Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Genetic and molecular analysis of the garnet eye colour gene of Drosophila melanogaster Lloyd, Vett 1995

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

Item Metadata


831-ubc_1995-060047.pdf [ 5.92MB ]
JSON: 831-1.0088352.json
JSON-LD: 831-1.0088352-ld.json
RDF/XML (Pretty): 831-1.0088352-rdf.xml
RDF/JSON: 831-1.0088352-rdf.json
Turtle: 831-1.0088352-turtle.txt
N-Triples: 831-1.0088352-rdf-ntriples.txt
Original Record: 831-1.0088352-source.json
Full Text

Full Text

GENETIC AND MOLECULAR ANALYSIS OF THE garnet EYE COLOUR GENEOF Drosophila melanogasterbyVETT LLOYDBSc., The University of British ColumbiaMSc., The University of GenevaA THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFDOCTOR OF PHILOSOPHYinTHE FACULTY OF GRADUATE STUDIESGenetics ProgrammeWe accept this thesis as conformingto the required standardTHE UNIVERSITY OF BRITISH COLUMBIAJuly 1995© Vett Lloyd, 1995In presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. it is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.(Signature)_______________Department of cO6YThe University of British ColumbiaVancouver, CanadaDate 4. /2DE-6 (2/88)Abstract.The garnet eye colour gene of Drosophila melanogaster is one the group ofgenes called the transport group of eye colour genes. The garnet generesembles other members of the transport group of eye colour genes in itsphenotype and shows extensive genetic interactions with them. The mostsignificant interaction is between garnet and a cryptic allele of the white gene,first identified as a mutation called enhancer of garnet (we(g)). The phenotypeof garnet mutations and the extreme sensitivity to decreased levels of the white+gene product suggest that garnet, as well as other members of the transportgroup of eye colour genes, act as positive regulators of the white gene. Thisinteraction may occur at the protein level. A simple model for the physicalinteractions between the gene products of garnet, white and other members ofthe transport group is proposed.A critical test of this model requires molecular cloning and analysis of theindividual members of the transport group of eye colour genes. Preliminarymolecular analysis of the garnet gene is reported in chapter two. The garnetgene is expressed in many different tissues at different stages in development.Two messages are produced from the garnet gene in wild type embryos.Conceptual translation of a 4kb c-DNA reveals a novel protein.In the final chapter I describe the use of the garnet gene to study an example ofepigenetic gene regulation. I have examined a mini-chromosome whichvariegates for the garnet gene. The variegation of this mini-chromosome isextremely unusual in that it depends on the sex of the fly transmitting the minichromosome. In this way it conforms to the conventional definition of parentalgenomic imprinting. I examined a number of possible mechanisms which mightbe responsible for the parental imprinting of the mini-chromosome The resultssuggest that heterochromatin formation is responsible for the somaticexpression of the genomic imprint, but a different system may operate toestablish the imprint.IIITABLE OF CONTENTSAbstract iiTable of contents ivList of Tables ViList of Figures xAcknowledgments xvDedication xviGeneral introduction 1Materials and Methods 39Chapter One Interactions between garnet andother eye colour genes.Introduction 45Results 47Discussion 102Chapter Two Analysis of the garnet geneIntroduction 116Results 118ivDiscussion 187Chapter Three Imprinting of a mini-chromosome inDrosophila melanogaster.Introduction 205Results 220Discussion 269Bibliography 287Appendix 1 Determination of eye pigment levels 306Appendix 2 Cloning of the garnet gene 312VIndex of Tables: pageTable 1. A list of genes which affect eye colour in Drosophilamelanogaster. 3Table 2. Genes proposed as members of the transport group ofeye colour genes. 32Table 3. Eye colour genes with known or proposed functions. 36Table 4. The Effect of 3C deficiencies on white and enhancerof garnet. 54Table 5. Complementation between e(g) and different whitealleles. 60Table 6. Rescue of the enhancer of garnet effect by white+transgenes. 64Table 7. Sensitivity of the enhancer of garnet effect togarnet dosage. 68Table 8. Effect of we(g) dosage on the enhancer of garneteffect. 71Table 9. The effect of the enhancer of garnet mutation onother eye colour mutants. 78viTable 10. Interactions between garnet and other eye colourgenes. 82Table 11. Summary of interactions between other eye colourgenes and enhancer of garnet and garnet. 84Table 12. Phenotypes of garnet double mutants with differentwhite alleles. 89Table 13. Epistatic interactions between the wa3 and theg2alleles. 91Table 14. Effect of zeste on garnet. 96Table 15. Effect of zeste modifiers on the zeste-garnetgenotype. 100Table 16. Pigment levels of various garnet alleles. 123Table 17. Quantitative assessment of pteridine pigmentsafter chromatographic separation of pigments. 128Table 18. Effect of various garnet alleles on colour ofmalpighian tubules. 131Table 19. Effect of various garnet alleles on testes sheathcolour. 134viiTable 20. The phenotype of various garnet alleles incombination with a deficiency. 137Table 21. Summary of lesions in different garnet alleles. 158Table 22. List of genes with sequence similarity to thegarnet gene. 179Table 23. List of published garent alleles. 189Table 24. List of other names given to imprinting phenomenon. 208Table 25. Imprinting and human medical conditions. 212Table 26. Parental effects in Drosophila melanogaster. 215Table 27. The effect of different garnet alleles on the imprint. 242Table 28. Maternal effect of the garnet gene. 245Table 29. Variegation of the Dp(1;f)LJ9 mini-chromosome:The effect of developmental temperature. 255Table 30. Variegation of the Dp(1;f)LJ9 mini-chromosome:The effect of sodium butryrate. 257viiiTable 31. Variegation of the Dp(1;f)LJ9 mini-chromosome:The effect of Y chromosome dosage: 259Table 32. Imprinting of the Dp(1;f)LJ9 mini-chromosome:The effect of developmental temperature. 262Table 33. Imprinting of the Dp(1;f)LJ9 mini-chromosome:The effect of sodium butryrate. 264Table 34. Imprinting of the Dp(1;f)LJ9 mini-chromosome:The effect of Y chromosome dosage. 267Table 35. The effect of attatched versus free sex chromosomeson imprinting of Dp(1;f)LJ9. 284ixIndex of Figures. pageFigure 1. Diagram of the eye, structure and organization ofdifferent cell types in the ommatidia and organization of the primaryand secondary pigment cells of Drosophila melanogaster. 11Figure 2. Diagram of the biosynthetic pathway of xanthommatinproduction in Drosophila melanogaster. 16Figure 3. A possible pathway of pteridine pigment biosynthesisin Drosophila melanogaster. 18Figure 4. Three different models for pteridine biosynthesis inDrosophila melanogaster. 20Figure 5. Phenotypes of three severe garnet alleles inconjunction with the enhancer of garnet mutation. 49Figure 6. Cytological localization of the enhancer of garnetmutation. 52Figure 7. Diagram of the structure of the white gene ofDrosophila melanogaster. 57Figure 8. Comparison between the severity of different whitealleles and their effect on garnet. 73xFigure 9. Analysis of garnet transcription by in situ hybridizationto rosy null tissues. 86Figure 10. The effect of zeste mutant alleles on garnetphenotype. 94Figure 11. The phenotype of zeste-garnet combinationsin females and males. 98Figure 12. Nolte’s model for the interaction between eyecolour genes. 110Figure 13. Model of the physical interactions between theproducts of white, brown, scarlet and the transport group ofeye colour genes, including garnet. 113Figure 14. Spectrum of eye colour phenotypes of differentalleles of the garnet gene. 120Figure 15. Chromatographic analysis of pteridine pigmentsof garnet alleles. 126Figure 16. Life span of wild type and various garnet mutants. 140Figure 17. Diagram of the mutation rate to garnet upon theS6-1 strain. 144xiFigure 18. Diagram of the mutation rate and phenotypes of derivativegarnet mutations derived from the original gP mutation. 146Figure 19. Southern transfer and hybridization analysis of thegP allele and gP derivative mutations. 150Figure 20. Southern transfer and hybridization analysis of thewild type garnet region and the g1 allele. 153Figure 21. Southern transfer and hybridization analysis ofgarnet mutants. 155Figure 22. Northern transfer and hybridization analysis ofwild type embryos and garnet mutant adults. 160Figure 23. Restriction fragment length analysis of lambdaphage clones containing garnet and flanking sequences. 163Figure 24. Restriction fragment map of garnet andsurrounding region. 165Figure 25. Strategy used to sequence the imaginal c-DNAclone of the garnet gene. 168Figure 26. Sequence and conceptual translation of theimaginal disc c-DNA clone of the garnet gene. 170XIIFigure 27. Analysis of garnet transcription by in situhybridization to various tissues. 183Figure 28. Southern analysis of regions of sequence similarityto the garnet gene in Drosophila melanogaster. 186Figure 29. Map of the garnet gene generated by intragenicrecombination. 193Figure 30. Diagram of the structure and origin of theDp(1;f)LJ9 mini-chromosome. 222Figure 31. Meiotic stability of the Dp(1;f)LJ9 mini-chromosome. 224Figure 32. The garnet phenotype in flies with maternallyand paternally derived Dp(1;f)LJ9 mini-chromosome. 228Figure 33. The garnet phenotype in malpighian tubules of flies withmaternally and paternally derived Dp(1;f)LJ9 mini-chromosome. 231Figure 34. Phenotype of narrow abdomen and tiny in flies bearingmaternally and paternally derived Dp(1;f)LJ9 mini-chromosomes. 234Figure 35. Variegation of narrow abdomen and tiny in genotypicallyidentical flies bearing either maternally or paternally derivedDp(1;f)LJ9 mini-chromosome. 236XIIIFigure 36. The Y chromosome does not cause the imprint. 240Figure 37. Test of the physiological compensation model. 249Figure 38. Under-representation of the garnet gene in theDp(1;f)LJ9 mini-chromosome. 252Figure 39. Model of parent-dependent spread of heterochromatinresponsible for the imprint of the Dp(1;f)LJ9 mini-chromosome. 273xivAcknowledgments:I would like to thank those who have helped me.It is not possible to list here all those who have given me support andinspiration, emotional, intellectual and financial, over the many years that thiswork was in progress. Nevertheless, those who have helped me know who theyare. They have my gratitude, and I hope, will allow me to opportunity toreciprocate.Although any work of science depends on the prior work of others, the specificcontributions of some must be mentioned.The garnet gene was cloned by Dr. D. Sinclair, working in the laboratory of Dr.G. Tener, biochemistry U.B.C. Dr. Sinclair initiated the molecular analysis of thegarnet gene, brought the parental effect of Dp(1;f)LJ9 to my attention andprovided me with advice and encouragement at every step. The gP allele,which facilitated cloning of the garnet gene was isolated by R. Wennberg, in thelaboratory of Dr. T. Grigliatti. The program used to analyze the pigment data waswritten by Dr. J. Berger. As with any genetic work, I am indebted to the fly stockcentres, at Bloomington, Indiana, Bowling Green and Umea, Sweden.A number of undergraduate students also contributed to this work in the courseof doing a directed studies project. D. Dyment and K Swanson contributed datato the pigment assay controls in appendix one. M. Maharaj made some of thestocks used to test the effect of the e(g) mutation on other eye colour genes. G.Mahon mobilized the CaSpeR element construct and characterized the newinsertion strains that were used to test the rescue of the enhancer of garneteffect. A. Rivers assisted in the sequencing of garnet genomic clones. B. Leecontributed to the test of the enhancer of garnet effect of different white alleles.L. Harvey tested the effect of some garnet alleles on longevity. The details of theexperiments are given in the appropriate section.In addition to these students, I also had the opportunity to supervise a number ofother undergraduate students, from whom I learned at least as much as theyfrom me. Although their work is unpublished they are: Melanie Klenk, KeyvanHyunda, Cara Warrington, Carol Lee, Gurdip Lalli, Lynn Ma, Chaucer Wong,and Mahnaz Kermati.Finally, I would like to thank my supervisor Dr. T. Grigliatti for allowing me therare privilege of complete freedom of research.xvQuotation:There is no answer. There has never been an answer, there will never be ananswer. That’s the answer.-Gertrude Stein.xviGeneral IntroductionThe garnet gene.1General introduction.The problem:The eye colour genes of Drosophila melanogaster have attracted the attentionof biologists for as long as the organism has been used as an experimentalsystem for genetics. Aside from their intrinsic beauty, eye colour genes haveplayed a role in the genesis of many important genetic concepts: sex linkage,position effects, pleiotropy, the chromosome theory of inheritance, autonomousand non-autonomous gene action, intra allelic complementation, taxonomy, theaction of genes through hormones and developmental regulatory cascades.One of the features of Drosophila eye colour mutants which first appealed tobiologists was the number of mutations which have an effect on eye colour.While an invaluable genetic resource, the remarkable number of eye colourgenes also posed practical and theoretical problems. The overriding problemposed by eye colour genes is the sheer number of them. Breme and Demerec(1942) list 45 eye colour genes. In 1950, 38 eye colour mutants were extant(Nolte 1950). By 1976, 51 eye colour genes were known (Phillips and Forrest1976). The list shown in Table 1 is derived from the list of all mutantsdescribed in Drosophila melanogaster up to 1992 (Lindsley and Zimm 1992). Itlists 110 genes; even if pattern and secondary effects are excluded, there arestill 85 genes whose primary effect is on eye colour. Nor are these 110 geneslikely to be all the genes which affect eye colour. Amongst eight eye colourmutations isolated from a cellar and a vineyard in Italy, in a study of naturallyoccurring eye colour variation, three defined novel eye colour genes(Calatayud, Jacobson and Ferré, 1989). A system allotted considerable geneticresources would be expected to be essential or at least important for the2Table 1. Genes which affect eye colour in Drosophila melanogaster.The first two columns list the name and gene designation. The eye colour of themutant is indicated in the third column. Where there is an allelic series thecolour of the first allele is given. This is only an approximate indication of theeye colour which varies with allele, age and often sex. The final column givesthe pigment group which is affected. Frequently the pigment or pigment groupaffected is not known and is only inferred from the eye colour. This instance isindicated by a question mark. In some cases the effect on eye pigment is asecondary effect of patterning defects which change cell fate specification as aresult of which pigment cells fail to differential. These instances are indicated by“pattern”. In other instances alteration in eye colour or pigmentation is asecondary result of other alterations such as decreased body size or increasedmelanization. These instances are indicated as “secondary”. This list wasderived from Lindsley and Zimm (1992).3Table 1. Genes which affect eye colour in Drosophila melanoqastermutant colour pigmentaffectedamy amethyst purple pteridines?bis bistre brown pteridines?bo bordeaux purple both?bos bordosteril brown pteridines?bre bright-eye orange ommochromes?bri bright orange ommochromes?buo burnt-orange orange ommochromes?bur burgundy brown ommochromesbw brown brown pteridinesca claret both bothcar carnation brown both?cast cast brown pteridines?cd cardinal bright bothcho chocolate brown pteridines?cm cinnamon brown ommochromesci clot dark pteridinescm carmine orange bothcmd carminoid orangish both?cml caramel brown pteridines?cn cinnabar orange ommochromescop copper brown pteridines?cr-3 cream-3 pale both?dcm dark carmine brown pteridines?diI--3 dilute-3 paler both?dk dark darker bothdke dark-eye brown patternDke-2 darkened eye darker pattern?dn doughnut pattern patterndor deep orange pale orange bothDr Drop darker patterndrb dark red brown darker both?dyb dusty body brown secondary?E(z) enhancer of zeste pale bothg garnet paler bothgi glass darker pattern?Hn Henna dark brown pattern?je jelly pink both?kar karmoisin orange pteridineskpn/awd killer of prune wild type pteridineslix little isoxanthopterin wild type pteridinesIt light pale bothltd Iightoid pale bothIxd low xanthopterin dehydrogenase wild type pteridines4ma maroon brown pteridinesmah mahogany brown pteridines?ma! maroon-like brown pteridinesman mandarin orange ommochromes?Me Moire both patternme! melanized darker secondarymk murky darker secondarymot-28 mottled 28 mottled patternmot-K mottled of K mottled patternmot-321 mottled 321 mottled patternmot-36c mottled 36 mottled patternmsd(gl) modifier of sexual dimorphism of gi darker secondarymtb matt brown brown secondary?mud! mudlike brown both?mur murrey purple both?nrs narrow scoop darker secondaryocr ochracea orange ommochromesor orange orange bothosh outshifted brown secondaryp pink pink bothpd purpleold dark pink bothPdr purpleolder rosy-like both?Pec Pupilla ecentrica pattern patternpers persimmon orange ommoch romes?pn prune brown pteridinespo pale occelli brighter pattern?port port pale both?port-b port-b browny pteridines?pr purple browny pteridinesPu Punch pale pteridinespur purplish ruby pteridinespw pink wing pink secondary?pw-c pink wing c lighter secondary?pwn pawn brown secondarypym/ade2 polymorph browny pteridinesral raisin brown pteridines?ras raspberry browny pteridinesrb ruby paler both?rdb reddish-brown brown both?red red malpighian tubules wild type both?rl rolled darker patternrm rimy brown pteridines?rs rose pink pteridines?rud ruddle browny pteridines?rv raven dark secondaryrwi red wine browny pteridines?ly rosy browny both?Sa Salmon browny both?sb soft brown browny pteridines?5se sepia very dark pteridinesst safranin brown pteridines?sf-3 safranin-3 brown pteridines?she sherry brown pteridines?som sombre dull brown secondarySt scarlet orange ommochromesswy swarthy darker secondarysyn syndrome brown secondaryte tenerchaetae dark secondarytrl translucent purple patternIt tilt lighter secondaryU Upturned mottled patternups upright dull, rough patternv vermilion orange ommochromesyin yin browny pteridines?w white white bothWe Washed eye paler pattern?z zeste lighter both6viability of the organism. Yet few eye colour genes affect the viability of the fly.The clearest example is that of the white gene. Flies with no pigments due tonull mutations in this gene are completely viable and fertile. In fact, the firstmutation isolated in Drosophila melanogaster was a such a complete loss offunction mutation for the white gene (Morgan in 1910 cited in Lindsley andZimm 1992). That it was isolated from a wild population suggests no great lossof biological vitality.The sheer number of mutations which alter eye colour was seen as a problemby early authors (e.g. Nolte 1 952b) and has not yet been adequately resolved.Even a cursory inspection of the two pigment biosynthetic pathways (Figure 2and Figure 3) reveals that there are approximately 15 enzymatic stepsrequired to produce the pigments found in the wild type eye. Thus, evenallowing for co-factors, there remains a considerable excess of genes involvedin the production of the pigments of the wild type eye. In addition to enzymesand co-factors, it is reasonable to suppose that some of this apparent “excess”of eye colour genes are concerned with transport, sequestration and control ofdeposition of the eye colour pigments. The study of any eye colour gene mustultimately address the question of pleiotropy and redundancy of these genes.Both the number of genes and the dispensability of many of them suggest adiverse range of functions, many of which may be shared by other genes.Despite their historical importance, the disposition and biogenesis of thepigments in the eye, the chemical structure of the pigments, the biosyntheticpathways responsible for the pigments, the genes involved and the complexphysiological, developmental and tissue interactions required to produce wildtype eye colour remain obscure and still subject to debate. In addressing these7questions, the physical context of pigments, the pigment cells of the eye, andthe complex developmental regulation of pigment biosynthesis, as revealed bythe intersection of genetic and biochemical studies on the pigments, must bebriefly summarized.The eye.The eye is the most thoroughly studied of the four pigmented structures inDrosophila melanogaster (excluding melanized structures). The structural anddevelopmental complexity of the eye is underscored by studies which show thatapproximately two thirds of randomly selected lethals have defects in the eye (ofthose two thirds, one third is responsible for general cell viability functions, theother third is eye specific, Thaker and Kankel 1992). The development andstructure of the compound eye have been the focus of extensive investigation. Iwill briefly summarize this work as it relates to pigment deposition. Both thesignals leading to specification of cell fate in the development of the eye, andthe nervous connections between the eye and the brain are subjects of intenseresearch. These subjects have been thoroughly and often reviewed(Meyerowitz and Kanker 1978, Renfranz and Benzer 1989, Pak and Grabowski1980, Zipursky et al 1984, Tomlinson 1985, Venkatesh, Zipursky and Benzer1985, Ready 1989, Zipursky 1989, Campus-Ortega 1988, Ranganathang,Harris and Zuker 1991) and will not be treated thoroughly here.Development of the eye. The compound eye of Drosophila melanogaster, likethat of many insects, is composed of hundreds of reiterated units, the ommatidia(Figure 1). The eye analage arises from approximately 20 cells whichinvaginate from the embryonic ectoderm and which eventually form the eyeimaginal disc (reviewed by Venkatesh, Zipursky and Benzer 1985, Zipursky81989, Ranganathang, Harris and Zuker 1991). During embryogenesis, first andsecond instar stages, the eye imaginal disc cells proliferate but remainundifferentiated. During the proliferation stage, the eye disc is attached to thebrain by the optic stalk. Later elaboration of the nervous system results inprecise spatial correspondence between individual ommatidia and theirconnections in the brain. During the third instar a dramatic wave ofmorphogenetic activity sweeps over the disc. This morphogenic furrow isassociated with differentiation of the various cell types of the eye. The pigmentand cone cells are among the last to differentiate and are recruited from theundifferentiated epithelium by underlying photoreceptor cells. As with other celltypes in the eye, cell fate is not clonally determined but is determined byposition-dependent induction. Finally, during pupation the eye imaginal discevaginates to form the adult compound eye.Structure of the eye. The compound eye of the adult Drosophila is composed of700-800 ommatidia. The structure of three adjacent ommatidia is shown inFigure lB. The distal end of each ommatidium consists of the corneal lens andpseudo cone which functions to gather and focus light. Proximal to the diopticapparatus are the 8 photoreceptor cells with central rhabdomeres whichtransduce light to nervous impulses. Each rhabdomere has different wavelengthspecificity and connects to the brain via a complex network of neuralconnections. Surrounding each ommatidium is a sleeve of pigment cells(arranged as shown in Figure 1C). The pigment cells act to regulate lightexposure and to optically isolate the ommatidia. The pigment granules withinthe pigment cells are not static. In bright light they move towards therhabdomere thus reducing the amount of light reaching the rhabdomeres9Figure 1. Diagram of the eye, structure and organization of different cell typesin the ommatidia and organization of the primary and secondary pigment cellsof Drosophila melanogaster.A. Schematic diagram of the head of Drosophila melanogaster.B. Diagram of a longitudinal section through the eye of Drosophilamelanogaster showing three adjacent ommatidia. The various cell types whichcompose the ommatidia are shown. Redrawn from Nolte 1950, figure 12.C. Schematic diagram of the organization of primary and secondary pigmentcells and the ommatidial lens. Redrawn from Nolte 1950, figure 11.10The eye of Drosophila melanogasterA Co cell nucleus• pigment granuleoutline of ommatidia lens— primary pigment cellB e secondary pigment cellcellmembraneo p /0 4—post retinal cells—layer of monopolar cells.external optic glomerulusexternal chiasma11whereas in low light they migrate to the periphery of the ommadium increasingthe light exposure but decreasing visual acuity.Pigment cells There are two types of pigment cells, the primary and secondarypigment cells. The primary pigment cells lie more distally in the eye directlysurrounding the pseudo cone, while the secondary pigment cells lie moreproximally, principally surrounding the photoreceptor cells. These two types ofpigment cells cooperate to completely encase each individual ommatidium(although each pigment cell is shared by adjacent ommatidia). The pigmentcells have an abundance of pigment granules in their cytoplasm, however, theyare not the only cells with pigment granules. The photoreceptor cells also havepigment granules in their lateral cytoplasm, although fewer than the pigmentcells. The post retinal and basal cells may also have some pigment granules(Mainz 1938, Nolte 1950, Reaume, Knecht and Chovnick 1991), although thishas been the source of some dispute.Pigment granules. Early light and electron microscopy work defined differencesin pigment granule morphology (Shultz 1935, Nolte 1950, Shoup 1966). Thereare two types of (normal) pigment granules, named type one and type two. Bothtypes of granules are ribosome-sized, multi-subunit (Hearl, Dorsett andJacobson 1983, Hearl and Jacobson 1984) membrane-bound organelles whichoriginate in close proximity to the golgi apparatus and are likely derived from it(Shoup 1966), although alternative origins have been proposed (Reaume,Knecht and Chovnick 1991). Type one pigment granules first appearapproximately two days before eclosion, which corresponds to the first time thatthe ommochrome pigment can be detected (Schultz 1935, Nolte 1950, 1954a).They are electron dense, grow in size after eclosion and contain only12ommochromes. This type is found in primary pigment cells and most probably inthe photoreceptor cells. Type two granules are complex membrane bound,fenestrated structures found only in the secondary pigment cells. They are thesite of pteridine pigment deposition. Development of these granules and the firstdetectable appearance of the pteridine pigments coincide at approximately oneday before eclosion. These granules do not change size in development but dobecome denser after eclosion. Change in both the size and morphology of typeone and two pigment granules continues for a few days after eclosion of thepharate adult, concomitant with increase in the amount of ommochrome andpteridine pigments. Eye colour mutants often show a variety of complexchanges in the colour and morphology of these pigment granules. For example,histological analysis of the g3 allele, the only garnet allele for whichhistological information has been recorded (Nolte 1950), shows normalnumbers, distribution, development and morphology of pigment granules. Theonly change from wild type was alteration in the intensity of colour of both thetype one and type two granules. This change presumably corresponds todeficiency for both the pteridine and ommochrome pigments.The discovery that pigments were associated with proteins (Schultz 1935) led tothe realization that the pigment granules were largely proteinatious. Coordinateappearance of the pigment granules and pigments in development (Schultz1935, Nolte 1950, 1954a), in conjunction with the proteinatious nature of thepigment granules led to speculation that pigment granules are not just passivesites of pigment deposition but are complexes composed of the biosyntheticpigment enzymes as well as the pigments themselves (Phillips, Forrest andKulkarni, 1973). This hypothesis has been contested (Sullivan, Grillo and Kitos1974) based on the finding that the enzymes in the xanthommatin pathway are13found in different subcellular compartments whereas others are free in thecytosol. In addition, there is evidence that suggests that the pleiotropic effect ofdifferent eye colour mutants on the various enzymes, on which the enzymecomplex model was based, is a result of failure to differentiate enzymatic andnon-enzymatic conversion of pigment intermediates (Wiley and Forrest 1981).Nevertheless, more recent work suggests that pigment granules contain at leastsome of the enzymes involved in pigment biosynthesis (Dorsett, Yim andJacobson 1978, Hearl, Dorsett and Jacobson 1983, Hearl and Jacobson 1984).Biosynthesis