhere in the fly genome. This was a possibility considering the compound nature of the Su(z;ar)3-9/eff2ylocus. However, BLAST searches (Altschul et al. 1997) of the recently completed Drosophila genome database (Adams et al. 2000) did not identify any other eIF2ygenes. The 336, P17rv9, and P17rvl2 alleles had reduced viability in combination with the P25 allele as described in Sections 3.1.3 and 3.3.1. The viability of the P25/P17rv9 and P25/P17rvl2 genotypes was 0.89 and 0.70 relative to their siblings, respectively. The 336 allele was different from the other alleles as it was recessive viable yet semilethal in combination with other alleles. As Table 7 shows, there was a decrease in the relative viability of the 336/322 and 336/P25 genotypes. There were many dead pupae indicating that this developmental stage is part of the lethal period of these genotypes. The surviving 336/322 and 336/P25 flies and to a lesser degree those of the 336/P17 flies were small, had disrupted tergite cuticle on their abdomens, and very short bristles. The first two defects can be attributed to the flies barely surviving development with 92 insufficient amounts of an essential protein (eIF2Y) but what of the bristle defects? In addition to recessive lethality, P25,322 and other mutant alleles of this locus share other mutant phenotypes with a class of mutations called the Minutes. Minutes are a group of fifty of more genes dispersed through the genome (reviewed in Lindsley and Zimm 1992; Lambertsson 1998). Loss of function mutations at these loci are unusual in that they have dominant phenotypes in addition to their recessive lethality. These phenotypes include reduced body size, viability, and fertility, small thin bristles, cuticle defects, rough eyes, prolonged larval development, and interactions with unrelated genes (Ferrus 1975; Sinclair et al. 1981; Sinclair et al. 1984). In general, Minute loci encode ribosomal proteins (Saeboe-Larssen et al. 1997; Seeboe-Larssen et al. 1998; and references therein). The pleiotropic mutant phenotypes that all the Minutes share are presumably due to a general reduction in protein biosynthesis which impairs many cellular processes. Do mutations in eIF2y, which presumably affect protein translation, cause the same phenotypes as mutations in ribosomal protein genes? Many alleles of Su(var)3-9 / eIF2yshow the most obvious Minute characteristic: short thin bristles (Ferrus 1975). To quantitate this phenotype, the longest bristles on the fly, the posterior scutellars on the thorax, were measured using an ocular micrometer (length=A). Table 11 shows the average length of the posterior scutellar bristles, x(A), in females of several genotypes. Males gave comparable results (data not shown). For all alleles, the flies used were the offspring of an outcross with 93 Table 11: The bristle lengths of Su(var)3-9/eIF2y alleles genotype N x(A) x(A/B) (all female) +/+ 13 447 ± 23 um 2.61 ± 0 . 1 5 Df(2L)M24F-B/+ 13 317 ± 14 pm 2.01 ±0 .06** P25/+ 12 309 ± 13 um 1.89 ± 0.13** P25rv2/+ 11 458 ± 16 um 2.59 ± 0.08 P17/+ 11 429 ± 1 4 um 2.17 ± 0.06** P17rvl2/+ 14 473 ± 8 um 2.48 ± 0.07* P17rvl0/+ 12 474 ± 7 um 2.47 ± 0.08* 318/+ 12 451 ± 21 um 2.56 ± 0.14 322/+ 14 383 ± 17 um 2.13 ± 0.09** 336/+ .13 486 ± 7 um 2.62 ± 0.06 N is the number of females scored (only one bristle was measured per fly). x(A) is the average length of the posterior scutellar bristles ± standard deviation. x(A/B) is the average ratio of [posterior scutellar bristle length / posterior to anterior scutellar bristle distance] ± standard deviation. Where a value was significantly smaller than the wild type results (as determined with Student's t test) it is indicated as ** for p « 0 . 0 0 0 5 and * for p<0.005. 94 Oregon-R. While wild type females had bristles that were 447 ± 23 um long, those of a typical Minute mutation, Df(2L)M24F-B, were only 317 ± 14 um. The P25, PI 7, and 322 females all had significantly shorter bristles than wild type females. Because the length of bristles is also dependent on the size of the fly, Table 11 also contains a value x(A/B). This ratio was calculated by dividing the length of each posterior scutellar bristle (length=A) by the distance between it and the adjacent anterior scutellar bristle (distance=B) for each fly. These corrected values give a better indication as to how compromised bristle formation was during pupation. As seen in the table, the bristle length to spacing ratio was significantly smaller in the P25/+, P17/+, and 322/+ genotypes ( p « 0 . 0 0 0 5 ) . Most of the other alleles had near wild type bristle length to spacing ratios. The partial and complete revertants of the P27 allele, P17rvl2 and P17rvl0 respectively, had significantly smaller ratios as well (p<0.005) although they in fact had slightly longer bristles than the wild type flies. These results confirm that certain alleles have significantly shorter bristles, a hallmark of Minute mutations. Another phenotype associated with Minute loci is the enhancement of the mutant phenotypes of unrelated mutations (see, for example, Sinclair et al. 1984). This enhancement is ascribed to the decrease in translation efficiency exacerbating the effects of any other mutation. To test whether alleles of the Su(var)3-9/eIF2ylocus shared this property, P25 flies were crossed to flies with mutations that have visible phenotypes known to be enhanced by Minute 95 mutations (Sinclair et al. 1984). While some interactions were found, they were not a severe as those caused by previously characterized Minute mutations. For example, alleles of the vestigial gene, vg1 and Df(2R)vg-B, were only mildly enhanced by certain Su(var)3-9/eIF2y alleles: Df(2R)vg-B /+;+/+ normal wings Df(2R)vg-B/ +; P25/ + wings have small notches at the distal tips Df(2R)vg-B/Df(2L)M24F-B;+/+ wing border absent The P25 allele was found to enhance the phenotype of the Delta3 mutation (wing vein defects) but not those of the mutations Jammed34e (wings are fluid filled) or clipped1 (wing margins missing). In addition to small bristles and enhancement of unrelated mutations, Minute mutations often have other dominant phenotypes including cuticle defects, rough eyes, prolonged development, reduced size, viability and fertility. Flies that were 336/P25,336/322 and 336/P17 also had these phenotypes except their eyes were normal (development time was not measured). However, as heterozygotes none of the Su(var)3-9 / eIF2y mutant strains had any of these defects when examined. 4.3 Summary The Northern analysis can be summed up as four points. Each must be qualified due to the limitations of Northern blot analysis as outlined above. First, 96 as the eIF2 y transcript was detected with shorter exposure times than the Su(var)3-9 transcript the former is likely to be more abundant in whole female flies. Second, there were neither aberrant transcripts detected in any of the mutant lines nor was a potential third transcript detected in the wild type lines. Third, there were less of both transcripts in the 322 and 336 strains than the wild type strains. For the P25 and P17 strains there was less of both transcripts in comparison to their complete revertant alleles, P25rv2 and P17rvl0, but the difference with the wild type strains was not as pronounced. The forth point is that the only mutation that altered the balance between the two transcripts was the 336 allele that appears to cause a preferential decrease in the abundance of the Su(var)3-9 transcript. In order to verify these statements, replicates of the Northern blots would have to be done. Better support would come from using a different technique such as Northern analysis using homozygous embryos, transcript quantification using RNase protection assays, or Western blot analysis using antibodies specific to each protein. The phenotypic analysis revealed that there were three primary phenotypes associated with the Su(var)3-9/eIF2ylocus. A l l mutant alleles aside from the partial revertants P17rvl2 and P17rv9 showed a dominant Su(var) phenotype. The PI 7, P25, and 322 alleles were also recessive lethal and had abnormally short and thin bristles. These mutations had Minute-like phenotypes though they were mild in comparison with previously examined Minute loci. As many Minute loci encode ribosomal proteins it was not unexpected that mutations that affected another component of the translation machinery, eIF2y, shared the same mutant phenotypes. 97 Let us assume for a moment that the working hypothesis is true and that the Su(var) phenotype is due exclusively to the reduction or loss of wild type Su(var)3-9 activity while the lethality and other Mmwte-like phenotypes are attributable to just the eIF2 y gene. The results of these Northern and phenotypic analyses could be then interpreted to divide the mutations into four categories: those specific to the Su(var)3-9 gene (322,318, and 330) and those that affect both genes strongly (P25, P17,322) or moderately (P17rvl2 and P17rv9) or the Su(var)3-9 gene preferentially (336). These results support the first hypothesis as there are four independent mutations, P25, P17,322, and 336, that compromise both genes in this compound locus. Aside from 336, each is a single mutation that in addition to being lethal when homozygous interferes with two cellular processes, chromatin structure and protein translation, when heterozgous as shown by the dominant Su(var) and Minute-like phenotypes. Conversely, these results provide no support for the second hypothesis. A complete separation between those phenotypes associated with the Su(var)3-9 gene and those associated with the eIF2 y gene would suggest that there is no overlap in the function of these genes. For this reason the proper attribution of phenotypes to genes is necessary to test the second hypothesis. There are two concerns that must be addressed if the second hypothesis is to be investigated in this way. First, it has not been ruled out whether loss of the Su(var)3-9 gene was also contributing to these Minute phenotypes. It is possible for chromatin proteins to influence protein biosynthesis. For example, the chromatin protein 98 Modulo is encoded by a Su(var) locus (Garzino et al. 1992). However, as it is preferentially found at the nucleolus it has been implicated in ribosome biogenesis (Perrin et al. 1998). Further support for a role in ribosome biogenesis for Modulo was the finding that somatic clones homozygous for a mutation in the modulo gene have the short bristle phenotype characteristic of Minute mutations (Perrin et al. 1998). If the Su(var)3-9 protein has a similar role as the Modulo protein in the nucleolus, then both proteins encoded by the Su(var)3-9 / eIF2ylocus would be involved in protein synthesis. In this situation, the lethality and bristle phenotypes could be due to both genes being affected. This would be an important finding with regards to the second hypothesis. A compound locus that encodes two dissimilar proteins both involved in protein translation would raise mteresting questions as to the origin and regulation of the locus. The second obstacle to the resolution of these phenotypes is the possibility that the lethality phenotype is partially or completely due to the disruption of a neighboring gene. The transposable element insertions that have caused each lethal allele are capable of acting on neighboring genes as well as the Su(var)3-9/eIF2y\ocus. P elements can impair genes that are up to 50 kb away (Hugh Brock, personal communication). Another problem is that there may be a second elF2ygene. Although a BLAST search of the recently completed fly genome did not reveal any other eIF2 y genes, at the time these experiments were done this was still a possibility. A second eZF2ygene could have been left behind by the events that generated the Su(var)3-9/eIF2y locus. If there is 99 another gene producing the eIF2y protein, the eIF2ygene in this compound locus may not be essential. The most parsimonious interpretation of this genetic data remains that the Su(var) phenotype is due to an alteration in Su(var)3-9 gene function while all the other phenotypes are due to disruption of the eJF2ygene. None of the mutations specific to the Su(var)3-9 gene (322,318,330) had any phenotypes beyond suppression of PEV. But it is necessary to prove that it was only the reduction of the eIF2y transcript that was responsible for the phenotypes of recessive lethality, sterility, and visible abnormalities described above. It is possible that some of these phenotypes were due to a decrease of both the Su(var)3-9 and the eIF2/transcripts simultaneously or that an unidentified neighboring gene was also involved. The resolution of these issues is the subject of the next chapter. 100 C H A P T E R 5: A S S I G N M E N T OF M U T A N T PHENOTYPES T O T H E G E N E S O F T H E SU(VAR)3-9/eIF2y L O C U S U S I N G A T R A N S F O R M E D e l F Z y G E N E The previous chapter has identified several mutant phenotypes associated with the Su(var)3-9/e/F2y locus. In order to test the second hypothesis it is necessary to identify any similarities in the function of these two genes. If the Su(var)3-9 and eIF2y proteins are involved in a similar cellular process the loss of either of them may cause the same type of defects. To determine if this is true it is necessary to establish the gene or genes responsible for each mutant phenotype. The goal of this chapter is to determine whether the three primary phenotypes, suppression of PEV, recessive lethality, and Minute bristles, are due to loss of just the Su(var)3-9 gene, just the eIF2y gene, both genes simultaneously, or another neighboring gene. Since there are no mutations that affect only exons specific to the eIF2y transcript, the phenotypes ascribed to a reduction of eIF2y activity are an inference based on the difference between those mutations affecting both transcripts and those specific to the Su(var)3-9 transcript. To determine which mutant phenotypes were due only to a reduction of the elFly transcript, flies were transformed with a construct containing an elFly gene. It was predicted that induction of an elFly transgene in Su(var)3-9 / eIF2y mutant flies would "rescue" the mutant phenotypes of recessive lethality and defective bristles but not affect PEV. 101 5.1 Germline transformation with the pP{hs-eIF2y.H} construct A total of 3355 isogenic white embryos were injected with the pP{hs-eIF2y.H} and pP-Turbo plasmids. 780 (23%) of these embryos survived to hatch as first instar larvae and 370 (11%) eclosed as adults. These survivors were individually mated with male or female white isogenic flies. The progeny from fertile survivors were examined for white+ eyes (which ranged in colour from pale yellow to red). 18 survivors produced one or more white+ offspring. In each case, a few white+ eyed progeny were collected and crossed back to white isogenic flies to maintain the lines. Usually all white+ siblings derived from a single survivor had the same level of white+ expression, suggesting that they all carried the same insert. However, in four of the 18 lines there was a noticeable difference in eye colour between the white+ siblings. From each of these four cases, two distinct transformed lines were established. Thus, 22 transformed lines were recovered in total. These numbers are similar to other transformation experiments where approximately 10% of adult survivors produce transformed offspring (Ashburner 1989). 5.2 Characterization of the P{hs-eIF2y.H} transformant lines The transformed lines were named P{hs-eIF2y.H}TlA through P{hs-eIF2y.H}T18 by the order in which they had been isolated (Table 12, column a). Those that were descendants of the same injected survivor were designated A and B. Segregation analysis was used to determine which chromosome the 102 Table 12: P{hs-eIF2y.H} transformed lines N a m e a Chromo Number Pigmentation Phenotype of -some*5 of inserts0 ofr/+malesd T/T homozy gotes e TIA n 1 30% sickly, short bristles, sterile TIB III 3 90% lethal T2 in 1 70% T3 II 1 40% lethal T5 II 1 60% T6A II 1 30% slightly sickly, female low fertility T6B X 1 70% T7A II 1 40% slightly sickly, male sterile T7B III 2 90% T8A ni 2 70% T8B ni 1 10% T9 in 1 20% T10 ii 1 30% TU ii 1 5% T12 II 1 20% male sterile, female low fertility T13 X 1 70% TU m 1 40% T15 n 1 20% lethal T16 in 1 20% wings curved, sterile T17 in 1 20% male sterile T18 ii 1 30% 103 inserts were on. The transformants T6B and T23 segregated with the X-chromosome, T4 was never successfully mapped, while the remaining 19 transformants were assigned to the second or third chromosome. As shown in Table 12, column b there are ten transformed lines with inserts on the second chromosome and nine with inserts on the third chromosome. Balanced stocks were made from all transformant lines except for T4. The approximate level of white* activity, as indicated by the amount of eye pigmentation shown in T/+ heterozygous males, was assayed visually. As shown in Table 12 column d, it ranged from the light yellow of the T11/+ males (-5% wild type pigmentation) to the dark red of T1B/+ males (-90%). Once in stocks, any surviving homozygotes were examined for visible phenotypes and tested for fertility. Several transformant lines had homozygous phenotypes ranging from physical defects (sickly) to sterility to complete lethality (Table 12, column e). However, as there was no consistent pattern, these phenotypes were likely the result of disruptions of various genes on the chromosomes caused by the inserts rather than an inherent property of the inserts themselves. To determine the number and structure of inserts in each transformed line required Southern blot analysis. D N A was isolated from T / + heterozgotes from each transformant line and digested with HmdIII or EcoRI. These Southern blots (three in total) were probed with the E probe, which recognizes both the endogenous and transformed copies of the eIF2ygene (see Figure 1). EcoRI cuts once in the middle of the insert, so each independent insert will show a different unique band containing 2.7 kb of the inserted construct and an unknown amount 104 of flanking genomic D N A . The E probe will recognize the endogenous elFly gene as a 5.4 kb band. As HmdHI cuts the insert at its extreme ends, an intact full length insert will appear as a 7.7 kb band, the endogenous eIF2y gene as a 19.2 kb band. Figure 10 shows one of these Southern blots. The lanes on the left contain D N A digested with EcoRI. As most lanes show only one band in addition to that of the endogenous gene, most transformed lines carry a single insert (Table 12, column c). The exceptions were the lines T7B and T8A which both carry two inserts on the same chromosome. T8A has one insert in common with T8B. The TIB line carries three inserts (data not shown). Those lines with multiple inserts have greater overall white* expression as seen by their darker eyes (Table 12, column d). Because all the transformant lines contain the 7.7 kb Hmdffl band on this and the other two Southern blots, all the constructs are intact. 5.3 Successful rescue of the dominant bristle phenotype The reason for generating the transgenic lines was to determine which mutant phenotypes were due to a reduction of just the eIF2/transcript. These phenotypes should be "rescued" in genotypes that have an induced transgene. As the transgene was under the control of the promoter from the Heat shock protein 70 (Hsp70) gene, elevated temperatures during development allowed its expression. The induction was done by rearing flies under three regimens; 25°C (continuously), 29°C (continuously), and 2 5 / 3 7 ° C (flies were reared at 25°C but were heat shocked for thirty minutes daily by partially immersing their vials in a 105 Figure 10: Southern blot of the P{hs-eIF2y.H} transformed lines D N A was isolated from wild type and eight transformed lines (all heterozygous) and digested with EcoRI and HmdIII. The membrane was hybridized with the E probe (eIF2/gene). The figure shows the resulting film. The distances traveled by the molecular weight markers (lambda HmdIII) are shown, as are the sizes (in kilobases) of pertinent bands. The endogenous gene appears as a 5.7 kb EcoRI restriction fragment and a 20.2 kb HmdIII restriction fragment. 106 EcoRI Hindl l l probe E wells • + + + + + + + \ ! > > I N C O C O O N r - l r - - l r - i + + + + o + < pa < pa v . N h CO CO ^ 4- f—' E™H h~* E™^ E™* + CO rH H 2 3 . 1 • 9.4 • 6.6 4.4 • 2 0 . 2 18.3 15.7 2 . 3 • 2 . 0 107 37°C water bath). While the latter treatment should produce enough eIF2y transcripts to be phenotypically noticeable, the other two may have been adequate as the Hsp70 promoter allows a low level of transcription at non-heat shock temperatures. The dominant bristle phenotype shown by P25 and other alleles was potentially a result of reduced eIF2y activity during pupation. The crosses done to rescue this phenotype are shown in Figure 11, line 1. They were more complex than necessary because other offspring from these crosses were used to set up the rescue experiments described in the next section. As shown on line 2, these crosses produced P25/ + heterozygotes both with and without a second chromosome insert (T). Because these crosses were done in a white mutant background, those flies with the insert were distinguishable as only they had pigmented eyes. A l l second chromosome transformed lines were tested for their ability to rescue the shortened bristles of P25 flies. Inserts on the second chromosome were used as they were the easiest to use for the subsequent rescue crosses. Visually scoring the P25/+ flies reared at 25°C and 29°C was inconclusive - flies with and without the transformed insert all had short bristles. However, P25/+ flies reared at 2 5 / 3 7 ° C did have noticeably longer bristles when they also carried an insert. The amount of rescue (that is, bristle length) varied somewhat between the different inserts. The four strongest, T1A, T5, T7A, and T18 were used in subsequent experiments. 