of the eye colour pigments: The eye colour of the wild-typeDrosophila eye is derived from the deposition of two biochemically distinct typesof pigment, the ommochromes and the pteridines. Both types of compounds arewidely present in nature in both the plant and animal kingdom. Theommochromes are fairly simple compounds derived from the amino acidtryptophane. The pteridine compounds, however, are remarkably complex. Thepteridines were first isolated by Hopkins (1889) from the wings of an Englishbrimstone butterfly but chemical identification of the red pigment in Drosophilaas a pteridine was not made until 1940 (Lehderer 1940). The structure of someof the pteridines is still subject to dispute.While the fundamental steps of the biosynthetic pathway for their productionappears conserved from the prokaryote, Escheria coil through Drosophila tohumans, the details are still under intense investigation (Phillips and Forrest1976, Pfleiderer 1993). Figures 2 and 3 show the biosynthetic pathways forthe production of these two compounds. While the biosynthetic pathway for theommochrome pigment, xanthommatin, has been determined, based in largepart on Drosophila genetics, the pathway for the pteridine pigments remains14Figure 2. Biosynthetic pathway of xanthommatin production in Drosophilamelanogaster.The structure of the various intermediates in xanthommatin biosynthesis and theenzymes responsible for their production are shown. Adapted from Phillips andForrest 1976.15BIOSYNTHESIS OF XANTHOMMATINTRYPTOPHANN-FORMYL KYNURENINEKYNURENINE3-HYDROXYKYNURENINEXANTHOMMATINDIHYDROXANTHOMMATINNH2•CH2/\COOHvermilion +cinnabar ÷tryptophan pyrrolaseNH2formamidaseCOOHHCI4H2 phenoxazinone synthetase16Figure 3. A possible pathway for pteridine pigment biosynthesis in Drosophilamelanogaster.The structure of the various intermediates involved in pteridine biosynthesis andsome of the enzymes and genes responsible for their production are shown.Adapted from Brown et al. 1978, Brown 1989, Calatayud, Jacobson and Ferré1990 and Pfleiderer 1993.170NH2NJXj>rj,-LPguanosine triphosphate+OHOHHPHA’d-dLcL-PH2NI1JN)Iidihydroneopterin-P3A possible pathway for pteridine pigment biosynthesis i nDrosophila melanogasterGTP cyclohydrolasePu + cm, ma-I,sepiapterin synthase Aprsepiapterin synthase BH4 pterin*F-12N-N6-acetyl-homopterinaka-pyrimidodiazepineH2N1X)OHH2 pterin00+H’H2Ndihydrobiopterin0*HH HHA()HH2N’ &HIbiopterinisoxanthopterinxanthopterinHH HdrosopterinH4 biopterin18Figure 4. Three models proposed for pteridine biosynthesis in Drosophilamelanogaster.Figure 4 shows three pathways proposed for pteridine biosynthesis. Thestructures of the pigments and intermediates are as shown in Figure 3.A. The pathway proposed by Calatayud, Jacobson and Ferré 1989. This is thepathway shown in Figure 3.B. The pathway proposed by Brown 1989.C. The pathway proposed by Ferré et al. 1983.19Three pathways proposed for pterindie biosynthesis I nDrosophila melanogasterA. pathway proposed by Calatayud, Jacobson and Ferre, 1989.GTPH2 neoperinP3pyruvoyl -144 pterinIactoyl-H4-pterinH4-ptennH2-riny-H2Lmopterin I spiapterinpterinisoxanttpterin‘terinsHtbiopterindrosoH4-biopterin biopterinB. Pathway proposed by Brown, 1989GTPH2 hydroneopterinH2-Pterin PyruvoyI-t4PterinLactoyl-H4 PterinH4 Biopterinosopterin H2Bitterin SIsoxanthopterin epiapterinC. Pathway proposed by Ferre et al., 1983.GTP44 H2-Neopterin H2-6-(V, 2’-dioxopropyl)-pterin?H2-BiotterinPterinI soxanthopterin “Drosoprins” p”Biopterin H4-Biopterin20somewhat speculative. Figure 4 shows three recent models of the pteridinepigment biosynthetic pathway.The first major advance in understanding the formation of pigments was therealization that the wild type pigmentation was composed of two pigment types;a red-orange pigment and a yellow-brown pigment. In 1924 Johannsen notedthat there were two types of pigment granules, one red and the other yellow(Johannsen 1924). This observation was related to two biochemically distinctpigment groups by Casteel (1929), Schultz (1935) and Mon (1937). Studies ofcombinations of eye colour mutants provided complementary evidence of twoindependent genetic systems contributing to the wild type eye colour ofDrosophila. In 1931, Wright and his genetics class crossed a variety of eyecolour mutants. They observed that white-eyed flies arose from a cross betweenbrown and scarlet and between brown and vermilion. (Wright 1931). Based onthese observations of synthetic white mutants he proposed that the wild typepigment was a compound of two independent genetic pathways. These, he alsorelated to the two types of pigment granules reported by Johannsen. A similarobservation led Glass to the same conclusion (Glass 1934), as did thebreakdown of a white-eyed mutant into two component eye colours (Nolte 1943and 1944). Mainz (1938) generated extensive combinations of eye colourmutants and examined their effects on the histology of the eye. His resultsgenerally supported the idea of two independent pigment pathways. Nolte,however, could not reproduce these results (Nolte 1950, 1 952a, 1 959a) anddisputed this conclusion as well as some of the morphological observations.Although the pathways leading to the red and brown pigments may bebiochemically, and genetically distinct, both pigments are associated with the21pigment granules of the pigment cells of the eye and thus might be expected tointeract at least at a physiological level. This association may be manifest byaltered pigment granule morphology in mutants with simple enzymatic lesionswhich should alter only one of the two pathways (Nolte 1950, 1952, Reaume,Knecht and Chovnick, 1991). The independence, or lack thereof, of the eyepigmentation pathways is still a subject of debate (Schwinck, 1975 1978, Ferréetal. 1983).Finally, the biosynthesis of eye pigments should be discussed in context of thecomplex inter-relation of the different tissues and developmental stagesinvolved (reviewed by Tearle 1991). For example, production of the oneommochrome pigment, xanthommatin involves four organs and twodevelopmental stages. During the larval stages tryptophan is absorbed by thefat body and malpighian tubules and the excess converted into 3-hydroxykynurenine and kynurenine, respectively. During metamorphosis theseorgans release these stored compounds which, in addition to tryptophanderived from protein catabolism during metamorphosis, is taken up by the eyeand occelli. The interplay between the various organs involved in pteridinebiosynthesis and metabolism is likely equally complex, but is unfortunatelylargely unknown. Analysis of the biosynthesis of the pigments is furthercomplicated by the fact that not all of the pigmented organs have all thenecessary biosynthetic enzymes. For example, all four of the enzymesnecessary for production of xanthommatin are found in the eye pigment cellswhereas the occelli has only the last two and nether organ synthesizes xanthinedyhydrogenase, the product of the rosy locus, which does nevertheless affectthe pigmentation of these organs.22Much of the impetus for investigation into these compounds stems from theirrole in nucleic-acid metabolism. For example tetrahydrobiopterin is an essentialcofactor in amino-acid metabolism, neu rotransmitter biosynthesis andmolybdoenyzymes in general (Bel and Ferré 1986 and McLean, Boswell andO’Donnell 1990). Since these compounds are derived from GTP, alterations intheir metabolism is of relatively dire consequence in humans. In this regardDrosophila has been an immensely useful research tool. Not only have certaineye colour mutations (Henna and Punch) been proposed as models formetabolic defects such as phenylketonuria (probably more as a conceptualthan physiologically accurate model) but the pigments provide an abundant,readily isolated source of an otherwise scarce and highly labile material. Thesehave proved to be both difficult to isolate from mammals, as would be expectedfor a cofactor and metabolic regulator, and to synthesize chemically (Brown andFan 1975). The central role of these compounds in cellular metabolism mayprovide the biological impetus for sequestering them in the specializeorganelles, the pigment granules. This function also makes the complex tissueand developmental regulation less surprising.History of the garnet geneEye colour genes of Drosophila hold an prominent place in the history ofgenetic analysis. The first mutant isolated in Drosophila was an allele of thewhite gene (Morgan in 1910 cited in Lindsley and Zimm 1992). The first mutantallele of garnetwas recorded not long after by Bridges in his pioneering workon the chromosomal theory of inheritance (Bridges 1916). Since then it hasfeatured in innumerable genetic studies. As a result of its long history aliterature review of genetic studies involving the garnet gene is essentially23tantamount to a survey of the history of genetic analysis in Drosophilamelanogaster. Consequently the following survey of genetic analysis involvingthe garnet eye color gene will necessarily be somewhat cursory.The first mutant allele of garnetwas isolated on February 19, 1915 as aspontaneous mutation in a sable mutant stock (Bridges 1916). This mutation(g1)was established as an allele of a sex linked gene. As a sex linked marker itwas an invaluable tool in a number of studies. It was instrumental in showingthat the behavior of genetic elements paralleled exactly the movements ofchromosomes in both regular and irregular meiosis. Thus garnet played apivotal role in the demonstration of the chromosome theory of inheritance. In thesame study garnetwas used to demonstrate that recombination occurs at thefour strand stage, that non-homologous chromosomes (the X and Y) will pairregularly and, by defining the mechanism of non-disjunction, emphasized theregularities of meioses. Finally, interestingly in context of chapter 3, in this samestudy the garnet gene was used to demonstrate the general rule of equivalenceof maternal and paternal genomes.The eye colour genes next resurfaced in the literature in context of the profoundproblem of how the actions of genes can determine the final phenotype of anorganism. Investigators seized upon the eye colour mutations of Drosophila asan amenable model system in “higher” eukaryotes for the one-gene-oneenzyme theory. It was initially expected that the relationship between the manygenes which affected eye colour could be defined genetically whilst theirproducts were studied biochemically. These studies were not as immediatelyfruitful as was hoped. Genetic studies revealed only that there were two largelyindependent pathways (Schultz 1935, Mainz 1938) and biochemical analysis24was hindered by the complex, colourless, highly labile and light sensitive natureof the many intermediate products of these biosynthetic pathways. In addition,the sheer number of genes affecting eye colour posed tactical as well asconceptual problems (Nolte 1952b, Lucchesi 1968).Spurred by Sturtevant’s pioneering work on the cellular autonomy of geneaction in somatic mosaics (Sturtevant 1932), between 1935 and 1937 Ephrussiand Beadle produced a flurry of reports (Ephrussi and Beadle 1 935a, Ephrussiand Beadle 1 935b Ephrussi and Beadle 1 935c, Beadle and Ephrussi 1 935a,Beadle and Ephrussi 1935b, Beadle and Ephrussi 1936, Beadle and Ephrussi1937, Ephrussi and Beadle 1 937a, Ephrussi and Beadle 1 937b, Beadle 1937)which focused on the relationship between gene action and development. Theyattempted to dissect genetic and biochemical interactions between eye colourgenes using a series of tissue transplants between donors and hosts of differentgenotypes. Their results showed that most of the eye colour genes, includinggarnet, behaved autonomously upon transplantation. These studies, amongothers, were instrumental in nucleating concepts of diffusable versus cellautonomous substances, the former providing a working model for the action ofhormones, tissue specific interactions (Beadle 1937), fluctuating levels ofsubstances in development (Hamly and Ephrussi 1937), conservation ofbiological functions across species (Howlan, Glancy and Sonnenbilick 1937)and pleiotropy (Hadorn and Mitchell 1951, Hadorn 1962). Although important inemphasizing many concepts in development, a contemporary review of thesestudies stated “Because of the rather heavy injection of new theoreticalconsiderations and laboratory symbolism, the uninitiated person is to bewarned that he may find himself bewildered in trying to follow theseinvestigators to their final conclusions or to comprehend what their conclusions25actually are.” (Blakeslee 1938), suggesting that this profusion of studies may nothave been met with unabated enthusiasm.While these studies were vital to the development of our current understandingof genetics, the role of the gamet gene was somewhat incidental. It was simplya convenient marker for the X chromosome. The garnet gene also featuredprominently in a debate between Chovnick and Hexter on the complexity ofgenetic loci. In contrast to the previous studies, an intrinsic physical property ofthe garnet gene, its relatively large size, was the basis of its role in theimbroglio.This debate centered around the question of whether or not garnetwas acomplex locus. The debate arose from the concept that genes were, bydefinition, indivisible by recombination. Genetic complexity was definedfunctionally; if recombination occurred between various alleles, defined as suchby similar position, phenotype and failure to complement, the alleles weretermed pseudo-alleles (a term coined by McClintock 1944, for another purposeentirely). A gene with pseudo-alleles was then termed a complex gene and wasproposed by Lewis (1951) to consist of a series of duplicated genes. This pointwas raised as an issue in assessing the universality of genetic material. It hadbeen established that alleles of prokaryotic genes could recombine (Benzer1955). If this was also true of a “typical” eukaryote, Drosophila melanogaster, animportant property of genetic material would be conserved between prokaryotesand eukaryotes.The core of the conflict revolved around the inability to recover double mutantchromosomes from intragenic recombination (Hexter 1956, 1 958a, 1 958b and26Chovnick 1956, 1957, 1 958a, 1 958b and 1961, reviewed rather more soberlyby Carlson 1959). The situation was further complicated by gene conversionevents, a phenomenon which at that time, was not documented in Drosophilamelanogaster. Although impossible to determine exactly the source of thisdiscrepancy, it might be posited that the double mutant chromosome arose froma conversion event in which the g53dallele was converted to g+, as suggestedby Hexter (1958). The four other garnet alleles used in these studies were g1’g2, g3 and g4. Data presented in chapter 2 (Table 21) may explain why it wasonly theg53dallele that might have converted. The g3 and g4 alleles areinsertions and thus would be expected to convert at a much lower frequencythan point mutants (Chovnick 1964). The failure to recover recombinants orconvertants between g1 and g3 is compatible with molecular data whichindicate that these insertions are highly similar and possibly identical (Figure20 and 21). In retrospect, both intragenic recombination, as established inprokaryotes (Benzer 1955) and conversion, shown in Drosophila virills(Demerec 1928) probably occurred at the garnet locus. Although these issueswere finally resolved by his studies of another eye colour gene, the rosy gene(Chovnick 1964), the debate concerning the complexity of the garnet gene isnotable for being the smallest pseudo allelic locus studied and for furnishing thefirst evidence of gene conversion in Drosophila melanogaster.The “transport” group of mutants:The garnet gene alters the levels of both the pteridine and ommochrome groupsof pigments. The garnet gene is not alone in this property. A number ofinvestigators have suggested that the eye colour mutants which alter both typesof pigments are functionally related. Based primarily on this phenotypic27criterion, Nolte (1955) grouped the mutants carmine, carnation, claret, garnet,light, maroon, pink, prune, purple, purploid, rosy and ruby together as thearbitrarily named ruby group. On the basis of genetic and histological analysisof these genes he further subdivided the groups (and their proposed function)into the sub-groups shown in Table 2. The difficulty with grouping thesemutants based on fine gradations of eye colour is that the groupings become arather arbitrary reflection of the allele used. The eye colour of the garnet alleles,for example, ranges from a pale orange which Nolte would likely have classedas a member of the light group, to weak alleles that would be classed asmembers of the red or dark groups. The hazard of this approach is illustrated byresults reported for interactions between white and carnation. These gave verydifferent results depending on the alleles used. Inspection of the data presentedin Nolte’s 1955 paper allows a slightly different grouping, which, while alsodependent on phenotype, is based on an effect on one versus both pigmentsystems and histology (Table 2). This grouping should be less dependent onthe choice of allele. Encouragingly, I show extensive genetic interactions withinthe first group identified by this criteria.The temptation to categorize eye colour mutations lured others. Schwinck(1975) again classed eye colour mutations phenotypically. Her criterion wasbased on relative concentrations of the drosopterin pigments. Her lesscolourfully named grouping, the group two mutations, is shown in the third rowof Table 2. These mutations were unique in that they were the only mutants torespond to implants of phenylalanine by increasing drosopterin production. Thismight indicate a common metabolic defect. Tearle (1991) also proposed afunctional grouping based on phenotype, this one based which organs were28Table 2. Genes proposed as members of the transport group of eye colourgenes.The first and second column shows the name given to the group and thereference. The third column shows the names of the genes proposed asmembers of that particular group.The ruby group proposed by Nolte (1956) was latter subdivided in to smallergroups (indicated here by parenthesis) for which he proposed slightly differentfunctions. The first group contains, rosy, pinkand purploid and is defined bynormal pigment granule morphology and a brown eye colour. The secondgroup contains claret and maroon, has normal pigment granule morphologyand a light brown eye colour. The third group consists of carnation, carmine,garnet, ruby and light and is defined by normal pigment granule morphologyand decreased levels of both red and brown pigments. The fourth groupconsists solely of purple and is defined by normal pigment granule morphology,less red pigment and elevated levels of brown pigment. The final group consistssolely of prune and has abnormal pigment granule morphology.I have regrouped the genes carnation, carmine, claret, garnet, light, pink andpossibly maroon and purploid as one subgroup defined as having normalpigment granule morphology and less of both classes of pigments. A secondsubgroup consists of rosy and possibly maroon and purploid with normalpigment granule morphology, less red pigments and normal levels of brownpigments. The last two groups are identical to Nolte’s last subgroups.29The group II mutants proposed by Schwink (1975) is defined by their increaseddrosopterin production in response to implants of phenylalanine. The transportgroup proposed by Sullivan and Sullivan (1975) consists of white, scarlet, claretand lightoid for which defects in keynurine transport were demonstrated, andcarnation, garnet, light, maroon and pink for which transport defects wereproposed. The group defined by their unusual response to transplantation ofeye discs consists of carnation, carmine, claret, garnet, pink, ruby, maroon-likeand rosy. The grouping proposed by Tearle (1991) is based on the organsaffected by mutations in these genes. Those he listed as group 1A affectpigmentation of all organs examined.Finally the group defined as participating in synthetic lethal interactionsincludes: purple-Purploider, prune-Killer of prune, Hennar3rosy6ry-SOD,light-carnation, deep orange-carnation, deep orange-rosy, deep orangepurploid, deep orange-cinnabar-brown, deep orange-fused. Some of thesecombinations may reflect special situations. prune is lethal when combined withKiller of prune (Sturtevant 1956). The latter gene is a dominant allele of the awd(abnormal wing disc) gene and is a nucleotide diphosphate kinase. The prunegene product has been identified as a ras GTPase activating protein. Thedominant Kpn allele may cause excessive stimulation of ras-like proteins (Teng,Engele and Venkatesh 1991). It would not be surprising if such an interactionwas lethal. The combination Hnr3 and iy6 has also been touted as a syntheticlethal combination (Taira 1960), however this interaction is specific to the ry6allele; it does not occur with other rosy alleles (Goldberg, Schalet and Chovnick1962). The gene fused disrupts pteridine metabolism but is not normallyconsidered an eye colour gene. The synthetic lethality of deep-orange with the30cinnabar brown double mutant chromosome is specific to this chromosome andis not seen with any combination of double mutants.31Membershipinthe“transport”groupofeyecolourmutations.namereferencemembersofgroupTherubygroupNolte(1956,1959)(rosy,pink,purploid)(claret,maroon)(carnation,carmine,garnet,ruby,light)(purple)(prune)Therubygroupthiswork(carnation,carmine,claret,garnet,light,pink,ruby,maroon?purploid?)revisited(rosy,maroon?,purploid?)(purple)(prune)GroupIISchwink(1975)carnation,garnet,maroon-like,orange,pink,rosy,rubyThetransportSullivanand(white,scarlet,claretlightoid)(carnation,garnet,light,maroon,pink)groupSullivan(1975)AberrantdiscBeadleandcarnation,carmine,claret,garnet,pink,ruby,maroon-like,rosytransplantationEphrussi(1936)32Schwink(1975)epistaticGreen(1955)whitea3,rubyandgarnetinteractionsReedyandbright,brown,clot,lightoid,cardinal,claret,mahogany,rosy,scarlet,sepiaCavalier(1971)SyntheticBridges(1922,purple-PurploiderlethalscitedinBridgesandBreme1944)Lucchesi(1968)deeporange-carnation,deeporange-rosy,deeporange-fuseddeeporange-purploid,deeporange-cinnibar,brownTiara(1960),Hennar3rosy6Goldberg,SchaletandChovnick(1962)Sturtevant(1956)prune-KillerofpruneNash(1971)light-carnationNickla,etal(1980)Hillikeretal.(1992)rosy-SODPigmentationofTearle(1991)claret,carmine,deeporange,garnet,light,lightoid,orange,pink,ruby,allorgansscarletwhiteaffected by mutations. His group 1A affects pigmentation of all organsexamined.At approximately the same time, Sullivan and Sullivan (1975) documenteddefects in metabolite transport in white and scarlet mutants. They alsodiscovered similar defects in claret and lightoid mutants. Although not examinedin this study, they proposed that the genes carnation, garnet light, maroon andpink were, in addition to claret and light, involved in metabolite transport.Other suggestions of functional similarly between these genes have arisen froma number of studies. Beadle and Ephrussi (1936) noted that while most mutantimaginal eye disc transplants were cell autonomous, transplants betweencarmine, carnation, claret, garnet, pink and ruby and either vermilion orcinnabar (the latter non-cell autonomous due to diffusion of the “hormone”- themetabolic ommochrome intermediate keynurenine) produced exceptionalresults. Eye discs from vermilion mutants transplanted into hosts mutant for mosteye colour genes, become wild type. Presumably, the mutant discs are able toscavenge enough kynurenine to generate wild type levels of pigment. Differentresults were seen with members of the “transport” group. Discs from vermiliondonors transplanted into carnation and garnet hosts developed an intermediatephenotype, whereas when transplanted into carmine, claret, pink and rubyhosts the vermilion discs develop a mutant phenotype. Minor differencesbetween mutant and intermediate phenotypes were likely due to thehypomorphic nature of the mutant used. For example the g2 allele used in theseexperiments is a moderate hypomorphic allele. These results suggest that thetransport group of mutations are physiologically distinct from other eye colourmutants. The authors implied that this difference could be due to differences in33levels of the “hormone”, although this is not consistent with their data. Schwink(1975) suggest that similar intermediate phenotypes seen in wild type eye discsimplanted into maroon-like or rosy hosts is due to negative feedback of thebiosynthetic pathways by excess products. However, the transport defectidentified by Sullivan and Sullivan (1975) seems a more appealing cause forthese aberrant results. Nevertheless, these results suggest a physiologicalequivalence between many of the members of the transport group.Finally, there have been a few other hints of functional similarity betweenvarious members of the transport group. Green (1955) found that some allelesof garnet and ruby were unique in showing no additive interactions incombination with certain white alleles. Certain members of these groups alsoshow highly specific synthetic dominant (Nolte 1 952a) and synthetic lethalinteraction (Luccessi 1968, Nash 1971, Nickla 1977).I shall use the term “transport group” genes to refer to members of this group asthey show extensive phenotypic similarity and share a number of physiologicalproperties, of which a role in transport is the best defined. A direct role intransport has, however, been shown for only one of these mutants, and this maynot be their primary role. Thus this term is adopted more for its descriptive valuethan as a claim of function.Summary:The study of eye colour genes is now less fashionable. But, many of theproblems noted by the early investigators still remain and the biological function34Table 3. Eye colour genes with known or proposed functions.The first two columns indicate the designation and name of those genes forwhich functions are known or proposed. A “+“ in the third column indicates thatthe gene has been cloned. The function proposed for the gene and thereference is given in the fourth and fifth columns, respectively. For those geneswhich encode an enzymatic function, the name of the enzyme is given in part Bon the next page.35A.EyecolourgeneswithknownfunctionsgeneclonedfunctionreferencebrownclaretcarnationcardinalcinnamonclotcinnabarDropgarnetglasskarmoisinkillerofprunelittleisoxanthopterinlightlightoidlowxanthinedehydrog.maroonlikepinkprunepurplePunchpolymorphrasberryrolledlyrosysesepiaStscarletvvermillionwwhitezzeste+transmembranechannel+transport?/nervoussysterm?-transport?/nervoussystem?-enzyme-ommochromepathway+molybdenumcontainingcofactor-enzyme-pteridinepathway+enzyme-ommochromepathway+cell-cellsignaling+transport?/nervoussystem?+transcriptionfactor+enzyme-ommochromepathway+regulationofGTPhydroxylase+enzyme-purinesynthesis+transport?/nervoussystem?-transport?/nervoussystem?+molybdenumcontainingenzymecofactor+molybdenumcontainingcofactor+transport?/nervoussystem?+regulationofGTPhydroxylase(Pu)+enyme-ommochromepathway+enyme-pteridinepathway+enzyme-purinepathway+enzyme-pteridinesynthesis+cell-cellsignaling-rhabdomereformation+enzyme-enzyme-pteridinepathway+transmembranechannel+enyme-ommochromepathway+transmembranechannel+transcriptionfactorDreesenJohnsonandHenikoff1988Yamamotaetal.1988SullivanandSullivanl975Sullivan,GrillandKitos,1974Kamdor,SheltonandFinnerty,1994Wiederrecht,PatonandBrown1984GhoshandForrest1967/HowellsinTearle1991RenfranzandBenzer,1989thisthesisMoses,EllisandRubin,1989Sullivan,GrilleandKitos,1974Teng,EngeleandVenkatesh,1991Ordono,SilvaandFerre,1988SullivanandSullivanl975/Devlinetal.1989McCarthyandNicklal980Schott,BaldwinandFinnety,1986Warner,WattsandFinnerty,1986Jones&Rowls1988Teng,EngeleandVenkatesh,1991Searle&Voelker1986MackayandO’Donnell,1983Henikoffetal.1986Nash-personalcommunicationZipurskyetal1993Keithetal1987WiederrechtandBrown,1984Tearleetal.1989Nissani1975/SearlesandVoelker1986Dressenetal.1988Pirrotta1988bw cacarcd cm cICnDr g gl karkpnlix It ltdlxdmal p pnpr Pupymrasri36B.genenzymese cidihydroneopterintriphosphatepyrimidodizepinecnkynurenine3-hydroxylasekarphenoxazeninesynthetasekpnnucleosidediphosphatesynthetaselixdihydropterinoxidaseprsepiapterinsynthasePuguanosinetriphospatecyclohydrolasepymformyiglycineamideribotideaminotransferaserasinosinemonophosphatedehydrogenaseryxanthinedehydrogeneasese6-pyrovoyltetrahydropterin2-amino4-oxo-b-acetyl-7,8dihydro-3H,9Hpyrimido[4,5,6]-[1,4]diazepinesyntetase37vtryptophanepyrrolaseof most remain unknown. Those eye colour genes which have been analyzed indetail have provided insight into a diversity of biological functions (Table 3).Not only does this diversity suggest a potential resolution of the historicalparadox of the apparently excessive number of eye colour genes, but it alsosuggests that the classical system of Drosophila eye pigmentation is a powerfulassay for a diverse array of biological functions. The sensitivity of this system issuch that it has the potential to allow the genetic detection even of those genessuch as the transport group which may have extensive functional redundancy.The crucial, if still undefined, functions of these genes can only be revealedthoughdetailed analysis of the genes involved.This thesis examines the genetic and molecular properties of one of the eyecolour genes, the garnet gene. Chapter one deals with the complex geneticinteractions between eye colour genes, specifically those of the transport group,which are likely involved in some aspect of inter or intracellular transport ofmetabolites. Membership of garnet gene in this group is indicated by extensivegenetic interactions between the garnet gene and other members of this group.Molecular analysis of the garnet gene is presented in chapter two. In the finalchapter the garnet gene is used as a genetic and molecular marker to examinethe question of epigenetic gene regulation. Historically, the garnet gene hasbeen an invaluable tool for those studying a variety of biological and geneticprocesses. Further analysis of the garnet gene and other members of thetransport group should lead to insights into the ubiquitous and essential role ofcellular communication in development.38MATERIALS AND METHODSAll crosses were performed at 220 C unless otherwise stated. Culture mediumwas standard cornmeal/sucrose media supplemented with antibiotics and0.04% tegosept as a mold inhibitor. Crosses were generally carried out in 8dram shell vials with groups of 3-5 virgin females crossed to an equal number ofmales. For each experiment 3-6 replicates were made. Crosses weresubcultured twice at 4-5 day intervals before the parents were discarded. Eachset of crosses was scored independently. The data from replicate crosses withina group were subsequently pooled.Mutant strains and chromosomes: The mutations and rearrangedvariegating chromosomes used in this study are described in Lindsey andZimm (1992). The g62 and gS3 alleles were generously provided by Dr. ASchalet via Dr. D. Sinclair. The Dp(1;f)LJ9 mini-chromosome is carriedbalanced against an attached XX and attached xy chromosome. As the mini-chromosome carries a region which is diplo-lethal in males the attached XYstock carries the deficiency g-I on the X chromosome. The mini-chromosomewas generously proved by Dr. G. Waring via Dr. D. Sinclair.Crosses:Details of the crosses will be described in the appropriate figure and tablelegends.Assays to quantitate eye pigment.Red pigments, five heads:39The amount of pigment deposited in the eye was measured separately for 25females and 25 males. Flies three to seven days post-eclosion weredecapitated by vigorously banging the frozen flies. Flies were held in the dark at700C for 4 hours-six months before decapitation. No appreciable loss ofpigment occurred during this time (data not shown). The heads were placed inwells of a microtitre plate and 30 uL of 0.25M 13-mercaptoethanol in 1%aqueous NH4OH per well was added. The eye pigment was released bysonication for three seconds and a five uL aliquot was removed from each welland applied to a piece of Whattman No. 3 filter paper. The amount of pigment inthe dried spot was determined fluorometrically using a MPS-1 Zeissmicroscope. The software program for processing the data from thephotomultiplier attached to the microscope was written by Dr. J. Berger. Fivegroups with five heads per group were measured for each genotype and sex. Ineach case, the amount of pigment in each of the five spots was averaged andthen expressed relative to wild type.Red pigments, single heads:The amount of pigment in single fly heads was quantified as above except asfollows: only one head was used per microtitre well; 2OuL of solution wasremoved after sonication and spotted onto filter paper; a minimum of 10individuals per genotype were measured.Red pigments, spectrophotometric assay:The spectrophotometric assay was adapted from the procedure of Real, Ferréand Mensua (1985). Five samples of five heads each, per genotype and sex,were placed in an eppendorf tube with 150 uL of 30% ethanol, acidified to pH 2and shaken on an orbital shaker for 24 ± 4 hours. The absorbance of the40pigment was then read at 480 nm.Brown pigments, spectrophotometric assay:This assay was also adapted from the procedure of Real, Ferré and Mensua(1985) and Euphrussi and Harold 1944. Three samples of twenty heads pergenotype and sex where placed in an eppendorf tube with 450 uL 2M HCI and0.66% sodium metabisulfite (wlv) and sonicated. 0.9 mL of n-butanol(equilibrated with H20 and 0.66% sodium metabisulfite) was then added andthe extract was shaken for 30 minutes on an orbital shaker. After centrifugation(5 mm. at full speed in a microcentrifuge) the aqueous layer was removed and500 uL of H20 and 0.66% sodium metabisulfite was added. The mixture wasagain shaken for 30 minutes, centrifuged and the aqueous layer removed. Thiswash was repeated twice. The absorbance of the organic layer was measuredat 492 nm.Quantification of separated red pigments by thin layer chromatography:The red pigments were separated and then quantified by a modification of themethod described in Hadorn and Mitchell (1951) and Hadorn (1962). Redpigments were extracted as described for the five head red pigment assayabove. An aliquot of 5 uL was spotted one inch from the bottom of a cellulosechromatography plate (Kodak) and allowed to dry. Further aliquots of the samesample were placed on the existing spot and allowed to dry, for a total of 6 x 5uL of solution. The pigments were then separated in a solution of n-butanol and1% ammonium hydroxide (2:1) for 3-4 hours. The pigments on the dried plateswere identified by their distinctive colour under UV light, position and bycomparison to the results of Hadorn and Mitchell (1951) and Hadorn (1962).The amount of pigment in each spot was quantitated by scanning the41chromatography plates with the fluorescent microscope (wavelength 500 nm).Using this procedure it was possible to identify 9 pigment spots.Assessment of malpighian tubule colour.Malpighian tubule colour was assessed essentially as described by Brehmeand Demerec (1942). Malpighian tubules where dissected from healthywandering third instar larvae in 0.7% NaCl. The isolated malpighian tubuleswere immediately viewed against a black background. Their colour wasassessed in comparison with wild type, carnation, carmine and white mutants.Assessment of testes sheath colour.Testes were dissected in 0.7% NaCl from freshly killed males and immediatelyexamined for colour when placed against a dark blue background. As withmalpighian tubule colour, the colour of the testes sheath fades rapidly in 0.7%NaCI, tap water or in frozen whole flies.Molecular analyses:Genomic DNA was extracted by the method of Jowett (1986) with minormodification. Following RNAse digestion the DNA was isolated by spooling fromthe ethanol-aqueous interface and dissolved in TE. Restriction endonucleasedigestion of the DNA was performed overnight, twice, with interveningphenol/chloroform extraction and ethanol precipitation.Isolation of RNA: RNA was isolated from the appropriate stages and genotypesusing either the “hot phenol” method of Jowett (1986) or the guanidine42thiocyanate method (Fluka). The latter technique was more effective.Isolation of lambda clones: lambda clones and DNA were isolated according tostandard procedures (Sambrook, Fritsch and Maniatis 1989).Isolation of plasmid DNA: Small scale plasmid DNA preparations were isolatedby the alkaline lysis method (Sambrook, Fritsch and Maniatis 1989) or using“magic mini-preps” (P romega) following manufacturer’s instructions. Largescale plasmid preparations were prepared by PEG precipitation (Sambrook,Fritsch and Maniatis 1989) or using “Qiagen” columns following themanufacturer’s instructions.Southern and Northern transfers and hybridization: Southern and Northerntransfers and hybridization were performed according to standard procedures(Sambrook, Fritsch and Maniatis 1989) and following manufacturer’s(Amersham) instructions.Labeling of DNA: DNA was radioactively labeled using 32P and either the nicktranslation kit (Amersham)or the random priming kit (Amersham) following themanufacturer’s instructions. For some experiments DNA was labeled with UTPflourescein using the ECL (enhancered chemioluminescence) random prime kit(Amersham) and detected using anti-flourescein antibody conjugated to HRPwhich catalyzed a light-emitting reaction, following the manufacturer’sinstructions.Sequencing: Short overlapping segments of DNA for sequence analysis weregenerated using the Exo Ill directed deletion kit following manufacturer’s43instructions. In some cases primers to previously determined sequences wereprepared by the UBC oligosynthesis laboratory and used for sequencing.Sequencing was done using the double stranded dideoxy chain terminationmethod using 35 labeled DNA. Initially the Sequenase (United StatesBiochemicals) kit was used. Subsequently 17 polymerase (Pharmacia) wassubstituted for the Sequenase enzyme and all reagents were made followingrecipes supplied with the T7 enzyme. The DNA fragments for sequencedetermination were separated on a 6% or 8% Acrylamide:bis acrylamide (40:1)gel or on a 6% “long ranger” (N-methly-acrylamide) (United StatesBiochemicals) gel in 0.5% TBE buffer. The genebank accession number for thegarnet c-DNA is U31351.Sequence similarity search: The search of sequence data bases was carriedout using the blastN and blastX algorithms provided by the national center ofbiotechnology information. This search included all eukaryotic sequencespresent in the Swiss protein, EMBL and genebank data banks.In situ hybridization: Hybridization to RNA in whole mount embryos and varioustissues was done essentially as described by Tautz and Pfeifle (1989) with theexception that hybridization was carried out at 55°C. DIG labeled RNA probeswere made following the manufacturer’s instructions (Boeringer Mannheim) anddetected with alkaline phosphatase-conjugated anti-DIG antibody and NBT/Xphosphate colour reaction.Photography: Flies and fly tissues were photographed using the Wildstereomicroscope adapter at 16 or 40 X magnification on a dark bluebackground. Kodak Eckatchrome Ti 60 colour slide film was used.44Chapter 1Interactions between garnet andother eye colour genes45Introduction-chapter 1. Interactions between garnet and other eye colour genes:The large number of genes which affect eye colour suggests that eyepigmentation is a powerful assay system, able to detect alterations in a widevariety of biological functions (Table 3). Study of these genes, while having along history is still in its early stages and further work on this group is certain tobe rewarding.Of the eye colour genes, those of the transport group are particularly intriguing.Membership in this group has been assigned somewhat haphazardly, based onan arbitrary assessment of phenotype and some suggestive histological andphysiological experiments. The garnet gene has been implicated as a memberof the transport group but detailed analysis of either the gene itself or ofinteractions between garnet and other members of this group is lacking. Thechoice of garnet as a representative member of the transport group, while inpart arbitrary, was promoted by the extensive previous work on this gene andthe existence of a second site mutant which enhanced the mutant eye colourphenotype of severe garnet alleles. This interaction provides an attractiveavenue for genetic analysis of the garnet gene.This chapter presents data on the nature of this second site mutation, theenhancer of garnet mutation (e(g)), its interaction with garnet and with othermembers of the transport group of eye colour mutants, and finally, interactionsbetween this latter group and garnet. This analysis supports the inclusion ofgarnet as a member of the transport group of eye colour genes and suggests asimple model for the biological role of this class of genes. The molecularanalysis presented in chapter two lays the ground work for testing of this model.46Results-Interactions between garnet and other eye colour genes.The enhancer of garnet mutation: Rather fortuitously, the P-element bearingstrain (S6-1) that generated the garnet allele, gP, which permitted the cloning ofthe garnet gene, also contained a second site mutation which made thephenotypes of severe garnet alleles appear more extreme. This mutation wascalled enhancer of garnet (e(g)). The enhancer of garnet mutation has noindependent phenotype. Individuals bearing only the enhancer of garnetmutation have no alteration in eye colour, viability or fertility (data not shown).However, in combination with severe alleles of the garnet gene the enhancer ofgarnet mutation reduces the amount of eye pigmentation by approximately half(Figure 5a, b and c and Table 16). This sex linked gene mapped, byrecombination, near the site of a previously isolated enhancer of garnetmutation (Payne and Denny, 1921) and is presumably allelic to this mutation.Direct test by complementation is not possible as the original enhancer ofgarnet mutation is no longer extant. Other than its phenotype in conjunction withgarnet, little information is available on this mutation.Cytological position of enhancer of garnet: As a first step in identifying thenature of the enhancer of garnet lesion, the cytological position of the enhancerof garnet mutation was determined. The enhancer of garnet mutation had beenmapped by recombination to approximately map position 4 (Wennberg 1988).This position corresponds roughly to division 3 on the cytological map. Threedeficiencies and one duplication encompassing most of division 3 wereobtained from the stock center (Bowling Green) and tested for complementationwith the enhancer of garnet mutation. None of the heterozygotes between thesedeficiencies and the double mutant e(g) g2 chromosome showed an eye colour47Figure 5. Phenotypes of three severe garnet alleles and these garnet allelesin conjunction with the enhancer of garnet mutation.The garnet alleles g53d,g5Oe and g2 are shown on the left of each panel fromtop to bottom, respectively. On the right are shown these same alleles inconjunction with the second site enhancer of garnet mutation.48. e(g) g53de(g) g5Oee(g) g249phenotype (data not shown), consistent with the absence of any detectablephenotype for enhancer of garnet homozygote. In contrast, when the g2 allelewas recombined onto these deficiencies and retested against a e(g) g2chromosome, all the deficiencies showed a strong enhancer of garnetphenotype. The deficiency/enhancer of garnet (Df e(g)g2/e(g) g2) phenotype ismore severe than the homozygote enhancer of garnet (e(g)g2/e(g) g2)phenotype indicating that the original enhancer of garnet mutation ishypomorphic. The presence of a duplication for this region rescued theenhancer of garnet mutation on the e(g) g2 chromosome reducing it to a g2phenotype (Figure 6 and Table 4). These results imply that the enhancer ofgarnet lesion lies within the area of overlap of these four rearrangements. Theonly area held in common by these deficiencies and duplication is region 3C3,although given the limits of cytological resolution the area of overlap mightextend to bands on either side as 3C2 and 3C3 are often difficult to distinguish.These data effectively place enhancer of garnet in cytological position 3C2 -3C4.The division 3 region has been extensively analyzed at the genetic andcytological level (Lindsley and Zimm 1992). No mutation corresponding toenhancer of garnet has been reported for this region but this is not unexpectedas the mutation has only an indirect phenotype. The position of enhancer ofgarnet is, however, extremely close to that of the white gene which is located atposition 3C2. Furthermore, there is complete congruence between removal ofthe white gene and the occurrence of the enhancer of garnet phenotype orrescue of these phenotypes (Figure 6B). Initial tests of allelism betweenenhancer of garnet and white indicated complementation; the phenotype of wg+/e(g) g2 flies is completely wild type. Nevertheless the proximity of the white50Figure 6. Cytological localization of the enhancer of garnet mutation.A. A diagram of the first part of polytene chromosome division 3 is shown(adapted from Lindsley and Zimm 1992). Above is indicated the position of thezeste and white genes and the location determined for the enhancer of garnetlesion. Below is shown the cytological limits of three deficiencies; Df(1)DF(1)N8, Df(1) wrJl and Df(1)wrJ3, and one duplication; Dp(1;3)N238 used tomap the enhancer of garnet mutation. The black bars indicate deficiencies, thestriped bar the duplication and the clear bars areas of uncertainty in thecytological determination of the extent of the deficiencies.B. The phenotype of the deficiency with respect to white and enhancer orgarnet is summarized. A “+“ indicates presence of wild type function for thatgene, a “-“ indicates absence of that gene function. See Table 4 for data andcrosses.51Cytological postion of e(g)A.white 3C2zeste e ) 3C2-319\T1Ths1E_• Ir1’IiiI . I — IIii II II_____I I4 3A b411 ‘4 •3D 0I II I11I’, Df(1)N81I II I Df(1) 3C2-3 - ‘3E3-4ni I ii II Df(i)w tI -h IDf(1) 3A1-2 - 3C2-3 I rJ3I Df(i)w IB. Complementation of white_______________and enhancer of garnet I DfcI) 3C3-3C1 2w e(g) IDf(1)N8 - I Dp(i;3)N238Df(1) rJ1rJ3 Dp(1;3) 3B2-3;3D6-7;8ODFDf(1) wDp(1,3)38 + +52Table 4. The Effect of 3C deficiencies on white and enhancer of garnet.The first column lists the three 3C deficiencies used to cytologically localize thee(g) lesion. The second and third column list the visual phenotype and redpigment values (as percent wild type) for the appropriate deficiency over a whitenull (w1). The fourlh and fifth columns list the phenotype and red pigmentvalues (as percent wild type) of the appropriate deficiency combined with the g2allele heterozygous with the e(g) g2 chromosome. The last row lists theequivalent values derived using the e(g) g2 chromosome for comparison.CROSSES:1. To generate the deficiency garnet chromosomes,P Df(1)*/In(1)dI49 ®e(g)g2/Yo’Fl Df(1)*/e(g) g2 0 In(1)d149/Yo’(sibs)F2 Df(1)*g/!n(1)d149, g4selected by garnet phenotype.pair matings with e(g)g2/YcJ sibs to isolate deficiency bearingchromosome (no e(g)g2cf progeny).2. To determine phenotypes:A. With respect to the white gene:P Df(1)*®w/YQFl Df(1)*/w4phenotype determined.B. With respect to enhancer of garnet phenotype.P Df(1)*g2/In(1)d149 0 e(g)g2/YcJ1Fl Df(1) *g2/e(g) g2phenotype determined.53Effect of 3C deficiencies on white and e(g)Deficiency Df/white Lg2/e(a) g2pigment phenotype pigment phenotypeDf(1)N8 0 completely white 11 + 1 pale yellowDf(1)WrJ1 0 completely white ND yellowDf(1)WrJ3 0 completely white ND pale yellowe(g) g2 88±9 wild type 21 + 3 orange54gene to the enhancer of garnet mutation led me to test for allelism in a lessdirect manner.enhancer of garnet is a subliminal allele of white. Two types of tests of allelismbetween white and enhancer of garnet were made. In the first series of tests,the enhanceable g2,g5Oe and the g53dmutations were recombined ontochromosomes bearing 11 different alleles of the white gene. The white allelesrepresent a variety of different types of lesions in the white gene (Figure 7).These recombinant chromosomes were then tested in trans with the appropriatee(g) g * mutation to see if the white alleles would complement the enhancer ofgarnet allele. Table 5 presents measurements of the levels of red eyepigments of these genotypes. In general, all the white alleles acted as extremeenhancers of the three different enhanceable garnet mutations. The oneexception was wsat in combination with theg5Oe allele. These individuals werevisually, and by pigment assay, indistinguishable fromg5Oe homozygotes. Thisresult was repeatable (data not shown). This suggests some allele specificity inthe interactions between white and garnet, which is borne out by the lack ofobvious correlation between the severity of the white allele and its effect ongarnet (Figure 8 and see below). Specifically, although the Sat allele is amore severe allele of white than the e(g) allele (column 3 of Table 5) it hasvirtually no effect on garnet expression. Also interesting, is the result with wa.The wa allele is a moderate hypomorphic allele of white, yet it has a very strongenhancer of garnet effect on all three garnet alleles. This effect is morepronounced than that of the white null alleles (w1 and w11 18).As a second type of test of allelism between white and enhancer of garnet thewhite+ transgene was tested for its ability to rescue the enhancer of garnet55Figure 7. Diagram of the structure of the white gene of Drosophilamelanogaster.The heavy line represents the molecular map of the white gene (from Levis andBingham 1985). The size, position and direction of transcription of this gene isshown above the line. The position and identity of inserts in the various whitealleles used to complement the enhancer of garnet mutation are shown aboveand below the line (adapted from Lindsley and Zimm 1992). The molecularlesions associated with sat and w are unknown. The two dashed lines belowthe heavy line represent the white DNA included in the P-element constructsP[(w,ry)]A4-3 (from Hazelrigg, Levis and Ruben 1984) and CaSpeR (fromPirrota et al 1985).56Structure of thewhite genew+ transcript3’ 44_A_A__A__ ____ 5’h chw+10P[(w,ry)]A4-3Sst___WIJ(R=EcoRiH = Hind IIIK = Kpn IS = SailCaSpeRw-5will8()RH K R H E57Table 5. Complementation between e(g) and different white alleles.The left half of the table shows the 11 alleles of white tested forcomplementation against e(g), their visible phenotype and total red pigmentlevels (expressed as a percent of wild type pigment levels). The right side of thetable gives pigment levels for the appropriate combinations of the genotype wg*/e(g) g*, where * indicates the given allele.The last two rows list the results obtained with the e(g) allele and wild typeallele for white for comparison.As an incidental note, these results show that when the amount of pigment inthe w g+ / e(g) g2 strain is assayed the conventionally recessive white allelesare slightly dominant. In contrast the garnet mutations are truly recessive. Thesedata agree with previously reported pigment levels of the white and garnetgenes. Nolte (1 959c) in a theoretical treatment of the significance of dominance,found w/w had slightly less pigment than wild type (w+/wj. In the same study,Nolte found garnet to be “truly” recessive. This has caused some difficulties inthe interpretation and significance of dominance of alleles but these problemspose conceptual rather than functional problems.CROSS:1. To generate recombinant w g* chromosomes:g*/g*® w*/YQFl + g*/ w’ + ® g*/y’f (sibs)1F2 w*g*/Y058(select by phenotype of test cross to g*/g* and w*/w* , balanced to makestock)2. To perform the complementation cross:P e(g)g*/e(g)g* ® w*g*/YoFl e(g) g*/w* g* assayed for pigment levels.Note: This complementation test was performed three times, once by BarneyLee as part of his directed studies course. The values of the individual testshave been averaged as they did not differ greatly.59white alleles5Oe 53d_____w* g*/e(a) g*11±1 6±1 2±113±1 ND ND5±1 5±1 2±110±1 7±1 3±19±1 3±1 3±112±2 8±1 ND12±1 8±1 4±111±1 7±1 ND9±1 3±1 ND10±1 22±1 4±16±1 8±1 NDe(g)g* 32±3 14±1 11±1w÷g* 38±2 26±1 13±2Complementation between e(g) and differentwhite allele g2w*/w*w1118< completely whitew1 <1 completely whitewa 3± .5 dull orangewbf 2±.5 faint yellowwBXW.7±.1 dull redwch >5 ± .1 pink-orangew’ ND dull redwe 4±.5 pinkyw1 .6±.1 faint pinksat 7± 1 browny orangew .6±.1 faint pink60effect on garnet. Two constructs were chosen to provide the white+ transgene.The first transgene construct contained the full white + gene coding region andapproximately 3kb 5’ and 3’ white regulatory sequence (along with rosy+)inserted within a P-element located at position 1 OOF on 3R (Hazelrigg, Levisand Ruben 1984). This particular construct shows an unusual pattern of white +gene expression. The anterior portion of the eye is normally pigmented (red)while a crescent at the posterior margin, approximately one quarter of the eye,remains unpigmented. This construct showed partial rescue of the enhancer ofgarnet effect (Table 6). The pattern of rescue was similar to the pattern ofwhite+ expression. The posterior portion of the eye in which white+ was notexpressed appeared more lightly coloured (enhanced) than the anterior. Thesecond white+ construct used was the CaSpeR P-element transformationvector. This construct contains all the coding region of the white gene but has adeletion of most of the large first intron and all but 300 bp of the 5’ regulatoryregion and 630 bp of the 3’ region (Pirrotta, Stellar and Buzzetti 1985). As thewhite gene in this construct lacks most of its 5’ regulatory sequences, theexpression of white tends to be weak and highly dependent on the site ofinsertion of the construct. Initial tests to determine if this construct would rescuethe enhancer of garnet phenotype were inconclusive (Table 6-top line). Theoriginal insert had marginally more pigment than its e(g) g2; TM3 siblings (26 ±2 vs 23 ± 3) but the difference was not significant. This failure to compensate forthe enhancer of garnet effect might have been due to the weak expression ofthe construct. In order to determine if this construct was able to rescue theenhancer of garnet effect when more strongly expressed, the element wasmobilized and over 30 lines that showed strong expression of the white+ genewere generated. Results of this experiment, for eight different insert lines, areshown in Table 6. In all eight lines the white construct was capable of61Table 6. Rescue of the enhancer of garnet effect by white+ transgenes.The first column gives the designation of the individual white transgene inserts.The first nine are different inserts of the CaSPeR insert, the first of which, 4-1, isthe original insert. These inserts also contain the cdc-2 homolog of Drosophilamelanogaster. The 4-3 transgene contains the complete white gene and mostof its 5’ and 3’ regulatory sequences. This insert was generously provided byBob Levis. The last two lines provide control values for e(g) g2 and g2 males.The e(g) g2 control is an internal control derived from the average values of thee(g) g2/Y; ÷/TM3 siblings of the experimental crosses.The second column indicates the strength of the white transgene expression bypigment assay (as percent wild type) and the last column indicates the ability torescue the enhancer of garnet effect in a e(g) g2 background also determinedby pigment assay.CROSSES:1. To generate the different inserts of the CaSpeR 4-1 transgene;P w67/w;casper 4-1/casper 4-1 0 w-/Y delta 2-3/TM3’.1. (transposase source)Fl w67/Y; casper/delta 2-3 c? 0w1/w;+/÷‘I,F2 w67/w1;casper/+scored for strong white+ expression and balanced as homozygotes stocks orover FM7.62To determine the chromosome on which the transgene was inserted;P w1/Y; casper insertJTM3 or w’/Y; casper insert/-i-; -i-/TM3cJ ®wm4/w CyO/TftSegregation of the white insert relative to the dominant markers indicates thechromosome into which the transgene is inserted.Of the 27 independent inserts all were located on the third chromosome. Nofurther effort was made to map them.2. To assess the strength of the white transgene expression;P w1/Y; insert/TM3cf 0w1/w;+/÷“Fl w1/w’orY;insert/+progeny assayed for pigment levels.3. To assess the ability of the white+ transgene insert to rescue the enhancer ofgarnet effect:P e(g) g2/e(g) g2; +/+ 0w1/Y; insertJTM3‘‘IF 1 e(g)g2/Y; insert‘ and e(g)g2/Y; TM3/+ a’progeny assayed for pigment levels.Generation of the different casper transgene inserts and determination of thechromosome into which the transgene had inserted was done by GwendoylnMahon as part of an undergraduate directed studies project.63Rescue of enhancer of garnet phenotype by white+ transgenesw+ insert effect on white effect on e(g)w-/Y:insert/÷ e(g) a2/Y:insert/+4-1 21±2 26±21-16 97±5 32±61-18 68±6 32±41-22 91±4 39±51-23 96±3 34±32-6 49±2 37±22-23 82±6 43±53-1 22±2 31±53-29 23±3 36±34-3 27±2e(g) g2IY; TM3/+ 23±3g2/Y 30±264rescuing the enhancer of garnet effect and restoring the full garnet phenotype.While all of these new insertions expressed the white + gene more strongly thanthe original insert, there was again no clear correlation between the strength ofwhite gene expression and rescue of the enhancer of garnet phenotype(Figure 8 and see below). The final note of interest arising from this latter set ofexperiments is that the interaction between garnet and enhancer of garnetappears to involve the coding region of the white gene. The white gene in theCaSpeR construct possesses little regulatory white+ sequences so it clearlycannot be providing additional white regulatory regions. Yet it is able to rescuethe enhancer of garnet phenotype. This suggests that interaction betweengarnet and enhancer of garnet is restricted to the coding region and is possiblya post-transcriptional process. Examination of white transcription in a garnetmutant background might resolve this question.In summary, the enhancer of garnet mutation cytologically maps to the sameposition as the white gene. Eleven white alleles, when combined with theenhanceable garnet alleles, show a severe enhancer of garnet phenotype, inconjunction with the enhancer of garnet