1G8 Figure 11: Rescue crosses using the P{hs-eIF2y.H} transgenes The crosses here were those done to rescue the bristle defect of the P25 allele (part A) and to rescue the lethality of P25/P25, P25/P17, and P25/322 flies (parts B and C). Shown are the X, second and third chromosomes. T refers to any second chromosome P{hs-eIF2y.H} transgene. TM3,Sb and TM6,Tb are third chromosome balancers which carry the Stubble and Tubby dominant markers, respectively. (A) Rescue of bristle phenotype (1) w T TM3,Sb g . w_ _ _+_ P25 Q y" ' T ' + | ~w ' T TM6,Tb (2) w T P25 versus w + P25 w/Y' T ' ~ w/Y + ' + (B) Rescue of the lethality phenotype with a single P{hs-eIF2y.H} (3) w T P25 w_ . _+_ P25 or PI7 or 322 ~Y~ ' "+* TM3,S& 1 w ' + TM3,Sb or TM6,Tb (4) it; T P25 or P17 or 322 w/Y ' + ' P25 do these survive? (C) Rescue of the lethality phenotype with multiple P{hs-eIF2y.H} (5) w_ _T P25 . 13 . P25 or P17 ~ Y ' + TM3,S& f w ' + TM3,Sfc * (6) w T P25 or P17 w/Y T/+ P25 do these survive? 110 The bristle length data obtained for the T5 insert is shown in Table 13. It shows the four genotypes tested (all female) under the three temperature regimens. The column labeled x(A) shows the average length of the posterior scutellar bristles (just as in Table 11). At 25°C the transgene had no effect, the two P25/ + genotypes had short bristles while the other genotypes had normal long bristles. However, at 2 5 / 3 7 ° C the T5/+ ; P25/+ females had bristles which were noticeably longer than their +/+; P25/+ siblings and almost as long as the wild type females. Since bristle length is also dependent on the size of each fly, the ratio of the length of the posterior scutellar bristle to the distance between the posterior and anterior scutellar bristles, x(A/B), was also calculated (Table 13, right most column). At 2 5 / 3 7 ° C , the T5/+ ; P25/+ females had a significantly larger bristle ratio (2.10 ( 0.073) than their +/+ ; P25/+ sibs (1.64 (0.083). The phenotypic rescue was not complete as the wild type and 75/ +;+/+ females had larger bristle ratios still. At 29°C the slight increase in bristle ratio due to T5 was not significant. In summary, these results demonstrate that induction of an eIF2y transgene was partially able to rescue the dominant bristle defect associated with the P25 mutation. 5.4 Successful rescue of the recessive lethality phenotype It would be expected that if the Su(var)3-9/eIF2y\ocus were the only source of the eIF2y protein in Drosophila then this eIF2 y gene would be essential. As shown in Tables 2 and 7, flies with any combination of the P25 (2.8 kb P I l l Table 13: Rescue of the Minute bristle phenotype by P{hs-eIF2y.H} Genotype Temp. N x(A) x(A/B) + P25 + + 25°C 29°C 2 5 / 3 7 ° C 9 14 14 338 ± 18 um 349 ± 23 um 339 ± 17 um 1.74 ± 0.117 1.86 ± 0.135 1.64 ± 0.083 T5 P25 + + 25°C 29°C 2 5 / 3 7 ° C 12 13 14 338 ± 28 um 380 ± 25 um 428 ± 14 um 1.72 ± 0.132 1.99 ± 0.148 2.10 ± 0.073 + + + + 25°C 29°C 2 5 / 3 7 ° C 17 15 13 473 ± 8 um 442 ± 12 pm 447 ± 13 pm 2.73 ± 0.131 2.64 ± 0.076 2.65 ± 0.098 T5 + + + 25°C 29°C 2 5 / 3 7 ° C 16 13 13 485 ± 11 pm 458 ± 11 pm 452 ± 18 u m 2.55 ± 0.092 2.57 ± 0.092 2.37 ± 0.048 Genotype lists the second and third chromosome constituents. Temp, is the temperature regimen the flies were reared under. N is the total number of flies scored. x(A) is the average length of the posterior scutellar bristles ± standard deviation. x(A/B) is the average ratio of [posterior scutellar bristle length / posterior to anterior scutellar bristle distance] ± standard deviation. 112 element insert/5'UTR), P17 (1.8 kb P element insert/first intron), and 322 (8.3 kb insert) alleles were inviable. The recessive lethality of these genotypes was likely due solely to insufficient eff2y mRNA during development. However, it was possible that the recessive lethal alleles affected another nearby gene or that there was a lethal interaction involving the Su(var)3-9 gene. To resolve this, Figure 11, line 3 shows the crosses done to rescue the lethality of the P25/P25, P25/P17, and P25/322 genotypes using a single transgene. Any surviving exceptional flies would be distinguishable as they would not have the phenotype of the Stubble mutation of the TM3,Sb balancer chromosome nor the Tubby mutation of the TM6,Tb balancer chromosome. Table 14 shows the results of these crosses that were done using the transformant lines T1A, T5, T7A, and T18 and at the three different temperatures. While no exceptional P25/P25 flies survived past pupation, P25/P17 and P25/322 flies were recovered under the 25/37°C temperature conditions. In all cases these flies also carried the transgene as seen by their ivhite+ eyes. Table 14 also lists the relative viability of all the surviving exceptional offspring. There were fewer T1A/+; P25/P17 and T5/+; P25/322 flies recovered than predicted based on the number of sibling flies in these crosses. Conversely, the crosses using the T7A and T18 transgenes did generate more than the expected number of P25/P17 and P25/322 flies. As these genotypes were more viable than their siblings, these transgenes may have allowed complete rescue of these genotypes. The P25/P17 and P25/322 survivors were healthy in appearance. They were large and had properly formed cuticle. The only visible defect was that all 113 Table 14: Rescue of recessive lethality by single copies of P{hs-eIF2y.H} + P25 TM3,Sb P25 TM6,Tb T1A 2 5 ° C 2 9 ° C 2 5 / 3 7 ° C 0/0/146 0/0/42 0/0/56 P 2 7 TM6,Tb 322 TM3,Sb o + 0/0/54 0/0/110 0/1/50 (0.08) not done (nd) nd nd T5 2 5 ° C 2 9 ° C 2 5 / 3 7 ° C 0/0/149 0/0/39 0/0/51 T7A 2 5 ° C 2 9 ° C 2 5 / 3 7 ° C 0/0/147 0/0/62 0/0/40 T18 2 5 ° C 29°C 2 5 / 3 7 ° C 0/0/159 0/0/53 0/0/36 0/0/111 0/0/90 13/10/94 (0.98) 0/0/208 nd 3/8/62 (0.71) 0/0/107 0/0/87 8/9/42 (1.62) nd nd 4/9/46 (1.13) 0/0/69 nd 0/0/72 nd 4/7/34 (1.29) nd Numbers of offspring from each cross are listed as: exceptional males / exceptional females / TM3,Sb or TM6,Tb offspring. The relative viability of exceptional offspring is shown in parenthesis. The expected number of exceptional offspring was calculated to be 1/4 of the number of TM3,Sb and TM6,Tb offspring. Assuming for complete viability, l / 6 th of the offspring would be Su(var)3-91Su(var)3-9 and have a transgene while 4/6th of the offspring would be Su(var)3-9/TM3,Sb or Su(var)3-9/TM6,Tb with or without a transgene. 114 the bristles were very short. For example, the average length of the posterior scutellar bristles of the T5/+; P25/P17 females was only 314 ± 20 um (x(A/B) was 1.73 ± 0.044; N=7). In comparison, T5/+ ; +/+ females had 452 ± 18 um long bristles when reared at 2 5 / 3 7 ° C (Table 13). The survivors also took noticeably longer to develop than their siblings. They did not begin to eclose until one to two days after their siblings had finished eclosing. When their fertility was tested by allowing survivor males and females to mate with wild type flies under the same 2 5 / 3 7 ° C regimen about half of both sexes were fertile. In addition to the above rescue crosses, more ambitious crosses were set up to rescue the lethality of P25/P25 and P25/P17 flies using two transgenes at once. As shown in Figure 11 line 5, the paternal parent carried a transgene (either T1A, T5, T7A, or T18 as before) while the maternal parent carried T5. Offspring from the cross may have carried up to two transgenes. As Table 15 shows, there were many P25/P17 exceptional flies recovered at the 2 5 / 3 7 ° C regimen. In all cases these flies also carried at least one transgene. There was a single P25/P25 fly recovered as well. As before, the P25/P17 flies were phenotypically healthy except for extremely shortened or absent bristles. In summary, the rescue of the recessive lethality phenotype required the presence of the eIF2/transgene and a daily heat shock to induce its transcription. Thus, eIF2yis the only essential gene lacking in all three of the mutant alleles: P25, P17, and 322. 115 Table 15: Rescue of recessive lethality by double copies of P{hs-eIF2y.H} T P25 T5 P25 _ • Cf <g) — ; + TM3,Sb + TM3,Sb o T5 P17 + TM3,Sb T1A 25°C 29°C 2 5 / 3 7 ° C 0/0/84 0/0/63 0/1/39 (0.07) 0/0/73 0/0/123 10/10/43 (1.24) T5 25°C 29°C 2 5 / 3 7 ° C 0/0/66 0/0/51 0/0/43 0/0/51 0/0/81 13/12/59 (1.13) T7A 25°C 29°C 2 5 / 3 7 ° C 0/0/85 0/0/80 0/0/40 0/0/96 0/0/83 11/20/37 (2.23) IIS 25°C 29°C 2 5 / 3 7 ° C 0/0/89 0/0/51 0/0/39 0/0/74 0/0/112 8/16/57 (1.12) Number of offspring from each cross are listed as: exceptional males / exceptional females / TM3,Sb offspring. The relative viability of exceptional offspring is shown in parenthesis. The expected exceptional offspring were 3/8th the number of TM3,Sb offspring (3/12th of the offspring would be Su(var)3-9/Su(var)3-9 and have at least one transgene while 8/12th of the offspring would be Su(var)3-9/TM3,Sb). 116 5.5 The eIF2 y transgene has no discernible influence on PEV If the reduction in eIF2/transcripts caused by certain Su(var)3-9 / eIF2y mutant alleles was partially responsible for their Su(var) phenotype, then induction of the P{hs-eIF2y.H) transgene may have the opposite effect. The transgene may enhance PEV by itself, or it may counteract the Su(var) phenotype of the P25 allele. The easiest variegating rearrangement to assay is wm4 but, as the transgene carries the white* minigene, the resulting eye pigmentation will not be exclusively due to the variegating white gene. To partially circumvent this, the third chromosome transgene, T8B, was chosen as it had a low level of white* expression {T8B/+ males had only 10% of normal pigmentation; see Table 12, column d). The effect of the Su(var)3-9 alleles on other variegating rearrangements such as bwVDe2 and B s c v is not strong (as discussed in Chapter 4) making them poor choices for testing. To test the effect of T8B, the following cross was done: w/Y; T8B/T8B males <8> wm4/wm4; P25/TM3,Sb females The cross was done at both 25°C (transgene -off) and 2 5 / 3 7 ° C (transgene on), and the level of pigmentation in the offspring was assessed visually. When raised at either temperature, the wm4/Y; T8B/P25 males had 90 to 100% pigmentation while the wm4/Y; T8B/TM3 Sb males always had light orange eyes (~20%) with flecks of darker pigmentation. As the heat shock treatment had no affect on eye colour, the transgene had no discernible influence on wm4. Had the transgene enhanced PEV, the males (especially the wm4/Y; T8B/P25) 117 would have had noticeably less eye pigmentation when raised at 2 5 / 3 7 ° C . Conversely if the transgene had suppressed PEV, the males (most noticeably the Wm4/Y • T8B/TM3,Sb) would have had more eye pigmentation when treated with heat shocks. The lack of an effect on PEV by the T8B transgene may have been due to a low level of expression of eIF2 y from this transgene. This was a possibility as there is only a low level of white gene activity from this transgene as indicated by the pale yellow eyes of w/Y; T8B/+ males (Table 12). To rectify this, other transgenes which had different levels of white gene expression were tested for their influence on wm4. T9 (20% wild type pigmentation, Table 12), TU (40%), and TIB (90%) were tested but they neither increased nor decreased the amount of pigmentation of wm4 and wm4; P25 flies when they were reared at 2 5 / 3 7 ° C . In summary, the eIF2 y transgene had no discernible effect on P E V by itself nor did it influence the Su(var) phenotype of the P25 allele. 5.6 Summary This chapter investigated three key phenotypes associated with the representative P25 allele, the Mmwte-like phenotypes of short thin bristles and recessive lethality as well as the Su(var) phenotype. To determine which phenotypes were due to a reduction in the amount of eIF2 y activity during development, twenty two transgenic fly lines were generated which carried one or more copies of the P{hs-eIF2y.Hj construct. Once these inserts were mapped to chromosomes they were established as stocks, tested for homozygous 118 phenotypes, and verified with Southern blot analysis. Daily heat shocks during development were used to drive expression of the eIF2 y transgene. The induction of the transgene allowed rescue of two of the three phenotypes tested. These conditions were sufficient to allow significant, though incomplete, rescue of the dominant short bristle defect of the P25 allele in P{hs-eIF2y.H}T5/ +; P25/ + females. A similar set of crosses showed that another phenotype associated with Minute mutations, recessive lethality, could also be rescued by the transgenes. The transgenes tested allowed survival to adulthood of P17/P25, 322/P25, and, in a single case, P25/P25 flies. These survivors were physically healthy but had delayed development, very short bristles, and were sometimes sterile. Four inserts were tested for an ability of the eIF2 y transgene to influence P E V but none caused a change in the pigment levels of wm4 or wm4; P25 flies when induced. The enhancement of the wing defects of vg1 and Df(2R)vg-B flies by the P25 mutation was also investigated, but the phenotypes were too mild to conclude whether or not the eIF2 y transgene was having an influence (data not shown). These results show that while the Minute bristle and lethality phenotypes were partially or completely rescued by induction of the eIF2 y transgene, PEV was unaffected. Therefore, these Minute phenotypes are caused by a scarcity of the eIF2 y transcript during development. It was necessary to assign phenotypes to genes in order to test both hypotheses. To test the first hypothesis, which states that compound loci are more vulnerable to mutations, it was necessary to determine which gene or genes are affected by each allele. The P25, P17, and 322 119 alleles have now been confirmed to affect both genes as they have mutant phenotypes associated with each. Each is a single mutation that by interfering with two genes has several phenotypic consequences, those attributable to the Su(var)3-9 gene (suppression of PEV) and to the eIF2ygene (recessive lethality and Minute bristles). This situation whereby single mutations compromise multiple cellular processes was proposed in Section 1.2 to be one way in which mutations of compound loci would be more severe than mutations of independent genes. The second hypothesis concerns the functional overlap of these two genes. It was suggested in Section 4.3 that the Su(var)3-9 gene may also be involved in translation initiation and may therefore be contributing to the Minute-like phenotypes of recessive lethality and dominant bristle defects. Because induction of the eIF2 y transgene was sufficient to allow otherwise non-viable genotypes to survive past pupation, the lethality phenotype is associated with this gene exclusively. The transgene was also able to significantly increase the bristle length of mutant flies linking the bristle phenotype and the eIF2ygene. The reason the transgene only partially rescued the bristle defects was likely due to insufficient induction of the transgene during a critical stage during pupation when the bristles are formed. This could be confirmed by repeating the experiment but increasing the number or duration of heatshocks during this developmental period. Because both the lethality and the Minute bristle phenotypes are associated with just the eIF2y gene there is no evidence that the Su(var)3-9 120 protein is also involved in protein translation. Until the role of the Su(var)3-9 protein is tested directly in Chapter 7 there is no support for an involvement of this protein in the same cellular process as the .eIF2.Y protein. Therefore, these results do not support the second hypothesis as there is no evidence of an overlap in the function of the two genes of the Su(var)3-9 / eIF2y compound locus. While both of these Minute phenotypes is due to just the eJF2ygene, the P E V phenotype remains slightly enigmatic. As expected, the transgene could not "rescue" the Su(var) phenotype as the entire Su(t7ar)3-9/eIF2ylocus could when it was transformed into flies (Tschiersch et al. 1994). While this implies that mutations in the e7F2ygene do not have a Su(var) phenotype, the eIF2ygene may influence PEV in another manner. As the next chapter investigates, mutations in just the eIF2y gene may cause an E(var) phenotype. None of four transgenes had any effect on PEV. However, it has not been ruled out that loss of the eIF2y gene has an effect on P E V that is masked by the dominant Su(var) phenotype caused by loss of the Su(var)3-9 gene in all of the existing alleles. 1 2 1 C H A P T E R 6: A P O T E N T I A L L Y N E W CLASS OF A L L E L E S O F T H E SU(VAR)3-9/eIF2y L O C U S A n interesting genetic interaction was discovered between certain alleles of the Su(var)3-9/eIF2ylocus and three nearby E(var) loci (Dorn et al. 1993c). E(var) mutations enhance the transcriptional repression associated with PEV. For example, a wm4; E(var) fly has white eyes because the variegation (inactivation) of the white* gene has been increased. E(var) mutations are either loss of function alleles of single genes (true E(var) loci) or they are duplications of Su(var)+ genes (Locke et al. 1988; Sass and Henikoff 1998). Certain Su(var) genes cause enhancement of P E V when they are triploid. That is, increasing the amount of a Su(var) protein by increasing its gene dose has the opposite phenotype as a decrease in its abundance or activity. For example, Duplication(3;3)E(var)88D8 (Dp(3;3)E8), a tandem duplication of the region between 88D4-6 and 88E4-F2, was isolated as a dominant E(var) mutation (Locke et al. 1988). Its E(var) phenotype is ascribed to it having duplicated the Su(var)3-9 gene at 88D10-E1 (Locke et al. 1988; Tschiersch et al. 1994). These three loci, named E(var)3-4, E(var)3-5, and E(var)3-6, have been mapped to lie within 1.5 c M of the Su(var)3-9 / eIF2ylocus (Dorn et al. 1993c). The mutant alleles that will be discussed below, E(var)3-409, E(var)3-501, and E(var)3-601 were isolated in a P element mutagenesis screen for dominant E(var) mutations (Dorn et al. 1993c). Complementation tests were used to determine 122 that although these mutations are lethal as homozygotes they are viable (though sterile) in heterozygous combinations (data not shown). There are three possible molecular explanations for these mutations. They may simply be duplications that include the 88D-E region as was found with the Dp(3;3)E8 mutation. This is consistent with their enhancement of PEV phenotype and their map position. However, the specific interactions described below argue against this explanation. Alternatively, all could be alleles of true E(var) loci located close to the Su(var)3-9 / eIF2y\ocus. A third possibility was that one or more of the E(var) mutations were lesions at the Su(var)3-9/eIF2y locus itself. The characterization of these E(var) mutations was carried out because if the third prospect was true, one or more of these mutations could represent a rare and valuable allele in just the eIF2 y gene. If the two genes have opposite effects on PEV there may be an overlap in their functions and perhaps their regulation (as proposed in the second hypothesis). Conversely, the E(var) phenotype may be the result of a new class of alleles that affect the genes differently than the previously described alleles. This would have important ramifications for the first hypothesis, that compound loci are relatively more vulnerable to mutations. Single mutations that affect both genes but not in the manner that the P25, P17, and 322 alleles do would support this hypothesis. 123 6.1 Genetic interactions with certain Su(var)3-9/eIF2y alleles Most flies which are wm4; Su(var)/E(var) have an intermediate level of pigmentation and no other phenotypes. For example, wm4; P25/Dp(3;3)E8 flies have 20-50% white* expression and are viable and fertile. However, when P25 was in combination with any of these three E(var) mutations, the heterozygous flies were physically defective and often sterile. Furthermore, the wm4; P25/E(var) flies had white eyes indicating that wm4: was completely enhanced. These phenotypes were only observed with these three E(var) loci; no interactions were found between P25 and 82 other E(var) mutations (data not shown) obtained from a variety of sources (personal unpublished results; Jim Whalen, unpublished results; Dorn et al. 1993c). Table 16 shows the results of genetic crosses between the three E(var) mutations and several alleles of the Su(var)3-9/eIF2 y locus. Flies that were genotypically 322/E(var) were reduced in number (relative to their siblings) indicating that these flies were semilethal. A l l the other genotypes were recovered at numbers indicating that they were as viable if not more viable than their siblings were. The 322/E(var) flies had several physical defects which included small dark bodies, malformed wings and tergites, and very short and often absent bristles. The P25/E(var) and to a lesser extent the P17/E(var) flies had milder visible phenotypes which included very small bristles and some flies with disrupted tergites. These flies also showed a homeotic transformation phenotype that is discussed later in this section. 124 Table 16: Genetic interactions with the E(var) mutants males females E(var)3-409 E(var)3-501 E(var)3-6^ Oregon-R 78/40/83 48/59/79 22/17/30 P25 18V20V67 (1.13) 24V32V93 (1.20) 17V20*/104 (0.71) P25rv2 65/65/110 52/50/69 55/74/105 P17 21V21V85 (0.99) TIT 128 (1.00) 6*116V20 (2.20) 311 32/38/51 28/19/41 23/22/39 330 55/46/77 28/48/63 73/56/125 322 1**/8*V55 (0.33) 1**/1*V41 (0.10) 2**/7**/51 (0.35) 336 54/67/82 38/54/71 74/82/121 Su(var)20501 14/16/60 16/18/96 29/22/83 Genotypes of parents: E(var)3-409/TM3,Sb ,E(var)3-501/TM3,Sb , E(var)3-601/TM3,Sb ,P25/TM3,Sb ,P25rv2/P25rv2 ,P17/TM3,Sb 311/311, 330/330,322/TM2,Ubx, 336/336, Su(var)205^/CyO. Offspring shown as: notable males/notable females/other genotypes. The phenotype of the offspring was normal except where indicated as: * very small bristles, homeotic transformations of ab 5 in males, some flies had malformed tergites ** body small and dark, extremely short bristles, homeotic transformations of the fifth abdomenal tergite in males, wings and tergites malformed, most abdominal bristles missing. Relative viability of notable offspring from certain crosses shown in brackets. 