and garnet mutations. The enhancer ofgarnet phenotype can be rescued by a white+ transgene containing white+coding region. Thus the enhancer of garnet mutation appears to be a crypticallele of the white gene.The nature and dose sensitivity of interaction between enhancer of garnet andgarnet: There are a number of systems of genes and specific modifier geneswhich have been described in Drosophila. Many of these consist of65transposable- or retro-element-induced mutations in one gene, the activity ofwhich is modified by a mutation in another gene. There are five such modifiergenes; Su(Hw) which modifies mutations produced by insertion of the gypsyelement (Modelell, Bender and Meselson 1983), su(f) which also modifiesgypsy-generated mutations (Parkhurst and Corces 1985), su(wa) whichmodifies copia expression (Bingham and Judd 1981), su(pr) and su(s) whichmodify mutations produced by the 412 element (Searles and Voelker 1986).Initially it seemed possible that the interaction between garnet and enhancer ofgarnet was of this nature. The first evidence that this was not the case camefrom the observation that the three enhanceable alleles, g2,g5Oe and g53d arealso the most extreme alleles (Table 16). In addition, none of these garnetalleles was associated with insertion of any foreign DNA (Figure 21). In orderto determine if sensitivity to the enhancer of garnet mutation was dependent onsome specific property of these three alleles (other than insertion of atransposable element) or due merely to the dose of functional garnet geneproduct, the enhancer of garnet mutation was recombined onto chromosomescarrying four of the weaker garnet alleles (g1, g3, g4 and gP). Dosage of thegarnet gene product was then manipulated by making these chromosomesheterozygous with a deficiency for the garnet gene. Table 7 shows thatalthough these alleles are not discernibly sensitive to the enhancer of garnetmutation as homozygotes, when the dose of the garnet gene is further reducedby a deficiency for the garnet gene, these alleles become dominantly sensitiveto the enhancer of garnet mutation. Dominant sensitivity to enhancer of garnetwas also observed when garnet gene dosage was reduced with the severeg53d allele instead of a deficiency (Table 7-legend). Thus the enhancer ofgarnet effect is highly sensitive to the dosage of the garnet gene and does notappear to be allele specific. As the weak garnet alleles used are hypomorphs66Table 7. Sensitivity of the enhancer of garnet effect to garnet dosageThe effect of the enhancer of garnet mutation on three strong garnet alleles, g2’g5Oe and g53d, and two weaker garnet alleles, g1 and gP, was determined bypigment assay. The first column lists the garnet allele. The next two columnsreport the eye pigment values for the garnet allele over a deficiency for thatregion and the same genotype heterozygous for enhancer of garnet,respectively. The last two columns list the eye pigment values of the appropriategarnet and enhancer of garnet and garnet homozygotes for comparison. Theenhancer of garnet effect was also sensitive to garnet dosage when the garnetgene dosage was reduced using the extreme g53d allele. All values areexpressed as percent wild type pteridine pigments.The e(g) g53d/÷ gP genotype has 24±4 % WT pteridine pigment levels.The e(g)53d/e(g) gP genotype has 13 ±2 % WT pteridine pigment levels.In comparison with 50 ± 5 for the gP homozygote.CROSS:To generate the deficiency genotypes,P Df(1)HA97/FM7 ® g/Yor e(g) g/Yo’‘I,Fl Df(1)HA97/g or e(g) g progeny assayedg*/g* c( and e(g) g*/ e(g) g* homozygotes are taken from stocks.67Sensitivity of garnet alleles to the enhancer of garnet effect and garnet dosagegarnet alleleg1g2g5Oeg53dqPg/Df(g)40±340±328±216±338±4e(g) g/+ Df(g)19±325±324±29+223±2e(g) g/e(g) g30±432±314±111±156±357±237±226± 113+150±568(Table 20), the sensitivity is presumably related to the amount of functioninggarnet gene product.To further examine the sensitivity of the enhancer of garnet and garnetinteraction to the amount of gene product, the expression of the weaker garnetalleles was tested in genotypes where the dose of the enhancer of garnetmutation was reduced using a deficiency for the enhancer of garnet region.Table 8 shows that only a single dose of mutant enhancer of garnet (DI e(g) g /e(g) g), enhances the phenotype of weak garnet alleles. In contrast to theprevious results, the interaction is not dominant, the deficiency for the enhancerof garnet region does not enhance the garnet phenotype in the presence of awild type allele of enhancer of garnet. Thus interaction between garnet andenhancer of garnet is very sensitive to the dose of the garnet gene but ratherless sensitive to the dose of enhancer of garnet.The complementation tests performed between different white alleles and theenhancer of garnet mutation provide a reasonable set of data to quantify thedose dependence between garnet and enhancer of garnet. Figure 8 shows acomparison between the strength of different white alleles and their effect ongarnet. Two points are immediately evident. There is no correlation between theseverity of the white allele and its enhancing effect on garnet. There is howevera good correlation between the strength of the enhancing effect and the severityof the garnet allele. Although this pattern might appear fortuitous, since onlythree garnet alleles were tested as opposed to the eleven white alleles, theresults generally support the findings described above. The interaction between69Table 8. Effect of we(g) dosage on the enhancer of garnet effect.The effect of reducing the dosage of the enhancer of garnet locus with adeficiency for that region, on different alleles of garnet is shown. The firstcolumn gives the garnet allele examined. The g2 allele is a strong,“enhancable” allele. The other three, g1’ g3 and gP are weaker alleles notnormally responsive to the enhancer of garnet mutation (as doublehomozygotes). The next two columns give pigment values (as percent wild typepigment) of these garnet alleles (shown as g*) in conjunction with a deficiency,heterozygous with either a wild type e(g) allele and g* or the e(g) mutation andg*, respectively. The last two columns show values for garnet and enhancer ofgarnet homozygotes for comparison.CROSSES:1. To generate Df(1)N8 garnet strains;P Df(1)N8, w-, N-/In(1)d149, Hw g4 0 g/YcJ1.Fl Df(1)N8+/+ g 0 In(1)dI49/Y’ (sibs)1F2 Df(1)N8 g/In(1)d149recombinants selected by garnet phenotype and notched wing.2. To generate e(g) g stocks;P ye(g)cvg/ye(g) ®yacscpn wvg1 f/YQ’or ecctvg3YcJor ycvvgPf/Y‘I,Fl y e(g) cvg2/y cv vg1 ®y e(g) cvg2cfy e(g) g* recombinants selected by appropriate markers.3. To generate the Df(1)N8 g/ + g or e(g) g individuals,P Df(1)N8 g*/In(1)d149 0 g/Yor e(g) g*ty“Fl Df(1)N8 g*/e(g) g*or70Effect of e(g) dosage on the enhancer of garnet effectDfe(g)g* Dfe(ci)g* e(cx)g*+ g* e(g) g* g* e(g)g*g2 28±5 7±1 37±2 32+313±2 5±1 57±2 NDg3 24±2 14±2 45±1 NDgP 16±1 8±1 50±5 ND71Figure 8. Comparison between the severity of different white alleles and theireffect on garnet.A. The effect of different white alleles (combined with three different garnetalleles, g53dg5Oe andg2(e.g. w*g*/e(g) g’ as heterozygotes with theenhancer of garnet mutation and the same garnet allele (relevant crosses andgenotypes given the legend to Table 5). The percent wild type pigment of thedifferent white alleles are shown on the X axis. The percent wild type pigment ofthe white garnet / e(g) garnet heterozygote are shown for the three garnetalleles, on the Y axis.B. The effect of different white+ transgenes in rescuing the enhancer of garneteffect. The crosses and relevant genotypes given in the legend to Table 6. Thegeneric genotype is y e(g) cvg2/Y; [white+]/+. The percent wild type pigment ofthe white+ transgene is shown on the X axis. The pigment level of the “rescued”genotype is shown on the Y axis.72Comparison between the strength of different white alleles andtheir complementation of enhancer of garnet.Comparison between strength of white transgene andrescue of enhancer of garnet effect.‘+z,-lb40-————— —I++I +30-— - -25---- - -200 102030405060708090100pigment of white transgene*g2g50e+g53dJ- - —20- —15- —10- — —rII5.; I__ — — — — — — —I In_ — —0 23456789pigment of white allele10CC.)-C•173enhancer of garnet and garnet is dramatically sensitive to the dose of thegarnet gene but the effect of the white gene is somewhat allele specific.Interactions between enhancer of garnet and other eye colour genes: Thegenetic interaction between enhancer of garnet and garnet points to somepreviously undetected interaction between their gene products. The white geneproduct has been proposed, based on both molecular and genetic studies, tofunction as transmembrane channels in conjunction with the products of thebrown and scarlet genes (Dressden, Johnson and Henikoff 1988). Thus thewhite gene product appears to interact with at least two other eye colour geneproducts. As the enhancer of garnet allele of white was an invaluable tool inexamining interactions between white and garnet, I decided to test other eyecolour genes for their sensitivity to the enhancer of garnet allele.The enhancer of garnet mutation was combined with 26 other eye colour genes.Since it was practical to test only one allele of each gene, and it is generally notknown which, if any, alleles of the different eye colour genes are amorphic,there is a possibility that some interactions would not be observed. In order toincrease the sensitivity of this test, the enhancer of garnet mutation was placedin trans to a deficiency for enhancer of garnet region. Table 9 shows theresults of this survey. Of the 26 eye colour genes examined, the mutantphenotype of nine, possibly ten, is enhanced by the enhancer of garnetmutation. These genes share with garnet the property of altering levels of bothbrown and red pigments. Significantly, brown and scarlet fail to show aninteraction with the enhancer of garnet mutation. Thus the interaction betweengarnet and enhancer of garnet appears to define a different type of interactionthan the structural role proposed for brown and scarlet. Interestingly, raspberry74Table 9. The effect of the enhancer of garnet mutation on other eye colourmutants.The e(g) mutation was tested in combination with 25 eye colour genes. The firsttwo columns list the symbol and name of the eye colour gene respectively. Thenext two columns list the amount of total pteridine pigments measured for thegiven eye colour mutant as a homozygote and for the same homozygousmutation in combination with the a hemizygous e(g) mutation, respectively. Allvalues are expressed as percent wild type pteridine pigment. The genericgenotypes are mutant/mutant vs. e(g) mutant! Df e(g) mutant or e(g)/Df e(g);mutant/mutant respectively. The values for the homozygous mutant genotypewere derived from internal controls for mutations on the second and thirdchromosome and external controls (mutant/mutant flies grown concurrently withthe test crosses) for those on the first chromosome. The next column gives thepercentage difference of the two measurements as hemizygouse(g) with mutantI mutant alone. The last column indicates whether or not this differenceconstitutes responsiveness to the enhancer of garnet mutation.The e(g) mutant phenotypes are heterozygous for the DF(1)N8 deficiency. Thisdeficiency has two effects on pigment levels unrelated to hemizygosity for thewhite locus. Flies bearing the N8 deficiency are generally smaller and have atransient, slightly brown eye phenotype shortly after eclosion. These effectshave been corrected for using DF(1)N8/ e(g); mutant/Balancer siblings wherepossible or an arbitrary cut off point of 75% of the control, mutant alone, value.In general there was little ambiguity.75CROSSES:1. For X-linked eye colour genes:A.P Df(1)N8/In(1)d149 ® mutant/Ycy1Fl Df(1)N8 +/ + mutant ® In(1)d149/Yo’ (sibs)‘IF2 Df(1)N8 mutant?/In(1)d149 0 mutant/Yo’(test cross individual virgin females to select double mutants)B.P y e(g) cvg2/y e(g) cvg2 0 mutant/Yo’Fl ye(g)cvg2÷/÷÷+÷mutant®ye(g)cvg2/Yf(sibs)‘I,F2 y e(g)? cv+ g+ mutant/Ye’ selected.Stocks were made by crossing to In(1)d149 balancer and the presence ofe(g), garnet and the eye colour mutant were confirmed by the appropriate testcross.The Df e(g) mutant chromosomes and the first two generations of the e(g)mutant crosses were made by Mitra Maharaj.C. The experimental cross:P Df(1)N8 mutant/In(1)d149 0 y e(g) mutant/Yo’“Fl Df e(g) mutant’ e(g) mutanttest genotype.2. For second chromosome mutants:A.P Df(1)N8/In(1)d149; GIa/ + 0 +/Y; mutant/mutant a’B.P y e(g) cvg2/y e(g) cvg2; Tft/CyO 0 +/Y, mutant/mutanto’76C.Fl Df(1)N8/÷; Gla/mutant ®y e(g) cvg2/Y; mutant/CyOcf‘I,F2 Df(1)N8/y e(g) cvg2;mutant/mutant (experimental genotype)Df(1)N8/y e(g) cvg2;mutant/CyO (NB effect control)+ /y e(g) cvg2; mutant/mutant (mutant control)+/y e(g) cvg2;mutant/CyQ (wild type control)3. For third chromosomes mutants:A.P Df(1)N8/In(1)d149; H/TM3 ® +/Y mutant/mutanto’B.P y e(g) cvg2/y e(g) cvg2; H/TM3 ® ÷/Y mutant/mutantc?C.Fl Df(1)N8/+; TM3/mutant 0 y e(g) cvg2/Y; mutant/TM3cIDf(1)N8/y e(g) cvg2; mutant/mutant (experimental genotype)Df(1)N8/y e(g) cvg2; mutantiTM3 (NB effect control)+ /y e(g) cvg2;mutant/mutant (mutant control)+/y e(g) cvg2;mutant/TM3 (wild type control)77The effect of enhancer of garnet on other eye colour mutantsphenotype ofgene mutant e(g) mutant difference effectbo bordeaux 74±5 108±2 144% NObw brown 1 .1±.2 1.5± .4 104% NOca claret 36±4 24±4 66% YEScar carnation 39±4 28±2 72 YES?Cd cardinal 97±4 113 ±5 106 NOcm carmine 59±10 10±1 13 YEScn cinnabar 103±3 96±2 93 NOdke-c dark eye 74±4 91 ± 16 123 NOdor deep orange 18±2 5± 1 29 YESIt light 37± 1 9± 1 25 YESltd lightoid 90± 1 17± 1 19 YESma! maroon-like 50±5 105±2 208 NOMe Moire 80±3 85±6 106 NOmel melanized 50±5 105±2 208 NOMot-KMottled of K 100 ±2 79 ±7 79 NOmw mottler of w 1.3± .3 1.6 ±.5 127 NOp pink 32±2 14±3 44 YESpn prune 49±3 49±2 100 NOpr purple 37±3 6±1 16 YESras raspberry 29±2 41±2 142 NOrb ruby 38±2 7± 1 19 YESry rosy 60±2 38±3 63 YESse sepia 111 ±3 132±2 119 NOSt scarlet 76±5 71±4 93 NOv vermilion 106 ±4 103 ±3 97 NOwo white 92 ±7 99±4 106 NOoccelli78and maroon-like, lesions in the pteridine pathway, appear suppressed by theenhancer of garnet mutation. The significance of this suppression is not clear.Thus, enhancement of mutant eye phenotype by the enhancer of garnetmutation appears to be a property shared by many eye colour genes in the“transport” class.Interactions between garnet and other eye colour genes: As the garnet gene isnot alone in its interaction with the enhancer of garnet allele, and as most ofthese belong to the transport group of eye colour mutants, it was of immediateinterest to test the garnet gene for interactions with other eye colour genes.Nolte (1 952b) found that while alleles of the genes garnet, carnation, carmineand ruby are all normally recessive, in pairwise combinations they showsynthetic dominance. I extended these analysis to include 14 other eye colourgenes (Table 10). Interactions between pairwise combinations of garnet andthe other eye colour genes were assessed by looking for synthetic dominanceof garnet (garnet/÷, mutant/mutant< +/+; mutant/mutant), synthetic dominance ofthe other gene (garnet/garnet or/Y; mutant/-i- <garnet/garnet or garnet/Y; +1+),synergistic or additive, versus epistatic effects on eye pigmentation of thedouble mutants, and effects on viability and fertility of double homozygotes. Ofthose mutations which interacted with enhancer of garnet (claret, carmine, deeporange, light, lightoid, pink, purple, ruby, rosy and probably carnation) all butpurple also interacted with garnet. In no case was an interaction with garnetobserved with a gene that failed to interact with enhancer of garnet. Takentogether with the data of Nolte (1 952b) there is considerable congruencebetween those genes which interact with enhancer of garnet and those whichinteract with garnet (Table 11). Finally it is interesting to note that, while thepartial sterility seen between garnet and other genes probably reflects a79Table 10. Interactions between garnet and other eye colour genes.Interactions between garnet and other eye colour genes was monitored in threeways:1. Epistatic versus additive or synergistic interactions (as monitored bothvisually and by pigment assay).2. Synthetic dominance.3. Synthetic lethality or sterility of the double homozygotes.The first two columns indicate the symbol and name of the eye colour genetested with in combination with garnet. The next four columns show pteridinepigment values (expressed as percent wild type levels) for; females and maleshomozygous or hemizygous for the other eye colour mutation and female andmale double mutant individuals. To determine if either garnet or the other eyecolour mutant acted in a dominant fashion in a mutant background of the othergene, the following genotypes were generated; ÷/g; mutant/mutant to determineif garnet acted in a dominant manner in a mutant background and g/g or g/Y;mutant/÷ to determine if the mutant became dominant in a garnet background.The next three columns show pteridine pigment values (expressed as percentwild type levels) for; individuals homozygous for the other eye colour mutationand heterozygous for garnet, and individuals homozygous or hemizygous forgarnet and heterozygous for the other eye colour mutant. The next two columnsindicate the fertility of the double homozygotes and notes on their phenotypes.The last column is a summary of the various models of interaction. A “+“indicates interaction, either of epistatic/synergistic, pigment interactions,synthetic dominance in either direction, synthetic sterility or extensive80secondary phenotypes generally indicative of cell death. One “+“ is given foreach type of interaction observed.CROSSES:1. To generate double homozygotes:For eye colour mutants on the X chromosome;P g2 mutant/In(1)d149 0 g mutant/YJ(double mutant stocks generated in cross 1 B described in legend toTable 8.For second and third chromosome mutants;P y zag53d/yza g53d; E(var) 18. l2Sp/CyOor E(var)303 ru h eg pP K1ffM30 In(1)IN(1)BM1/Y;mutant/mutantcfFl In(l)!N(1)BM1/y zag53d,. mutant/CyO or TM3 y za g53d/y,.mutant/CyO orTM3’2. Genotypes used to examine synthetic dominance;For X-linked eye colour genes,p g53d/g53d 0 g2 mutant/Yo’andP mutant/mutant 0 g2 mutant/Yo’For second and third chromosome eye colour mutants the genotype g53d/g53dor g53diy; mutant/Balancer and In(1)!N(1)BM1/g53;mutant/Balancer weregenerated as siblings of the crosses described above to make doublehomozygotes. Cross 2 and 3 described in the legend of Table 9 generateequivalent phenotypes with the g2 allele and second and third chromosomemutations. In all cases the conclusions derived from these crosses were thesame as with the g53dallele.81InteractionsbetweengarnetandothereyecolourgenespigmentsyntheticfertilitySUMMARYsinglemutantdoublemutantdominancefemalemalefemalemale÷7q;m/mj/q;+/mq/Y÷/mgenebwbrowncaclaretcdcardinalcncinnabarItlightltdlightoidMeMoireMotKMottledppinkprpurpleiyrosysesepiaStscarletwowhiteoccelli1+.236±295±3103±337±178±1080±3100±234±230±1063±3111±387±10ND1±.10043±29±16±1101±4ND5±1108±2NDND42±52±11±.568±60082±3NDND101±3NDND45±3ND5±140±104±17±164±10ND6±1112±5NDND81±1010±27±292±7NDND822±15±.53±.535±38±16±1101±3NDND106±1NDND20±17±17±150±28±17±179±5NDND95±1NDND25±2NDND37±4NDND54±3ND16±2113±4NDND69±513±216±294±4NDNDf&mfertilemalesemi-sterilemalefert.f.NDNDf&mfertilemalesemi-sterileNDNDmalesemi-sterilef&mfertilef.sterileNDf&mfertileND-9 + +++++++Table 11. Summary of interactions between other eye colour mutations andboth garnet and enhancer of garnet.The first two columns give the symbol and the name of the eye colour gene. Thethird list the pigment pathway effected or the general nature of the defect (fromTable 1). The final two columns list whether there was an interaction witheither enhancer of garnet or garnet respectively. A gene was considered tointeract with garnet if it showed any one or more of epistatic/synergisticinteractions, synthetic dominance, sterility or unusual phenotypes of doublehete rozygotes.83Summary of eye colour genes interacting with e(g) and garnet.mutant pigment interaction interactionaffected with e(g) with garnetca claret both YES YEScar carnation both YES NDcd cardinal both NO NOcm carmine both YES NOdor deep orange both YES NDg garnet both YES NAIt light both YES YESltd lightoid both YES YESp pink both YES YESpd purpleoid both ND NO?rb ruby both? YES NDiy rosy both? YES YESw white both NA YES*cn cinnabar ommochrome NO NOSt scarlet ommochrome NO NO?v vermilion ommochrome NO? NDbo bordeaux pteridines? NO NDbw brown pteridines NO NOpn prune pteridines NO? NDpr purple pteridines YES NOras raspberry pteridines NO NDse sepia pteridines NO NOmal maroon-like pteridines NO NDdke dark-eye pattern NO NDMe Moire pattern NO NOMot-K Mottled of K pattern NO NOrI rolled pattern NO? NDmel melanized secondary NO NDpw-c pink wing c secondary? NO NDz zeste regulatory ? YES*84Figure 9. Analysis of garnet transcription by in situ hybridization in a rosy nullgenetic background.Panel A shows garnet message in brain (b), salivary gland (sg), eye antennaldiscs (e), leg discs (I), wing discs (w), fat body (f) and trachea, of ry506/rythird instar larvae.Panel B shows garnet message in leg discs (I), a wing disc (w), and possibly ahaltere disc (h?) of ry506/t third instar larvae at higher magnification.Panel C shows two leg discs of ty506/rythird instar larvae hybridized withsense probe as controls.8538r I 6sccq4•t....•4Vsecondary effect of the small, weak, poorly viable males that died a few daysafter eclosion, only the rosy gene, when combined with garnet, shows bothcomplete sterility and female sterility. The interaction between garnet and rosywas investigated in more detail by examining garnet transcription in imaginaldiscs, brain and other tissues derived from rosy null (ry506) third instar larvae.Figure 9 shows these results. Although this is not a quantitative assay, it wouldseem that garnet transcription is not noticeably altered in the imaginal discs ofrosy null larvae. Interestingly, however, there appears to be no garnettranscription in the brain of rosy null third instar larvae, a tissue which showsgarnet expression in wild type larvae (Figure 27).Interaction between garnet and white: Two of the garnet interactions deserveadditional comments. The interactions between garnet and enhancer of garnethave already been detailed at some length. As the enhancer of garnet mutationappears to be an allele of white it is clear that garnet and white show a fairlycomplex genetic interaction. It is only by virtue of being a subliminal allele ofwhite that it was possible to investigate these interactions. The conventionalalleles of white (necessarily) effect eye pigmentation so that the spectrum ofinteractions seen with other eye colour mutations is quite limited.Nolte (1 952b) noted that g3 interacted additively with the white alleles, wCh,w0, wa, wsat, wbl and w00. This analysis was extended to include 7 additionalwhite alleles, two extreme (w1 andw1118 completely white), one weak (wBVlfXwhich produces a dull red colour), and four moderate alleles (wbf, we, w1 andwt, pinky-yellow eyes) in combination with three garnet alleles, g2,g5Oe andg53d(Table 12). The results are consistent with those reported by Nolte. Thewhite and garnet alleles act additively to reduce eye pigmentation. One87Table 12. Phenotypes of garnet double mutants with different white alleles.The phenotype of 11 different white alleles alone and combined with g2’g5Oeand g53d is shown. The first column shows the white allele, the next columnlists its visual phenotype and below, the value obtained from pigment assay forthe white allele alone (as percent wild type pteridine pigment). The last columnlists the phenotype and pigment value of the appropriate white allele and theg53dallele of the genotype w* g53d/y where *“ indicates the given whiteallele. There was in general no great difference between the three garnetalleles in combination with the white alleles.88Additive interactions between white and garnet alleleswhite allele phenotype white and garnet doublemutant phenotypew1 178 completely white completely white<1 <1w1 completely white completely white<1 <1wa dull orange-red completely white3±1 <1wbf faint yellow completely white1±.5 <1Bwx dull red pale pink-orange1±.2 NDwch faint pink/orange completely white<1w00 dull red faint yellow6±1 2±1we pink/orange completely white3±.5 <1w1 faint pink completely white<1sat browny-purple pale orange8±1 NDw faint pink completely white.6±1 <189Table 13. Epistatic interaction between g2 and wa3.The first row gives the genotype of males with either single or double mutantcombination of g2 and wa3. The second row gives the values for pteridine eyepigments (as percent pteridine wild type pigment) for each of these genotypes.90Epistatic interaction between g2 and a3genotype g2/Y wa3/Y a3g2/Ypigment 38±3 25 ±3 28 ±...291exception has been noted to this rule. Green (1959) noted, while attempting tofunctionally distinguish white alleles, that a2 and wa3 in combination with g2can not be distinguished from the single mutant. Table 13 shows pigmentdeterminations for a2, g2 and the double mutant combination. These resultsconfirm that, in contrast to the situation with all other combinations, there is noadditive interaction between these alleles. (Green also claims this effect is alsofound for another member of the “transport” group of eye colour mutations, rubyand vva3.) These results further emphasize the allele specificity of the white-garnet interaction. Further study of this aberrant epistatic interaction might helpreveal the nature of the physical nature of the white-garnet interaction.Interaction between garnet and zeste. The zeste gene was originally identifiedas a modifier of the white gene. In the presence of a mutant zeste gene, pairedcopies of the white gene function less effectively, decreasing pigmentation.Thus zeste acts as a pairing-dependent positive regulator of white. Molecularanalysis of the zeste gene has shown it to be a protein which acts as atranscription factor which enhances the expression of white and many othergenes (Pirrota 1988). Mutations in the zeste gene have a dramatic effect ongarnet expression. Figure 10 shows g2,g5Oe and g53dflies without and withthe z1 and the za alleles. Table 14 shows the corresponding pigment levels.Clearly mutants in the zeste gene further reduce the amount of pigmentation ofmutant garnet individuals. The zeste-garnet interaction appears sensitive tochromosome pairing. While the z1 allele does reduce garnet expression inmales, with a single X chromosome, in females where the two X chromosomesmay pair, eye pigmentation is essentially abolished (Figure 11). Thus both thezeste gene and the enhancer of garnet mutation reduce garnet expression.Both the cytological mapping of enhancer of garnet (Figure 6) and by92Figure 10. The effect of zeste mutant alleles on the garnet phenotype.The effect of the za and z1 alleles on the phenotype of flies mutant for the threegarnet alleles, g53d,g5Oe and g2 (from top to bottom respectively). Eachphotograph shows the eye colour of the garnet allele alone, the same garnetalleles when combined with the za allele (and also the z1 allele with theg5Qegarnet allele) and the appropriate e(g)-garnet combination for comparison.93e(g) g53dz1 g5Oezag5Oee(g) g2e(g)g294Table 14. Effect of zeste on garnet.The effect of zeste on garnet expression was determined by visual inspectionand by determination of pteridine pigment levels (expressed as percent wildtype levels) for two alleles of zeste (z1 and za) and the three most severe allelesof garnet (g2,g500 and g53d).Cross: The zeste and garnet mutations were combined as follows:g*/g* ®yz*/Yc/Fl yz’ ÷/÷ + g* ® g*fyf(sjs)F2 yz*g/Ycfprogeny selected as yellow, hence probably zeste, and garnetindividuals. These were balanced over the ln(1)d149 balancer togenerate stocks.95Effect of zeste on garnetgarnet allele zeste genotypeyzgg53d/g53d 16± 1g53d/y 19±1g5Oe/gQe 29±2g5Oe/y 33±2g2/g 31±2g2/Y 38±2yzag10 + 112±218±319±514 + 117+1yz1g<1 ± 110±1<1 ± 110±3<1 ± 121±296Figure 11. The phenotype of zeste-garnet combinations in females and males.The phenotypes of three garnet alleles, g53d, g5Oe and g2 (from top to bottom)alone and in combination with the z1 allele are shown. The phenotypes ofhomozygous and hemizygous z’- garnet flies differ noticeably. The females(where pairing between the two X chromosomes is possible) have essentiallywhite eyes whereas the males show a less dramatically enhanced garnetmutant phenotype.97z1g53d/yz1g5Oe/Yzlg53d/zl g53d121g5Oe/zl g5Oez1 g2IYz1g2/z1Table 15. Effect of zeste modifiers on the zeste-garnet genotype.The effect of two modifiers of zeste, E(z) 1 and Su(z) 302 was determined ontheyza g2 and yz1g53dgenotypes.Cross:P yzi g*/In(1)d149; +1+ oryza g*/Jn(1)d149; ÷/÷ 0 +/Y E(’z,)1 orSu(z)302/TM3a’where * indicated either g53d or g2.The g4 allele occurs in the complex inversion In(1)In(1)d149.The progeny of this cross were scored visually and by determination ofpteridine pigment levels (values expressed as percent wild type).99Effect of modifiers of zeste on garnetgarnet genotype E(z)1 Su(z)302yzlg53d/÷;Mod.