125 The bristle defects noted above were more severe than the dominant Minute bristle defects previously ascribed to the 322, P25, and PI 7 alleles (Chapters 4 and 5). While the E(var) mutations have normal bristles when heterozygous, the E(var) / E(var) survivors did have short bristles (data not shown). The short thin bristle phenotype shared by these different genotypes is consistent with the E(var) mutations disrupting the eIF2 y gene. None of the other alleles tested showed a substantial decrease in viability or had any visible or fertility phenotypes in combination with these three E(var) mutations. A mutation in a different Su(var) gene, Su(var)20501 was also tested but Su(var)20501 /+; E(var)/+ flies were phenotypically normal as well. Thus it appears that this interaction is specific to just those alleles of the Su(var)3-9/eIF2y locus which disrupt both genes: P25 (P element insert/5'UTR), P17 (P element insert/first intron), and 322 (8.3 kb insert). A n unusual phenotype shared by these three E(var) mutations is a mild homeotic transformation of the fifth abdominal segment to an anterior segment identity (Dorn et al. 1993a). This is readily observable in males because normally their fifth tergite (dorsal cuticle plate) is entirely black, while their fourth tergite is mostly tan and only black along its posterior edge. A partially transformed fifth abdominal tergite has tan patches. This phenotype is characteristic of mutations in the trithorax-group genes (trx-G). The trx-G proteins are positive regulators of the homeotic genes that control body segment identity (reviewed by Kennison 1995; Paro and Harte 1996). This particular phenotype is ascribed to a reduction 126 of the homeotic gene product Abdominal-B in the fifth abdominal segment (discussed in Breen 1999). Mutations that both enhance PEV and cause these homeotic transformations have been described before. The genes E(var)3-93D (Dorn et al. 1993a) and Trithorax-like (Farkas et al. 1994) both encode chromatin proteins which have been implicated in the creation of transcriptionally active chromatin. Mutations in either gene decrease the activity of the Abdominal-B+ gene in all flies and the variegating white* gene in wm4 flies causing the homeotic transformation and enhancement of PEV phenotypes, respectively. There is a trx-G gene in the vicinity of Su(z;ar)3-9/eiF2ylocus named moira (located at 3-58.1, 89B2-3). However, none of the three E(var) mutations is an allele of moira as E(var)/moira flies were viable and fertile when complementation tests were performed. The 322, P25 and P17 alleles enhanced the homeotic transformation phenotype of the E(var) mutations in the flies shown in Table 16. The expressivity was quantified according to Dorn et al. (1993a) by classifying males on a scale from zero (no patches of transformed tergite) to six (completely transformed tergite). The E(var)/+ males listed in Table 16 showed a low average level of transformation of between 0.00 and 0.43, depending on the genetic background. However, E(var)/P25 males averaged 3.00 or more while E(var)/P17 males were 1.00 or less. 127 6.2 Recombination mapping the E(var) mutations It is possible that one or more of the E(var) mutations is a disruption of just the eZF2y gene. This would explain many of the genetic interactions described above. It would also imply that mutations specific to e7F2yhave the opposite effect on PEV (enhancement) compared to those mutations that affect only the Su(var)3-9 gene such as the 311, 318, and 330 alleles. This would not be completely unexpected as other conditions which delay development, such as cold temperatures, enhance P E V (Michailidis et al. 1988). In this situation, mutations in the e/F2ygene, which slow protein biosynthesis, would also delay development with the consequence of enhancement of PEV. To test whether the E(var) mutations were novel alleles of the Sw(uar)3-9/eIF27locus, they were recombination mapped relative to two closely flanking genetic markers. For each E(var), the enhancer phenotype was mapped relative to recessive alleles of crossveinless-c (at 3-54.1) and stubbloid (3-58.2). As a control, the Su(var) phenotype of the 336 allele (hobo element insert/first intron) was also mapped. The map positions determined were: 336 3-56.7 N (total E(var)3-409 3-55.8 N=1868 E(var)3-501 3-56.8 N=1230 E(var)3-601 3-56.6 N=752 offspring scored)=752 These data show that each E(var) mutation maps near the Su(var)3-9 gene at 3-56.7. If any of the E(var) mutations had been mapped further away from 128 Su(var)3-9 it could have been discounted as being an allele of this gene. Because these calculated map positions are similar it is not possible to distinguish whether each E(var) mutation is closely linked to the Su(var)3-9 gene or is actually an allele of the Su(var)3-9 / eIF2ylocus. To further examine the possibility of allelism, the E(var) mutants were also mapped relative to the 336 allele itself. If an E(var) mutation was not in the Su(var)3-9/eIF2y\ocus it might be possible to detect recombinant events occurring between the two loci in heterozygous females: E( var) + . + 336 The genetic crosses were, by necessity, rather complex. Simply put, for each E(var), females were generated with the genotype shown above except that the 336 chromosome also carried the cv-c and sbd markers. Thus the females were: + E(var) + + . cv-c + 336 sbd Recombinant male offspring were isolated which were either cv-c sbd+ or cv-c* sbd. These offspring must have been the product of a crossover in the vicinity of the Su(var)3-9 gene. Each of these flies was individually progeny tested to determine if the crossover had occurred between the E(var) and the 336 mutations. These flies would be either E(var) 336 or E(var)+ 336+. For E(var)3-409 five recombinant male offspring were tested, for E(var)3-501 nineteen were tested, and for E(var)3-601 there were thirteen tested. However, all of these offspring were found to be the result of a crossover that had occurred distal or proximal to the location of the E(var) and the 336 mutation but never between 129 them. This negative result does not allow any of the three E(var) mutations from being discounted as an allele of the Su(var)3-9 / eIF2y\ocus. 6.3 Southern blot analysis of the E(var) mutations These E(var) mutations were originally generated in a P element mutagenesis screen and may represent insertions of P elements or deletions (Dorn et al. 1993c). Therefore, if they do alter the Sw(i?ar)3-9/eIF2y locus they may cause detectable restriction fragment length changes. To address this, D N A was isolated from each of the E(var) strains (outcrossed first with Oregon-R), including a second allele of E(var)3-5, E(var)3-502. The D N A was digested with BamHI, EcoRI, or PuuII. Figure 12 shows the resulting Southern blot probed with the A+B probe. As all strains, mutant and wild type, had the same restriction fragments detected with this probe there were no detectable alteration of the 5' half of the locus. The only differences were in the E(var)501 / + lanes which, as seen in the ethidium bromide stained gel (not shown), had run slowly in all digests and were under-loaded in the PvuU lane. Figure 13 shows the same Southern blot hybridized with probe E , which is specific for the 3' end of the locus. Again, all the E(var) mutations had the same pattern of bands as the Oregon-R wild type strain. This combination of probes and restriction enzymes (see Figure 1) showed that there were no detectable changes in the Su(var)3-9 / eIF2 y locus from 3.3 kb upstream to 2.0 kb downstream of the transcribed region in any of these E(var) mutations. 130 Figure 12: Southern blot of the E(var) mutations using the A+B probe D N A was isolated from wild type (Oregon-R) and mutant flies (from an outcross of E(var)/TM3,Sb males ® Oregon-R females) and digested with BamHl, EcoRI, or Pvull. The membrane was hybridized with the E probe (3' end of the locus, see Figure 1), stripped, and hybridized with the A+B probe (5' end of the locus). The ethidium bromide stained gel (not shown) showed the E(var)3-501 /+ D N A was retarded in all digests and was also under-loaded in the Pvull lane. Band sizes are indicated in kilobases. The 3.9 kb BamHl restriction fragment from the 3' end of the locus is faintly visible. 131 B a m H l EcoRI Pvull + + + + + + + + + + + + + O N 1—I C O o I C N r H o o to to r o CO C O C O S-H S H ra cs > > H TO > S H C S > w w w w + + O N r H O O S H C S > C N rH o o HH LTj L O \ 0 I I I I C O r o CO C O > S H > S H C O > W gq w pq O N rH C N r—i o o o o T$< L O L O to I I I I c o C O CO C O S H V H S H C S C O C S > > > SH C S > + W W W W probe A+B 4.2 4.0 2.5 • 132 Figure 13: Southern blot of the E(var) mutations using the E probe The Southern blot shown is the same membrane as described in Figure 12. Band sizes are indicated in kilobases. 133 BamHI EcoRI Pvull + + + + + + + + + + + + probe E ON r H CN r H 2 2 2i 2; in i n vo I I I I co co co co ^ > H 'ZT * H I (8 (5 nj + w w w w + OA r H o o I L O CM r H I I C O CO C O C O J H > > 03 > > H > + w pq pq w + ON r H C N r H o o o o N H in in i i i i C O C O CO C O U l-t u <Tj <Tj (Tj > > > I H > + w w w w 134 6.4 Can the P{hs-eIF2y.H} transgene rescue the E(var) mutations? If the E(var) phenotypes were due to a reduction of the eIF2 y transcript during development they may be rescuable by the P{hs-eIF2y.H} transgene. The ability of the P{hs-eIF2y.H}T5 transgene to rescue the phenotypes of the P25/E(var) genotypes was tested with the three E(var) mutations. Typical were the results seen with E(var)3-501. Under the 2 5 / 3 7 ° C temperature regimen, there were more T5/+; P25/E(var)3-501 than +/+; P25/E(var)3-501 offspring observed (39 and 15 flies, respectively). The T5/+ flies appeared larger and had more offspring when fertility tested. Furthermore, they had a lesser degree of the abdominal five homeotic transformations in males (2.25 versus 4.40). However, this phenotypic rescue was due at least partially to the already established ability of the P{hs-eIF2y.H} transgene to rescue P25 (Chapter 5). A better test to see whether the eIF2y transgene could rescue the E(var) mutations was conducted in the crosses shown in Table 17. There were more T5/+;E(var)/E(var) than + / +; E(var) / E(var) offspring recovered in all the crosses combined (97 and 70, respectively). The T5/+; E(var)/'E(var) males also had a lower expressivity of the abdominal five homeotic transformation in three of the four crosses. However, neither of these trends was convincing evidence that the E(var) mutations had been rescued by the eIF2y transgene. 135 Table 17: Rescue of the E(var) mutations using P{hs-eIF2y.Hj (A) Genetic crosses performed w T5 E(var) •* w . + E(var) — / — / o 09 — / — / ~ Y + TM6,Tb w + TM3,Sb (B) Offspring obtained parents3: offspringb: male female T5 E(var) + E(var) other + ' E(var) + ' E(var) E(var)3-409 E(var)3-4^9 0 / 0 0 / 0 37 E(var)3-501 20 (ab 2.75)c / 30 13 (ab 4.10) / 7 195 E(var)3-601 5 (ab 3.75) / 6 3 (ab 5.00) / 4 42 E(var)3-501 E(var)3-4^ 7 (ab 2.33)/14 12 (ab 3.78) / 15 128 E(var)3-501 0 / 0 0 / 0 17 E(var)3-601 10 (ab 4.40)/5 8 (ab 4.20)/8 54 a = Al l crosses were done under the 2 5 / 3 7 ° C conditions, b = The offspring are shown as males / females. c = "ab" is the average abdominal five homeotic transformation in males. 136 6.5 Northern blot analysis of the E(var) mutations The E(var) mutations may be lesions in the Su(var)3-9/eIF2ylocus, duplications of the locus, or mutations in nearby genes. These possibilities predict a decrease, an increase, or no change, respectively, in the abundance of the Su(var)3-9 and eJF2ymRNA. A Northern blot was used to determine whether any of the E(var) mutations altered the steady state amount of either transcript (Figure 14). m R N A was isolated from young adult E(var)/+ females (obtained from an outcross with Oregon-R). The Northern blot was hybridized three times with the probes E (eIF2y), C (Su(var)3-9), and RP49 (RP49 loading control). Because the eIF2/transcript was detected first, its bands are prominent in the film shown after hybridization with the C probe. The Su(var)3-9 transcript is indicated with an arrowhead over the eIF2/band. A visual inspection of these and other films showed no noticeable difference in the amount of elF2y transcript (relative to RP49 ) between the mutants and Oregon-R. However, there does appear to be an increase in the relative abundance of the Su(var)3-9 transcript in each of the mutants. The transcripts in the E(var)3-501 /+ lane in one of the Northern blots shown in Figure 9 were quantitated (Table 10). As described in Chapter 4, the optical density of each band was divided by the optical density of the RP49 band to control for unequal sample loading in each lane. While the adjusted amount of the eIF2y transcript in the E(var)3-501/+ lane (0.638) was comparable to that of wild type 137 Figure 14: Northern blot of the Su(var)3-9 and eIF2y transcripts in the E(var) mutant flies The figure shows a Northern blot prepared with m R N A isolated from wild type (Oregon-R) and mutant (the offspring of an outcross between E(var)/TM3,Sb males ® Oregon-R females) adult females. The membrane was hybridized with probe E (eIF2y), then probe C (Su(var)3-9), and finally probe RP49 (RP49). After each hybridization the film was exposed for an optimal length of time; probe E (8 hours), probe C (2 days), and probe RP49 (2 hours). Because the more abundant eIF2 y transcript was detected first, its bands are also visible on the film exposed after hybridization with the C probe. The location of the Su(var)3-9 transcript is indicated with the arrowhead. 138 + + + O N O i CO u C3 w + LO I CO J-l >^ W + CN O LO I CO ca >^ W + I CO probe C (Su(var)3-9) i|2 probe E (eIF2y) probe RP49 139 flies (0.719) there did appear to be an increase in the adjusted amount of the Su(var)3-9 transcript in the E(var)3-501 lane (1.109 versus 0.598). As with all Northern blot quantitative analysis, these results are preliminary at best. If true, they suggest that the E(var)3-501 mutation (Figures 9 and 14) as well as the other three E(var) mutations (Figure 14), have an unanticipated affect on the Su(var)3-9/eIF2y\ocus. Each E(var) mutation directly or indirectly causes an increase in the amount of just the Su(var)3-9 transcript. This contrasts with the results obtained with the Dp(3;3)E8/ + strain which, as also seen in Table 10, appeared to cause just a small increase in both transcripts. 6.6 Summary These E(var) mutations were investigated because they may have represented alleles of the Sw(i>ar)3-9/eIF2y locus that either compromised just the eIF2y gene or that affected both genes in a manner different than the P25, PI7, and 322 alleles. If any of the E(var) mutations had been identified as an allele specific to the eIF2ygene it would have meant that the two genes had opposite effects on P E V and may therefore have an overlap in their function (pertinent to the second hypothesis). Likewise, any E(var) mutant that directly affected both genes would represent a new way in which mutations can compromise compound loci (pertinent to the first hypothesis). There were two alternative explanations for these E(var) mutations that had to be ruled out, namely that they were duplications of the locus or were mutations at another locus. 140 None of these E(var) mutations could be identified as a lesion in the eIF2ygene. The only evidence that supported this proposition was the phenotypic interactions seen between each E(var) mutation and those alleles of the Su(var)3-9/eIF2y\ocus that affected the eJF2ygene. These alleles, P25, P17, and 322, were semilethal or had visible phenotypes in combination with the E(var) mutations, while the other alleles, 311,318,330 (all missense), and 336 (hobo element insert) were viable and fertile with the E(var) mutations. Yet if the lack of the dF2ygene was all that was wrong with the E(var) mutant flies, there should have been more rescue of the phenotypes of the E(var) /'E(var) flies by the P{hs.eIF2y.H}T5 transgene. The recombination mapping confirmed the close proximity of each of the E(var) mutations to the Su(var)3-9 / eIF2ylocus when each was mapped relative to two flanking markers. Mapping the E(var) mutations relative to the 336 allele was ambiguous. The absence of detectable recombination between each E(var) and the 336 allele may be indicative of allelism, the very close linkage of distinct loci, or a problem with these mapping crosses. The Southern and Northern blot analysis implied that there were no alterations specific to the e/F2ygene in any of the E(var) mutations. The Southern blot did not reveal any detectable restriction fragment length abnormalities in any of the mutations. Surprisingly, the Northern blot analysis suggested that while the eIF2_ m R N A was present at near normal levels, the amount of the Su(var)3-9 transcript may have been elevated. This change in transcript abundance in the E(var)3-501/+ flies did not resemble that of a duplication of the 141 region, Dp(3;3)E8, w h i c h appeared to cause just a slight increase i n b o t h transcripts. If E(var)3-501 is i n fact a dupl ica t ion that includes the Su(var)3-9 gene, this discrepancy may be due to it being m u c h smaller than the large dupl ica ted reg ion o n the Dp(3;3)E8 chromosome. These results do not a l l o w the E(var) mutat ions to be easi ly categor ized as alleles of the eIF2ygene, duplicat ions of the locus, or mutat ions i n unrela ted genes. W h a t fol lows is a possible, though theoretical, explanat ion u s i n g the E(var)3-501 muta t ion as an example. If the N o r t h e r n results accurately represent the actual consequences of this muta t ion , there is a large increase i n the amount of the Su(var)3-9 transcript and a sl ight decrease i n the eIF2y t ranscript i n E(var)3-501 / + flies. A n increase i n the abundance of the Su(var)3-9 pro te in caused the E(var) phenotype just as dupl ica t ions of the Su(var)3-9 gene are k n o w n to do. A decrease i n the eIF2y pro te in caused the reduced v i ab i l i t y of the E(var)3-5011322 flies. The reduct ion i n the eIF2y pro te in also caused the extremely short bristle phenotype found i n several genotypes, for example the E(var)3-501/P25 flies. The eIF2 y transgene w as able to par t ia l ly rescue the E(var)3-501 flies but , as there is another essential gene muta ted o n the E(var)3-501 chromosome, c o u l d not a l l o w recovery of E(var)3-501 / E(var)3-501 homozygotes . C o n t i n u i n g this explanation, the E(var)3-501 mu ta t i on m a y i n fact be a n allele of the Su(var)3-9/ eJF2ylocus that alters the n o r m a l transcript processing. This w o u l d cause the transcript imbalance a n d the resul t ing phenotypes as hypothes ized above. A muta t ion i n a n i n t r o n spl ice site, for example , w o u l d be 142 consistent w i t h bo th the recombinat ion mapp ing data and the lack of altered restriction fragments i n the Southern blot. The only diff iculty w i t h this m o d e l is exp la in ing the homeotic transformation phenotype. It m a y be due to the overabundance of the Su(var)3-9 pro te in or perhaps to this overabundance i n combina t ion w i t h a decrease i n the eIF2y protein. If further results c o u l d conf i rm this m o d e l it w o u l d be an interesting example whereby a single muta t ion has compromised two genes. This f i nd ing w o u l d suppor t the first hypothesis as since this situation is unique to c o m p o u n d loci it illustrates their vu lnerab i l i ty to mutat ions. 143 C H A P T E R 7: T H E F U N C T I O N O F T H E S U ( V A R ) 3 - 9 P R O T E I N The previous chapters have not supported the second hypothesis. There does not appear to be any connection between the Su(var)3-9 and the eIF2y proteins other than their genes are part of a common locus. If the function of the Su(var)3-9 protein could be determined it may allow the relationship between its function and the function of the eIF2y protein to be resolved. Recently our lab made an exciting discovery (Sarb Ner, personal communication). Antibodies to the Su(var)3-9 protein showed that it was, as anticipated, a chromatin associated protein. What was unusual was that, unlike the polytene chromocentre associated HP1 (the product of the Su(var)205 gene; James et al. 1989) and Su(var)3-7 proteins (Cleard et al. 1997), Su(var)3-9 had a more specialized localization. It was found primarily at a single euchromatic band at the base of the 2L chromosome (S. Ner, unpublished data). This site, 39D3 to El-2, is the location of the Histone gene cluster (HIS-C) (Pardue et al. 1977). To confirm this localization, unsquashed salivary gland nuclei were simultaneously probed (by S. Ner) with the Su(var)3-9 antibodies and labeled Histone D N A . The resulting confocal microscopy images showed the two signals co-localized at a single site on the polytene chromosomes. The HIS-C locus contains 100 or more copies of a 5 (or 4.8) kb repeat that encodes the five major Histone genes (Lifton et al. 1978). The HIS-C locus is also an example of intercalary heterochromatin in Drosophila. Intercalary heterochromatic regions are underreplicated in polytene tissues and are thus 144 constricted, poorly banded, and subject to chromosomal breaks and ectopic pairing (Lamb and Laird 1987; Belyaeva et al. 1998). Thus the Drosophila HIS-C differs from "normal euchromatin" in two respects, it is a tandem array rather than single copy D N A and it is cytologically distinctive. The presence of the Su(var)3-9 protein at the HIS-C raises two important questions: is this where the Su(var)3-9 protein fulfills its primary (or only) role and what exactly does it do at this locus? If the Su(var)3-9 protein operates at the HIS-C locus rather than at, for example, the nucleolus, it would be unlikely that it shares any function in common with the cytoplasmic eIF2y protein, thus disproving the second hypothesis. The interaction between the Su(var)3-9 gene and the HIS-C locus was investigated using two methods: genetic interactions and transcript abundance analysis 7.1 Genetic interactions between Su(var) mutations and the Histone gene cluster The function fulfilled by the Su(var)3-9 protein at the HIS-C is not necessary for viability and fertility of flies. Lethal alleles of the Su(var)3-9/eIF2y locus were demonstrated to be due to the reduction in eIF2y transcription exclusively. That said, there might be genetic interactions between Su(var)3-9 alleles and deficiencies of the entire HIS-C locus. Three of these deficiencies were tested, Df(2L)DS5 (38C7-10 to 39D3-E1), Df(2L)DS6 (38F5 to 39E7-F1), and Df(2L)TW65 (37F5-A1 to 39E2-F1) (Moore et al 1983). If the Su(var)3-9 protein is partially responsible for the normal functioning of the Histone genes, flies that 145 are Df(HIS-C) / +; P25/ + and DfiHIS-C) / +; 318 / + may have additive phenotypes. As controls for these crosses, alleles of the Su(var)205 and Su(var)3-7 genes were included as they would not be expected to interact with the HIS-C. The mutation Suppressor of under replication (Su(UR)ES) was also tested. As its name suggests, this mutation increases the copy number and banding pattern of underreplicated regions in the polytene chromosomes including the HIS-C (Belyaeva et al. 1998). Table 18 shows the results of the genetic crosses performed with the mutants listed above. While it was anticipated that the Df(HIS-C) / +; P25/+ and Df (HIS-C)/ + ;318/+ genotypes would have reduced viability this was not found. Instead these genotypes had greater than expected viability (based on the numbers of their sibling flies) and were strikingly healthy in appearance and fertile. Unexpectedly, it was the other mutations tested that interacted with the deficiencies of the HIS-C. Flies which were Df(HIS-C)/Su(var)205, Df(HIS-C)/+ ; Su(var)3-7/+, and Df(HIS-C)/Su(UR)ES were small in size and were sometimes reduced in viability relative to their siblings. To quantitate the sizes of some of these genotypes, different strains were mated with Df(2L)DS5/CyO flies and the offspring were collected and weighed. While Df(2L)DS5/+ females weighed an average of 0.882 mg (N=76), Df(2L)DS5/+; 318/+ females were slightly heavier (1.042 mg, N=48) while Df(2L)DS5/+; Su(var)3-7/+ females were slightly lighter (0.828 mg, N=29). This relationship was also true when the weights of their siblings were taken into account to control for environmental differences between culture 146 Table 18: Genetic interactions involving the HIS-C locus females3 males 3 Su(UR)ES Df(2L)DS5 Df(2L)DS6 Df(2L)TW65 P25 33/30/42b 5/12/49 22/18/73 22/17/81 1.50c 1.04 1.64 1.44 318 Su(var)205 56/65/- 26/24/43 34/44/60 21/45/51 (1.00) 1.16 1.30 1.29 36/29/76 13/24/66 22/29/112 8/18/63 0.86 1.12 0.91 0.83 Su(var)3-7 15/16/33 16/13/110 7/4/64 18/13/122 0.94 0.79 0.52 0.76 Su(UR)ES (1.00) 31/29/85 21/15/57 35/25/64 0.71 0.63 0.94 Df(2L)DS5 36/36/81 0/0/119 0/0/96 0/0/105 0.89 0.00 0.00 0.00 a = Genotypes of parents with second chromosome mutations were: Su(var)20501/CyO, Df(2L)DS5/CyO, Df(2L)DS6/CyO, and Df(2L)TW65/CyO. third chromosome mutations: P25/TM3, 318/318, Su(UR)ES/Su(UR)ES, Df(3R)AceHDl /MKRS (Su(var)3-701). b = Offspring are listed as: notable males/notable females/other genotypes, c = Viability of the notable flies, calculated relative to the number of siblings. 147 vials. The genotypes listed above weighed 83.6%, 90.9%, and 80.3% as much as their CyO siblings, respectively. Thus it seems that the physical state of flies which carried half the normal number of Histone genes may have been improved by mutations in the Su(var)3-9 gene yet worsened by mutations in the other genes tested. 7.2 Histone transcript levels may be altered by certain Su(var)3-9 alleles The binding of the Su(var)3-9 protein to the HIS-C locus may have a direct or indirect influence on the transcription of the Histone genes. This was tested with a series of Northern blots using total R N A (Histone transcripts are not polyadenylated) from many alleles of Su(var)3-9. These blots were hybridized with probes for the Hisl, Hislb, His4, and RP49 transcripts. Preliminary results suggested that there might be an increase in the amounts of these Histone transcripts in the P25, PI7, P17rvl2, and 322 strains (data not shown). To verify whether there is an increase in Histone transcript abundance in certain Su(var)3-9 mutant lines, a Northern blot was prepared using R N A from two wild type strains and the four homozygous viable alleles; 311,318,330 (missense mutations) and 336 (hobo element insert). The use of only homozygous strains eliminates the problem of interpreting results when mutations are in the presence of wild type alleles. As shown in Figure 15, the Northern blot was hybridized with probes which recognize the 312 nucleotide His4 transcript, the 780 nucleotide Hisl transcript, and the 650 nucleotide RP49 transcript. The figure shows that, despite the variability between the lanes, there 148 Figure 15: Northern blot of the Hisl and His4 transcripts in the mutant alleles The Northern blot was prepared using total R N A from the homozygous strains indicated. The blot was hybridized with the RP49 probe, then the H4 probe (His4), and finally the H I probe (Hisl). The films were exposed for 3.25 hours (RP49), 4.5 hours (H4), and 24 hours (HI). The RP49 bands were cropped out of the figures for probes H I and H4. + + + + CO CO 00 r H co co r H co o co co o CO CO vO CO co NO CO CO probe HI (Hisl) probe H4 (His4) probe RP49 (RP49) 150 appears to be an increase in the relative amount of both Histone transcripts in the 336 strain. Table 19 shows the bands quantitated as was done before (Chapter 4, Table 10). The ratios of Hisl to RP49 and His4 to RP49 are shown for each genotype. The values are also shown relative to the wild type averages. The only notable deviation from the adjusted amounts of the Histone transcripts were in the 336/336 lane which showed increased amounts of both the Hisl (1.45) and His4 (1.39) transcripts. The missense mutations did not appear to possess altered amounts of either transcript. 7.3 Summary The results from these two approaches are consistent with a model whereby mutations at the Su(var)3-9 / eIF2ylocus cause an increase in Histone gene expression. In contrast to the other Su(var) mutations tested, alleles of Su(var)3-9 appear to be beneficial to flies carrying deletions of the Histone gene cluster. This was revealed in the relative abundance of these genotypes, and their physical state and relatively large size. The Northern results, if they can be verified, show that at least one mutant allele of the Su(var)3-9 gene may cause an increase in at least two of the Histone transcripts. These results suggest that the product of the Su(var)3-9 gene normally has a negative influence on the Histone genes. The possible function of the Su(var)3-9 protein at the HIS-C is the subject of the next section. 151 Table 19: Quant i f ied amounts of the H i s l and His4 transcripts i n Su(var)3-9/eIF2y mutants G e n o t y p e Hisl His4 +/+ 0.856 (1.05) 0.623 (0.83) +/+ 0.778 (0.95) 0.870 (1.17) 311/311 0.881 (1.08) 0.817 (1.09) 318/318 0.670 (0.82) 0.654 (0.88) 330/330 0.917 (1.12) 0.696 (0.93) 336/336 1.187 (1.45) 1.038(1.39) The data presented is the ratio O D H l s l / O D R P 4 9 a n d O D ^ ^ / O D ^ ^ calculated from each lane us ing N I H Image. The values are also s h o w n relat ive to the average w i l d type ratios of Hisl (0.817) and His4 (0.746). 152 C H A P T E R 8: DISCUSSION This thesis has described the genetic dissection of the Su(var)3-9 / eIF2y compound locus of Drosophila melanogaster. This allowed the testing of two hypotheses; are compound loci such as this more vulnerable to mutations than independent genes and are the two genes in this locus functionally related. The previous chapters have introduced a variety of mutant alleles that affect the Su(var)3-9 and the eJF2y genes to differing degrees. These alleles, together with eIF2y transgenes, were used to address these hypotheses as well as to investigate the role of the Su(var)3-9 and eIF2y proteins in Drosophila. This chapter is divided into five sections dealing with the function of the Su(var)3-9 gene, the first hypothesis and the nature of each mutation, the function of the eIF2y gene, the second hypothesis and the origin of this locus, and finally a summary of the major findings of this thesis. 8.1 The Su(var)3-9 gene of Drosophila 8.1.1 The role of the Su(var)3-9 protein at the HIS-C locus What do these results reveal about a possible function for the Su(var)3-9 protein? Its cytological distribution suggests it has a major function at the HIS-C locus. Although it is associated with two other euchromatic sites in some polytene chromosome preparations, the locations of these have not yet been determined (S. Ner, personal communication). To understand why the Su(var)3-9 protein may bind the HIS-C it is necessary to consider two properties Of the HIS-C locus. As mentioned before, it is underrepHeated in polytene 153 chromosomes and therefore has the cytological properties of intercallary heterochromatin (Lamb and Laird 1987; Belyaeva et al. 1998). A second unusual characteristic of the HIS-C is that, despite being gene rich, it is comprised of tandem repeats. The hundred or more copies of the 5.0 kb or 4.8 kb Histone repeat are in eight to twelve arrays that are separated by intervening sequences (Liftonefa/. 1978). Tandem repeats are often not conducive to transcription. Repeats are either transcriptionally silent, such as those comprising the pericentric heterochromatin (Hennig 1999; Locke et al. 1999; and others), or silenced, as with repeated transgenes (Dobie et al. 1997; Henikoff 1998; Wolffe 1998). For example, Dorer and Henikoff (1994) showed that arrays of three or more white* transgenes were transcriptionally repressed in a manner resembling PEV. Thus, the Drosophila HIS-C locus differs from "normal euchromatin" in two respects; it is a tandem array of genes rather than single copy D N A and it is cytologically distinctive. The Su(var)3-9 protein may be at the HIS-C to regulate the transcription of the Histone genes. In this scenario, Su(var)3-9 mutant strains would have unregulated Histone transcription with the resultant increase in the Hisl, His2b, and His4 transcripts, as was observed. In support of this model there is a detectable increase in the amount of H I and its isoforms in Su(var)3-9 mutant strains as shown in Western blots (S. Ner, personal communication). 154 The unusual nature of the HIS-C may imply that it requires a specialized form of chromatin for normal transcriptional activity. A n alternative role for the Su(var)3-9 protein may be to create this permissive chromatin structure. Without the "intervention" of the Su(var)3-9 protein and others (since Su(var)3-9 is by itself not an essential), the tandem repeats which comprise the HIS-C may target it for silencing and /or heterochromatinization. Another experiment done by S. Ner supports this hypothesis. There was a change in the micrococcal nuclease digestion pattern at the HIS-C locus in Su(var)3-9 mutants (S. Ner, personal communication). In the Su(var)3-9 mutant flies the nucleosome spacing at the HIS-C was more regular. This finding is similar to that of Wallrath and Elgin (1995) who showed that transgenes that had inserted into heterochromatin (and were therefore subject to PEV) had a more ordered nucleosome spacing pattern than those that had inserted into euchromatin. There are different ways in which permissive chromatin at the HIS-C may be achieved. For example, Su(var)3-9 may repel heterochromatin factors and thus insulate the HIS-C from their effects. One such factor may be FLP1; it is a chromocentre-associated protein but it is also associated with tandem repeats of euchromatic transgenes (Fanti et al. 1998). Alternatively, Su(var)3-9 might modify the action of these heterochromatic factors to produce a distinctive chromatin state. There is evidence for interactions between the HP1 and Su(var)3-9 proteins of Drosophila (Mike O'Grady and S. Ner, personal communication), S. potnbe and mammals. As discussed below, in both fission yeast and mammals the HP1 homologue acts as a "typical" component of silent 155 chromatin (yeast) or heterochromatin (mammals) while the Su(var)3-9 homologue has a more specific role in chromatin structure. There is a relationship between the Su(var)3-9 and HP1 homologues in S. pombe. These are the Clr4 protein (Ivanova et al. 1998) and the Swi6 protein (Lorentz et al. 1994), respectively. In wild type fission yeast Swi6p is found at three transcriptionally silent regions, most prominently at the centromere cluster with minor spots at the telomeres and the mating loci (Ekwall et al. 1995). In clr4 mutant cells, Swi6p is dispersed in the nucleus (Ekwall et al. 1996). The Clr4 protein has a different nuclear distribution, it is found in the prominent nucleolus and the mitotic spindle (Sawin and Nurse 1996). (Note that Sawin and Nurse referred to the protein as nuclear marker S26; this was later identified as Clr4p by Ivanova et al. 1998). Swi6p is normally not found in the nucleolus yet it is present in this location in clr4 mutants (Ekwall et al. 1996). The clr4 gene was originally identified by mutations that caused derepression of the mating loci and of reporter genes inserted into silent regions of the S. pombe chromosomes (Ekwall and Ruusala 1994; Thon et al. 1994; Ekwall et al. 1996). However, as noted above, the Clr4 protein has a different nuclear distribution than the Swi6 protein that is located at these sites. This implies that Clr4p is more likely involved in the establishment of silent chromatin rather than a structural component of it. The data presented above show that not only is the Clr4 protein responsible for the proper localization of Swi6p to its proper sites (centromeres, 156 telomeres, and mating loci) it may also be involved in preventing its improper distribution (nucleolus). How it achieves this is not known. Sawin and Nurse (1996) speculated that during mitosis, Clr4p present at the mitotic spindle might organize the silent chromatin at these sites. During interphase it may just be sequestered in the nucleolus. Alternatively, Clr4p may actively exclude Swi6p from the nucleolus. This second scenario may parallel the situation in Drosophila where HP1 is present at silent chromatin (centromeric heterochromatin) while Su(var)3-9 is found elsewhere. The proposal that the Su(var)3-9 protein is present at the HIS-C to prevent binding of HP1 may be analogous to the Clr4p being present in the nucleolus to prevent the localization of Swi6p there. In mammals, there are distinctions between the spatial and temporal distribution of heterochromatic proteins such as M31 (the HP1 homologue) and SUV39H1/Suv39hl (the human and murine homologues of Su(var)3-9, respectively). There is significant but not complete overlap in the cytological distribution of the mouse Suv39hl and M31 proteins at heterochromatic sites (Aagaard et al. 1999). The human SUV39H1 and M31 proteins co-immunoprecipitate and have overlapping sedimentation profiles suggesting that they are potentially components of a common protein complex. A second paper from the Jenuwein lab (Aagaard et al. 2000) described how SUV39H1 was present at centromeres but only during prometaphase and metaphase. The protein is detectable throughout the cell cycle although there are two phosphorylated isoforms present only during mitosis. Both papers concluded that the likely role of the SUV39H1 and Suv39hl proteins is in establishing the appropriate chromatin structure at mammalian centromeres. 1 5 7 In fission yeast and mammals, the Su(var)3-9 homologues are involved in establishing transcriptionally silent, though functional, chromatin at the centromeres and the yeast mating loci. The model stated earlier proposes that the Su(var)3-9 protein of Drosophila is also involved in the establishment of a specific form of chromatin, that being intercallary heterochromatin. Intercalary heterochromatin is a "compromise" state, a transcriptionally competent form of chromatin in a region that would otherwise be silenced by its inherent repeated structure. The Histone genes of invertebrates are tandemly repeated to allow maximal transcript accumulation during S phase of the cell cycle (reviewed by Osley 1991). Though speculative, the amplification of the Histone genes and perhaps other genes as well has necessitated structural changes in their local chromatin to prevent these new repeats from being silenced. The presence of the Su(var)3-9 protein at the HIS-C is in conflict with recent findings of Reuter and colleagues. They have generated transgenic fly lines that express a Su(var)3-9-Green Fluorescent Protein (Su(var)3-9-EGFP) fusion protein under the control of either the native Su(var)3-9 /eIF2 y promoter or a heat shock promoter (Schotta and Reuter 2000). When the fusion protein was expressed from either class of transformed lines it was preferentially associated with the polytene chromocentre in intact salivary gland nuclei. As the paper did not show the distribution of the fusion protein on fixed polytene chromosomes and only showed the presence of the fusion protein in a portion of two intact nuclei it is not known whether the Su(var)3-9-EGFP fusion protein is also found at the HIS-C. It is possible that the antibodies made by S. Ner recognize epitopes that are hidden when Su(var)3-9 is present in the 158 chromocentre. Conversely, the fusion protein may have a different distribution than the endogenous Su(var)3-9 protein. If Su(var)3-9 was present at both the HIS-C and the chromocentre what might that imply about its function? In this scenario, Su(var)3-9 would be, like HP1 and Su(var)3-7, a heterochromatin associated protein. Its presence at the chromocentre would suggest that it plays a role in the establishment and structure of the dense packaging or compartmentalization of heterochromatic D N A . If this is true, its presence at the HIS-C may be to create permissive chromatin as speculated above. Alternatively, Su(var)3-9 may bind the HIS-C due to its repetitive structure and cause a degree of transcriptional repression. This is consistent with the increase in the amounts of Histone transcripts caused by certain Su(var)3-9 mutants. This would also explain the vitality of the Df(HIS-C)/+;P25/+ and Df(HIS-C)/+; 318/+ genotypes. This would imply that the Su(var)3-9 protein might be causing a certain amount of transcriptional silencing at the HIS-C rather than functioning to prevent this silencing. Only by determining the true localization of the endogenous Su(var)3-9 protein can its role at the HIS-C be resolved. 