*/÷ 19±1 101±2yz1 g53d/+; TM3/+ 96±2 101 ± 1y z1g53d/Y; Mod./+ 1 ± .1 15± 1yzlg5d/Y;TM3/÷ 11±1 15±1y zag2/+; Mod./+ 97±2 103± 1yzag/+;TM3/÷ 94±2 100±1yzag/Y;Mod./+ 10±3 25±2y zag2/Y; TM3/+ 9±1 18± 1yzg4/+; Mod.!-,- 102± 1 104±2y z+ g4/÷; TM3/+ 97±2 101 ±3yzg/Y;Mod./÷ 29±3 38±4yz+g4/Y;TM3/÷ 14±1 35±1* Mod. = Modifier of zeste, either E(z) 1 or Su(z)302.100complementation between zeste and enhancer of garnet (data not shown)indicate that the enhancer of garnet mutation is not an allele of zeste. Thisinteraction is sensitive to modifiers of zeste. The enhancer of zeste mutationfurther reduces pigment of z1g53d in males and is, furthermore, capable ofinducing dominant garnet expression in z1 g53d/÷ + females (Table 15). Theenhancement of theg53dallele by z1 occurs in males as well as femalesindicating that in this case, pairing of X chromosomes is not necessary.Interestingly the eye phenotype of these flies is a tine grained mosaic of brownspots on an orange background which becomes more pronounced with age.Genetically the interaction between zeste and white resembles that betweengarnet and white. At least superficially, both appear to be necessary for fullwhite gene expression. Although the zeste gene has been extensivelyinvestigated and is known to modify the transcription of a number of genes,there is no mention in the literature of an interaction between zeste and garnet.This interaction could occur either directly, if zeste acts as a transcriptionalenhancer for the garnet gene, or indirectly, if it occurs due to compromising thefunction of the white+ gene. The latter may be more likely as there are no zestebinding sites at garnet. These possibilities could be distinguished genetically byproviding another wild type copy of the white gene as a transgene which wouldbe insensitive to pairing-mediated zeste effects.101Discussion-Interactions between garnet and other eye colour genes.The enhancer of garnet mutation:The first clue to the extensive genetic interactions between garnet and othergenes was the discovery of the enhancer of garnet mutation. The chance offinding not only a mutant white allele, but a cryptic white allele, in the strain inwhich the gP mutation was induced is truly remarkable. This mutation waspivotal in unraveling the complex network of interactions between white andmany other eye colour mutations. These interactions could not have beendiscovered with a conventional allele of white where the phenotype of the whiteallele would mask that of the interaction.The enhancer of garnet mutation is a cryptic allele of white:Three lines of evidence indicate that the enhancer of garnet mutation is anallele of the white gene. Firstly, their cytological positions coincide. Secondly 11different alleles of white acted as strong enhancer of garnet mutations incombination with the enhancer of garnet mutant and the three alleles of garnetwhose phenotypes are sensitive to the enhancer of garnet mutation. Finallywhite+ transgenes were able to rescue the enhancer of garnet effect. Thus theenhancer of garnet mutation should properly be designated we(g)Dosage studies with deficiencies for the white gene indicate that the originalenhancer of garnet mutation was hypomorphic. In conventionalcomplementation tests with different white alleles and white deficiencies, theenhancer of garnet mutation complements white mutations. Mutations in the102white gene, however, fail to complement the enhancer of garnet phenotype.Thus the enhancer of garnet mutation is a cryptic allele of the white gene. As acryptic allele it has no independent phenotype.The we(g) mutation reveals interactions between white and members of thetransport group.The we(g) mutation was the key tool in the exploration of the interactionbetween garnet and white. Combinations between we(g) and mutations in 26other eye colour genes were examined in the hope that e(g) would revealsimilar interactions between the white gene and other eye colour mutations. Ofthese 26 combinations, nine or possibly ten showed interactions which mimicthe interaction seen between garnet and enhancer of garnet. More importantly,only those eye colour genes identified as members of the transport groupshowed this interaction. Neither brown nor scarlet, likely structural componentsof the transmembrane pore complex, nor genes encoding enzymes in theommochrome biosynthetic pathway (vermilion and cinnabar) or pteridinebiosynthetic pathway (maroon-like, raspberry and sepia), showed thisinteraction.These results imply that the white gene interacts not only with garnet but alsowith claret, carnation, carmine, deep orange, light, lightoid, pink, purple, rosyand ruby. These genes are largely the same as those described as the rubygroup by Nolte (1955), group two by Schwink (1975) and the transport group bySullivan and Sullivan (1975). (Two of the eye colour genes which have beenimplicated as members of this group, orange and maroon, were not testedbecause mutant alleles were not available).103Conceptually, this interaction could arise from a situation where the white geneproduct interacted separately with all these genes. Alternatively, the white geneproduct, and the products of the transport group of eye colour genes mightinteract as members of a large macromolecular complex. The second modelpredicts that the transport group of genes would interact with each other as wellas with the white gene. This was found to be the case. Pairwise combinationsbetween garnet and many of these genes showed a variety of novelphenotypes, including synthetic sterility, cell death phenotypes, syntheticdominance and synergistic/additive effects on eye pigmentation. Although thesegenes have been previously linked by phenotypic, physiological andhistological analysis, genetic analysis of interactions between these genes hasbeen haphazard. There is remarkable congruence between the genes whichinteract with enhancer of garnet and those which interact with garnet. Thisstrongly suggests that the white gene interacts in the same fashion with all ofthese gene products.One member of the transport group deserves additional comment; the rosygene. The rosy gene is exceptional in this group because it is the only onewhich is known to encode a defined enzymatic function. The rosy gene encodesxanthine dehydrogenase. Unlike other eye colour genes which encodeenzymes, no role for xanthine dehydrogenase has been confirmed in either theommochrome or pteridine biosynthetic pathways, (Reaume, Knecht andChovnick, 1991). Interactions between rosy and other members of the transportgroup might suggest a structural rather than enzymatic role for this geneproduct. While structural enzymes are hardly unprecedented, enzymatic activityappears necessary for both normal eye pigmentation and correct localization of104xanthine dyhydrogenase in the pigment granules (Reaume, Knecht andChovnick, 1991). It is also possible that rosy has a general function onlyindirectly related to eye pigmentation (Hilliker et a!. 1992). It may be important tonote in this context that the interaction between garnet and rosy was unusual inthat it was the only combination that resulted in complete female sterility. Furtherinvestigation may well reveal that the rosy gene plays a slightly different, andpossibly a key role in the function of this group of mutations.The nature of the interaction between garnet and white:Null mutations in the white gene show no pigment deposition in any tissue.Thus the white gene product is necessary for the correct localization of both thepteridine and ommochrome pigments. Early models advanced the white geneas a common element which interacted with the product of the brown andvermilion, cinnabar and scarlet genes to deposit both red and brown pigments(Nolte 1952b). The role of the white gene in pigment deposition has beenreformulated based on its similarity to a number of transmembrane channelcomplex proteins such as bacterial permeases, mammalian multiple drugresistance genes and the cystic fibrosis gene product (Dressen, Johnson andHenikoff 1988). These proteins function as transmembrane pores or partsthereof. One model for its function in eye pigmentation is based on a physicalinteraction with the products of the brown and scarlet genes, respectively, toform the principal component of the transmembrane channel complexes whichallow transport of pteridines and ommochrome pigments, pigment intermediatesand metabolites into the pigment cells. The role of the garnet gene in thisprocess is unknown.105The extensive complementation testing between garnet and different whitealleles allowed me to examine the sensitivity of the enhancer of garnet effect tothe dosage of both the white and the garnet genes. The enhancer of garneteffect appears to be very sensitive to the dosage of the garnet gene. Theinteraction is sufficiently sensitive that if the dosage of the garnet gene isseverely reduced, such as when a mutant garnet allele is heterozygous with adeficiency for the garnet region (e(g) g / + Df(g)), the enhancer of garnetmutation acts in a dominant manner. In contrast, the enhancer of garnet effectappears to be sensitive to the specific allele of white rather than dosage of thewhite gene. The allele specificity may reflect the mechanism of the interaction.In at least one case, that of the wa allele, the lesion at the white gene isassociated with aberrant transcripts. If these altered transcripts are translated,the strong enhancer of garnet effect might indicate a protein-protein interaction,conceivably via titration of active garnet gene product. Characterization of thelesion at the white locus in the wa3 allele, which shows an epistatic interactionwith garnet might be illuminating. A further suggestion that the interactionbetween white and garnet may occur post-translationally can be inferred fromthe rescue of the enhancer of garnet effect by white+ transgenes withessentially only white coding region. Regardless of the nature of theinteraction, genetic interactions between white and garnet clearly indicate afunctional link between these genes.Severe garnet mutations and weak white mutations are very similar inphenotype. In this context, it is interesting to note that the chromatographyprofile of a garnet allele (g2) is indistinguishable from that of a hypomorph ofwhite (Hadorn and Mitchel 1951). Furthermore, Nolte (1959b) noted thatchange in red and brown pigments of the wbl allele produced by temperature106change parallels the changes in these pigments between different garnetalleles. This similarity in phenotype, in conjunction with the proposed structuralrole of white in forming the transmembrane channel complex suggests thatgarnet is a positive regulator of white function. Specifically if garnet is nonfunctional, the white gene functions, but much less efficiently. If there is anyfurther compromise of this system, such as with the we(g) or other white mutantallele, pigment deposition, which in the presence of a functional garnet geneproduct would be unaffected or only slightly diminished, is essentiallyeliminated.Models of gene interaction in the eye:The most complete model proposed for the role of the many eye colour genes inthe structure and pigmentation of the eye was proposed by Nolte (1952b,1959),based on data obtained from his decade of histological and genetic work onthese genes. Nolte proposed that the genes which altered levels of both the redand brown pigments, the “transport group”, were involved in general aspects ofprotein metabolism, specifically protein catabolism. The massive rise in proteincatabolism that accompanies metamorphosis and the inability to excrete toxicby-products could result in the evolution of pathways whereby toxic by-productswere converted into pigments. Genes which regulate or participate in thesefunctions might alter the amounts of pigment precursors and thus final pigmentlevels. Nolte (1 952b) derived a genetic pathway that invoked two partiallyindependent systems, controlled by brown and scarlet, which interacted with thewhite gene. In this scheme garnet, along with ruby, carmine and carnationoperated in separate but parallel pathways to provide appropriate levels ofprecursors. Later he included the genes claret, maroon, pink, purple, purploid,107prune and rosy (Nolte 1955) and light (Nolte 1 954b) as other members withsimilar functions. The genetic interactions, the synthetic dominance andsynergistic interactions, between these genes were the impetus for this modeland they are certainly consistent with the proposal of an interrelated network ofsomewhat redundant pathways where mutation in one would lower thethreshold for others. Such a system would also explain why lesions in such ageneral and important function were not lethal. Figure 12 shows the functionsderived from genetic interactions as proposed by Nolte (1952b). Nolte’s later(1959b) variation of the model included available information on tissuespecificity, developmental control of pigment formation and multiple feedbackloops.Little more can be added conceptually to this model. The analysis of some eyecolour genes over the past thirty years has, however, suggested more specificroles. This information, in conjunction with the results of this work, allows me topostulate the model shown in Figure 13. The essential features of this modelare:1. The white, brown and scarlet genes are structural components of atransmembrane pore complex.2. The white gene product interacts separately with the scarlet and brown geneproducts.3. The transport group of eye colour genes also interact with the white geneproduct.4. This interaction is different from that of brown and scarlet, likely fulfilling anon-structural role.5. Interaction between the transport group and the white gene is proposed tooccur at the protein level.108Figure 12. Nolte’s model for interactions between eye colour genes.Figure 12 shows a simplified models of interactions between eye colour genesadapted from models proposed by Nolte (1952b and 1959). The lower circleindicates biochemical reactions proposed to occur in the body whereas theupper portion of the figure indicates reactions proposed to occur in the eye.Roles of a number of eye colour genes are indicated. The genes garnet,carmine, carnation and ruby are shown at the bottom performing similar,redundant functions involved with metabolism of eye pigment precursors. Thewhite gene is shown interacting with the products of both the brown and scarletgene to produce the red and brown pigments respectively. The cinnabar andvermilion genes are indicated as enzymatic lesions.109Nolte’s model of eye colour gene interactionsprersorsv+ftryptophang+cmar4rb +precursorst• reactionsineyereactionsin bodyRedchromoproteinchromoproteinbrown protein rec’pigment pigment(s)precursorscn +fprecu1sorsprecursors/1106. The products of the transport group genes are proposed to form a complex.7. The rosy gene may have a different and possibly key role in this process.8. The function of this complex is to enhance the function of the transmembranepore complex.9. These gene products may associate in various combinations to performsimilar functions in other tissues, other locations in the cell, or at other times indevelopmentThese points are discussed in more detail below.The white, brown and scarlet genes appear to encode transmembrane transportproteins. The phenotypes of mutations in these genes, the interactions betweenthese genes, and their physical properties lead to a simple model in which thewhite gene product associates with either the brown or the scarlet geneproducts to form the transmembrane channel complexes responsible fortransport of the pteridine and ommochrome pigments or pigment intermediates.The role of the “transport” group of mutants appears somewhat different. Thegarnet, light (Devlin et al. 1990) and pink (Jones and Rawls 1988) genes havebeen cloned. The garnet gene does not show a strong similarity totransmembrane proteins, nor has such similarity been published for the lightgene, however, A.J. Howells (cited in Tearle 1991) reports that the pink genemay have some sequence similarity to the white gene. Genetically, neitherbrown nor scarlet interact with the we(g) mutation, suggesting that the transportgroup gene products participate in a different type of interaction with the whitegene product from that hypothesized for the brown and scarlet gene products.Rescue of the enhancer of garnet effect by the coding region of white leads meto postulate that this interaction involves the protein products of these genes,although the data are inferential for the garnet gene and there are no equivalent111Figure 13. Model of the physical interactions between the products of thewhite, brown, scarlet and the transport group of eye colour genes, includinggarnet.The product of the white gene is proposed to interact with the products of thebrown and scarlet genes, separately, to form transmembrane channels in thecell membrane to allow the ingress of the pteridine and ommochrome pigments(or intermediates), respectively. The products of the garnet, carnation, carmine,light, lightoid and pink genes are shown interacting in some ill-definedmacromolecular complex, the function of which is to enhance the activity of thewhite gene product. The product of the rosy gene is shown as participating bysome independent mechanism.1120 3 9 0 C) 0 3 CDD -‘ 0 a CD 0 -‘ -h C•) 0 = 0 CD CD D CD a C,p- C)O3.0Z0data concerning other members of the transport group. The effect of e(g)other members of the transport group of eye colour genes parallels that of we(g)on garnet. Furthermore, those members of the transport group of eye colourmutations which are sensitive to we(g) also show genetic interactions withgarnet. This suggests that these “transport” group of genes are functionallylinked. I have interpreted this functional interaction as a macromolecularcomplex. Transmembrane permeases are typically large multi-subunitcomplexes but other possibilities for the physical nature of the interaction exist.The rosy gene shows a slightly different spectrum of interactions with garnet,and is the only member of this group which is known to encode an enzyme. Forthis reason, as well as its extensive participation in synthetic lethal interactions, Ihave postulated that the rosy gene performs a different, and possiblycontrolling, function. This function might conceivably be involved in metabolicfeedback. The phenotypes of null and extreme members of all of the transportgroup of genes resemble moderate mutations of the white gene. Thisphenotype suggests the function of these genes, or of the complex, is toenhance the activity of the transmembrane pore complex. I have indicated thatthis involves interactions between the gene products. It is, however, possiblethat these genes all act as transcriptional enhancers of the white gene. Finally,the action of this proposed complex may not be restricted to the plasmamembrane of the eye pigment cells. At least one of the enzymes necessary forommochrome pigment synthesis is found in the mitochondria (Sullivan, Gruband Kitos, 1974). Hence the resulting pigment precursor must be transportedacross the mitochondrial membrane. The accumulation of pigment in themitochondria of rosy mutants also implies the need for transport across thismembrane (Bonse 1967). And, as the pigment granules themselves are alsomembrane bound, transport across this membrane would also be necessary for114normal pigmentation. The synthetic lethal interactions between certainmembers of the transport group, as well as the synthetic sterility or garnet androsy, suggests that different members of the complex may associate in othertissues or at other times in development to perform slightly different functions.This transport system may in fact have a fairly ubiquitous distribution. The factthat the sole phenotype of mutants in the system is an alteration in eyepigmentation may be the results of functional redundancy within the transportgroup of eye colour genes and the requirement for massive amounts of pigmentand pigment intermediate transports, within a short time, for normal eyepigmentation. Thus while the eye cells may place the greatest demand on thistransport system, it does not mean that this system is confined to that tissue.The expression of the white gene is extensively regulated. Pairing dependentregulation of the white gene by the zeste gene product is well documented(Pirrotta 1988). Additional regulators have been described which appear toregulate not only white but also the brown and scarlet genes, also proposed toencode structural components of the transmembrane channel complex(Rabinow et a!. 1991, Birchler et al. 1994). These regulators all appear to act atthe transcriptional level. If the interaction between garnet and white involvespost-transcriptional regulation this will describe a novel form of regulation of thewhite gene. Testing of this model will require molecular analysis of the transportgroup of eye colour genes. As an initial step in this process, chapter 2 presentsphenotypic characterization of the different garnet mutants and preliminarymolecular characterization of this gene.115Chapter 2.Analysis of the garnet gene116Introduction-Chapter 2. Analysis of the garnet gene.Any attempt to test a model of the biological role of eye colour genes must restupon the detailed analysis of individual genes. The garnet gene is a member ofa particularly interesting class, the transport group of eye colour genes. Thesegenes are defined by alterations in the levels of both the ommochrome and thepteridine pigments. This phenotype likely stems from lesions in inter- and intracellular transport and communication. Chapter 1 presented genetic analysis ofthe interactions between the garnet gene and other eye colour genes, which ledto a simple model for the function of these genes. This chapter presents detailedphenotypic analysis of a number of garnet alleles, evidence for the cloning ofthis gene and preliminary molecular analysis. This work lays the necessarygroundwork to determine the function of garnet, to test the model presented inchapter 1, and leads to further studies on the garnet gene presented in chapterthree.117Results: Characterization of the garnet eye colour gene.Phenotype of garnet mutants:The garnet gene was originally described as an eye colour mutation and this isthe most obvious phenotype of the garnet mutants. Figure 14 shows thespectrum of eye colours found in a variety of garnet alleles. Eyes of garnetmutant flies range from the slightly brown colour of weak alleles through thebrowny-orange of the intermediate alleles to the pale orange of the extremeg53dallele. This latter allele is unusual that in that it displays a fine grainedmottling of red spots on a pale background as the flies age. Although thevarious garnet alleles have been used for many years, quantitative informationof the amounts of red and brown pigments exists for only a few garnet alleles(g1, g2, g3 and g4, Nolte 1959). Table 16 gives a quantitative estimate of thelevels of pteridine (red) pigments and the ommochrome (brown) pigmentxanthommatin for 18 garnet alleles using four different methods (comparisonsbetween these methods is discussed in more detail in Appendix 1). Nolte(1950, 1952b and 1959) gives values for the relative amount of red and brownpigments in the g1, g2, g3 and g4 alleles as 38, 15, 23 and 57 % (red pigments)and 56, 32, 47 and 23 % (brown pigment) respectively. He used a sequentialpigment extraction technique to obtain these values. It should be noted thatEphrussi and Harold (1944), as well as the controls presented in both thesepapers show that this technique is problematic. Nevertheless these data agreereasonably well with those shown in Table 16.118Figure 14. Spectrum of eye colour phenotypes of different alleles of the garnetgene.The top portion of the figure shows the phenotypes of wild-type (Canton S) andtwo garnet alleles, gP and g53d. The two garnet alleles shown represent twoextremes of the spectrum of eye colours due to mutations in the garnet gene.The g53d allele is the most extreme of the garnet alleles and has a pale orangeeye. The gP allele is a very weak allele. Flies bearing the gP mutation haveslightly browner eyes than wild type upon eclosion which then darken to wildtype within three days.The lower portion of the figure shows the spectrum of eye colours seen in thegarnet alleles, g1, g2, g3, g4, g53d, g5Oe, g61, gS3 and gP. A wild type male(Canton S) is shown in the center of the figure for comparison.119oIDetermination of pigment levels in garnet mutants.Table 16 presents values for total relative absorbance of the pteridine and theommochrome pigments, all expressed as percent of the appropriate wild typepigment levels. As there is only one ommochrome pigment in Drosophilamelanogaster, dihydroxanthommatin, the “brown pigment” column gives anadequate estimate of the amount of xanthommatin in these mutant strains.There are however, at least 28 different pteridine pigments and intermediates(Ferré et al. 1983, 1986). All of the pteridine pigment procedures used togenerate the values given in Table 16 measure principally the content of thedrosopterin pigment (Appendix 1). In order to assess the effect of thesemutants on at least some of the other pteridine pigments, the red pigments wereseparated by thin layer chromatography and the relative amount of eachpigment was determined. Adequate discrimination was obtained for only 9pigments. These results are shown in Table 17 and Figure 15. The garnetalleles tested generally decrease the amount of all the pigments although thevarious alleles have variable effects on the individual pigments. Quantitativeassessment of pigment levels was meaningful only for the drosopterin pigments(see appendix 1) but visual assessment of pigment levels generally supportthis conclusion. The g53d allele appears to be the most severely affected. Nonovel pigments or intermediates are detected in any of the garnet mutations.Ferré et al. (1983, 1986) examined pigments for 52 eye colour genes, the g1allele among them. The values they reported for the amount of these nine121Table 16. Pigment levels of various garnet alleles.The first column shows the garnet allele and sex of the mutant assayed. Thenext four columns give quantitative determination of the red and brown pigmentlevels. The first three values are measurements of total pteridine levels by threedifferent methods. The last column is a measurement of the brown pigmentxanthommatin. Pterins (5 heads) refers to a measurement of pteridine levels bythe 5-head microflourimeter method. Pterins (1 head) is measurement by thesingle head microflourimeter method and pterins (spec.) refers to measurementby spectoflourimeter. Details of the experimental method are given in thematerials and methods. Comparison between the methods for red pigmentdetermination is given in appendix 1.122Pigment levels of garnet allelesallele pterins pterins pterins ommochromes(5 heads) (1 head) (spec.) (xanthommatin)57±2 4±1 9±1 25±3g1 72±4 9±1 15±3 66±1g2 32±2 2±1 3±.3 NDg2 37±2 2±1 57±2 15±1g3 45±1 17±2 18±3 70±8cca 50±3 21±2 56±18 52±5g4 25±1 7±1 10±1 28±5g4 38±3 10±1 12±3 37±3g5Oe 24±2 1±1 3±4 NDg5Oe 33±3 3±1 35±4 19±4g53d 15±2 12±1 4±.5 13±2g53d 21±3 3±1 19±4 7±4gS3 43±4 23±3 26±3 41±5gS3 44±3 26±1 29±7 31±3e(g)g2 29±2 ND ND 15±9e(g)g2cf 32±3 1±1 34±3 4±3e(g)g50e 12±1 1±1 2±.3 13±2e(g)g5OecJ 15±1 1±1 17±2 4±2e(g)g53° 10±1 2±1 3±.3 6±2e(g)g53dcJ 12±1 2±1 12±2 4±2123S6-1 96±2 47±3 43±5 75±5S6-1 c? 97±6 75±5 82±18 77±6gP[P] 50±5 3±1 10±3 ND*gP-Rev1c 93±4 99±4 60±6 101±11gPReV2 93±4 78±5 86 ±22 NDgP-Rev3cJ 87±6 ND 88 ±20 78±9gPReV4 89±5 72±7 72 ±20 98±6gPO2 3±1 1±1 2±.5 9±1gPdcc? 60±3 44±3 54± 10 31 ±2gPX2’ 60±2 48±3 49±10 62±4124Figure 15. Chromatographic analysis of pteridine pigments of garnet alleles.Figure 15 shows some of the separated pteridine pigments. The various garnetmutants from which the pigments were isolated are shown below thechromatograms. The names of the pigments are shown on the left of the figure.In addition to the seven pigments identified, there were three unidentifiedpigments. Two of these migrate slowly in the solvent and fluoresce blue in ultraviolet light, the third has a greater mobility and fluoresces yellow. Theseproducts might also be degradation products of the pteridine pigmentsgenerated during isolation and chromatography.125blopter2-amino-4-hydroxypteridiflelsepiapterirunknownyellowspotMxanthopterin-N)isoxanthopterinunknownbluespot2unknownbluespot1drosopterinsresidueI......•.