8.1.2 The dominant Su(var) phenotype of the Su(var)3-9 gene There are different mechanisms by which mutations at the Su(var)3-9 gene may cause the dominant suppression of position effect variegation phenotype. As discussed in Chapter 1, the pairing-looping model as proposed by Sabl and Henikoff (1996) is the currently favored model to explain PEV. Mutations in the Su(var) loci that encode the chromocentre associated proteins 159 HP1 and Su(var)3-7 have been hypothesized to interfere with the recruitment of variegating euchromatic genes into a heterochromatic compartment in the nucleus (see, for example, Sass and Henikoff, 1998). Suppression of PEV would therefore be due to susceptible genes being less often confined to these transcriptionally repressive areas. If the Su(var)3-9 protein is also normally present at the chromocentre it may be involved in this pairing-looping process as well. The Su(var)3-9 protein may designate chromatin as being destined for a heterochromatic environment, to propose one possible function. A reduction in the amount of the Su(var)3-9 protein may allow variegating genes to escape this specification at a higher frequency. When PEV is suppressed in only a minority of cells is the variegating gene sequestered into the transcriptionally repressive domains of the nucleus. Because the variegating gene represents only a tiny portion of the total amount of D N A that has to be labeled as heterochromatin, a 50% reduction of one of the proteins responsible, Su(var)3-9, is enough to allow it to be "overlooked" most of the time. It is also important to consider that as the variegating gene is only adjacent to a true heterochromatic region it is not a preferred target for the action of the Su(var)3-9 protein. These reasons account for the dominance of the Su(var) phenotype within the pairing looping model. If the Su(var)3-9 protein is present at both the HIS-C locus and the chromocentre, or just the HIS-C locus, the way (or ways) in which mutations at the Su(var)3-9 gene suppress PEV is less clear. The reason is that chromosomal deletions that include the HIS-C locus are dominant suppressors of PEV 160 themselves - though not as strong as alleles of Su(var)3-9 and other Su(var) loci (Moore et al. 1983). Since changes in the dose of the Histone genes can alter the extent of P E V it is possible that changes in the dose of a HIS-C associated protein could also affect PEV indirectly. It is odd that alleles of Su(var)3-9 that appear to upregulate the Histone genes have the same phenotype (dominant suppression of PEV) as does hemizygosity for the HIS-C. This could be explained if the dominant Su(var) phenotype associated with mutations in the Su(var)3-9 gene and hemizygosity for the HIS-C was not due to global increases or decreases in Histone abundance. Instead, the phenotype might be caused by imbalances in the amounts of certain Histone proteins and their isoforms. These changes may be caused by alterations in the timing of Histone gene transcription or the relative transcription rates of each Histone gene repeat and each type of Histone gene. In support of this model there is evidence that Su(var)3-9 mutants show an increase in the less abundant H I isoforms (S. Ner, personal communication). One connection between the Histone proteins and heterochromatin was the finding that while most acetylated forms of H4 are underrepresented in the polytene chromocentre, H4 which is acetylated at Lysine 12 is preferentially found there (Turner et al. 1992). It is plausible that the Su(var)3-9 and the Df(HIS-C) mutations could alter the production, deposition, or modification of these different acetylated forms of H4 and thus influence the extent of heterochromatin in the nucleus. Subtle changes in the nucleus could have a large effect at the variegating region for the reasons mentioned above, the small size 161 of the variegating gene compared with the total nuclear complement of heterochromatin and the gene's location adjacent to but not within a normally heterochromatic region. This would account for the dominance of the Su(var)3-9 and the Df(HIS-C) mutations on PEV. There is disagreement in the literature as to whether alleles of Su(var)3-9 are strong or only moderate suppressors of different variegating rearrangements. Locke et al. (1993) found 322 and 336 to be strong suppressors of BarStoneVarie8ator. These results could not be replicated (Section 4.2). Tschiersch et al. (1994) reported that certain Su(var)3-9 alleles were strong suppressors of the variegation of the brown gene in the In(2LR)bwVS2S rearrangement and the scute gene on the In(l)sc8 chromosome. Sass and Henikoff (1998) tested two of these same alleles, Su(var)3-901, which has a frameshift in the openreading frame, and Su(var)3-905, which is yet uncharacterized (G. Reuter, personal communication). They found that neither had a significant effect on Byron (a brown gene variegator) or p{wAR}B133 (a white gene variegator). Based on these and unpublished results Sass and Henikoff (1998) concluded that Su(var)3-901 was only a weak suppressor of PEV. Reuter's lab proposed that despite being a non-essential gene, Su(var)3-9 plays an important role in gene inactivation and heterochromatin formation (Tschiersch et al. 1994). This was based on two properties they ascribed to their Su(var)3-9 alleles, the strong suppressor phenotypes discussed above and also to the strong effect that Su(var)3-9 alleles had in combination with various E(var) 162 mutations. They found that the suppressor phenotype of their Su(var)3-9 alleles dominated the enhancer phenotype of over forty E(var) mutations, though the data was not presented (Tschiersch et al. 1994). In the course of defining the interactions with the specific E(var)3-4, E(var)3-5 and E(var)3-6 mutations, eighty two E(var) mutations from Reuter's lab and our lab were tested for interactions with Su(var)3-9 alleles (Chapter 6). It was found that most E(var) alleles from both labs had a negligible effect on zvm4 PEV. For this reason, most flies that were genotypically zvm4; E(var) ; Su(var)3-9 had a Su(var) phenotype. Any claim that alleles of a Su(var) locus dominate over a number of E(var) mutations must be treated with caution as it may reflect the weakness of the E(var) mutations rather than the strength of the Su(var) alleles. Because of these conflicting results it is unwarranted to conclude that Su(var)3-9 is an important component of heterochromatin based on the existing genetic data. Indeed, Sass and Henikoff (1998) concluded that based on their findings (most of which were not published) the main role of the Su(var)3-9 protein was unlikely to be in heterochromatin. The genetic results are in agreement with a role for the protein at the HIS-C locus, as suggested by the immunolocalization work of S. Ner and the genetic and Northern blot analysis presented in Chapter 7. If true, the weakness of the suppressor phenotype relative to other Su(var) loci would imply that Su(var)3-9 has a different function than heterochromatic proteins. 163 8.2 The first hypothesis and the nature of each mutation This section will interpret the information obtained for each mutant allele. Each has its own unique D N A lesion, change in the abundance of each transcript, and pattern of mutant phenotypes. These results will be discussed in terms of how the Sufaar,)3-9/eff2ylocus functions in wild type and mutant flies. Many of these results support the first hypothesis, which stated that compound loci such as the Su(var)3-9/eIF2 y locus are more vulnerable to mutations. The following section (Section 8.3) will interpret these results in terms of what they reveal about the eIF2/gene specifically. 8.2.1 The 322 allele (8.3 kb insert of unknown D N A / 5 ' end of locus) The 322 allele was generated in a P element screen for dominant Su(var) mutations yet this strain did not contain any P element sequences (Locke et al. 1993). In situ hybridization of polytene chromosomes showed that there was a hobo element in the vicinity of the locus at 88D (Locke et al. 1993). The Southern analysis showed that there is likely an insertion of about 8.3 kb of unknown D N A into the 5' end of the locus. As this insert is not associated with hobo element sequences, the presence of a hobo element nearby may be coincidence. However, as hobo element transposition frequently causes chromosome rearrangements, this hobo element may have been left behind by the events that generated the insert (these types of mutational events reviewed by Lim and Simmons 1994), Alternatively the insert may be of a retrotransposable element. If a retrotransposon had inserted into the Su(var)3-9/eIF2y\ocus just prior to the mutation screen it could have been recovered several times in what would 164 appear to be independent alleles, namely 306,310,322,328, and 340. A retrotransposon would also account for the restriction fragment patterns seen in the Southern blots (Figures 5,6, and 7). 322 is the strongest allele of the Su(var)3-9 / eIF2y\ocus obtained. It was lethal with itself, P25 and P17 (Table 7), semilethal with the E(var) mutations (Table 16), and had the lowest frequency of survivors in combination with 336 (Table 7). 322/+ flies had a decrease in the amount of the Su(var)3-9 and eIF2y transcripts in the Northern blot analysis presented in Figure 9 and Table 10, though this has not been verified. This allele is likely a null mutation as the insertion separates the promoter and exon A from the remainder of the genes. However, if the insert is in the first intron and the exons and promoter remain intact, the insert may be spliced out of a small number of mature transcripts. There were no aberrant polyadenylated transcripts produced in these flies that were detectable with the C probe or the E probe. 8.2.2 The P25 allele (2.8 kb P element insertion/5' UTR) P25 contains a non-autonomous 2.8 kb P element at the beginning of the 5' untranslated region. The P element excised at a high frequency, 20%, during the germline reversion crosses (Table 4). Because it has inserted into a non-coding region, the revertants may retain a small portion of the P element or have a small deletion in the first exon. However, the Southern blot of the P25rv2 and P25rv3 revertants did not contain any aberrant bands (data not shown). The Northern blot analysis suggested that P25/+ flies had decreased levels of both the Su(var)3-9 and eIF2y transcripts when compared with the P25rv2 revertant 165 strain (Figure 9, Table 10). No abnormally large transcripts were detected which might have contained all or a portion of the P element. These transcripts may actually be produced but subsequently destroyed by the nonsense mediated m R N A decay pathway (reviewed by Culbertson 1999). If there are detectable levels of transcripts produced by the P25 mutant locus, the P element must be partially or wholly spliced out. The D N A and R N A analysis suggest that the P25 allele is either a strong hypomorph or a null. The P25 allele had a stronger phenotype in some of the genetic assays than did the 322 allele. For example, the P25/+ females had shorter bristles than did 322/+ females (309 ± 13 pm versus 383 ± pm, Table 11). During the P{hs-eIF2y.H} rescue crosses, 322/P25 flies were recovered at a high frequency (Table 14) yet only a single P25/P25 fly survived past pupation (Table 15). This may have been due to other factors on the P25 chromosome that reduced the viability of the homozygotes. However, as four of five revertants of P25 were homozygous viable and fertile (Table 4) there is unlikely to be a second-site mutation on the P25 chromosome. Conversely, the relative viability of the P25/336 flies was twice that of the 3221336 flies (Table 7) which is consistent with the 322 allele being the stronger of the two. As there was no clear pattern in these results, it is possible that both P25 and 322 are null alleles and any differences were due to random variability in these genetic assays. A different explanation could be true if normal transcripts were being produced from either of the mutant alleles. These transcripts may lessen any mutant phenotypes. The efficiency at which these normal transcripts 166 could be produced may vary between different tissues and developmental stages. Perhaps the differences between the P25 and 322 alleles were due to variability in the limited production of normal transcripts from the mutant alleles. One speculative explanation is that one of the mutant alleles occasionally produced aberrant transcripts that had a detrimental effect, perhaps by encoding a defective protein with antimorphic properties. Characterization of a true null allele such as a deficiency would help to resolve this question. 8.2.3 The P17, P17rvl2, and P17rv9 alleles (1.8,1.3, and 0.4 kb P element insertions, respectively/first intron) The PI 7 allele contains a non-autonomous, 1.8 kb P element that has inserted into the beginning of the first intron. It is a hypomorphic allele because, although it was recessive lethal (Table 2) and lethal with P25 and 322 (Table 7), it was completely viable with 336 (Table 7). P17/336 flies were phenotypically abnormal but not as severely as 322/336 or P25/336 flies. In P17/+ females the Su(var)3-9 and the eIF2/transcripts appeared to be reduced relative to the complete revertant allele, PI7rvl0 (Figure 9, Table 10). Because no larger aberrant transcripts were detected it seems that the P element can be spliced out of the intron it occupies. P17rvl2 and P17rv9 are partial revertants of P17 which retain 1.3 and 0.4 kb of the P element respectively. As all P17 revertants are homozygous lethal due to a second site mutation, the relative viability of these alleles was determined by crossing them to the P25 allele. P17/P25 flies were lethal (Table 2) while P17wl2/P25 and P17rv9/P25 flies were semi-lethal. The relative viability 167 of these genotypes were 0.70 and 0.89, respectively. These partial revertants were also moderate Su(var)'s with intermediate levels of pigmentation in a wmi background. It is notable that these alleles retain phenotypes indicating that both the Su(var)3-9 gene (intermediate Su(var) phenotype) and the eIF2ygene (semi-lethality with the P25 allele) were still partially compromised. This series of alleles, P27, P17rvl2, and P17rv9, have progressively smaller P element inserts in their first intron and correspondingly less severe mutant phenotypes. It is the size of the insert that appears to be the critical factor rather than which portion of the insert remained (for example, cryptic splice sites or the P element's own promoter). These mutations may have their effect on the locus by interfering with transcription or the splicing out of the first intron. This situation is similar to an allele of the white gene, whitettPricot which is the due to a copia element insertion in an intron (Levis et al. 1984). Most transcripts terminate at a polyadenylation site inside the copia element but sometimes there is transcription readthrough followed by the complete splicing out of the intron containing the copia element. The low levels of normal transcripts produce a hypomorphic orange eye phenotype. This illustrates one manner in which the P17, P17rvl2, and P17rv9 alleles could behave as hypomorphic mutations. No matter the way this series of P element inserts compromises the Su(var)3-9/eIF2ylocus, the mutations do not appear to affect one gene more than the other. 168 8.2.4 The 336 allele (hobo element insert/first intron) The 336 allele was generated in the same screen for dominant Su(var) mutations as the 322 allele (Locke et al. 1993). There is a hobo element at the cytological location of the Su(var)3-9 / elF2y\ocus (Locke et al. 1993) that was demonstrated to have inserted into the 3' end of the first intron (Section 3.3.3). Although the 336 allele is maintained as a homozygous stock, the viability of homozygotes was determined to be 87% of expected by Locke et al. (1993). 336 is semi-lethal with P25 and 322 and the survivors were small and had disrupted tergite cuticle and Very short bristles. O n the basis of these results, the 336 allele would be best classified as a hypomorphic allele that is less severe than P17. The complementation analysis revealed an interesting point. The viability of the 336/P25 and 336/322 genotypes were much less when the maternal parent had been 336/336 (Table 7). This difference in viability may be due to a greater reduction of eIF2ytranscripts in the 336/336 versus the P25/+ or 322/+ mothers. However, the preliminary Northern blots did not show any differences in the amount of eIF2 y transcripts between the 336/336, P25/+, and 322/+ females (Table 10). It is possible that 336/336 females deposit less eIF2y m R N A or protein into their oocytes and this caused the decrease in viability. The 336 mutation is an unusual allele in that it did not affect the Su(var)3-9 and eIF2ygenes equivalently. The previous section detailed how the P17 allele and its partial revertants seemed to affect both genes to the same extent. The 336/P25 genotype had a relative viability of 0.82 when the male parent had been 336/336. This was about the same as obtained with the P17rvl2 and P17rv9 169 alleles (0.70 and 0.89, respectively). Yet the 336 allele was a much stronger Su(var) than either of these alleles. This suggests that the 336 allele affects the Su(var)3-9 gene (suppression of wm4:) more than it does the eff2ygene (viability). This model is consistent with the preliminary Northern blot results (Table 10) that showed that the 336/336 females had a greater reduction in the Su(var)3-9 transcript than the eIF2ytranscript (54% and 86% wild type levels, respectively). If 336/336 flies do in fact have a large decrease in the abundance of the Su(var)3-9 protein but only a small decrease in the eIF2y protein it would account for the phenotypes seen. To examine why a mutation in the first intron might affect the production of one transcript more than the other requires an examination of how alternative splicing of the pre-mRNA from this locus normally produces the two transcripts. When the locus is being transcribed a choice is made at the end of exon C , whether to use the polyadenylation sequence at the end of this exon or continue transcription into exon D (for reviews of the regulation of polyadenylation see Keller 1995; Virtanen 1995). If transcription stops at the completion of exon C , then the splice donor site at the end of exon B is used to remove just the second intron to form the Su(var)3-9 m R N A . If transcription continues through exon D then this same splice donor site is used to remove the second and third introns as well as exon C to produce the eIF2y m R N A . To only produce the two functional m R N A molecules, it is essential that the appropriate section of the nascent transcript be removed. One mechanism to achieve this would be to always splice out the second intron but for a consensus 170 splice donor sequence to be present at the new junction between exons B and C . This way the splice donor site at the end of exon B would still be capable of being spliced to the splice acceptor site at the beginning of exon D and thus cause the removal of exon C and the third intron during formation of the eIF2 y mRNA. A stepwise removal of introns such as this is used during the processing of the very long Drosophila Ultrabithorax transcript (Hatton et al. 1998). However, the exon B / C junction ( G G / A A C G C T ) does not match the consensus Drosophila splice donor sequence (ag/GTaagt; Mount 1993) so this mechanism cannot be used during transcript processing at the Su(var)3-9/eIF2ylocus. Another mechanism to remove the appropriate section of the nascent transcript would be if the splice acceptor site at the beginning of exon D were stronger than that at the beginning of exon C. Transcripts that contain exons C and D would be processed using the strong splice acceptor site and exon B would be joined to exon D during the generation of the eJF2ymRNA. Transcripts that end at exon C would use the default splice acceptor site and exons B and C would be joined to form the Su(var)3-9 mRNA. Returning to the 336 allele, for a mutation to decrease the Su(var)3-9 transcript preferentially it must influence the decision to terminate transcription at the end of exon C or not. The transcripts from the 336 allele would be abnormally large until the hobo containing first intron is removed. Somehow this additional sequence would have to interfere with the transcription termination decision occurring nearby. If this model were true, why didn't the PI 7 allele, which also has an insert into this intron, affect this transcription termination 171 choice? It may be due to the smaller size of the P element insert (1.8 kb versus 3.0 kb) or its position closer to the beginning of the first intron. A third possibility is there may be differences in the secondary structure of the nascent transcripts depending on which insert they contain. 8.2.5 The 321,318, and 330 alleles (G521D, S616L, D536N) The 311,318, and 330 mutations are EMS-induced, homozygous viable alleles. None was associated with mutant phenotypes other than suppression of PEV and interactions with the HIS-C. They were viable with the P25 allele and had no dominant visible phenotypes such as shortened bristles (Table 11). A l l were found to have missense mutations in the carboxyl-portion of the Su(var)3-9 open reading frame, in or adjacent to the SET domain. 311 and 330 are within the SET domain while the 318 mutation is in the small carboxyl-tail of the protein, 14 residues beyond the core SET domain (Jenuwein et al. 1998). The SET domain and carboxyl-terminus comprise less than 25% of the amino acid sequence of the Su(var)3-9 protein and yet all three EMS induced Su(var)3-9 alleles were found here. Most of the Su(var)3-9 mutations isolated have been mapped to the SET domain. Reuter's lab has characterized twelve Su(var)3-9 alleles (Tschiersch et al. 1994; G. Reuter, personal communication). Five of these mutations were in the SET domain, five were frameshift mutations that were upstream of it but would have been unable to produce a protein with the SET domain, while only two were independent of the SET domain. Similar results were obtained with the S. pombe homologue, clr4, by Ivanova et al. (1998). A l l four alleles of clr4 have 172 missense mutations in the SET domain. Using site directed mutagenesis they created two mutations in conserved residues of the chromo domain of Clr4p. These mutants showed derepression of an ura4+ reporter gene inserted into the silent mating loci, yet unlike with the other alleles, there was no derepression of the mating loci themselves (Ivanova et al. 1998). The predominance of Su(var)3-9 mutations in the SET domain could be due to a bias favoring either the generation or the recovery of these mutations. Some mutagens, most notable P elements, have preferential targets in genes as they require accessible chromatin to insert. A chemical mutagen such as EMS would not be expected to have a preference for a portion of exon C though. The more likely hypothesis is that the SET domain is more sensitive to changes than other parts of the protein, most notably the chromo domain. Alternatively, mutations in the SET domain might produce a stronger dominant Su(var) phenotype and be preferentially isolated in mutagenesis screens. This question may be resolved when the functions of the chromo and SET domains are determined. There was a difference between the 311,318, and 330 alleles and the 336 allele beyond the latter's subtle effects on the eIF2ygene. Though all are suppressors of PEV, only 336/336 flies were observed to increase the levels of the Hisl and His4 transcripts in a preliminary Northern blot analysis (Figure 15, Table 19). Without knowing the mechanism behind PEV, the complete distribution of the Su(var)3-9 protein, or the function of the Su(var)3-9 protein at the HIS-C it is not possible to resolve this. The mutant Su(var)3-9 proteins that 173 have these amino acid substitutions may be partially functional and able to fulfill their function at the HIS-C but not elsewhere in the nucleus such that only a Su(var) phenotype is generated. However, the 318 mutation was able to increase the size and physical state of flies carrying deficiencies of the HIS-C so there must be some change at the HIS-C in flies that possess this particular mutant Su(var)3-9 protein. The 336/336 flies are predicted to have less Su(var)3-9 protein overall. This reduction compromises the role or roles of the Su(var)3-9 protein which result in both suppression of PEV and alterations at the HIS-C. 8.2.6 The vulnerability of compound genes to mutations The first hypothesis states that compound loci are more vulnerable to mutations than single genes. One of the reasons by which compound loci are non-ideal is that a single mutation can compromise two genes and therefore interfere with two cellular processes. This was especially true of the Su(var)3-9 / eIF2ylocus as both genes have dominant phenotypes. The Su(var)3-9 gene has the Su(var) phenotype and the HIS-C interactions discussed in Section 8.1 while the eIF2ygene has the Minute phenotypes discussed next in Section 8.3. The 322, P25, P17, and to a lesser extent 336 alleles affected both loci and had both categories of mutant phenotypes. Compound loci such as Su(var)3-9 / eIF2y, the cholinergic locus (Kitamoto et al. 1998), and DUb80/IP259 (Mottus et al. 1997) have the overlap in their two genes at the 5' end of the locus. Drosophila transposable elements preferentially insert into the 5' end of genes (Ashburner 1989). Because of this, mutations of compound loci due to transposable element insertions would be more likely to 174 compromise both genes rather than just one. This was supported by these results. The P element inserts P25 and PI 7 compromised both genes as did the 322 allele, which is likely due to a retrotransposon insertion. The 336 allele preferentially affected the Su(var)3-9 gene although there was evidence that the eiF2ygene was also affected. The only other allele of this type is Su(var)3-914, which contains a P element in the middle of exon C (Tschiersch et al. 1994). This allele only affected the Su(var)3-9 gene (data not shown). As four of the five known transposable element induced alleles of this locus inserted into the 5' end and damaged both genes, the first hypothesis is supported. Not only can single mutations affect two genes in compound loci, this occurs more often than not because the shared region of the two genes is the preferential target for transposable elements. 