__________________________________________________________________________________________________________________________0cJ-1CDCDCD——CDcCDCDCDCDCDCD)CDQCi)CD((C)CDCDCDCD(C)Ci)0CI)-jJJD)C)CDCDCDCDDDtJDC))IIIIIIIII-‘——-CDCDCDCD4Q01-+Q(71I)F’3Q1ciCA)(-CA)0QQCD‘(D0-40-+oq-+040-+0isosepiapterin.ETable 17. Quantitative assessment of pteridine pigments afterchromatographic separation of pigments.The first column indicates the genotype of various garnet mutants. Theremaining columns show the percent of wild type levels of the indicatedpteridine pigment.127Effectofgarnetallelesonpteridinepigmentspigmentresiduedroso-mysterymysteryisoxanthyellowxantho-sepia-2-NH4-biopterinisosepiapterinspot1spot2o-pterinspotpterinpterin4-OH-pterinpteridinealleleg15800000033015g1f24180000000023g31646000202000054g3cf26330000000015g438000000000g4f1380000000015g5Oe0014016743129267167086g5Oe01000001330057g53d23000147110037100121128g53d0100000330058gS32653000000008gS3f18520000000023gP,’189000000000S6-11754000200500073S6-11061080002020500064gP-R1200920004060500027gP-R3531170000000038gP-R4112970004020500036gP-O2010000000036gP-dc1627600000000145gP-X2f71770050206000018pigments generally agree with the results here, although their method mayprovide for greater accuracy.Malpighian tubule phenotype of garnet alleles.There are four pigmented structures in the adult fly body, the eye, ommatidium,the malpighian tubules and the testes sheath. Beadle (1937a, 1937b) hasdescribed the colour of g2 malpighian tubules. Breme and Demerec (1942)examined the malpighian tubule colour of 25 different eye colour mutations.Their survey included four garnet alleles, g’, g2, g3 and g4. Table 18 showsthe results of a similar survey extended to include an additional 13 garnetalleles. The effect of the two garnet alleles surveyed by Breme and Demerec(1942) on pigmentation agree with my estimate of pigmentation, however, theresults from the control mutations used to compare colour suggest that theywere somewhat more discriminating than I in detecting very low levels ofpigmentation. All of the garnet mutations reduce pigment deposition inmalpighian tubules and, in general, the reduction in colour is more extremethan observed in the eye. It is not possible to distinguish the colour of themalpighian tubules of most of the more extreme alleles, as they all appearessentially colourless even though their phenotypes in the eye are readilydistinguishable. In contrast, the subliminal allele gX, which is associated with aninversion, shows a weak variegated pigment phenotype in the malpighiantubules although the eye phenotype is visually indistinguishable from wild type(data not shown). It is also interesting that the gP revertants (discussed below)do not appear to be complete revertants based on malpighian tubulephenotype.129Table 18. Effect of various garnet alleles on colour of malpighian tubules.The first column shows the garnet allele. The second and third give the colour ofthe malpighian tubules as determined visually or where reported in theliterature (Breme and Demerec 1942).130Survey of malpighian tubule colour of garnet allelesmutant colour reported colourCS orange bright yellowcar very pale/clear pale yellowcm colourless very pale yelloww colourless colourlessg1 colourless very pale yellowg2 very pale/clear very pale yellowg3 colourlessvery pale/clearg5Oe very pale/clearg53d colourlessg61 very pale/clearg271 yellowgEMS pale yellowgS3 colourlessgX orange/variegated?e(g)g2 colourlesse(g)g5Oe colourlesse(g) g53d colourlessgP pale yellow131gPReV1 yellowgPReV2 yellow/orangegPReV3 yellowgPReV4 yellowS6-1 orangeO.R orange132Table 19. Effect of various garnet alleles on testes sheath colour.The first column shows the garnet alleles. The second indicates the colour ofthe testes sheath immediately after dissection.133Survey of testes sheath colour of garnet allelesmutant colourCS bright yellowg1 pale yellowg2 pale yellowg3 yellowg4 yellowg5Oe very pale yellowg53d bright yellowg61 very pale yellowg271 bright yellowgEMS pale yellowgS3 pale yellowgX bright yellowe(g) g2 pale yellowe(g)g5Qe very pale yellowe(g) g53d yellowgP yellowgPReV1 bright yellowgPReV2 bright yellowgPReV3 bright yellowgPReV4 bright yellowS6-1 bright yellowOR bright yellow134Testes sheath phenotype of garnet alleles.In general, all the mutants decreased the pigmentation of the testes sheath(Table 19). The effect of the different alleles on testes sheath colour wasgenerally more severe than the effect on the eye pigmentation but less severethan the effect on malpighian tubule colour. Theg53dallele provides the oneexception. Three different strains carrying the g53d mutation all showedessentially wild-type levels of testes sheath pigmentation. This suggests atissue specific pattern of expression in this allele.The search for a garnet null allele.As observed above, the g53dallele confers the least pigmentation of all thegarnet alleles. In order to determine whether theg53dor any other allele mightbe a null allele of the garnet gene, I quantitated the effect on eye pigmentationfor different garnet alleles when heterozygous with a deficiency. Table 20shows the results of red pigment levels of 10 garnet alleles when combined witha deficiency that includes the garnet locus. The phenotype of the g53d allele isthe most severe. Comparison between the pigment level of a homozygous g53dstrain (Table 16) and the hemizygous g53d condition (Table 20) shows nopronounced difference. Thus, by the criterion of Muller (1936), the g53d allelebehaves as an amorphic allele with regard to eye pigmentation (this conclusionis also supported by Northern analysis of the g53d allele - Figure 22).Comparison of the pigment levels of other garnet alleles as homozygotes orhemizygotes suggest that these alleles are all hypomorphic lesions.135Table 20. The phenotype of various garnet alleles in combination with adeficiency.The first column shows the garnet allele. The next two columns show thepteridine pigment values of the appropriate hemizygous garnet allele, asdetermined by microflourimetric assay (5 heads) and by spectrophotimetricassay respectively.CROSS:P Df(1)HA97/FM7 ® g*/jFF“Fl Df(1)HA97/g*progeny assayed.where g* indicates the given allele of garnet.136Pigment levels of various garnet alleles in combination with a deficiencyallele percent wild type red pigmentmicroflourimeter spectrophotometerg1 40±3 8±4g2 33±3 4±1g3 40±3 9±3g4 32±4 6±1g5Oe 28±2 5±2g53d 16±3 3±1g61 34±3 7±3gP 38±4 8±3gS2 34±4 8±1gS3 38±3 10±2137Aging and garnet.Shephard et a!. (1989) have reported that a null mutation for the rosy gene hasa decreased adulte life span. Furthermore, Hilliker et al. (1992) have reportedthat mutants of the rosy and maroon-like genes show hypersensitivity to oxygenstress in addition to a decreased life span and Humphreys, Duyf, Hilliker andPhillips have isolated an allele of the pink gene, another member of thetransport group of eye colour mutations, in a screen for mutations that arehypersensitive to the free radical generating chemical paraquat (J. Humphreys,personal communication). Since, longevity should be a sensitive assay forbiochemically uncharacterized lesions that effect cell viability, I tested thelongevity of three of the most severe garnet alleles, with and without the e(g)mutation, at two temperatures. Figure 16 A and B shows the longevity (%survival) of wild type and different garnet genotypes at 22° and 29°. Some ofthe garnet mutant strains do die earlier than wild type. But reduced vigor is notparticularly surprising in a highly inbred population of flies. More significantly,the relative longevity of the various garnet mutant alleles differs at 22° and 29°and the presence of the e(g) mutation, which profoundly reduces the level ofpigmentation, has no appreciable effect on the time of death or shape of thedeath curve. Finally flies mutant for the garnet gene lived longer than the wildtype strains in some instances (at 29° g53d and at 22°, all butg53doutlivedthe wild type flies of the appropriate sex). Thus there is no evidence that garnetaffects longevity.138Figure 16. Adult life span of wild type and various garnet mutants.A. Longevity of wild type (Canton S) and g2, g5Oe and g53d females and malesis shown at 22° . Minimum population size is 90 for each genotype.B. Longevity of wild type (Canton S) and g2, e(g) g2, g5Oe e(g) g5Oe, g53d ande(g) g53d females and males is shown for at 29°. Minimum population size is30 for each genotype.The key for the different genotypes is shown on the graph.The 29° aging experiment and the first portion of the 22° experiment was doneby Layne Harvey as part of a directed studies project.139day—.--e—6--H--CSfemaleCSmaleg53dfemaleg53dmaleg5oefemaleg5Oemaleg2femaleg2maleLifespanofgarnetmutants22°1I CC > U) C G) a) 0.—•—CSfern.-R-CSrn.—rne-—g53df.—a--g53dm—e—g5Oefg5Oem—K—g2f—-g2m—v—e(g)g53df—0—e(g)g53dm—+--e(g)g5Oef—+—e(g)g5Oem—+--e(g)g2f—.*—e(g)g2mLifespanofgarnetmutants29°> 0 C)dayCloning of the garnet gene:Isolation of unstable garnet mutants from dysgenic crosses.The original gP allele was isolated from crosses involving a natural P-elementbearing strain, S6-1 (Wennberg 1988). This strain has two P-elements insection 12 of the polytene chromosome, the cytological location of the garnetgene (Wennberg 1988). Weak garnet mutations arise frequently in this strain(Figure 17). The relatively high frequency of garnet mutations arising from thisstrain is likely due to the proximity of P-elements to the garnet gene (Towers eta!. 1993). These weak P-element induced mutations, as well as the original gPallele, remain active and will further mutate to a variety of garnet phenotypesupon out crossing to non-P bearing strains. Figure 18 shows the frequencywith which new, secondary and tertiary garnet mutations occur. The secondarygarnet mutations are also subject to further mutation upon outcrossing to non-Pbearing strains, although at a much reduced rate. Figure 18B summarizesextensive lineage and out crossing experiments with these strains. Only theoriginal gP mutation, four wild-type and four secondary mutations which werederived from the original gP allele, have been analyzed at the molecular level.Cloning of the garnet gene:Cloning of the garnet gene was facilitated by the original gP mutation isolatedby R. Wennberg (Wennberg 1988). The garnet gene was cloned by Dr. D.Sinclair. Details of the cloning are provided in Appendix 2. Briefly, P-elementcontaining clones were isolated from a size fractionated library made from thegP mutation. One such clone hybridized to position 1 2C, the cytological position142Figure 17. Diagram of the mutation rate to garnet upon outcrossing the S6-1strain.A. The top part of this figure shows the lineage of garnet mutants generated byoutcrossing the P-element containing strain S6-1. The number and phenotypesof the progeny are shown for five generation of outcrossing. The phenotypesare given relative to the garnet phenotype, g indicates wild type eye colour, gPindicates a weak garnet mutation similar to the original gP allele and gM orgMOd indicates a moderate garnet eye phenotype.B. The rate of conversion and interconversion to the various garnet phenotypesis summarized.143Generation of garnet mutants in the S6-1 background24 “+“18 “+“A. sterileterileF4 F5_..-31 •‘+ (gP) ll.-40 gP/ terile (3.2.2+,4l_409+ all—30g+(437÷ (gP) afl.-4OgP31 -i-_______/ A 44 •• ‘all—4OgP - all.-4OgPsterile /,‘all.-40 gP.p(3.1—aP( __43 2÷Y3lgPg+,1 gModl-4OgPterile3 gP 66 gP .afl-4O gP(3.219?)(4a94.+ gMod/8i 9p (g.Ste Og112g÷allgPI / /\. 1/42 gP 69 gP ‘i_’all_gMod 1 g÷aIl.-gMod30 ÷/ 4 40 +‘• all—gModI, / “‘ ‘ sterile \\all..gMod_Øi.-lgMod lg÷/—4OgP, / 39 ÷ Øll.-40 gP Z*sterilep./ •4( 33+. 0 +/—4OgP allgP37S6-1 C(1)DX [MI V..V24 +“6p.p24gP(437 gPstock — 44 +, g÷/—40 gPFl ‘Zç38 “+“ 48“÷. +“ g+/.40 gPall -i-’ \‘1J2S9P 26 +(4 26‘+ sterilec22 ÷“S... 33\ 45+\ 49 .,+\ \33+‘, \ 48+\cl/5gX—‘.sterile6+48÷4sterite(8.1=gP)43+\c26 ÷ gP = weak garnet mutationsterile—0—48 ÷ gM = moderate garnet mutation_p-84 +50 “+(4 42 +‘ “+ garnet revertantcsterilesterileB Summary of garnet mutation rate in S6-1 background.0.6-2%S6-1 g”P4P g+gM144Figure 18. Diagram of the mutation rate and phenotypes of garnet mutationsderived from the original gP allele.A. The top part of this figure shows the lineage of garnet mutants generated byoutcrossing the gP mutation. The number and phenotypes of the progeny areshown for six generation of outcrossing. The phenotypes are againoperationally defined relative to the garnet phenotype, g indicates wild typeeye colour, gP indicates a weak garnet mutation similar to the original gP allele,gM indicates a moderate garnet eye phenotype and gX indicates an extreme orstrong garnet phenotype.B. The rate of conversion and interconversion to the various garnet phenotypesis summarized.145Generation of new garnet mutants from the original gP alleleF3 F41 gXI io—P Øall 43X2gXL_L...iE. sterile2gX/A.___________I27gPF2 439pJ 32 gP all40+gXsterile — _,all 40+9PI,,,/19X lgMod/ 21.r2gX,—2OgP,--2Og+3 gf’ 2gX/40+gPall 40-fgX,I I ‘- 1 gL3 .all 40+gXsterilec40-allI / steril*terilalI 40+gP/ I38gP40÷g÷Fl 24gP ,1 2gMod/ 34j4.g45 gP 10+9+! 32 gI-’9P c(x M] 49 gP / 10+9+!most6 9P / / sterile___som4 gP 37 gP 40+aM0 gP 54 gP — _240+g+C(1)DX g+ 09P\[P1 some gX ‘4OgP’’(stock) (discarded)33gf\\ 3g+48g9gp 1X1g+/40+gP‘ I 4o+gP2g lgXI4all36 gP steri e —20g+,-.2OgPI’ ‘10+M/ 309 all 40+gX ..20g+....2OgP“ ‘ 40+9+ ‘: steriLsterilesterile ,.- all 40+gM stenreX140+ciM,io+gif ,all 40+gMF2 ‘ all 4oi-gM10+g+.10+gMI 40+9M38 yP Il 40+gP all 40+gM36 gP lr4o+9p all 40+gM9.2 (+ 3g+/ 39 gIll 40+gP all 40+gM\479ll4O+aP 19X/40+9M9.3 (+) 59 gP gX!40+çP ,1gXJ40+gPlg+,-.2OgX,-.2OgP2aXI 35 gall 40+gPB.40+gPSummary of interconversion rates 33 g1gX,2gP/40+g+1 g+!40+gPof garnet mutants 45sterile .-20g+,-2OgP1 3 gP 1 X3g+/40 all 40+gP! 0 30/ a? 4b+gP sterile20a+ -.2OgPg+ . 0 all 40+gP gP = weak garnet alleleall 40’-i-gPgP3 2% gM\all4O+gP1%all 40+gPall 40+gP gM= moderate garent allelesterile gX = extreme garent allele9+ = garnet revertantgX146F5F3 F4 F6•_JpralI40+gX__________________________—P all 43X ——i- 1 gM/50+clX‘ sterl______________sterile all 40+9X — all 50+gXi5ja.1grn3all40+gM .. 2gP,9gX/50+OM2gX2g!,gMall 40+gX ________________ all 50+gM27 0P terile II 40+43 10 ,49X/40+OP II 40+gigX 1g+/40+OPcsterile II 40÷9Pall 50+ gXall 50+ 9P—p-4o-Xall 50+ gXsterile 2gX -.20 P -20+lgM/50+gXigall 50+ gX-b40+q* ..40+(JX ,ll 40-i gX______________________all 50+ gX40+sliall 40+gP -/50 i-9Xall 50+ gX2gX1g40+gX O+gXall 50+ gX\38 gPall 50+ Ox2gMod/Psterile1 0-i-g+/_all 50+ gMall 50+ 9M1O-i-g-i437 0P_ 40+i+ 1140+92:gP+____all 50+ gMc::: _zE39 gall 50+ gP36 gterile all 40agX,10g40+gall 50÷ 0P2g+/50+gPic X/40+9PqX/40+yMall 50+ çjX0+g+/ 4AP I 40+rjMall 40+0P-_..10irJ+ 10+M/ cjPc-*.41 9P40+rJMMul40.IgtVi40+gMM140÷0X all.,0÷gXX/40+g 0X140+9P gM2g+/40+gX38 gP 1140UPll40+pP1_______36pP 40+ all 40+M all 40+9M 0M/50+9XcSO9P X/40+c40÷ Ox all 50+ Ox39 g II 40-i-gP all 40-i-gM33 OP 10X,2gP/40+ + lg+/40+9P 1l40+gPall 40+gP2Xl4xallzi+gP-gxg÷/ 22 9Psterile -.20i-,—20 aall 50-i-gMall aI63all 40+gPall 40-i-gP_ _ __ _all 40+9 Pall 40-i-9P iOgP/50+g+all 40+gPsterile147of the garnet gene. Unique sequence DNA flanking the P element was used toisolate lambda clones containing inserts of wild type genomic DNA. Theseinserts shared a 6.6 Eco RI fragment. This fragment was subcloned and used toisolate c-DNA clones from an imaginal disc library.Four lines of evidence indicate that this 6.6 Eco RI fragment identifies at leastpart of the garnet gene: The spontaneous gP mutant has an insertion into thisfragment, the size of which is altered in revertants. The P-element in the gPmutant is inserted into a region corresponding to a 3’ intron of a transcript fromthis region. Nine, out of fourteen garnet mutants examined, have alterations,detectable by Southern analysis, in this fragment. Finally, of the two garnetmutants examined by Northern analysis both show alteration in the twotranscripts derived from this region. These data are discussed below.Molecular analysis of the gP allele and its derivatives. The gP mutation isassociated with a P-element insertion into the garnet gene. The P-element thatis inserted in the gP allele is approximately 2Kb in length (Figure 19). Thus itis not a complete P-factor. The four revertants of the gP allele all showalterations in the size of the P element inserted as do the spontaneous garnetP-extreme mutations (Figure 19). Neither the g-revertants nor the gP...extreme alleles are associated with complete loss of the original P-elementinsert. Thus while the revertants are wild type (based on a visual assessment ofeye colour) the P-element insert is not completely removed in any of them. Thisincomplete molecular reversion is consistent with incomplete phenotypicreversion as assessed by pigmentation of the malpighian tubules. Since the2kb insert in the gP allele causes only a moderate reduction in the function of148Figure 19. Southern analysis of the gP allele and gP derivative mutations.The top portion of this figure shows the results of Southern analysis of the gPmutation and four phenotypically wild type revertants. DNA from wild type(Canton S), the gP allele and four revertants was isolated, restricted with Eco RI,separated on a 0.8% agarose gel, transferred to nylon membrane and probedwith the 6.6 Eco RI putative garnet fragment. The wild type DNA shows a bandat 6.6 kb as expected, as well as a bands at higher molecular weight likely dueto incomplete digestion. The size of the 6.6 kb band is increased to 8.2 kb in thegP mutation. Similar analysis with DNA restricted with Bam HI and Hind III showthat this change in mobility is not due to polymorphism for an Eco RI site (datanot shown). That this change in mobility is caused by insertion of a P-element isshown by hybridization of this band to P-element DNA (data not shown). Thesize of the 6.6 kb Eco RI band is also altered in each of the gP revertants. Inevery case the size of the insert in the gP allele is diminished but in no case isthe insert completely removed.The two lower figures show the results of Southern analysis of morespontaneous derivatives of the gP allele. In this case the membrane wasprobed with the 4 kb imaginal disc c-DNA clone. The gP’X2, gPX3, gPX5 andgPO2alleles are moderate or strong garnet alleles. The gPX2dC allele is asubliminal garnet allele (males and homozygous females appear wild type butfemales heterozygous with a strong garnet allele show a weak mutantphenotype) which occurred spontaneously in the gPX2strain. With thepossible exception of the gPX2 allele each of these gP derivatives showalterations in the size of the gP insert. The gPX5and gPO2alleles mightrepresent small deletions in the garnet gene due to imprecise excision of the Pelement.14982kb66k*Cs gP gP—Rev2 gP-Rev3 gPRev4 gPRevlCs gP 9P-X2 gP-X3gP.X5 gP-dC gP-02—PX2-dc865kb 65-150the garnet gene (based on the weak mutant phenotype of this allele) it ispossible that internal deletions of the P-element might relieve whateverimpediment this original insertion produced.Sequence extending outwards from the termini of the P-element in the clonedgarnet region of the gP mutation identified the position of the insert ascorresponding to the most 3’ intron at position 3234 of the imaginal disc c-DNAfrom the garnet region (see below).Southern analysis of other garnet alleles.Southern analysis was performed on wild type and fourteen garnet mutants todetermine the presence and nature of lesions in the 6.6 kb Eco RI putativegarnet fragment. Genomic DNA from wild type (C.S) and g1 mutants wasdigested with Eco RI, Barn HI and Hind Ill, size fractionated and probed with the6.6kb Eco RI fragment (Figure 20) and the 4kb imaginal disc c-DNA (data notshown). Southern blots of wild type DNA digested with Eco RI and probed withthe 6.6 kb Eco RI putative garnet fragment yield the expected 6.6kb fragment.Barn HI digestion of wild type DNA gives two fragments one roughly 8kb and theother approximately 10 kb. Only the smaller of these is detected by the c-DNAprobe (data not shown) indicating that the c-DNA is encoded exclusively by thesequences to the left of the Barn HI site of the 6.6kb Eco RI fragment. Hind IIIdigestion of wild type DNA gives two fragments detected by the 6.6kb Eco RIsequence as a probe, one of 4kb and one of 2.5 kb. Only this 2.5 kb fragment isdetected using the c-DNA as a probe (data not shown). Alterations in the size ofrestriction fragments indicates that the g1 mutation is due to an insertion ofapproximately 2kb into this smaller 2.5kb Hind III fragment. Less detailed151Figure 20. Southern analysis of the wild type garnet region and the g’ alleleWild type (Canton S) DNA from the garnet region was restricted with Eco RI,Barn Hi and Hind lii and probed with a portion of the 6.6 Eco RI fragment.Similar analysis was performed on the g1 allele. In the g1 allele mobility of aband is diminished in each of the digests indicating the presence of an insertionin this mutation. Of note, the mobility of the smaller Hind III fragment, whichcorresponds to the 3’ end of the transcribed region, is altered in the g’ allele.A schematic diagram of the restriction map of the garnet gene and segmentused as a probe is shown in below.152g1 CsHE B H E B‘p.41248.47.44.43g1 (&gqB H RR\H HABL1 ii\AiiRHI I5’1 kb probeB153Figure 21. Southern analysis of garnet mutants.The top portion of this figure shows Southern analysis of gl, g2, g3, g4,g53d,g5Oe and gS3. The lower portion of the figure shows Southern analysis of g61,gEMS, gim, gS2, gXand T(1;Y)B166. DNA from these flies was isolated,restricted with Eco RI, separated on a 1% agarose gel, transferred to nylonmembrane and probed with the 6.6 Eco RI garnet fragment. The g1 and g3alleles appear to have identical insertions into this fragment. The g4,g61,gEMS and gim also have insertions into the 6.6 Eco RI fragment. The g53 alleleappear to have a small deletion in the 6.6 Eco RI fragment. Theg53d,g5Oe,gS2and gX alleles show no change in the mobility of the 6.6 Eco RI fragment.The g2 allele shows altered mobility of the Eco RI fragment. That this is due to arestriction fragment length polymorphism of the 3’ Eco RI site, is shown byrestriction analysis of the g2 allele with other restriction enzymes, such as BamHI, shown at the lower right. Finally, the translocation B 166, shown as aheterozygote with g5Oe, is broken within the 6.6 Eco RI fragment.154a) V— Ipco—i’VU,Co ciT41 4‘CoCo14Co U’0.Co C’,analysis was pertormed to identify the nature of 12 other garnet alleles (Figure21).Genomic DNA from g2, g3, g4,g5Oe, g53d, g61, gEMS, gim, gS2, gS3 andT(1;y)B 166 mutants was restricted with Eco RI, separated by electrophoresis,Southern blotted and probed with the 6.6 kb Eco RI fragment. Of the 13 garnetalleles examined in this way, six show insertions into this fragment, one has adeletion, one a translocation break and five show no structural alterationsdetectable by Southern analysis. Interestingly, Eco RI digest of the g’ and g3alleles generated apparently identically sized fragments. This implies that thesealleles possess identically sized insertions into the same region of the garnetgene sequence. It is also interesting that the g2,g5Oe and g53dalleles whichare the most extreme alleles and also those sensitive to the enhancer of garnetmutation, showed no structural alterations in garnet. These data aresummarized in Table 21.Analysis of transcripts arising from the 6.6kb Eco RI putative garnet region.This 6.6Kb. Eco RI fragment was used as a probe to investigate the size andabundance of messages transcribed from this region. Figure 22 showsmessages detected by this probe from wild type (Oregon R) embryos and g3and g53d newly eclosed adults. Two messages are detected with this probe inwild type embryos, one approximately 3.5 and the other 4 kb in length. Both ofthese messages are absent from adults of the severe allele, g53d, and onlyone, possibly slightly larger than the 4kb transcript is present, at reduced levelsin the g3 adults. The absence of both messages from the g53d allele suggeststhat both messages are derived from the garnet gene and that no other m-RNAsare transcribed from this region in embryos or adults.156Table 21. Summary of lesions in the garnet locus in different garnet alleles.The first column indicates the garnet allele. The second and third columnsshow, respectively, the type of lesion and where relevant, the approximate size.(NA = not applicable.)157Summary of lesions at the garnet locus in different garnet allelesallele type of lesion sizeg1 insertion 4kbg2 point mutation? NAg3 insertion 4kbinsertion 1 kbg5Oe point mutation? NAg53d point mutation? NAg61 insertion 3kbgEMS insertion 2kbgim insertion 1 kbgP insertion 2kbgPReV1 insertion .5kbgPReV2 insertion 1.5kbgPReV3 insertion 1 kbgPReV4 insertion 1.5kbgS2 point mutation? NAgS3 deletion 0.5kbgX point mutation?/position effect?T(1;Y)B 166 translocation breakpoint in garnet158Figure 22. Northern analysis of wild type embryos and garnet mutant adults.RNA from adult (0-3 days) g3 and g53dindividuals and wild type (Oregon R)embryos was isolated, electrophoretically separated, transferred to nylonmembrane and probed with the garnet imaginal disc c-DNA. There are twomessages in wild type embryos. One message, perhaps slightly greater thanthe wild type 4 kb message is seen in g3 adults. No messages are seen in g53dadults. The lower figure shows the same four lanes probed with the RP 49 geneas a control for equal loading and RNA integrity.e=embryoa=adult159ICo 0.(DIPcDAA0•DIn summary, the presence of a P-element insert in the 6.6 Eco RI fragment in thegP mutation, which is altered in garnet-P revertants, the number of garnetmutants with alterations occurring in this same fragment and the disruption ofthe transcripts from this region in garnet mutants all support the contention thatthis fragment encodes at least part of the garnet gene.Molecular analysis of the garnet gene.Restriction map of garnet and the genomic region encompassing garnet.The unique sequence DNA flanking the insertion site of the P-element of the gPmutant was used to isolate lambda phage containing inserts of the homologousgenomic regions of wild type Drosophila melanogaster. Five different lambdaphage were isolated and subjected to restriction analysis using Barn HI, EcoRl,Hind Ill and Sal I to generate a crude map of the region surrounding the garnetgene (Figure 23). A summary of these restriction maps is shown in Figure24. The map includes approximately 30 kb 5’ and 15 kb 3’ to the garnet gene.The phage all held a 6.6 Kb Eco RI fragment in common. This fragment wassubcloned and subjected to more detailed restriction and sequence analysis.Detailed restriction map of the garnet region:A restriction map of the 6.6Kb Eco RI fragment using the enzymes, Eco RI, BarnHI, Sal I, Hind III and Kpn I is shown in Figure 24. Below this figure is shown arestriction map of the imaginal disc c-DNA clone isolated using this 6.6Kb EcoRI fragment as a probe. The restriction analysis of the c-DNA also include therestriction enzymes Sst I, Pst I, Sph I, and CIa I.161Figure 23. Restriction fragment analysis of lambda phage clones containingthe garnet gene and flanking sequences.Restriction analysis of five lambda phages which were isolated from a genomiclibrary screened with single copy sequences adjacent to the P-elementresponsible for the gP mutation. The five lambda clones, g212, g?13, g?14,g22O and gA21 were digested with the restriction enzymes, Barn HI, Eco RI,Hind Ill and Sal I, single and in pairwise combinations. The DNA fragmentswere electrophoretically separated, transferred to nylon membrane and probedwith the same g?. DNA. The information derived from these restriction digests aswell as from probing these blots with portions of the garnet gene (data notshown) was used to generate the restriction map shown in Figure 24.16248..(0 r%3II • •I.• .,*•11*1(0 .