8.3 The eIF2 y gene of Drosophila 8.3.1 The Drosophila eIF2ygene is an essential locus One of the conclusions reached by this thesis is that eIF2 y is an essential gene in Drosophila. Alleles such as 322, P25, and P17 that significantly compromise the Su(var)3-9/elF2ylocus were recessive lethal in all homozygous and trans-heterozygous combinations. This lethality was due exclusively to the reduction of the eIF2ytranscript because induction of the P{hs-eIF2y.H} transgene was sufficient to rescue these genotypes. It should be noted that it is not known whether the decrease in the eIF2ytranscript in the mutant lines is uniform or more pronounced in certain tissues. A brief daily heat shock was sufficient to rescue the recessive lethality phenotype of the P17/P25,322/P25, and, in one 175 case, P25/P25 flies. This suggests that the transcript or the protein or both have a long enough half life to allow the fly to survive between heat shock treatments. In the presence of the T5, T7A, and T18 transgenes there were usually more P25/P17 and P25/322 flies obtained than expected based on the number of siblings. This implies that not only was there complete rescue of the exceptional genotypes (in that they all survived development) but these genotypes were slightly more viable than there TM3,Sb and TM6,Tb siblings. Some of these genetic crosses provide insight into which developmental periods require large amounts of the eIF2y protein. The lethal period of the eIF2y gene includes pupation as seen by the dead P25/336 and 322/336 pupae (Table 7). It also includes earlier stages as no dead pupae were seen of other lethal genotypes such as P25/322. Two T5/ +; P25/P17 male adult survivors were set aside after being grown to adulthood at 2 5 / 3 7 ° C . Despite the cessation of heat shock treatments, they survived over two months at 25°C. This suggests that either adults have no need for additional eIF2 y transcripts or protein or that there is an adequate production of the eIF2 y transcript from the endogenous mutant and transformed genes to sustain the low levels of eIF2y needed. The sensitive periods and lethal phases of the e/F2y alleles could be determined more precisely by repeating the rescue experiments but heat shocking the flies only during certain developmental stages. 8.3.2 Mutations in the Drosophila eIF2yger\e have Minute phenotypes A l l the mutant phenotypes of the Su(var)3-9 / eIF2y\ocns that were not attributed to the Su(var)3-9 gene (suppression of PEV, the effects at the HIS-C) 176 resemble those of the Minute mutations (Sinclair et al. 1984; Lambertsson 1998). The recessive lethality and dominant bristle defects phenotypes were due to the reduction of eIF2y transcripts as proven by the P{hs-eIF2y.H} rescue experiments. This means that mutations in a gene encoding a translation initiation factor have the same dosage sensitive phenotypes as mutations in genes encoding ribosomal proteins. There are only a few other exceptions where non-ribosomal protein encoding genes mutate to give a Minute phenotype. These include the r D N A gene arrays bobbed and mini, a gene involved in polyamine synthesis named SAMDC (all reviewed in Lambertsson 1998), and the nucleolar proteins modulo (Perrin et al. 1998) and minifly (Giordano et al. 1999) Mutant alleles of the eIF2ygene have mild Minute phenotypes. Saeboe-Larssen et al. (1998) classified Minute phenotypes of different alleles of the M(3)95A gene, which encodes the ribosomal protein S3. The phenotypes were moderate, strong, and extreme if they had bristles which were reduced by 20,40, and 60% and had larval development increased by 22,51, and 75 hours, respectively. The length of the larval development of the Su(var)3-9/ eIF2y alleles was not measured due to the lack of a true null allele. The bristle lengths of the 322 and P25 alleles were reduced by 14 and 31%, respectively (Table 11). By the bristle criterion, these alleles have a moderate Minute phenotype. Because this phenotype is dominant the growth of bristles during pupation must be particularly sensitive to a decrease in the efficiency of protein biosynthesis. Though P25/+ flies had bristles as short as those of a typical Minute mutation, Df(2L)M24F-B (Table 11), the P25 allele did not enhance the mutant phenotypes of unrelated mutations as much as Minute mutations did (Section 4.2). 177 Assuming that the P25 allele is a strong hypomorph if not a null allele, it seems that eIF2y activity is not as sensitive to its gene's dose as ribosomal proteins are. This could be due to a natural excess of eIF2y in wild type cells, so that heterozygous fly strains that under produce eIF2y still produce amounts above a critical threshold. Alternatively, the amount of eIF2y could be regulated in an unknown manner so that even mutant lines possess near wild type amounts, provided they have one wild type allele. Controlling the amount of eIF2yby regulating its protein turnover is one possibility. During the P{hs-eIF2y.H} rescue experiments the P25/P17 and P25/322 survivors had an overall healthy appearance (Section 5.4). However, these survivors took noticeably longer to develop, had very short bristles, and were often sterile despite continued heat shock treatments. The amount of eIF2y produced in these flies must have been enough to allow them to develop normally, albeit at a slower rate. There was insufficient eIF2y in the bristle primordia during pupation and the tissues involved in reproduction (germline, reproductive structures, and libido). The processes involved in producing bristles and a functional germline must be most sensitive to the reduced and/or intermittent supply of the eIF2/transcript during these experiments. No previously identified Minute loci are in the vicinity of the Su(var)3-9/ eIF2ylocus at 88D-E (Tschiersch et al. 1994). The closest are M(3)86D proximal to the locus and M(3)95A distal to it (the names denote their cytological position; Lambertsson 1998). The eIF2ygene was never isolated as a Minute locus despite this class of mutation having been first identified in 1919 by Bridges and studied 178 ever since (reviewed by Lambertsson 1998). Possible reasons for this are the relatively moderate Minute phenotypes of alleles of this locus and the lack of chromosome deletions that include this region. Most Minute loci were originally identified by deletions that showed the characteristic haplo-insufficient effects -most noticeably the short bristles. As discussed in Chapter 3 there are no existing deletions that include this locus, though some have been created but subsequently lost (U. Gaul, personal communication). There are no known haplo-insufficient loci with lethal or sterile phenotypes in the area that would have prevented the recovery and maintenance of deletion stocks (Ashburner 1989). However, there are two loci at 88F, Tropomyosin and Actin 88F, that cause a flightless phenotype when either is hemizygous (Ashburner 1989). Any deletions that included 88F would have been less viable because of this. The P element mobilization screen for new lethal alleles of eIF2y (Section 3.4) may have failed to generate deletions for this reason. The Drosophila gene that encodes the eIF4A translation initiation factor, which binds the 5' end of the mRNA, has also been the subject of mutational analysis (Dorn et al. 1993b). Mutant alleles were recessive lethal or semilethal, consistent with a disruption of the only Drosophila gene that encodes this essential protein. Flies with certain combinations of alleles had what was described as a weak Minute effect; presumably meaning that they had small thin bristles. In addition, certain alleles (they were not identified) had the dominant enhancement of PEV and homeotic transformation phenotypes (the abdominal 179 five to four transformation that was discussed in Section 6.1). The authors argued that the weak Minute and homeotic phenotypes were the result of a general decrease in protein biosynthesis (Dorn et al. 1993b). They maintained that the E(var) phenotype was due to an unidentified gene that was also compromised by their mutations. It should be noted that at the time, E(var) loci were optimistically thought to encode structural components of euchromatin (Dorn et al. 1993b). Today they are recognized to be a diverse collection of genes that often have only indirect effects on chromatin structure (personal observation; Sass and Henikoff 1998). Mutations in the vicinity of the eIF4A locus have four phenotypes: recessive lethality, a weak Minute bristle phenotype, the E(var) phenotype, and homeotic transformations (Dorn et al. 1993b). Mutations in eIF2/have two of these phenotypes: recessive lethality and a moderate Minute bristle phenotype. The E(var) mutations described in Chapter 6 have three of these phenotypes - all but the Minute phenotype. A l l four phenotypes may be attributed to a general reduction in protein production during development. A decrease in translation may have several detrimental consequences; for example a prolonged cell cycle, decreases in the concentration of important proteins, imbalances in regulatory proteins, and so on. Because it was not possible to determine whether the E(var) mutations directly or indirectly affected the Su(z;arj3-9/dF2ylocus it is hard to resolve these observations. A close examination of the chromosomes in the E(var) strains may reveal if they have a disruption that includes the Su(var)3-9/ eIF2ylocus. 180 8.4 The second hypothesis and the origin of the Su(var)3-9leIF2y\oc\is As discussed in Chapter 1, the Su(uar)3-9/eIF27locus has a very unusual gene organization whereby two genes share a common promoter and two exons. The second hypothesis states that this gene arrangement involves genes that have overlapping functions. The genetic analysis discussed in the preceding three sections did not support this proposition. Each gene possessed a distinctive set of phenotypes. There was no evidence that the proteins were involved in the same or similar processes. This would have allowed the genes at the Su(var)3-9/ eIF2 7 locus to possibly benefit from coordinated regulation. The Su(var)3-9 protein performs a function at the HIS-C that is not obviously related to protein biosynthesis. There is no evidence for the presence of the Su(var)3-9 protein in the nucleolus (S. Ner, personal communication). Mutations affecting only the Su(var)3-9 gene did not possess mutant phenotypes that have been assigned to the eIF2ygene. Because of these points it is unlikely that the Su(var)3-9 protein is involved in the same process, protein translation, as eIF2y. While none of the data presented in this thesis supports an overlap in the function or functions performed by these proteins this hypothesis cannot be completely dismissed. Although the eIF2 protein complex has a well defined role in the cytoplasm there is recent evidence that it has another function in the nucleus (Kimball 1999). While the mammalian eIF2 is mostly found in cytoplasmic areas rich in ribosomes it is also present in the nucleus in both 181 nucleolar and extranucleolar areas (Lobo et al. 1997). A hint as to what it might do there came from work on the DNA-dependent protein kinase (DNA-PK) and the DNA-binding protein K u which are necessary for D N A repair in humans. Their association with each other and with D N A is stabilized by five polypeptides, three of which were identified as the a, p\ and y subunits of eIF2 (Ting et al. 1998). The role of eIF2 in the nucleus remains a mystery but Ting et al. (1998) did point out that as eIF2p can bind R N A it may be able to bind D N A as well. Because eIF2y may perform a similar function in the nucleus as the Su(var)3-9 protein it is not possible to completely dismiss the second hypothesis, that there is a functional overlap in the genes at the Su(w)3-9/eIF2ylocus. As there is no obvious reason why these two genes benefit from being together it seems that this locus is the product of a joining of two unrelated genes. The remainder of this section will examine how this locus may have been generated and what were the implications of this event on the functioning of the Su(var)3-9 gene. How this compound locus was created can be understood by examining the exons that are shared by the two transcripts. Exons A and B encode an 80 amino acid region that is common to both the eIF2y and Su(var)3-9 conceptual proteins as proposed by Reuter's group (Tschiersch et al. 1994). However, this protein region may not be an essential part of both proteins. To address this requires an examination of the eIF2y protein. eIF2y is the subunit of eIF2 that, by binding either G T P or G D P , determines the active (eIF2-GTP) or inactive (eIF2-GDP) state of the complex during the translation initiation cycle (reviewed by Kimball 1999; Kozak 1999). Most of the amino portion of 182 eIF2y proteins consists of a GTP binding domain. As reviewed by Kjeldgaard et al. (1996), four conserved sequence motifs interact directly with the bound nucleotide. GxxxxGK(S/T), is the "phosphate binding loop" and attaches to the a- and B-phosphates. DxxG is the part of the "effector loop" that binds the y-phosphate if GTP is in the pocket. The conformation of this loop, and other parts of the protein, differ depending on which nucleotide is in the pocket. N K x D and the less conserved (C/S)Axx determine the specificity for guanine by binding to the base itself. Returning to the origin of this locus; do exons A and B contain an essential portion of either the eIF2y or Su(var)3-9 open reading frames? Exon A encodes 23 amino acids. There is little conservation at the amino terminus among the characterized eIF2y proteins, but 10 of these 23 residues are identical with the human eIF2y protein (Erickson et al. 1997). Of the 57 amino acids encoded by exon B, 44 are identical with the S. pombe, S. cerevisiae, and human eIF2y proteins (Erickson et al. 1997). Thus the open reading frame in exon B and perhaps exon A would appear to encode a conserved and therefore integral part of the eIF2y protein. This 80 amino acid region comprises the non-conserved amino terminus of eIF2y and the GxxxxGk(S / T) motif of the GTP binding domain. Indeed, BLAST searches of the GenBank database (Altschul et al. 1997) with this 80 amino acid region of the Drosophila open reading frame recover both eIF2y (eukaryote) and EF-Tu (prokaryotic) homologues. The Su(var)3-9 conceptual protein would therefore possess the non-conserved amino terminus and the GxxxxGK(S/T) motif of the eIF2y protein. It 183 lacks the remaining three motifs of the GTP binding domain. The first 138 residues unique to the Su(var)3-9 protein (up to the chromo domain) do not contain any G T P binding protein motifs, and, save for a potential nuclear targeting signal, no other distinctive motifs (Tschiersch et al. 1994). Although the GxxxxGK(S/T) motif is found in other purine nucleotide binding proteins without the other motifs (Kjeldgaard et al. 1996) it is unlikely that it functions as such in Su(var)3-9. Despite this, the region is referred to as a GTP-binding domain in the literature (Aagaard et al. 1999; Jenuwein et al. 1998). Had a mutation in this region been isolated, the relevance of this region to the Su(var)3-9 protein would have been resolved. Conversely, transgenic flies could be created that carry Su(var)3-9 genes with this region mutated. As exons A and B are an integral part of the eIF2ygene, the Su(var)3-9 gene must be the newcomer. Thus, this compound gene arrangement could have arisen as a simple insertion of a D N A segment containing the Su(var)3-9 gene (that is, exon C) into the second intron of the dF2y gene. A s the locus exists today, there is about 700 bp upstream and 190 bp downstream of exon C , the second and third introns respectively. Of course the original insertion may have been much larger and have contracted with time. A n alternatively spliced form of the endogenous eIF2 y gene would create a transcript containing the Su(var)3-9 open reading frame. This modified Su(var)3-9 transcript would now be under the control of a new, though constitutive, promoter. This arrangement, where a single promoter drives the transcription of two dissimilar transcripts, demonstrates the flexibility of the transcription and splicing machinery. There 184 may be some variability in the relative amounts of each transcript produced at different developmental stages and in different tissues though. There are two ways that an insertion of D N A containing exon C could have come to be in the second intron of the eIF2 y gene. As explained above, there may have been an insertion of a piece of genomic D N A during a chromosome rearrangement. Another method would have involved a retroposition event. Some new genes are created when an R N A molecule that has been reverse transcribed into D N A becomes inserted into the genome (discussed by Brosius 1991). Occasionally, these retrosequences are in a position whereby an existing promoter can be recruited to allow their transcription. This movement of genes to new chromosome sites by an R N A intermediate resembles the spread of retrotransposable elements. It has been proposed that exon shuffling may occur when portions of genes are inappropriately included in retrotransposable elements during their mobilization (Eickbush 1999). One such example is tinejingwei locus of Drosophila teissieri and Drosophila yakuba that was created by an insertion of an Alcohol dehydrogenase (Adh) retrosequence (Long and Langley 1993). This Adh sequence has recruited the promoter and three exons from a gene called yellow-emperor to form the chimerical jingwei gene (Long et al. 1999). A s the Adh sequence originated from a mRNA, it is lacking the introns present in the endogenous Adh gene. The Su(var)3-9/eIF2ylocus may have also been created by a retroposition event. The entire Su(var)3-9 specific region is contained in a single exon, 185 consistent w i t h it be ing der ived from an m R N A transcript. If the Su(var)3-9 gene cou ld be isolated from another species that d iverged w i t h D . melanogaster before the creat ion of the Su(par)3-9/eIF2ycompound locus it may reveal h o w this c o m p o u n d locus was created. For example, i f the ancestral gene contains in t rons it w o u l d suppor t this retroposit ion mechanism. A t h i r d possible o r i g i n for this locus cou ld have occurred i f the ancestral eIF2yand Su(var)3-9 genes were beside each other o n the chromosome. If there was a dup l i ca t ion of the reg ion that i nc luded bo th genes and then loss of parts of each eZF2ygene, the Su(var)3-9 gene (exon C) c o u l d have ended u p i n the m i d d l e of the r ema in ing elFl y gene. The successive events are s h o w n here where A , B , D , E , and F represent the exons of the efF2ygene and C is the s ingle exon of the Su(var)3-9 gene. ancestral arrangement A B D E F C duplicat ion A B D E F C A B D E F C l o s s o f e x o n s A B A B D E F C D E F C loss of exons D E F A B C D E F C current arrangement A B C D E F (C) A n y intermediate stage w o u l d have to retain a fu l l y funct ional eIF2ygene. The Su(var)3-9 gene w o u l d not have been unde r this pressure as there w o u l d a lways be the second u n i n v o l v e d Su(var)3-9 gene that c o u l d con t inue to funct ion (as indicated by the exon C o n the right). A process s imi l a r to this has been 186 proposed in the creation of the Sdic gene of Drosophila (Nurminsky et al. 1998) and the SNURF-SNRPN locus of mice (Gray et al. 1999). Of the three mechanisms by which the Su(var)3-9/eIF2ylocus may have arisen, the retroposition explanation is the most parsimonious. Regardless of which mechanism is correct, there may have once been an endogenous Su(var)3-9 gene elsewhere in the genome. This gene may have degenerated or it may still be functioning, perhaps in a diverged role. If it is still functioning, this may explain why the Su(var)3-9 gene in the Su(var)3-9 /eIF2/locus is not essential. However, if this other Su(var)3-9 gene does exist it was not detectable by low stringency Southern blots (data not shown) nor Western blots (S. Ner, personal communication). A search of the recently compiled Drosophila genome database (Adams et al. 2000) did not reveal any Su(var)3-9 like proteins (those with both a chromo domain and a SET domain) nor any regions of D N A sequence similarity. Clues to the nature of the ancestral Su(var)3-9 gene before this gene rearrangement can be obtained from the mammalian and S. pombe homologous proteins (Aagaard et al. 1999; Ivanova et al. 1998). They have much shorter amino-terminal ends (forty and four residues before their conserved chromo domains, respectively). The Su(var)3-9 gene has 134 residues encoded by exon C before the start of its chromo domain. If the ancestral Su(var)3-9 protein also had a short ammo-terminus, it is possible that its translation start site might have been in what is now the beginning of exon C . There are three inframe A T G codons in exon C upstream of the chromo domain. One of these, Met207, is in a 187 reasonably good fit to the translation start site consensus sequence (Cavener and Cavener 1993). There is some evidence that the actual Su(var)3-9 protein is smaller than its conceptual product (S. Ner, personal communication). This may be achieved by using a downstream translation start codon as explained above or by post-translational removal of the "unneeded" amino terminal region. If either mechanism is actually occurring, it would indicate that the Su(var)3-9 gene has managed to eliminate the "compromising" 80 amino acids of an incorrect protein from the amino terminus of its protein. 