••fl•4II0IjISfI II941SI *Io1—•,.J•Figure 24. Restriction map of garnet and surrounding region.A. The top of this figure shows the extent of the five lambda phage clones whichencompasses the garnet region. Below is a crude restriction map of this region.The 6.6 Eco RI fragment which includes the garnet transcript is indicated by aheavier bar.B. The 6.6 Eco RI fragment which identifies the garnet gene is shown ingreater detail. The site of the P-element insertion in the gP mutant, as well asthe location of three introns, the site of the putative hydrophobic domains and ofthe polyglutamine repeats is shown. The arrow below indicates the location anddirection of transcription for the 4 kb c-DNA clone isolated from the imaginaldisc library as well as additional restriction enzyme sites. The sequence of themost 3’ intron into which the P-element is inserted in the gP allele isTATGCCGCGATNTTTGNANNATCGAAGAGTATG*TTCCAG where theindicates the insertion point of the P-element.164A.Restriction map of garnet and surrounding regionR R Sst SsçxbaH S SstI I II I I II5’ R=EcoRI CIa=CialH=HindIIl Sst=SstIB=BamHI Ssp=SsplS=SaIl Xba=XbaIgx21 gM4 —g2O —gM3 —RB B BRH HBBRII I III II U5kb///R/Inn///gx129-SRRH BSHSR Ffl-IS1BB RHBHSRBHRSRHBSR BBRliii II 11111 III 11111 I I liii liii liii I III// \/////// \/ \/\\\\\\HB B RELiBIII I/////HSI I= I •I IBHOClaCla CIa RIII H3’hydrophobic domainsintronspolyglutamine region1kb165Sequence analysis of the garnet gene. The DNA sequence was determined fora 1.2 kb DNA fragment isolated from an embryonic library, a 4kb imaginal discc-DNA, and part of the 6.6Kb EcoRl genomic fragment which encompasses thebulk of the garnet gene. The sequencing strategy used to sequence theimaginal c-DNA and the equivalent section of genomic DNA is presented inFigure 25.Sequence analysis of the approximately 1.2 kb DNA fragment derived from theembryonic c-DNA library showed that, relative to the imaginal disc c-DNAsequence, the 3’ end of this DNA segment is in the intron at position 3234 andthe 5’ end, marked by a poly C tail, occurs 3’ to the disc c-DNA, presumably inthe genomic region flanking the coding region (data not shown). As this DNAfragment encompasses both intronic and genomic sequences, lacks asubstantial open reading frame, and is in the opposite orientation to theimaginal disc-derived c-DNA, it seems likely that this fragment is a genomiccontaminant in the c-DNA library. This explanation is consistent with theabsence of a 1kb message in embryos (Figure 22).The sequence of the 4kb c-DNA isolated from a 3rd instar imaginal disc libraryis shown in Figure 26. The sequence is 3825 bp in length. There is an ATG atposition 297 which could initiate a potential polypeptide of 1054 amino acids.No poly A tail is present in this c-DNA clone however the size indicates that itmight nevertheless be complete.Comparison between the sequence of the genomic and imaginal c-DNArevealed that the c-DNA extends 221 bp 5’ to the 6.6 Eco RI genomic fragment.Thus the genomic fragment does not contain all of the imaginal c-DNA.166Figure 25. Strategy used to sequence the imaginal c-DNA clone of the garnetgene.This figure diagrams the segments of the imaginal c-DNA clone that weresequenced. The thin scale bars represent the size, in base pairs, of the imaginalc-DNA sequence. The underlying bars represent independent sequencedetermination where the location of the bars indicates which section wassequenced, the arrow indicates the strand sequenced and the thickness of thebar represents the number of independent sequence determinations of thissegment. The solid bars indicate c-DNA sequence, the stippled bars indicatesequences from the corresponding genomic region. Introns are shown astriangles above the scale bar, in the appropriate location.167Sequencing stratagy for the garnet geneI liii III II I 111111111 I liii I I I II0 100 200I 111111111 I300 jooI ‘‘‘I ‘‘I IJ liii III I400 500IF 1 I I - [ 11J700 800800 900I I 11111100011111 11111111111100 120041111111111 III1200 1300I III II I liii1400liii I liii1500—II I I I I- 16001111111111 III2000 210011111 111111111 I2300 2400r ri i iii i i i2800 2900I liii III II3000I III I 11111 I3100 3200I 1111111113200 - np 34n0I I III 11111 I::spo 600‘ I3900 — lx—— genomic — 2X— c-DNA 3XI I I I I - - I - I600*1IIIIIIIIII 1111111111 IIIIIIIII•I 111111111 I1600 1700 1800 1900 2000I III II I 1111122001111111111 III2400 2500E;.*IIøI11111 —I1111111 I600I liii III I I I2700 2800I till ill II I till ‘‘‘ I I I I III3600 3700 3800168Figure 26. Sequence and conceptual translation of the imaginal disc c-DNAclone of the garnet gene.The sequence of the imaginal c-DNA clone from the garnet gene is shown. Thesequence is 3825 bp long. Below the sequence is shown the conceptualtranslation of the imaginal c-DNA clone. A long open reading frame starts at anATG at position 297 which could encode a polypeptide of 1054 amino acids.The direction of transcription was determined by tissue in situ hybridization withRNA probes (shown in Figure 27). This putative polypeptide has threeindifferent hydrophobic domains and a stretch of polyglutamine residues. Aschematic summary of the sequence motifs found in this putative polypeptide isshown in Figure 24. The genebank accession number for this sequence isU31 351.169garnet Sequence10 20 30 40 50CGGAAAACCC GAAATACGCT CAAGTCACCG GCTTTAAATC AATCATCCGGTTTAATGTAG CTGAACGTTT TGCCTAACGC TAATGACAAA GAATCGCAAGCACCAAGCAC CAAAAAAAAG TTAATGATTG GACTCGTGAT TAATATGAGTGTTCTAGGTA TTCTTCCGGA ATTCGTACTG TAAGGAGATT TATAAAAATACAATTAGCCT TTTTTAATTC TGCGCCAGCT TTCAGCTCGT TGATTACATGM100150200250300TCCAATTCAASNSMTGTCCAGGCTSRLCGATGCGGATDADCATTTGGATCHLDRGTATTTGCCAICH350TCCTTGTCCTPCPCGCTGTCCGARCPTCATGGAGACAWRHCTTAGGGCATLGHGTCCAGCGTGVQRG400GTGTTCACAAVHNTGTGCAATGG TTTCGGTTCAVQW FRFTCCTTCCGATTFRFCGGAGCGCTGGAL450TTGTATGCCALYATCTTTATTTTT GGACCTTTTTL FL OLEGCTCTTTTTAAL FTCCCTTTTTATLFM500GCTTCTTCTT GCCATCCTTG CTGCCCTGCG CCGCCTGCTG TTCCTGCAGGL L L A I L A ALR RLL FLQV550TACTTATCCGLIRACCGTTTGGT GATGCCCACGPFG DAHGGCCAGAGCGGQSGCCACACCCTCHTL6001234567890AAATTCCGCG1234567890 1234567890ACCCCGTCCT CTCGGGAATC1234567890CGTCATAGCT1234567890GAGTATTCAC 50170garnet Sequence10 20 30 40 501234567890 1234567890 1234567890 1234567890 1234567890CATGTCCAGT GGCAGCTCCG TAATGGGTAT GTCATCAATA TTGTCATACT 650HVQW QLR NGY VINI VILGATCCGCATT CGAGGCACCA CGTCCGTGGG GCGTGGACTT GAGGTAGTGA 700IRI RGTT SVG RGL EVVRGGATTGTTCG ACTGCTCAAT GAGACGAGCC ATCTGGCCGC TCCAGTTGCT 750IVR LLN ETSH LAA PVACGGGGCGTGA GCTCCCAGAT GCTGCGCAAC CAGCTGTCCG CGTCGACCGA 800RGVS SQM LRN QLSA STDTGCCATGGCC ATGGACACGA CGACGGAGGG CGGCATTCCG GTGGCCATTG 850AMA MDTT TEG GIP VAlEAGATTGTCCA GGAGATGACG CTGCTGTTCA CAGGCGAGCT AATACCGGTG 900IVQ EMT LLFT GEL IPVGCGCCCAAGG AACCGGCATG CCCCTGCCAG ATGGTCTCGA TCTGGA(i IF 950APKE PAC PCQ MVSI WTLGAGTGGATTA ATGCACCGCC ACCGGAAGAT GCGGCAACAG AGTTCATCCT 1000SGL MHRH RKM RQQ SSSSCGGAGCACGA CAAGGACGCA GCTATTCGGT TAGTGCCACC CGAGGCAGGA 1050EHD KDA AIRL VPP EAGACGGGGAGCT GCATAGCGCG AAATTACCGC CCGTTTGGTG ATGCCCACCC 1100TGSC IAR NYR PFGD AHPGCAGACGGCN CACACCCTCC ATGTCCAGTG GCAGCTCCGT AATGGGTATT 1150QTA HTLH VQW QLR NGYFTCGATCAATA TTGTCTACTT ACTGATCCGC ATTTCGAGGC ACCACGTCGC 1200DQY CLL TDPH FEA PRR171garnet Sequence10 20 30 40 501234567890 1234567890 12467R9 1234SE7RqO 1234567890CGTGGGCTGC GTACTTTGAA GGTAGTGAGG ATTGTTCGAC TGCTCAATGARGLR TLK VVR IVRL LNE1250GACGAGCCAT CGCCGCTCCA GTTGCTCGGG CGTGAGCTCG AGACTCTGTC 1300TSH RRSS CSG VSS RLCRGCNCTTTTCG CCGCCATCAG CTCCCGTTCC TGGCCTTTCG TGGCACTAAA 1350XFR RHQ LPFL AFR GTKCGAATAGCTC GTCCTTGGTC GTGCTCCGAG GATCAACTGC TTGCCGCATC 1400RIAR PWS CSE DQLL AASTTCCGGTGGC GGTGCATTAA TCCACTCGTC CAGATCCGAG ACCATCTGGA 1450SGG GALl HSS RSE TIWTCGGGGCACCT TCCGTTGCCT TGGGGCGCCA CCGGTATTAG CTCGCCTGTG 1500GHL PLP WGAT GIS SPyAACAGCAGCG TCATCTCCTG GACAATCTCA ATGCGCACGC GGAATGCCGC 1550NSSV ISW TIS MRTR NAACCTCCGTCGT CGTGTCCATG GCCATGGCAT CGGTCGACGT GGACAGCTGG 1600LRR RVHG HGI GRR GQLVTTGCGCAGCA TCTCGATCAG CATGCAAGCG GAATTGGCTC GCTCTTGCAC 1650AQH LDQ HASG IGS LLHCTCAATGTCG CTGGAACCAT TGAAGTGCTG AAGCTTGTCC AGCACATGGT 1700LNVA GTI EVL KLVQ HMVCGCAGAGCTA CCAATGATAA TGGTAATACT CAGAATACAA TCATTTGAAA 1750AEL PMIM VIL RIQ SFENATACGTATAG AGAGAGTCAC TTACTGTCAC CAAGCCAGGC AGATCCTGAA 1800TYR ESH LLSP SQA DPE172garnet Sequence10 20 30 40 501234567890 1234567890 1234567890 123457R90 1234S7RøGCTCTAAACA CGTGGTGGCC AGGCGAGCGA ACAGCTTCAT CAGCTTCTGCALNT WWP GER TASS ASA1850ACATAGACAC CTGGATGTGA CCGGGCCAGT AGCTTCGGAC GAGCAGATGT 1900HRH LDVT GPV ASD EQMLTAAGCGTCTC TCGCATCCTC CAGTTCGCCA GCAAACTCGC CAACGATCCA 1950SVS RIL QFAS KLA NDPGGCGGCGGCG TAGAGCACCT CGTACATGGA ATTACTTTGC GCCGAAACGG 2000GGGV EHL VHG ITLR RNGTGAACGTGTC AGCATGATTG GTCATCTCGT TGACGGCAAA TTGCCGGACC 2050ERV SMIG HLV DGK LPDHACAGGCACTG CAATGGCCNT NGTTNTCGCT ATAGGCGCCG TGCGTGGTNG 2100RHC NGX XXRY RRR AWXAGCTCGACGA CGACAGTTAG ATACCACTCG AAGTTGGTCA CATACAGATA 2150SSTT TVR YHS KLVT YRYCGAACTCTGC GCGCAATATC TCGATCACCT TGTAAAGCCA ATTCGTCCCG 2200ELC AQYL DHL VKP IRPDATAGGCCGAA CCCTCTGACC CGCTGCCATG TGGCCCAGCA ATCGCTTAAC 2250RPN PLT RCHV AQQ SLNAATCTCCATA AGGTTCTTCT TCGAGACCAT GCCGTAGAGC AGGTCCAGGG 2300NLHK VLL RDH AVEQ VQGCGCGCAGACG TATCGATTCG TCCTTGTCGT CCAGGCAGGC GAGTATGAGA 2350AQT YRFV LVV QAG EYEITCCTTGTGCG CCTGCACACT CTTCGGGTGC GTCTTCAGGA TTTTCGACAT 2400LVR LHT LRVR LQD FRH173garnet Sequence1234567890 1234567890 1234567890 1234567890 1234567890GGCCAACAGT CCCAGATATT TCAAGTTCTG GTCCGCGTCC TCGATGAGGAGQQS QIF QVL VRVL DEDTGCGCAGCTT CTGCACGCAG AGCTGAATGG AGGCACTGTG GTTGGGCATGAQL LHAE LNG GTV VGHACCGCTGCTAA TGCTGATAAC GACCGCGATG ACCGTGTTGA TGCACTCATAAAN ADN DRDD RVD ALlCAGAGACTCA TGGCGAGTGC TGTGATAGTT GGGTTTGGTT TTTGGTTCGTQRLM ASA VIV GFGF WFVI I iGi iii GT1 I ICATTT TTTGTTGGTG TTCGGTTAGT TTTAGTAATTFCF VFIF CWC SVS FSNFI I I I TTGTTT [Ti [TI F I TT GTGATTTGTA TTAGTTTGAG AAAAAAAAAAFCF FEE VICI SLR KKKGTAGGAACAT GTATTAAGAG TAAGCAGCGG GTAAAACGCA CAACATTGCAVGTC IKS KQR VKRT TLHCTACAACAAT AGCAATAGCA AATTAGGCAA TGGCAACAAC AACGGTAGCAYNN SNSK LGN GNN NGSITATCGATAGC ATACTATACT ATACATATAC TACTAGCTAC TTGCGGTACGSIA YYT IHIL LAT CGTAAGGTAACAA TTAGCGATTA TTGCGATAGA CATTGGGGAA AGAGAACATTKVTI SOY COR HWGK RTLGCAAAGCAAC GCAGCAATGG CAACCAAAAA GAAAAACGTC ACTAAACAACQSN AAMA TKK KNV TKQQAGCAACAACA ACAACAACTG CAACGGCTCG TCATTTGTTT TTCTTTCGCTQQQ QQL QRLV IC.F SFA10 20 30 40 50245025002550260026502700275028002850290029503000174garnet Sequence101234567890 1234567890 1234567890 1234567890 1234567890TCATCCGATT TGACTAGTTT AGAACTTTGG ATCTCAATGA GTGTGCTCGASSDL TSL ELW ISMS VLD20 30 40 50TAAGCAAAAA ATCGATAGGC AAACGATGAA TTATAGAAAC AAAGACAAAC 3100KQK IDRQ TMN YRN KDKLTTAAGCAGTA TGGCGACAGT CATAAGTTGA GCGAGTGGGA GAGAGAAAGA 3150KQY GDS HKLS EWE RERGATAGACAGA GAGAGAGAGA GAGAGAGAGT ACGCTAGAGC TAGAGAATTG 3200DRQR ERE RES TLEL ENCTACAGTAAAT GATATAACGA ATATATCCAG TCACACGACA ATCATCGAGC 3250TVN DITN ISS HTT IIEQAGCTTCAATT ATCGATCATT GATATCGACC TTTTAATCGG TCACTTTCGA 3300LQL 511 DIDL LIG HFRTTTGATTTTT CGAATTTTTT CTTTGCTTTC GCCTTGCTTT GTTGCAATCG 3350NFF FAF ALLCTTTTCCACAC ATTCTTGGGA AATCGTATCG TATTTTACAT TTTCAGTTCA 3400FPH ILGK SYR ILH FQFSGTTCAGTTGA TTTGTATTTG TATTTTTGTT TTGTTTTGTT TTGTTTGTTT 3450SVD LYL YFCF VLF CLFGTTTTGCAAT GATTTTAAGA CTTGCCTATG AATTAGATTT GTGAGTGGTT 3500VLQ.CTATTAATTT CTTTCCTAGC CGGGGTTCTA AGGGGGTTAA AGCGCCAAAC 3550TGCAAATGCA AAGAGAAAAA GAAACAAGCA AGAAATTATA AATTACATAC 36003050FDFS CNR175garnet Sequence1234567890 1234567890 1234567890AATCGAAATC ATCAACGTCC TTATCCACAA1234567890 1234567890CTAAAACTAG AACTAAAACTAAAGCTAAAA CCGAAAACGA AACTAGAAAA TGAAGTGTTC AAGAAAATGGTAAACTGGAA CTGGAATACT CTAAGAAGTA ATTTAACTTT CTCTTAAACTGGTCCTGGTC CTTTTCCATT CGGATCGAAT CTCTTGATGA TATGTCTATATATTTTTGTG TATCTCAGGG TGGAA370037503800382510 20 30 40 503650176Comparison between the c-DNA sequence and genomic sequence reveals atleast four introns, the most 3’ of which is the site of the P-element insert in thegP mutation.Conceptual translation of the garnet c-DNA. The direction of transcription wasdetermined by in situ hybridization of digoxigenin labeled RNA probes toimaginal discs of third instar larvae. The garnet c-DNA from the Hind Ill site at1720 and the Sst I site at 2143 was subcloned into pBS KS. Sense andantisense RNA was produced by transcription of this garnet fragment by T3 andT7 polymerase respectively. The T7 transcribed probe detected messagewhereas the T3 transcribed probe did not. Thus the direction of transcription isas shown in Figure 26.There is a reasonable open reading frame in this orientation, starting at position297 encoding a potential polypeptide of 1054 amino acids. Conceptualtranslation of the long open reading frame yields the protein shown in Figure26. This protein has a number of motifs. There is a poly glutamine stretch (8repeats) at position 3145. There are three reasonable hydrophobic domains.Comparison between this sequence and the EMBL and Swiss protein databanks reveals no informative similarities (Table 22). The first two sequencesshow essentially exact correspondence with the garnet gene sequence. Theslight discrepancies in sequence can be ascribed to sequencing errors. As theyare sequence tagged sites from the European genome mapping project thisdata will assist in aligning the genetic and molecular maps for the Xchromosome but is not otherwise informative. The other sequences show short177regions of similarity restricted to repeated sequence motifs. Thus it wouldappear that if garnet -homologous sequences exist in other organisms, (garnetTable 22. Genes with sequence similarity to garnet.The first column gives the name and accession code of the gene. The secondand third columns give the region of sequence similarity, relative to the garnetimaginal c-DNA sequence and the other gene, respectively. The next columnsgive the percent similarity within this region and an indication if the similarity isrestricted to a repeated sequence motif.178Geneswithsequencesimilaritytogarnet.generegionofsimilarityregionofsimilarity%similaritymotifingarnetinothergene1. Drosophila2352-25141-22195%non-repeatmelanogasterSTSembZ31953DM189B8T2. Drosophila363-428&428-4682-68&130-17197%non-repeatmelanogasterSTSembZ32301DM7C5T3. Rat5.5kbDNAfragment3096-31751188-125378%GAGArepeatcontainingrepetitiveDNA3395-34463717-376585%Trepeats179embXl3424RN55REP2600-26703721-376876%Trepeats3534-35724262-430074%Arepeats4. C.grisousdhfr3092-31795833-592575%GAGArepeatoriginofreplication2575-26702222-231965%TrepeatsembX520341CGDHFRORI5. RatMHCClassIAggene3094-31903429-353175%GAGArepeatRT1-uhaplotype2572-26541728-180865%TrepeatsgbM64795RATMHRT13395-34253745-377563%Trepeats6. Mousebeta-globincomplex3059-318651988-5193370%GAGArepeatbho,bhl,bl,b2,bh3andbh32558-2669584-653&45125-4626070%TrepeatsembX14061MMBGCXD3383-343533732-3378570%GUTrepeat3770-379820477-2050279%ATrepeat7. RatNa+,K+ATPase3094-3161910-97779%GAGArepeatalpha2subunitgene2600-26872834-292165%Trepeatand5’flankseq.2565-26372839-291167%GUTrepeat180embD90049RNATPA25492-5168043-806784%GTTrepeat8. MouseHox3.1,3.23092-3188577-663&6399-651873%GAGArepeatgenesandintergenicregion2583-266679845-991870%TrepeatsgbM35603musHOXMAA9. Mousemyoglobinexon13111-3196687-75179%GAGArepeatandflankingregions3376-3435&2587-2646180-25865%GUTrepeatembX04405MMMYOGG110.Rathepaticsteroidhydroxylase3091-321617713-1783470%GAGArepeathAlgene2566-26414891-496667%GTTTrepeatgbM33312RATCYP2A1is present in sibling species of Drosophila melanogaster (Sturtevant et al. 1925,SturLevant and Novitski 1941)) these have not been entered into the data bank.Tissue distribution of garnet transcripts. The tissue specificity of garnettranscription was investigated by in situ hybridization to embryos and variousorgans present in third instar larvae and adults. Figure 27 presents theseresults. The garnet gene is clearly transcribed in the eye-antennal imaginaldisc, the anlage of the adult eye. Interestingly it is also present, in other imaginaldiscs as well as other tissues such as the larval brain and ovarioles. It is alsoabundantly expressed in embryos. Thus garnet is expressed in a variety oftissues during at least three stages of development.Do sequences similar to garnet exist in Drosophila melanogaster?Genetic evidence suggests that the garnet gene is one of a group of genes thatare functionally redundant. Functionally redundant loci might have sequencesimilarity. To test this possibility, genomic DNA from wild-type flies was probedwith the 6.6 Eco RI fragment under conditions of reduced stringency. No extrabands were seen under conditions where sequences sharing approximately66% identity should be detected (Figure 28). Thus, there do not seem to beany sequences in the Drosophila melanogaster genome which are highlysimilar to the garnet gene.181Figure 27. Analysis of garnet transcription by in situ hybridization to varioustissues.A. The first three images (viewed top to bottom, left to right) show garnetmessage in blastula, gastrula and neurula stages of Drosophila melanogasterembryos.B. The next two pictures show garnet message in the leg and eye-antennalimaginal discs.C. Below is shown garnet transcription in an ovariole.D. The three pictures on the lower right show the same tissues hybridized withthe sense probe (T3 probe) after 90 minutes of staining.E. garnet transcription in the brain and the eye-antennal imaginal disc (forcomparison) of wild type third instar larvae.The embryos and the larval brain were stained for ten minutes before thestaining reaction was stopped. The imaginal discs and ovarioles were stainedfor 90 minutes.182-aa3 VFigure 28. Southern analysis of regions of sequence similarity of the garnetgene in Drosophila melanogaster.DNA from wild type flies was isolated, restricted with Kpn I and Xho I (which donot cut within the 6.6 Eco RI fragment which contains the garnet gene),separated on a 0.8% agarose gel, transferred to nylon membrane and probedwith the 6.6 Eco RI garnet fragment with reduced stringency. Under theconditions used sequences with approximately two thirds similarity should havebeen detected. With the exception of a high molecular smear evident in the XhoI lane, likely due to incomplete digestion, no additional bands were seen. Thegel was overloaded with DNA and then the filter overexposed to allow detectionof any weakly hybridizing bands.185981‘IIL19IudiIoqDiscussion-garnetPhenotype of the garnet alleles.The garnet alleles affect both pteridine and ommochrome pigments. This is nota phenotype that can easily be explained as a simple enzymatic deficiency. Asthe two pigment biosynthetic pathways are biochemically distinct, an enzymaticlesion should alter only one group of pigments. In addition there is no evidencefor novel pigments, or pigment intermediates, that might be expected toaccumulate in a blocked pigment biosynthetic pathway.The garnet mutant alleles alter pigmentation of all of the three major pigmentedstructures in Drosophila melanogaster, the eye, the malpighian tubules and thetestes sheath. Pigmentation of the malpighian tubules is generally moreseverally altered than that of the testes sheath, which in turn is more severelycompromised than that of the eye. This difference could reflect the timing ofdifferentiation, the type or amount of pigment in these organs.Theg53dallele presents an important exception to this pattern of decreasedpigment levels. Although this allele is the most severe mutant allele, andgenetically behaves as an amorph, based on dosage tests in the eye, it exhibitsessentially wild-type levels of pigmentation in the testes sheath. Interestingly,Tearle (1991) reports that this allele does not have a pronounced effect onpigmentation of the occelli either. The obvious conclusion is that this particularallele is not a typical amorph or hypomorph but has an alteration in tissuespecific regulation of garnet expression. Many alleles of the garnet gene havebeen isolated and described (Table 23). Mutant alleles have been isolated asspontaneous mutations, and induced by treatment with X-rays, gamma rays,187Table 23. Published alleles of garnet.The first column gives the allele designation. The second column gives theinducing agent, when known. The third and fourth columns gives the name ofthe investigator responsible for isolating the allele and the reference. This listwas adapted from Lindsley and Zimm (1992).188Published garnet allelesallele origin discoverer referenceg1 spontaneous Bridges Bridges 1916g2 spontaneous Bridges Lindsley and Zimm 1992g2” unknown unknown Lindsley and Zimm 1992g3 spontaneous Bridges Lindsley and Zimm 1992g4 X-rays Glass Lindsley and Zimm 1992gl7B X-rays Valencia Valencia 1966g261° X-rays Sobels Lindsley and Zimm 1992g2615 SMS Sobels Lindsley and Zimm 1992g2641 SMS Sobels Lindsley and Zimm 1992g2810 SMS Sobels Lindsley and Zimm 1992g284° SMS Sobels Lindsley and Zimm 1992g29h spontaneous Wallace Lindsley and Zimm 1992g3Od spontaneous Bridges Lindsley and Zimm 1992g32d spontaneous Emerson Lindsley and Zimm 1992g33J spontaneous Bridges Lindsley and Zimm 1992g33’ spontaneous Ives Lindsley and Zimm 1992g34e spontaneous Duncan Lindsley and Zimm 1992spontaneous Mossige Lindsley and Zimm 1992g37f spontaneous Ecken Lindsley and Zimm 1992g37k spontaneous Mather Lindsley and Zimm 1992g38b spontaneous Bridges Lindsley and Zimm 1992g42a X-rays Green Lindsley and Zimm 1992g49h unknown King King 1950g5Oe unknown unknown Lindsley and Zimm 1992g53d spontaneous Hexter Hexter 1958g55k spontaneous Williams Lindsley and Zimm 1992g64b X-rays Ives Lindsley and Zimm 1992g68d EMS Maddorn Hayman, Madden 1967g7Ok spontaneous Schwinck Schwinck, Schwinckl 972g79’ spontaneous Najera Najera 1985g2712 X-rays Demerec Lindsley and Zimm 1992g2716 X-rays Demerec Lindsley and Zimm 1992g2719 X-rays Hoover Lindsley and Zimm 19922711° X-rays Hoover Lindsley and Zimm 1992ge spontaneous Gottschewski Lindsley and Zimm 1992gEMS EMS unknown Lindsley and Zimm 1992gF spontaneous Waddle Lindsley and Zimm 1992189gim unknown unknown Lindsley and Zimm 1992g’ unknown unknown Lindsley and Zimm 1992gSl spontaneous Schalet Chovnick 1961gS2 spontaneous Schalet Schalet 1986gS3 spontaneous Schalet Schalet 1986gtUh4 spontaneous Kuhn Kuhn 1972gtUh2 spontaneous Kuhn Kuhn 1972gW spontaneous Muller Lindsley and Zimm 1992gX X-rays Muller Lindsley and Zimm 1992190chemical mutagenesis (Table 23) and P-elements (Wennberg, 1988 and thiswork). It is interesting that of the 56 published alleles and the many P-elementderived alleles described in this work, none appear to be true null mutations, Ofthe over 300 mutants described for the white gene, which has a similarly sizedtranscript, approximately a third 98/344) are phenotypic nulls. This mightsuggest that a null mutation of the garnet gene has a phenotype other thanreduced eye pigmentation, although the paucity of null garnet mutants may notbe beyond the bounds of bad luck.Genetic and molecular limits of the garnet gene:Perhaps the most useful issue of the Chovnick/Hexter debate on the complexityof the garnet locus was a recombination fine structure genetic map of the garnetgene (Figure 29). This map was generated by selecting rare wild type intraallelic recombinants between different garnet alleles. Each of these studiesinvolved visually scoring more than a hundred thousand (Hexter 1958-583,416, Chovnick 1958-762,429 and Chovnick 1961 - 176,526) flies for rarewild type recombinants in a background of brown-red eye mutants, thephenotypes of some of which approach wild type with age. The value of workingon a gene where such extensive work has been performed, by others, cannotbe overstated.The intra-allelic genetic map shown in Figure 29 should, in principle, providea basis for correlating the genetic and molecular limits of the garnet gene.Unfortunately, the gSl allele which defines the right-most limit of the gene is nolonger extant. The molecular lesions responsible for the three left-most allelesare not known. They are not associated with deletions or insertions. The g1 andg3 alleles are however associated with seemingly identical insertions which191Figure 29. Map of the garnet gene generated by intragenic recombination.The top line indicates the genetic limits of the garnet gene. Below is shown theorder of six garnet alleles, derived from Chovnick 1958, Hexter 1958 andChovnick 1961. If the g1 and g3 lesions are in the 3’ region of garnet assuggested by the results of Figure 21 then the direction of transcription can beoriented relative to the genetic map as shown.192RECOMBINATIONAL MAP OF THEgarnet LOCUSI I I Ig53d g2 g5Oe g1 gSl5’ 3•TELOMERE CENTROMERE193have occurred at apparently the same point in the garnet gene. These allelesare genetically inseparable. This restriction fragment is near the 3’ end of thegene and is consistent with their recombination position in the rightmost third ofthe map. It is also interesting that the 3 leftmost, or possibly 5’, alleles are themost extreme. The g53d allele might be a lesion in the 5’ regulatory region. Ifthis were so it would indicate that the orientation of the garnet gene relative tothe centromere is with transcription away from the telomere.Evidence that the cloned region corresponds to the garnet gene:The weak P-element induced gP allele was used to clone the garnet gene.There are three lines of evidence that indicate that the cloned region,specifically the 6.6 kb Eco RI fragment does in fact contain the garnet gene.First, a partial P-element interrupts this fragment in the gP allele. This P-elementis altered in both revertants and extreme secondary derivatives of the gPmutation. Sequence analysis of wild type garnet DNA, the equivalent region ofthe gP mutation and a large c-DNA arising from this region places the Pelement in a 3’ intron. Secondly, garnet mutations have a high frequency ofalterations in this fragment. Of the 14 garnet alleles examined by Southernanalysis, 9 had either insertions or deletions in this 6.6 Eco RI fragment. Finally,analysis of transcripts from this region identify two transcripts from wild typeflies, both of which are absent in the extreme g53d mutant and one of which isabsent and the other possibly altered in the hypomorphic g3 allele. Thus itseems reasonable to propose that the garnet gene is located within the 6.6 EcoRI fragment and that the c-DNA which comprises approximately two thirds of thisregion, and corresponds to the only major open reading frame within this DNAsegment identifies the garnet gene product. Final proof will, however, requirerescue of the garnet mutant phenotype by P-element mediated transformation.194Expression pattern of the garnet gene:Tissue in situ analysis indicates that the garnet gene is transcribed at manystages in development including embryos, third instar larvae and adults. It is nothighly transcribed at any developmental stage. Pigment deposition in the eyeand testes sheath starts two to three days before eclosion, about midwaythrough the pupal stage, and darkening continues up to a week after eclosion(Schultz 1935). Pigmentation of the malpighian tubules can be detected fromthe first instar larvae onwards (Breme and Demerec 1942). The presence ofgarnet m-RNA in third instar larvae and young adults is consistent with a role ineye pigmentation. However, the ubiquitous, low levels of garnet transcription inembryos is less obviously related to pigmentation. Interestingly, this pattern ofexpression is also seen for the light gene (Devlin, Bingham and Wakimoto1990), another member of the transport group. The tissue distribution of thegarnet transcript is also somewhat surprising. As expected the garnet gene canbe detected in the eye-antennal disc of third instar larvae. It is, however, alsopresent, albeit at somewhat reduced levels, in the leg and wing imaginal discs,larval brain as well as in ovarioles and embryos. In addition, garnet mutationscause diminished pigmentation of the adult eye, malpighian tubules, fat body,occelli, and testes sheath, suggesting that garnet is also expressed in thesetissues.Sequence analysis of the garnet gene.Sequence analysis of the garnet gene has not been particularly revealing.Conceptual translation of the 4kb imaginal disc c-DNA yields a putative proteinof 1054 amino acid residues. This putative protein has three indifferent195hydrophobic domains and a polyglutamine stretch. Polyglutamine repeats havebeen found in a number of neurogenic genes but are not necessarilydiagnostic. They have also been implicated in parental imprinting (Green 1993).In these instances the polyglutamine repeats tend to be long and are notnecessarily either translated or transcribed. Polyglutamine repeats have alsobeen implicated in protein-protein interactions. This function might be related tothe proposed interactions between the garnet gene product and those of theother eye colour genes.Function of the garnet gene.The considerable genetic resources devoted to the production, developmentaland tissue specific regulation of eye pigments raises the question of theirfunction. The function of the garnet gene product has not been explicitlyaddressed by any author. Nevertheless, there is extensive genetic evidence,detailed in chapter 1, that the product of the garnet gene interacts with not onlythe white gene, but with other members of the transport group of eye colourgenes. Various investigators have proposed functions for this group of genes.The function of pigments in optically isolating the ommatidia and providing forlight adaptation seems incontestable, Interestingly, a similar function in bothshort and long term light adaptation has been proposed for pteridines found inthe mammalian retina (Cremer-Bartels 1975), although the evidence for this isnot compelling. Nevertheless, many authors have sought additional roles for theeye colour genes.196garnet and cell metabolism: The first conceptual approach to studying thefunction of eye colour genes derived, appropriately, from attempts to resolve theparadox of the apparently excessive number of eye colour genes. The first, andobvious, attempt to deal with this problem was to group the mutations by similarphenotypes. Nolte (1 954a) proposed that there were in fact only 6 groups ofeye colour mutations, the vermilion group, the light group, the dark group, thered group, the variegating group and the ruby group, of which garnet is amember. Detailed studies of the red and brown pigments of these groups (Nolte1 954b, 1955, 1959) led to the realization that while the vermilion group mightbe united in generally disrupting ommochrome synthesis (Nolte 1954a), the restof these groups did not represent biochemically or functionally related genes.Nolte (1 954b, 1 959b) then proposed that the genes of the ruby group representlesions in general aspects of cell physiology. Specifically, he proposed that theyare involved with the protein catabolism that generates precursors for pigmentproduction. This hypothesis is mirrored in various forms in most of the ensuingproposals for the function of the transport group of genes.garnet as a transport gene. Sullivan, Grillo and Kitos (1974) and Sullivan andSullivan (1975) provided data for a specific variation of Nolte’s hypothesis. Theyproposed that this group of genes encoded products responsible for metabolitetransport. Based on the results of a series of experiments where isolatedorgans, eye discs and malpighian tubules, were cultured in vitro in labeledkynurenine, an intermediate in the ommochrome pathway, they