8.5 Summary Loci such as Su(var)3-9 / eIF2ythat are comprised of two dissimilar genes are unusual. Because compound loci represent a seemingly non-ideal gene arrangement it was proposed that they are relatively more vulnerable to induced and spontaneous mutations. To test this hypothesis, the locus was genetically dissected using a variety of different mutant lesions. A s shown in Figure 8, these were insertions of P elements (P25, P17, P17rvl2, and P17rv9), a hobo element (336), a retrotransposon or unknown D N A (322), and chemically induced missense mutations (311,318,330). The mutations could be divided into two categories: transposable element insertions into the common 5' region of the locus and missense mutations in or near the SET domain of the Su(var)3-9 gene. Analysis of the former group of 188 mutations provided supported for the first hypothesis in two ways. Four independent alleles (322, P25, P17, and 336) were demonstrated to be single lesions which compromised both genes. These alleles had a strong (322 and P25) or moderate (P17) effect on both genes or affected one of the two preferentially (336). This illustrated how the close proximity of multiple genes in compound loci allows single mutations to damage more than one gene and more than one cellular process. Three of these mutations showed multiple dominant phenotypes due to each gene being affected. Because transposable elements preferentially insert into the 5' end of genes they are more likely to disrupt the shared region of compound loci and compromise both genes. As stated previously, not only can single mutations affect two genes in compound loci, this is more likely to occur because the shared region of the two genes is the preferential target for transposable element insertion. These two points support the first hypothesis, which stated that compound loci are more vulnerable to spontaneous and induced mutations. Because many examples of compound loci involve ancestrally or functionally related genes it was decided to test if this held true with the Su(var)3-9/eIF2ylocus (the second hypothesis). Structurally the genes share a common promoter and two exons that encode the first 80 amino acids of both of their proteins. While these residues constitute an essential part of the erF2y protein they are likely dispensable for the Su(var)3-9 protein, if they are part of the mature protein at all. Functionally the genes are independent. Determining the phenotypes associated with the mutant alleles and a transformed <?IF27gene 189 This was assessed by determining the phenotypes associated with the mutant alleles and a transformed e7F2ygene. Because each phenotype could be assigned to either the Su(var)3-9 or the eIF2y gene there is no evidence for overlap in the functions of these genes. A direct analysis of the function of one of these proteins supported this conclusion as well. It was a possibility that the Su(var)3-9 protein performed a function in the nucleolus and was therefore involved in protein translation as is eIF2ybut this was not the case. The converse situation, whereby the Drosophila eIF2y protein performs a role in the nucleus similar to Su(var)3-9 is an intriguing possibility that remains to be investigated. Because none of the results obtained supported a connection between these two proteins at a regulatory or functional level, the second hypothesis was not supported. Despite many other compound loci being comprised of ancestrally duplicated genes or genes whose proteins were functionally related this is not the case with the Su(var)3-9 / eIF2y locus. The Su(var)3-9/eIF2ylocus arose by an insertion of a D N A fragment containing most of the Su(var)3-9 open reading frame into the second intron of the eIF2y gene. Alternative splicing allows a single promoter to produce two transcripts. The most parsimonious mechanism that would explain this compound gene is a retroposition insertion of a Su(var)3-9 transcript into the elFly gene. Most of the transposable element insert alleles compromised the functioning of both genes proportionally. The 336 hobo insert was the exception in that it appeared to preferentially affect the Su(var)3-9 gene in genetic and preliminary Northern blot assays. This could be due to the inclusion of hobo 190 element sequences in the pre-mRNA influencing the choice to terminate transcription at the end of exon C. The genetic analysis used to address these two hypotheses has also revealed much about the eIF2yand the Su(var)3-9 genes in Drosophila. The eIF2y gene is essential as one phenotype associated with this locus is recessive lethality. Three alleles, 322, P25, and P17, were recessive lethal in all combinations. This recessive lethality was due exclusively to a reduction in the amount of eIF2y activity as induction of the eIF2 ytransgene was sufficient to allow survival of otherwise lethal genotypes. The eIF2y gene possesses the same mutant phenotypes that define Minute loci. The most striking visible defect characteristic to Minute mutations are the smaller and thinner sensory bristles. Each of the recessive lethal alleles, 322, P25, and P17, had significantly shorter bristles than wild type flies. The shorter bristles of P25/+ flies were due to a reduction of the eIF2ygene during development as the transgenes were able to significantly increase the length of bristles of these flies. Another characteristic of Minute mutations is the enhancement of unrelated mutations. Although the P25 allele did enhance certain wing mutations, the effect was not pronounced. The P25/P17 and P25/322 survivors from the transgene rescue crosses did show delayed development and reduced fertility, two other hallmarks of Minute loci. In conclusion, fly strains that have reduced eIF2yactivity possessed four phenotypes associated with Minute mutations: recessive lethality, short bristles, delayed development, and reduced fertility. These results show that mutations 191 in a translation initiation factor have the same dosage sensitive phenotypes as do mutations in ribosomal proteins. Mutations in the Su(var)3-9 gene cause a dominant suppression of P E V phenotype. A l l alleles of the Sw(W)3-9/dF2ylocus were isolated by this criterion. With the exception of the partial revertants, P17rvl2 and P17rv9, each allele had a strong effect on the variegating white gene in the wm4 rearrangement. However, even the strongest alleles had only a moderate effect on a brown gene variegator, bwVDe2, and a Bar gene variegator, B s c v . It was argued that the Su(var)3-9 protein has a more specialized role than other Su(var) encoded proteins and this accounts for the relatively weak phenotype of the Su(var)3-9 alleles. The Su(var)3-9 protein is a chromatin protein found at the Histone gene cluster (S. Ner, personal communication). Surprisingly, flies that carried both deletions of the HIS-C and mutations in the Su(var)3-9 gene were viable, physically robust, and fertile. Preliminary Northern blot analysis indicated that the 336 allele which presumably decreases the amount of Su(var)3-9 protein caused an increase in at least two of the Histone transcripts. Both experiments suggest that the mutations in the Su(var)3-9 gene cause upregulation of the Histone genes. As the HIS-C locus is comprised of tandem repeats it may require a specialized form of chromatin to remain transcriptionally competent. The Su(var)3-9 protein may function in the formation, regulation, or maintenance of this chromatin. 192 C H A P T E R 9: C O N C L U S I O N S (1) The first hypothesis, compound genes are relatively more vulnerable to mutations, was supported. A l l four independent transposable element insertions into this locus affected both genes. Three of these mutations showed dominant phenotypes due to each gene being compromised. The 5' ends of genes are preferentially targeted by transposable elements as was the case with these alleles. Because the two genes in compound loci overlap at the 5' end, transposable element insertions are more likely to damage both genes and therefore affect multiple cellular processes. (2) The second hypothesis, the two genes in the Su(var)3-9/eIF2ylocus are structurally or functionally related, was not supported. The Su(var)3-9 gene uses a common promoter and two exons of the elFly gene but is otherwise structurally distinct. The functions of each gene, as determined by genetic analysis, are independent as well. The likely origin of this locus was a retroposition event that inserted most of the Su(var)3-9 openreading frame into the second intron of the eZF2ygene (3) Mutations in the eIF2ygene are recessive lethal and have other phenotypes associated with Minute mutations: short slender sensory bristles, delayed development, and reduced fertility. Therefore, mutations in a gene encoding a translation initiation factor have the same dosage sensitive phenotypes as mutations in genes encoding ribosomal proteins. 193 (4) Mutations in the Su(var)3-9 gene, though not recessive lethal, cause a dominant suppressor of PEV phenotype and interact with the Histone gene cluster. The Su(var)3-9 protein has a more specialized role in chromatin structure than other Su(var) encoded proteins. 194 C H A P T E R 10: F U T U R E R E S E A R C H In terms of strengthening the conclusions reached by this thesis the Northern analysis must either be expanded or supplemented with protein work. It is not possible to conclude that either the P25 or the 322 alleles are null mutations based on the existing data. This could be remedied by repeating the Northern analysis but using replicates and a dilution series from each strain and then quantifying the intensity of the bands from a timecourse of phosphoimager exposures. A better source of m R N A than adults would also improve the interpretation of the results. It may be possible to collect homozygous embryos or larvae from 322 /TM3 or P25/TM3 parents. Barring maternal contribution, there would be expected to be little or none of either transcript present. A different method would be to determine the abundance of each protein directly in embryos, larvae, or adults using antibodies specific to the Su(var)3-9 and the eIF2y proteins. A n EMS mutagenesis screen for alleles specific for the elF2 y gene could have been done. (As preparation, an isogenic fly strain was generated that carried ebony, a recessive marker on the third chromosome, but there was not time to start the screen itself.) These alleles would have been useful to test both hypotheses as well as to allow further characterization of the dF2ygene. The function of the first 80 amino acids of the Su(var)3-9 protein are encoded by the common exons yet do not appear to belong to this protein as discussed in Section 8.4. If any missense mutations could be obtained which map in this region (which would presumably disrupt the eIF2ygene and cause lethality) the 195 effect they have on the Su(var)3-9 gene would be interesting. If they disrupted Su(var)3-9 as well it would support both hypotheses. Because the region of the locus that is essential to both genes is larger than first thought the first hypothesis would be supported. The second hypothesis would be supported as the structural overlap between the two proteins that is a consequence of the formation of this locus has allowed a functional overlap as well. While Table 16 listed all alleles besides 322 as completely viable in combination with the E(var)3-4, E(var)3-5, and E(var)3-6 mutations, during early experiments most E(var)/P17, E(var)/P25, and E(var)/322 combinations were non-viable. Because of this, the screen may not work as new mutations that are lethal with the P25 or 322 allele may actually be new alleles of these E(var) loci and not lesions in the Su(uar,)3-9/eIF2ylocus. This was one reason that P element mutagenesis was used in the screens for new eIF2yalleles described in Section 3.4. Any mutations that were lethal with the Su(var)3-9/ eIF2ylocus could be quickly mapped and characterized if they were due to P element insertions and excisions rather than EMS induced missense mutations. This problem has actually occurred to other researchers. Victor Corces' lab, which also works on chromatin structure, is investigating a gene that they mapped to the vicinity of the Su(var)3-9/eIF2ylocus. This gene, as defined by EMS alleles, was an excellent candidate for being one of these E(var) loci (V. Corces and Chi-Yun Pai, personal communication). In order to clone it they generated P element induced alleles. Unfortunately, all three of their new alleles 196 were actually P element insertions into the promoter of the Su(var)3-9/eIF2y locus (C.-Y. Pai, personal communication, June, 2000). The lethal interaction between the Su(var)3-9/dF2y locus and this gene could be due to their close proximity on the chromosome such that single mutations are capable of disrupting both loci. If the loci are separated then the reason for the lethality of double heterozygous flies depends on whether the genes are involved in a common cellular process or are functionally independent. The flies die because either one cellular function is badly compromised or two different functions are moderately damaged. The resolution of this situation may have implications for the regulation of the Su(pflr)3-9/(3lF2ylocus and the functions fulfilled by its genes. O n a larger scale this inter-loci interaction may reveal just how independent different genes are in terms of their local chromosome architecture. The true nature of all three E(var) mutations is not known. It was proposed in Section 6.6 that they may affect the Su(uar)3-9/eJF2ylocus directly, perhaps by interfering with the normal alternative splicing pattern. Sequencing the locus in these strains would allow this to be determined. If one or more of the E(var) mutations were a defect at the Su(var)3-9/eIF2ylocus it would strengthen the first hypothesis. This would be another example of how one mutation can compromise two genes when the genes are overlapping. The mechanism by which the E(var) mutation generated its diverse array of phenotypes would be interesting from the perspective of the regulation of alternative splicing. 197 The 336 allele causes a strange situation, certainly in comparison with the superficially similar P27 allele, whereby damage in the common region of this locus preferentially compromised the Su(var)3-9 gene. First this result would have to be verified as discussed above. Then a reverse transcriptase-PCR approach would be used to determine if any pre-mRNA or mature transcripts contain part or the entire hobo element insert. This would reveal if the Su(var)3-9 transcript is preferentially burdened with this additional region and perhaps other parts of the first intron as well. As the nonsense-mediated m R N A decay (NMD) pathway may prevent the accumulation of irregular transcripts, N M D mutants (if they exist) would be tested to determine if they alter the abundance of the Su(var)3-9 transcripts in 336 flies. Two minor experiments would clarify the 322 allele and one of the Minute phenotypes. The insert in the 322 could be confirmed (or not) as a retrotransposable element insert. A PCR strategy using degenerate primers specific to conserved regions of Drosophila retrotransposons would be the fastest way to achieve this. The one Minute phenotype that was not examined was the delayed larval development. The existing alleles could be tested for a dominant lengthening of this period, though the results would have to be interpreted cautiously. This is because all of these alleles also disrupt the Su(var)3-9 gene. As Su(var)3-9 mutations appeared to make Df(HIS-C) flies healthier it is possible that they could also make wild type flies more robust and thus decrease the developmental time. For this reason it may be necessary to have alleles specific to the eIF2 y gene to perform this experiment. 198 The first hypothesis could be strengthened by a parallel investigation of another Drosophila compound locus, DUb80/IP259 (Mottus et al. 1997). There are several alleles of this locus, most of which were generated in our lab, yet a genetic analysis of this locus has not done. Only two categories of mutations of the Su(var)3-9/eIF2y\ocus were isolated, transposable element inserts into the 5' end of the locus and missense mutations in one of the two open reading frames. Some of the existing DUb80/IP259 alleles may be polar in their action or they may affect the alternative splicing of these transcripts. Because the two genes at this locus only share a promoter and a single non-coding exon it would be interesting to determine whether correspondingly fewer mutations affect both genes. The second hypothesis, though not supported by this thesis, could still be true if eIF2y is revealed to have a role in the Drosophila nucleus as it may have in mammals. Two methods could test this. The location of the eIF2y protein could be determined using either eIF2y specific antibodies (those generated against the mammalian protein may also recognize the fly protein) or by generating transgenic flies which express a tagged eIF2y protein. The other method would be to check for genetic interactions between alleles of eIF2y and genes involved in the same processes as the mammalian D N A - P K and K u proteins. (In mammals these proteins require eIF2 for their stability with D N A ; Ting et al. 1998.) 199 R E F E R E N C E S Aagaard, L . , G. Laible, P. Selenko, M . Schmid, R. Dorn, G . Schotta, S. Kuhfittig, A . Wolf, A . Lebersorger, P. B. Singh, G. Reuter, and T. Jenuwein, 1999, Functional mammalian homologues of the Drosophila PEV-modifier Su(var)3-9 encode centromere-associated proteins which complex with the heterochromatin component M31. E M B O J18:1923-1938. Aagaard, L . , M . Schmid, P. Warburton, T. Jenuwein, 2000, Mitotic phosphorylation of SUV39H1, a novel component of active centromeres, coincides with transient accumulation at mammalian centromeres. J Cell Sci 113: 817-829. Adams, M . D., et al., 2000, The genome sequence of Drosophila melanogaster. Science 287:2185-2195. Albalat, R., and R. Gonzalez-Duarte, 1993, Adh and Adh-dup sequences of Drosophila lebanonensis and D. immigrans: interspecies comparisons. Gene 126: 171-178. Alfonso, A . , K. Grundahl, J. R. McManus, J. M . Asbury and J. B. Rand, 1994, Alternative splicing leads to two cholinergic proteins in Caenorhabditis elegans. J Mol Biol 241:627-630. Altschul, S. F., T. L. Madden, A . A . Schaffer, J. Zhang, Z . Zhang, W. Miller, and D. J. Lipman, 1997, Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res 25:3389-3402. Andrews, J., M . Smith, J. Merakovsky, M . Coulson, F. Hannan, and L . E. Kelly, 1996, The stoned locus of Drosophila melanogaster produces a dicistronic transcript and encodes two distinct polypeptides. Genetics 143:1699-1711. 200 Arkhipova, I. R., N . V. Lyubomirskaya, and Y. V. Ilyin, 1995, Drosophila retrotransposons. R. G. Landes Company, Austin, Texas. Ashburner, M . , 1989, Drosophila: a laboratory handbook. Cold Spring Harbor Laboratory Press, Cold Spring Harbor. Belyaeva, E. S., I. F. Zhimulev, E. I. Volkova, A . A . Alekseyenko, Y. M . Moshkin, and D. E. Koryakov, 1998, Su(UR)ES: a gene suppressing D N A underreplication in intercalary and pericentric heterochromatin of Drosophila melanogaster polytene chromosomes. Proc Natl Acad Sci U S A 95: 7532-7537. Blumenthal, T., 1998, Gene clusters and polycistronic transcription in eukaryotes. BioEssays 20:480-487. Blumenthal, T., and J. Spieth, 1996, Gene structure and organization in Caenorhabditis elegans. Curr Opin Genet Dev 6: 692-698. Breen, T. R., 1999, Mutant alleles of the Drosophila trithorax gene produce common and unusual homeotic and other developmental phenotypes. Genetics 152:319-344. Brogna, S., and M . Ashburner, 1997, The Adh-related gene of Drosophila melanogaster is expressed as a functional dicistronic messenger RNA: multigenic transcription in higher organisms. E M B O J16: 2023-2031. Brosius, J., 1991, Retroposons-seeds of evolution. Science 251: 753. Capy, P., C. Bazin, D. Higuet, and T. Langin, 1998, Dynamics and evolution of transposable elements. Molecular Biology Intelligence Unit, Austin Texas 201 Cavener, D. R v and B. A. Cavener, 1993, Translation start sites and m R N A leaders., pp. 359-377 in An atlas of Drosophila genes, edited by G. Maroni. Oxford University Press, Oxford. Cleard, R , M . Delattre and P. Spierer, 1997, SU(VAR)3-7, a Drosophila heterochromatin-associated protein and companion of HP1 in the genomic silencing of position-effect variegation. E M B O J16:5280-5288. Culbertson, M . R., 1999, R N A surveillance: unforeseen consequences for gene expression, inherited genetic disorders and cancer. Trends Genet 15: 74-80 De Haro, C. , R. Mendez and J. Santoyo, 1996, The eIF-2oc kinases and the control of protein synthesis. FASEB J10:1378-1387. Dobie, K. , M . Mehtali, M . McClenaghan and R. Lathe, 1997, Variegated gene expression in mice. Trends Genet 13:127-130. Dorer, D., and S. Henikoff, 1994, Expansions of transgene repeats cause heterochromatin formation and gene silencing in Drosophila. Cell 77: 993-1002. Dorn, R., V. Krauss and H . Saumweber, 1993a, The enhancer of position-effect variegation of Drosophila, E(var)3-93D, codes for a chromatin protein containing a conserved domain common to several transcriptional regulators. Proc Natl Acad Sci USA 90:11376-11380. Dorn, R., H . Morawietz, G . Reuter and H . Saumweber, 1993b, Identification of an essential Drosophila gene that is homologous to the translation initiation factor eIF-4A of yeast and mouse. M o l Gen Genet 237: 233-240. 202 Dorn, R., J. Szidonya, G. Korge, M . Sehnert, H . Taubert, E. Archoukieh, B. Tschiersch, H . Morawietz, G. Wustmann, G. Hoffmann, and G . Reuter, 1993c, P transposon-induced dominant enhancer mutations of position-effect variegation in Drosophila melanogaster. Genetics 133:279-290. Dorris, D. R., F. L. Erickson and E. M . Hannig, 1995, Mutations in GCD11, the structural gene for eIF-2gamma in yeast, alter translational regulation of G C N 4 and the selection of the start site for protein synthesis. E M B O J 14: 2239-2249. Duncan, R. F., D. R. Cavener and S. Qu, 1995, Heat shock effects on phosphorylation of protein synthesis initiation factor proteins eIF-4E and eIF-2alpha in Drosophila. Biochemistry 34:2985-2997. Ehrmann, I. E. , P. S. Ellis, S. Mazeyrat, S. Duthie, N . Brockdorff, M . G . Mattel, M . A . Gavin, N . A . Affara, G. M . Brown, E. Simpson, M . J. Mitchell, and D. M . Scott, 1998, Characterization of genes encoding translation initiation factor eIF-2y in mouse and human: sex chromosome localization, escape from X-inactivation and evolution. Human Mol Genet 7:1725-1737. Eickbush, T., 1999, Exon shuffling in retrospect. Science 283:1465-1466 Eiden, L . E . , 1998, The cholinergic gene locus. J of Neurochem 70:2227-2240. Eissenberg, J. C , T. C . James, D. M . Foster-Hartnett, T. Hartnett, V. Ngan, and S. C. R. Elgin, 1990, Mutation in a heterochromatin-specific chromosomal protein is associated with a suppressor of position-effect variegation in Drosophila melanogaster. Proc Natl Acad Sci U S A 87:9923-9927. Ekwall, K, J.-P. Javerzat, A . Lorentz, H . Schmidt, G . Cranston, and R. Allshire, 1995, The chromodomain protein Swi6: a key component at fission yeast centromeres. Science 269:1429-1431. 203 Ekwall, K. , E . R. Nimmo, J.-P. Javerzat, B. Borgstrom, R. Egel, G . Cranston, and R. Allshire, 1996, Mutations in the fission yeast silencing factors clr4+ and rikl+ disrupt the localization of the chromodomain protein Swi6p and impair centromere function. J Cell Sci 109:2637-2648. Ekwall, K. , and T. Ruusala, 1994, Mutations in rikl, clrl and clr4 genes asymmetrically derepress the silent mating-type loci in fission yeast. Genetics 136: 53-64. Engels, W. R., W. K. Benz, C. R. Preston, P. L. Graham, R. W. Phillis, and H . M . Robertson, 1987, Somatic effects of P element activity in Drosophila melanogaster. pupal lethality. Genetics 117: 745-757. Erickson, F. L . , and E. M . Hannig, 1996, Ligand interactions with eukaryotic translation initiation factor 2: role of the gamma-subunit. E M B O J15: 6311-6320. Erickson, F. L . , L . D. Harding, D. R. Dorris and E. M . Hannig, 1997, Functional analysis of homologs of translation initiation factor 2gamma in yeast. M o l Gen Genet 253:711-719. Fanti, L . , D. R. Dorer, M . Berloco, S. Henikoff and S. Pimpinelli, 1998, Heterochromatin protein 1 binds transgene arrays. Chromosoma 107: 286-292. Farkas, G. , J. Gausz, M . Galloni, G . Reuter, H . Gyurkovics, and F. Karch, 1994, The Trithorax-like gene encodes the Drosophila G A G A factor. Nature 371:806-808. Farrell, R. E. , 1993, RNA Methodologies: A laboratory guide for isolation and characterization. Academic Press, Toronto. Fergestad, T , W. S. Davis, and K. Broadie, 1999, The stoned proteins regulate synaptic vesicle recycling in the presynaptic terminal. J Neurosci 19:5847-5860. 204 Ferrus, A . , 1975, Parameters of mitotic recombination in Minute mutants of Drosophila melanogaster. Genetics 79:589-599. Garzino, V . , A . Pereira, P. Laurenti, Y. Graba, R. W. Levis, Y. L . Parco, and J. Pradel, 1992, Cell lineage-specific expression of modulo, a dose-dependent modifier of variegation in Drosophila. E M B O J11:4471-4479. Gaspar, N . J., T. G. Kinzy, B. J. Scherer, M . Hiimbelin, J. W. B. Hershey, and W. C. Merricks, 1994, Translation initiation factor eIF-2: cloning and expression of the human c D N A encoding the gamma-subunit. J Biol Chem 269:3415-3422. Giordano, E . , I. Peluso, S. Senger, M . Furia, 1999, minifly, a Drosophila gene required for ribosome biogenesis. J Cell Biol 144:1123-1133. Gray, T. A . , S. Saitoh and R. D. Nicholls, 1999, A n imprinted, mammalian bicistronic transcript encodes two independent proteins. Proc Natl Acad Sci USA 96:5616-5621. Grigliatti, T., 1991, Position-effect variegation-An assay for nonhistone chromosomal proteins and chromatin assembly and modifying factors, pp. 587-627 in Methods in Cell Biology, Academic Press. Grigliatti, T. A. , 1986, Mutagenesis., pp. 39-58 in Drosophila: a practical approach, edited by D. B. Roberts. IRL Press, Oxford. Hannig, E. M . , A . M . Cigan, B. A . Freeman and T. G . Kinzy, 1993, GCD11, a negative regulator of GCN4 expression, encodes the gamma subunit of eIF-2 in Saccharomyces cerevisiae. Mol Cell Biol 13:506-520. Hatton, A . R., V. Subramaniam and A . J. Lopez, 1998, Generation of alternative Ultrabithorax isoforms and stepwise removal of a large intron by resplicing at exon-exon junctions. Mol Cell 2: 787-796. 205 Henikoff, S., 1996, Position-effect variegation in Drosophila: recent progress., pp. in Epigenetic mechanisms of gene regulation, edited by V. E. A . Russo, R. A . Martienssen and A . D. Riggs. Cold Spring Harbor Laboratory Press, New York. Henikoff, S., 1998, Conspiracy of silence among repeated transgenes. BioEssays 20:532-535. Hennig, W. , 1999, Heterochromatin. Chromosoma 108:1-9. Huang, H . , H . Yoon, E. M . Hannig and T. F. Donahue, 1997, GTP hydrolysis controls stringent selection of the A U G start codon during translation initiation in Saccharomyces cerevisiae. Genes Dev 11:2396-2413. Ivanova, A . V. , M . J. Bonaduce, S. V. Ivanov and A . J. S. Klar, 1998, The chromo and SET domains of the Clr4 protein are essential for silencing in fission yeast. Nature Genet 19:192-195. James, C . T., J. C. Eissenberg, C . Craig, V. Dietrich, A . Hobson, and S. C . R. Elgin, 1989, Distribution patterns of HP1, a heterochromatin-associated nonhistone chromosomal protein of Drosophila. Europe J Cell Biol 50:170-180. Jeffs, P. S., E. C . Holmes and M . Ashburner, 1994, The molecular evolution of the alcohol dehydrogenase and alcohol dehydrogenase-related genes in the Drosophila melanogaster species subgroup. Mol Biol Evol 11:287-304. Jenuwein, T., G. Laible, R. Dorn and G . Reuter, 1998, SET domain proteins modulate chromatin domains in eu- and heterochromatin. Cell M o l Life Sci 54: 80-93. Keeling, P. J., N . M . Fast and G. I. McFadden, 1998, Evolutionary relationship between translation initiation factor eIF-2y and selenocysteine-specific elongation factor SELB: change of function in translation factors. J M o l Evol 47:649-655. 206 Keller, W. , 1995,3'-end cleavage and polyadenylation of nuclear messenger R N A precursors., pp. 113-134 in Pre-mRNA Processing, edited by A . I. Lamond. R.G. Landes Company. Kennison, J. A . , 1995, The Polycomb and trithorax group proteins of Drosophila: trans-regulators of homeotic gene function. A n n Rev Genet 29: 289-303. Kimball, S. R., 1999, Eukaryotic initiation factor eIF2. Int J Biochem Cell Biol 31: 25-29. King, R. C . and W. D. Stansfield, 1997, A dictionary of genetics. Oxford University Press, New York Kitamoto, T., W. Wang and P. M . Salvaterra, 1998, Structure and organization of the Drosophila cholinergic locus. J Biol Chem 273: 2706-2713. Kjeldgaard, M . , J. Nyborg and B. F. C . Clark, 1996, The GTP binding motif: variations on a theme. FASEB J10:1347-1368. Kozak, M . , 1999, Initiation of translation in prokaryotes and eukaryotes. Gene 234:187-208. Krieg, D. R., 1963, Ethyl methanesulfonate-induced reversion of bacteriophage T4rII mutants. Genetics 48:561-580. Lamb, M . M . , and C. D. Laird, 1987, Three euchromatic D N A sequences under-replicated in polytene chromosomes of Drosophila are localized in constrictions and ectopic fibers. Chromosoma 95: 227-235. Lambertsson, A . , 1998, The Minute genes in Drosophila and their molecular functions. Adv Genet 38:69-134. 207 Levis, R., K. O'Hare, and G. M . Rubin, 1984, Effects of transposable element insertions on R N A encoded by the white gene of Drosophila. Cell 38:471-481. Lifton, R. P., M . L. Goldberg, R. W.'Karp and D. S. Hogness, 1978, The organization of the histone genes in Drosophila melanogaster: functional and evolutionary implications. Cold Spring Harb Symp Quant Biol 42:1047-1051. Lim, J. K., and M . J. Simmons, 1994, Gross chromosome rearrangements mediated by transposable elements in Drosophila melanogaster. BioEssays 16:269-275. Lindsley, D. L. , and G. G. Zimm, 1992, The genome of Drosophila melanogaster. Academic Press, San Diego. Liu , H . , J. K. Jang, J. Graham, K. Nycz, and K. S. McKim, 2000, Two genes required for meiotic recombination in Drosophila are expressed from a dicistronic message. Genetics 154:1735-1746 Lobo, M . V. T., F. J. M . Alonso, S. Rodriguez, A . Alcazar, E. Martin, F. Munoz, R. G-Santander, M . Salinas, and J. L . Fando, 1997, Localization of eukaryotic initiation factor 2 in neuron primary cultures and established cell lines. Histochem J 29:453-467. Locke, J., S. Hanna and D. Kong, 1993, Evidence for mobilization of hobo transposons in a P-element mutagenesis screen. Genome 36:1138-1147. Locke, J., L. T. Howard, N . Aippersbach, L . Podemski and R. B. Hodgetts, 1999, The characterization of DINE-1, a short, interspersed repetitive element present on chromosome and in the centric heterochromatin of Drosophila melanogaster. Chromosoma 108: 356-366. 208 Locke, J., M . A . Kotarski and K. D. Tartof, 1988, Dosage-dependent modifiers of position effect variegation in Drosophila and a mass action model that explains their effect. Genetics 120:181-198. Long, M . , and C. H . Langley, 1993, Natural selection and the origin oijingwei, a chimeric processed functional gene in Drosophila. Science 260:91-95. Long, M . , W. Wang and J. Zhang, 1999, Origin of new genes and source for N -terminal domain of the chimerical gene, jingwei, in Drosophila. Gene 238:135-141. Lopez, A . J., 1998, Alternative splicing of pre-mRNA: developmental consequences and mechanisms of regulation. Annu Rev Genet 32:279-305. Lorentz, A . , K Ostermann, O. Fleck, and H . Schmidt, 1994, Switching gene swi6, involved in repression of silent mating-type loci in fission yeast, encodes a homologue of chromatin-associated proteins from Drosophila and mammals. Gene 143:139-143. McKim, K. S., C. Matheson, M . A . Marra, M . F. Wakarchuk and D. L . Baillie, 1994, The Caenorhabditis elegans unc-60 gene encodes proteins homologous to a family of actin-binding proteins. Mol Gen Genet 242:346-357. Merrick, W. C. , and J. W. B. Hershey, 1996, The pathway and mechanism of eukaryotic protein synthesis, pp. 31-69 in Translational Control, edited by J. W. B. Hershey, M . B. Mathews and N . Sonenberg. Cold Spring Harbor Laboratory Press, USA. Messmer, S., A . Franke and R. Paro, 1992, Analysis of the functional role of the Polycomb chromo domain in Drosophila melanogaster. Genes Dev 6:1241-1254. 209 Michailidis, J., N . D. Murray and J. A. Marshall Graves, 1988, A correlation between development time and variegated position effect in Drosophila melanogaster. Genet Res 52:119-123. Moore, G . D., D. A . Sinclair and T. A . Grigliatti, 1983, Histone gene multiplicity and position effect variegation in Drosophila melanogaster. Genetics 105:327-344. Mottus, R. C . , I. P. Whitehead, M . O'Grady, R. E. Sobel, R. H . L . Burr, G. B. Spiegelman, and T. A . Grigliatti, 1997, Unique gene organization: alternative splicing in Drosophila produces two structurally unrelated proteins. Gene 198:229-236. Mount, S. M . , 1993, Messenger R N A splicing signals in Drosophila genes., pp. 333-358 in An Atlas of Drosophila genes, edited by G. Maroni. Oxford University Press, New York. Muller, H . J., 1930, Types of visible variations induced by X-rays in Drosophila. J Genet 22: 299-334. Naranda, T., I. Sirangelo, B. J. Fabbri and J. W. B. Hershey, 1995, Mutations in the N K X D consensus element indicate that G T P binds to the y-subunit of translation initiation factor eIF2. FEBS Lett 372:249-252. Nurminsky, D. I., M . V. Nurminskaya, D. De Aguiar and D. L . Hartl, 1998, Selective sweep of a newly evolved sperm-specific gene in Drosophila. Nature 396: 572-575. O'Hare, K., and G . M . Rubin, 1983, Structures of P transposable elements and their sites of insertion and excision in the Drosophila melanogaster genome. Cell 34: 25-35. 210 Olsen, D. S v B. Jordan, D. Chen, R. C. Wek and D. R. Cavener, 1998, Isolation of the gene encoding the Drosophila melanogaster homolog of the Saccharomyces cerevisiae GCN2 eIF-2a kinase. Genetics 149:1495-1509. Ono, S., and G . M . Benian, 1998, Two Caenorhabditis elegans actin depolymerizing factor/cofilin proteins, encoded by the unc-60 gene, differentially regulate actin filament dynamics. J Biol Chem 6:3778-3783. Osley, M . A . , 1991, The regulation of histone synthesis in the cell cycle. A n n Rev Biochem 60:827-861. Pardue, M . L . , L . H . Kedes, E. S. Weinberg and M . L . Birnstiel, 1977, Localization of sequences coding for histone messenger R N A in the chromosomes of Drosophila melanogaster. Chromosoma 63:135-151. Paro, R., and P. J. Harte, 1996, The role of Poly comb group and Trithorax group chromatin complexes in the maintenance of determined cell states., pp. in Epigenetic Mechanisms of Gene Regulation, edited by V. E. A . Russo, R. A . Martienssen and A . D. Riggs. Cold Spring Harbor Laboratory Press, New York. Paro, R., and D. S. Hogness, 1991, The Polycomb protein shares a homologous domain with a heterochromatin-associated protein of Drosophila. Proc Natl Acad Sci USA 88:263-267. Perrin, L . , O. Demakova, L . Fanti, S. Kallenbach, S. Saingery, N . I. Mal'ceva, S. Pimpinelli, I. Zhimulev, and J. Pradel, 1998, Dynamics of the sub-nuclear distribution of Modulo and the regulation of position-effect variegation by nucleolus in Drosophila. J Cell Sci 111: 2753-2761. Pirrotta, V. , 1998, Polycombing the genome: PcG, trxG, and chromatin silencing. Cell 93:333-336. 2 1 1 Platero, J. S., T. Hartnett and J. C. Eissenberg, 1995, Functional analysis of the chromo domain of HP1. E M B O J14:3977-3986. Powers, J. A . , and J. C . Eissenberg, 1993, Overlapping domains of the heterochromatin-associated protein HP1 mediate nuclear localization and heterochromatin binding. J Cell Biol 120: 291-299. Qu, S., and D. R. Cavener, 1994, Isolation and characterization of the Drosophila melanogaster eIF-2a gene encoding the alpha subunit of translation initiation factor eIF-2. Gene 140: 239-242. Qu, S., S. E . Perlaky, E. L . Organ, D. Crawford and D. R. Cavener, 1997, Mutations at the Ser50 residue of translation factor eIF-2a dominantly affect development rate, body weight, and viability of Drosophila melanogaster. Gene Expr 6: 349-360. Rand, J. B., 1989, Genetic analysis of the cha-1 unc-17 gene complex in Caenorhabditis. Genetics 122: 73-80. Reuter, G. , R. Dorn, G. Wustmann, B. Friede and G . Rauh, 1986, Third chromosome suppressor of position-effect variegation loci in Drosophila melanogaster. Mol Gen Genet 202:481-487. Reuter, G. , M . Giarre, J. Farah, J. Gausz, A . Spierer, and P. Spierer, 1990, Dependence of position-effect variegation in Drosophila on dose of a gene encoding an unusual zinc-finger protein. Nature 344:219-223. Reuter, G. , H.-J. Hoffmann and I. Wolff, 1983, Genetic study of position-effect variegation in Drosophila melanogaster: In(l)zvm4 as a standard rearrangement for the isolation and characterization of suppressor and enhancer mutants. Biologisches Zentralblatt 102:281-298. 212 Robertson, H . M . , C . R. Preston, R. W. Phillis, D. M . Johnson-Schlitz, W. K. Benz, W. R. Engels, 1988, A stable genomic source of P element transposase in Drosophila melanogaster. Genetics 118:461-470. Rubin, G . M . , and A . C. Spradling, 1982, Genetic transformation of Drosophila with transposable element vectors. Science 218: 348-353. Sabl, J. F., and S. Henikoff, 1996, Copy number and orientation determine the susceptibility of a gene to silencing by nearby heterochromatin in Drosophila. Genetics 142:447-458. Saeb0e-Larssen, S., M . Lyamouri, J. Merriam, M . P. Oksvold and A . Lambertsson, 1998, Ribosomal protein insufficiency and the Minute syndrome in Drosophila: a dose-response relationship. Genetics 148:1215-1224. Saeboe-Larssen, S., B. Urbanczyk Mohebi and A . Lambertsson, 1997, The Drosophila ribosomal protein L14-encoding gene, identified by a novel Minute mutation in a dense cluster of previously undescribed genes in cytogenetic region 66D. Mol Gen Genet 255:141-151. Sambrook, J., E. F. Fritsch and T. Maniatis, 1989, Molecular cloning: a laboratory manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor. Santoyo, J., J. Alcalde, R. Mendez, D. Pulido and C . de Haro, 1997, Cloning and characterization of a c D N A encoding a protein synthesis initiation factor-2a (eIF-2a) kinase from Drosophila melanogaster. J Biol Chem 272:12544-12550. Sass, G . L. , and S. Henikoff, 1998, Comparative analysis of position-effect variegation mutations in Drosophila melanogaster delineates the targets of modifiers. Genetics 148:733-741. 213 Saville, K. J., W. D. Warren, P. W. Atkinson and D. A . O'Brochta, 1999, Integration specificity of the hobo element of Drosophila melanogaster is dependent on sequences flanking the intergration site. Genetica 105:133-147. Sawin, K. E . , and P. Nurse, 1996, Identification of fission yeast nuclear markers using random polypeptide fusions with green fluorescent protein. Proc Natl Acad Sci U S A 93:15146-15151. Schmucker, D. , H . Jackie and U . Gaul, 1997, Genetic analysis of the larval optic nerve projection in Drosophila. Development 124: 937-948. Schotta, G . , and G. Reuter, 2000, Controlled expression of tagged proteins in Drosophila using a new modular P-element vector system. Mol Gen Genet 262: 916-920. Sinclair, D. A . R., T. A . Grigliatti and T. C. Kaufman, 1984, Effects of a temperature-sensitive Minute mutation on gene expression in Drosophila melanogaster. Genet Res 43:257-275. Sinclair, D. A . R., R. C. Mottus and T. A . Grigliatti, 1983, Genes which suppress position-effect variegation in Drosophila melanogaster are clustered. M o l Gen Genet 191: 326-333. Sinclair, D. A . R., D. T. Suzuki and T. A . Grigliatti, 1981, Genetic and developmental analysis of a temperature-sensitive Minute mutation of Drosophila melanogaster. Genetics 97:581-606. Streck, R. D., J. E . MacGaffey and S. K. Beckendorf, 1986, The structure of hobo transposable elements and their insertion sites. E M B O J 5:3615-3623. 214 Talbert, P. B., and S. Henikoff, 2000, A reexamination of spreading of position-effect variegation in the white-roughest region of Drosophila melanogaster. Genetics 154:259-272. Tartof, K. D., C. Hobbs and M . Jones, 1984, A structural basis for variegating position effects. Cell 37:869-878. Thon, G . , A . Cohen and A . J. Klar, 1994, Three additional linkage groups that repress transcription and meiotic recombination in the mating-type region of Schizosaccharomyces pombe. Genetics 138: 29-38. Thummel, C. S., and V. Pirrotta, 1991, Technical notes: new pCaSpeR P-element vectors. Dros Info News Ting, N . S. Y., P. N . Kao, D. W. Chan, L . G . Lintott, and S. P. Lees-Miller, 1998, DNA-dependent protein kinase interacts with antigen receptor response element binding proteins NF90 and NF45. J Biol Chem 273:2136-2145. Trachsel, H . , 1996, Binding of initiator methionyl-tRNA to ribosomes., pp. 113-138 in Translational Control, edited by J. W. B. Hershey, M . B. Mathews and N . Sonenberg. Cold Spring Harbor Laboratory Press, USA. Tschiersch, B., A . Hofmann, V. Krauss, R. Dorn, G. Korge, and G . Reuter, 1994, The protein encoded by the Drosophila position-effect variegation suppressor gene Su(var)3-9 combines domains of antagonistic regulators of homeotic gene complexes. E M B O J13:3822-3831. Turner, B. M . , A . J. Birley, J. Lavender, 1992, Histone H4 isoforms acetylated at specific lysine residues define individual chromosomes and chromatin domains in Drosophila polytene nuclei. Cell 69:375-384. 215 Virtanen, A . , 1995, Strategies for regulating nuclear pre-mRNA polyadenylation., pp. 135-149 in Pre-mRNA processing, edited by A. I. Lamond. R.G. Landes Company. Wallrath, L . L . , 1998, Unfolding the mysteries of heterochromatin. Curr Opin Genet Dev 8:147-153. Wallrath, L . L . , and S. C. R. Elgin, 1995, Position effect variegation in Drosophila is associated with an altered chromatin structure. Genes Dev 9:1263-1277. Weiler, K. S., and B. T. Wakimoto, 1995, Heterochromatin and gene expression in Drosophila. A n n Rev Genet 29: 577-605. Wolffe, A . P., 1998, When more is less. Nature Genet 18:5-6. Wreggett, K. A. , F. Hill , P. S. James, A . Hutchings, G . W. Butcher, and P. B. Singh, 1994, A mammalian homologue of Drosophila heterochromatin protein 1 (HP1) is a component of constitutive heterochromatin. Cytogenet Cell Genet 66: 99-103. Wustmann, G. , J. Szidonya, H . Taubert and G . Reuter, 1989, The genetics of position-effect variegation modifying loci in Drosophila melanogaster. M o l Gen Genet 217: 520-527. Ye, X., and D. R. Cavener, 1994, Isolation and characterization of the Drosophila melanogaster gene encoding translation-initiation factor eIF-2p\ Gene 142:271-274. Zorin, I. D., T. I. Gerasimova, and V. G . Corces, 1999, The lawc gene is a new member of the trithorax-group that affects the function of the gypsy insulator of Drosophila. Genetics 152:1045-1055. 


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



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"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items