proposed thatmany of the eye colour mutants were defective in transport of metabolites andpigment intermediates.197The data presented by Sullivan and Sullivan on possible transport defects ineye colour mutations did not include the garnet gene. They did however,propose a list of criteria that would identify a gene primarily concerned withpigment transport. These criteria are: cellular autonomy, effects on both pigmentpathways, effects on all of the pigmented organs, and diminished pigmentlevels in conjunction with normal biosynthetic enzyme activity. Cell autonomywould be expected for a membrane based gene product such as atransmembrane channel protein. Alterations in both ommochrome and pteridinepigments suggests that the transport apparatus handles more than onemetabolite or compound. This is not unprecedented (Christensen 1973). Oneexample, possibly quite relevant to the transport of pigments in Drosophila, isprovided by the mouse pallid locus. Defects in this gene are associated withdiminished transport of tryptophane, L-dopamine and Manganese ions(discussed by Wiley and Forrest 1981). Alterations in the pigmentation ofdifferent organs would suggest reasonably ubiquitous use of the transportmechanism. The genes white, brown, scarlet, lightoid, claret, carnation, light,maroon and pink fulfill all of these criteria. The garnet gene fulfills the first threeof these criteria; the last has not been fully tested (although Glassmann (1956)reported that the g2 allele has normal levels of the enzyme kynurenineformanidase). Nevertheless, an alteration in the transport of kynurenine or anyother compound remains to be shown for garnet.In retrospect, in light of the complex tissue and developmental interactions, theknown excretory function of the malpighian tubules and fat body, the cellautonomous nature of many mutations and importance of transport in cellfunction, a role in metabolite transport for some of the eye colour genes mayseem obvious. Nonetheless, these authors were the first to furnish data for the198role of transport mechanism in the final production of wild type eyepigmentation. Finally, it should be noted that transport probably involves themitochondrial membrane, the pigment granule membrane and possibly thegolgi body membrane, as well as the plasma membrane. Defects in thistransport may have a variety of consequences, such as alterations inintracellular storage or movement of precursors and excretion of wastecompounds, as well as diminished accumulation of pigments.garnet in the brain. More recently, McCarthy and Nickla (1980) have proposedthat the genes carnation and light, both members of the transport group, areinvolved in a variety of (unspecified) functions and have an essential role in thedevelopment and function of the nervous system. Their studies were based onextensive genetic and histological examination of double mutant light-carnationindividuals. Flies homozygous for either one of these mutations survive,whereas, flies homozygous for both die. The lethal phase of the double mutantis protracted and depends on dosage and activity of light (Nickla, 1977). Thesynthetic lethal focus maps to the ventral blastoderm, site of the presumptiveventral nervous system (Nickla, Lilly and McCarthy, 1980), and double mutantindividuals display abnormal brain morphology (McCarthy and Nickla, 1980).These authors propose that in addition to a role in pigmentation, carnation andlight perform an essential function in neural development. The garnettranscription seen in the larval brain, as well as the apparent absence of thistranscription in rosy null larvae, might suggest a similar function for the garnetgene in this tissue, however further genetic and histological analysis isrequired. It should be noted that unlike the carnation-light double mutant, therosy-garnet double homozygotes are viable, and other than female sterility,display no obvious behavioral or physical defects.199The suggestion that groups of eye colour genes represent an essential andredundant function is supported by findings that certain pairwise combinationsof these genes behaved as synthetic lethals. Synthetic lethal combinationshave been known in Drosophila for some time but are not common and offer apowerful tool to identify functional identity between redundant genes. Althoughsynthetic lethal combinations not involving eye colour genes exist, therepresentation of not only eye colour mutations but specifically of mutants of thetransport group of eye colour genes (Table 2) is intriguing. Notwithstanding, asystematic search for synthetic lethal combinations amongst pairwisecombinations of eye colour genes has never been done and is not a task to beundertaken casually; the number of eye colour mutations, even discounting thenumerous alleles of most, and the need to examine multiple allele combinationswould make this an onerous task. A search for interactions amongst the smallerset of the transport group of eye colour genes might, however, prove revealing.The garnet gene does not display lethal interactions with at least prune, light,rosy or deep orange (other members of the transport group have not beentested). It does however display a full spectrum of other interactions with thesegenes, including female sterility, synthetic dominance and a variable spectrumof cell death phenotypes. These other types of interactions may suggestspecialized roles for the garnet gene.garnet and intracellular transport Very recently the g2 allele has beenidentified as an enhancer of the quartet mutation (C. Cheney- personalcommunication). The quartet mutation is a female sterile mutation which seemsto be involved in localization of the nanos posterior determinant in eggs. Theabnormal phenotype of the quartet mutation is proposed to result from a defect200in intracellular transport. The action of the g2 allele as an enhancer of thisdefect might implicate the garnet gene as being involved in intra- as well asintercellular transport. Preliminary results indicate that nanos transcription maybe severely reduced in the female sterile e(g) g53dhomozygote (S. Gorski,personal communication). While the females sterile phenotype could be due toa general and non-specific physiological effect, examination of nanoslocalization in the female sterile garnet and rosy-garnet double mutants mayprove to be informative and is being pursued by the Cheney laboratoryThe metabolic role of some of the pigments and intermediates is an area ofintense research. In the earliest work on the chemical nature of thesecompounds, Schultz (1935) speculated that as they were highly susceptible tooxidation-reduction reactions, their function related somehow to this property.More recently Hilliker et a!. (1992) have observed that mutants for the rosy genehave increased sensitivity to oxygen stress and a reduced life span. Thissuggests an evolutionary impetus for the development of pigments might stemfrom an alternative use of metabolic waste products. Metamorphosis is ametabolically active portion of the insect life cycle. In holometabolic insects,during this period, excretion of all but gaseous waste products is restricted.Conversion of toxic by-products of amino acid and nucleic acid catabolism intostable, non-toxic molecules which could be deposited in high concentrations indifferent organs such as the eye, would be useful. If these products served orenhance some other functions in these organs, there should be considerableevolutionary impetus to develop such a system. While this may be true of asubset of eye colour genes this function is unlikely to be restricted to this classand there is no evidence that mutations in the garnet gene alter longevity.201While the suggestion of intercellular transport, cell communication and neuralfunction are certainly compatible, these studies have yet to do more than implyan important, but undefined, redundant biological role for this group of eyecolour mutants. Resolution of the biological role of garnet gene will requirefurther study.Testing the model of the function of the garnet gene: In summary, the biologicalfunction of the “transport” group of eye colour genes, including garnet remainsspeculative. Their proposed functions in general cellular metabolism,intercellular transport, intracellular transport and a role in neural formation andfunction are certainly not mutually exclusive. Genetic analysis suggests that thisgroup of genes possess a ubiquitous essential and redundant, if unknownfunction. Based on the similarity between the phenotypes of mutants in thetransport group of genes and white hypomorphic mutants, similar effects onpigment accumulation and hypersensitivity to the cryptic we(g) allele describedin chapter 1, I suggest that all of these gene products associate with the productof the white gene. The unusual epistatic interaction found between the a2 andthe g2 allele, as well as the similar interaction reported for wa3 and ruby mightprovide an avenue to investigate the physical nature of this interaction. Thefunctional redundancy of the transport group of genes, implied by the pleiotropicgenetic interactions compared with their rather weak phenotypes as single locimutations, suggests that they may coordinately perform some essentialfunction(s). In contrast, the phenotype of other combinations, such as light andcarnation, suggest that they may perform other more specialized functions. Theabsence of synthetic lethal combinations involving the garnet gene suggeststhat this gene may operate only with many other gene products. However, thefemale sterile phenotype of the double mutant rosy-garnet combination202intimates that this pair of genes might have a more unique role in the femalegerm line. One may envision a complex involved generally in aspects oftransmembrane transport, various members of this complex associating indifferent cell types, in different subcellular compartments, and at different stagesof development to perform variations of this function. These proposed functionsremain completely speculative and await further investigation. The cloning ofthe garnet gene should lead to further definition of the biological role of not onlythe garnet gene but possibly also that of the “transport” group of eye colourmutations. Direct physical proof of the existence of a macromolecular complexwhich regulates a transmembrane pore as proposed above, must await directanalysis of the garnet gene product. Antibodies to the garnet gene productcould be used to examine the subcellular location of the garnet gene product.Co-localization and co-immunoprecipitation with anti-garnet and anti- whiteantibodies or use of the yeast dihybrid selection system would offer a directmeans of ascertaining if the genetic interaction between these genes wasmirrored by a structural association. In the interim, the genetic and molecularanalysis of the garnet gene makes this gene a useful tool to investigate otherphenomenon.In the next chapter, I describe a system where the garnet gene is used to studygenomic imprinting in Drosophila melanogaster. Genomic imprinting hasrecently attracted attention by the association between imprinting and somehuman genetic syndromes but it has been described in insects and is a welldefined phenomenon the sciarids and coccids. In the final chapter I describe amini-chromosome which is imprinted. The imprinting is manifest as parentdependent expression of the garnet gene. The ease of examining eye colour tomonitor imprinting as well as the sophisticated genetics of Drosophila203melanogaster have allowed tests of a number of possible mechanisms ofimprinting. The garnet gene has a long history of being used as a tool toexamine other interesting biological phenomenon. The next chapter continuesin this tradition.204Chapter 3.Imprinting of a mini-chromosomein Drosophila melanogaster205Introduction-imprinting of a mini-chromosome in Drosophila melanogasterThe phenomenon of genomic imprinting1 encompasses a number of processeswhereby a gene or a region of a chromosome is reversibly modified so that it retains a“memory” of its genetic history. The term imprinting was originally coined by Crouse(1960) to refer to the complex behavior of the X-chromosome in the dipteran insectSciara. She defined imprinting as the “differential behavior of the members of a pair ofhomologous chromosome which is predetermined several to many cell generationsbefore the stage in development at which resulting behavioral differences becomeobvious.” This definition is fairly broad. As a result, the term imprinting has beenapplied to a vast number of exceptions to normal Mendelian segregation of traits.Conversely, for historical and traditional reasons, phenomena, that would be definedas imprinting by contemporary criteria, have been given a variety of other names.Table 24 provides a list of these terms, the organism with which they have been usedand the probable type of imprinting. A working definition of imprinting has beenproposed by Reik (1992) as a process whereby “epigenetic information is introducedinto chromosomes and is stabley replicated together with the chromosomes as cellsdivide”. But even this definition is sufficiently broad that it encompasses a number ofbiological oddities which are undoubtedly mechanistically distinct. This problem wasfirst addressed by Monk (1990) who proposed subgroups to encompass four generalclasses of phenomenon which have been collectively called genomic imprinting.These include species-specific imprinting, differentiation, epimutation and parentalimprinting.1 The term “imprinting” properly refers to a phenomenon in behavioral psychologywhereby an immature animal learns appropriate behavior from adults. For thepurposes of this thesis it will be understood that the term imprinting refers to thephenomenon of genomic imprinting.206Table 24. Terms used for genomic imprinting.The first column gives the term used to describe the phenomenon of genomicimprinting. The second column gives the general group of organisms with which thisterm has been used and the third gives the type of imprinting as defined by Monk(1990). With the exception of the phrase “Non-Mendelian ratios” (Hall 1990) andparental effects (Baker 1963) all these terms are explained in greater detail in HeslopHarrison (1990).207Other terms used for genomic imprintingName of phenomenon organism group type of imprintingParental effects Insects (Drosophila) parentalBlock transference of characters plants speciesGenetic affinity plants speciesSuppression plants speciesSelectivity of expression plants speciesCryptic structural differentiation plants speciesSkewed back cross ratios plants speciesHomeosis plants speciesCharacter pseudo-linkage plants speciesNon-Mendelian ratios humans parental208Strain or species specific imprintingThe term imprinting has been used to refer to the variable phenotypes of the hybridsresulting from crosses between different subspecies or strains. This type of imprintinghas been called species or strain-specific imprinting (Monk 1990) as the hybridphenotype appears influenced by a “memory” of the parental species. The differencebetween a mule (horse mother, donkey father) and a hinney (donkey mother, horsefather) is the classical example of this type of imprinting. This type of imprinting likelyreflects the preferential action of maternally deposited activators (Castro-Sierra andOhno 1968) or repressors (Schmidtke, KuhI and Engel 1976) on the subtly differentregulatory regions of the genes of the two subspecies or strains. As such, species-specific imprinting is dependent of different information encoded in the DNA of the twospecies and is not an epigenetic process.Somatic imprinting or differentiation.The term imprinting has also been used to refer to the processes of determination anddifferentiation whereby the developmental potential of a mitotic clone is restricted(Paro 1990, KIar 1987, 1990). From both a mechanistic and theoretical point of viewthe processes involved in somatic versus germ line “memory” are expected to differ.Mechanistically, the packaging of DNA is grossly different between the germ line andsoma. More importantly, meiotic products must remain totipotent in order to producethe complete spectrum of cell types present in the next generation. Somatic cells donot face such demands so that loss of totipotence implied by parental imprintingposses no conceptual difficulties. While determination remains a central question indevelopmental biology, renaming this process “imprinting” adds nothing to ourunderstanding of the processes involved.209Permanent imprinting.The term imprinting has also been used to describe parent-specific, permanentchanges in gene activity. This phenomenon is manifest as a permanent alteration ingene activity after passage though one parent or genetic background (Hadchouel et al1987, Reuter 1985, Dorn et al. 1993) and has been called epimutation by Holliday(1987) and allele-specific imprinting by Monk (1990). The mechanism where by apermanent alteration is produced in one parent and then maintained remainssomewhat obscure. In the situation described by Hadchouel et al (1987) the geneinactivation was associated with methylation of multiple CpG islands of the hepatitis Bsurface antigen transgene, but only after passage through the female germ line. Theseinvestigators proposed that if methylation was reversed at only a low frequency inmales, possibly due to chromatin remodeling in spermatogenesis, then the extensivemethylation in females would constitute a virtually permanent alteration. Themechanism responsible for this phenomenon in Drosophila remains unknownalthough it has been implicitly associated with changes in chromatin structure.Parental imprintingIn parental imprinting the activity of the imprinted gene is determined by the sex of theparent transmitting that gene. This type of imprinting has also been termed gametespecific imprinting (Monk 1990). Parental imprinting is distinguished by its transientnature, specifically, the imprint is completely reversed in one generation by passagethrough the other sex. The result of parental imprinting is the functionalnonequivalence of the maternal and paternal genome.In mammals the consequence of this non-equivalence is drastic. Embryos (eitherparthenogenic, gynogenic or androgenic) or tissues (such as ovarian tumours and210Table 25. Human diseases in which imprinting has been implicated.The first column gives the name of the disease or condition in which imprinting hasbeen implicated. The second column gives the reference.211Human diseases and conditions in which imprinting has been implicatedCondition or syndrome ReferenceAngelman/Prader Willis syndromeCarmillia De Lange syndromeDuchenne muscular dystrophyFamilial glomus tumoursFloating harbour syndromeFragile X mental retardationHuntington’s choreaHydatidiform molesMyotonic dystrophyNarcolepsyNeurofibromatosisOsteogenic sarcomaOvarian tumoursPhiladelphia chromosomeRetinoblastomaRhabdomyosa rcomaRubinstein-Taybi syndromeRussel-Silver dwarfismSotos syndromeWeaver syndromeWiedemann-Beckman syndromeWilm’s tumourWolf-Hirschor syndromeNicholls et al 1989/Hall 1990Clarke 1990/Hall 1990Hall 1990Clarke 1990Clarke 1990Laird 1987Laird 1990/Sapienza 1990Kajii and Ohama 1977Clarke 1990/Hall 1990Hall 1990Clarke 1990Toguchida et al. 1989Linder et al 1975Haas, Argyriou and Lion 1992Clarke 1990Dryja et al 1989Clarke 1990/Hall 1990Hall 1990Clarke 1990Clarke 1990Hall 1990Clarke 1990Clarke 1990212hydratidiform moles) that are derived from two complete maternal or paternal genomesare not viable even though they have the full complement of genetic information. Thislethality is thought to be due to the cumulative effect of many imprinted genes, at leastsome of which are growth regulators acting early in development. Parental imprintinghas been most thoroughly described in mammals and is currently being intenselyinvestigated. Much of the impetus for these investigations comes from the implicationof imprinting as the underlying cause of a number of human syndromes. Table 25lists diseases and medical conditions which are associated with either aberrantimprinting or the aberrant transmission of imprinted regions. The evidence forinvolvement of imprinting in these conditions is well defined in some cases (e.g. fragileX mental retardation) and considerably more inferential in others.Parent-dependent gene expression or parental imprinting is, however, found in manyother eukaryotes, including plants (Kermicle and Alleman 1990), C. elegans (Gilchristand Moerman 1992), and a variety of insects including the Homopteran Coccids,mealy bugs and other amoured scale insects (Chandra and Brown 1975), theHymenopteran wasp, Nasonia vitripennis (Nur et aL 1988), the Coleopteran coffeeberry borer beetle, Hypothenemus hampei (Brun et a!. 1995), and the Dipterans, thefungus gnat Sciara (Crouse 1960) and the genetically well characterized fruit flyDrosophila melanogaster (see below). The consequence of parental imprinting is farless drastic in these groups of organisms. For example, gynogenic Drosophilamelanogaster are completely viable and fertile (Fuyama 1984) as are, apparentlyandrogenic flies (Muller 1958).Parental imprinting in Drosophila:Imprinting phenomena have been recognized and studied in Drosophila, albeit underthe name of parental effects, for more than 50 years. All of the reported parental213Table 26. Imprinting (parental effects) in Drosophila.Six examples of parental effects published for Drosophila are shown. The first columngive the direction of the imprint. Following the terminology of Reik (1992) the parenttransmitting the inactivated allele or chromosome is indicated. The second and thirdcolumns give the name of the rearrangement which shows the parental effect and thereference. The last example occurs in D. hyde!, all the others are found in D.melanogaster.214Imprinting (parental effects) in DrosophilaDirection of imprint Chromosome ReferencePaternal T(1;2)dorvar7 Demakova and Belyaeva1988Paternal Dp(1;4)wm254.58aBaker and Spofford 1959Spofford 1959Hesser 1961Spofford 1961Baker 1963Maternal Dp(1;3)wVCQ Khesin and Bashkirov1978Maternal Khesin and Bashkirov1978Paternal In(1)sc8and Prokofyeva-BelgovskayaDp(1;f)1187 1947Karpen and Spradling1990Paternal Dp(1;f)LJ9 this workMaternal In(1)wm2 Hess 1970215(imprinting) effects (Table 26) involve chromosome rearrangements that exhibitposition effect variegation. Position effect variegation is a process whereby a fullyfunctional gene becomes inactivated due to its relocation adjacent to a brokensegment of constitutive (Spofford 1976) or facultative (Cattanach 1970)heterochromatin. As the gene inactivation is correlated with the adoption ofheterochromatic morphology in the appropriate section of the salivary glandchromosomes (Hartmann-Goldstein 1967), it is thought that the genetic inactivation isdue to the spread of heterochromatin which packages and condenses the normallyeuchromatic region in such a way that necessary transcription factors cannot accessthe gene. The result is variable genetic inactivity.Imprinting as seen in Drosophila shows many intriguing similarities to the imprintingdescribed in mammals. The key features of genomic imprinting have been defined byMetz (1938), based on his work with the Dipteran insect Sciara, but are true ofimprinting phenomenon in all organisms. The primary criterion of imprinting is that theprocess affects genetically identical DNA, and is thus a strictly epigeneticphenomenon. The second criterion is that the imprint persists for only one generation.Thus the imprint is reversed by passage through meiosis. Genetically this is evidencedby parental effects but no grandparental effects. This latter feature may be aconceptual definition rather than a genuine mechanistic distinction. Long term effectssuch as seen in epimutation (Darn et al 1993), or medium term effects with diminishinggrandparental effect (such as seen with wmVC0; Khesin and Bashkirov 1978) may bemerely mechanistic variations on a theme. The third criterion for parental imprinting isthat the imprint is mitotically stable. This stability generates functionally distinct clonalregions. These clones are readily apparent in Figure 32 and have been noted asintrinsic features of imprinting in mice (Allen et al. 1988), maize (Kermicle and Aliman1990) and insects (Nur 1990). The clonal nature of the gene expression results in cell216to cell variability within a tissue. This variability suggests a stoichiometric processoperates in imprinting and may provide a clue to the mechanism responsible forimprinting. A fourth criterion is the physical continuity and extent of the imprint. Theimprinted region frequently affects entire regions of a chromosome and mayencompasses more than one gene. Such an effect is self-evident in coccids where theimprint encompasses an entire chromosome (Nur 1970) but is also seen in mammals.For example, the closely linked genes thought to be involved in the pathology ofPrader-Willis syndrome, SNRPN, SNF127, PAR-i and PAR-5, are coordinatelytranscribed, replicated and maternally imprinted (Gunaratne et al. 1995). Theneighboring genes H19 and !GF2 (Reik 1992) are also imprinted, although in thiscase in opposite directions. Cohen (1962) found that the closely linked genes white,split, notchiod, facet and roughest were all maternally imprinted in the wm254.58arearrangement. Imprinting of contiguous genes is also seen with the mini-chromosomeexamined here, the garnet gene and the nearby genes narrow abdomen and tiny(Figure 35) are paternally imprinted. It should be noted, however, that most examplesof parental imprinting in mammals have been reported for isolated genes. The relativerarity of mammalian imprints which encompass more than one gene may reflect thelower resolution of genetic studies of the imprint in mammals or possibly the greatergenome size of mammals as opposed to Drosophila melanogaster which might act tolimit a domain to one gene. Finally, the formation of aberrant chromatin structuresappears to be the most universal aspect of imprinting. Heterochromatin formation islikely responsible for the altered gene expression seen in position effect variegationwhich is involved in all of the reported cases of imprinted effects in Drosophila.Likewise imprinting seen in Coccids and Sciara (Chandra and Brown 1975) clearlyinvolve large segments of cytologically visible heterochromatin. Cytological studies ofthe imprinted heterochromatic chromosomes of coccids have shown that the imprintcan spread and can be somewhat variable (Nur 1970). Heterochromatin is also clearly217involved in mammalian X-chromosome inactivation (Lyon 1993) which is imprinted inboth marsupials and some tissues of eutherian mammals. Paramutation (anepigenetic phenomenon probably closely related to imprinting) seems to beassociated with altered chromatin structure in maize (Patterson, Thorpe and Chandler1993). Aberrant chromatin structure has been proposed as the causal feature leadingto mis-expression of the fmrl gene responsible for fragile X mental retardation (Laird1987) and a number of other recent studies have implicated chromatin bindingproteins in the establishment of the imprint in mice (Sasaki et al 1992, Bartolomei et al1993, Brandeis et al 1993, Ferguson-Smith et al 1993, Stoger et al 1993, Chaillet et al.1995) and humans (Monk 1988, Barlow 1994, Varmusa and Mann 1994).This chapter describes a mini-chromosome which shows parent specific imprinting.The imprint is manifest as clonally repressed expression of the wild type garnet genewhen the mini-chromosome is inherited from the father. The expression of at least twoother genes on this mini-chromosome is also imprinted. The immediate cause of therepressed garnet expression is position effect variegation. This variegation isunconventional in that it is strictly dependent on the sex of the transmitting parent.Thus it would appear that chromatin packaging in this mini-chromosome is imprinted.Using the cloned garnet gene sequence to analyze this mini-chromosome I havetested and eliminated a number of factors which might cause the imprint. To assessthe role of heterochromatin in the imprinting process, I have tested the stability of theimprint in this region using chemical, environmental and genetic modifiers of positioneffect variegation which are thought to alter heterochromatic formation. Thesemodifiers of heterochromatin formation and integrity alter the expression of the imprintbut not the initial decision of whether or not to imprint. This implies that whileheterochromatin is involved in the maintenance and somatic expression of the imprint,218it probably does not establish the imprint. Thus the imprinting decision may be underindependent genetic control.219RESULTS-imprintingThe Dp(1;f)LJ9 mini-chromosome.Origin:The Dp(1;f)LJ9 mini-chromosome is derived from the In(1)sc29chromosome. TheIn(1)sc29chromosome is an inversion between the tip of the X chromosome (1 B) andthe region adjacent to the garnet gene (13A2-5) which places this region near the tipof the X chromosome. The Dp(1;f)LJ9 mini-chromosome was induced by X-raytreatment and is a deletion of most the X chromosome euchromatin from In(1)sc29(Hardy et al 1984). Only the euchromatic bands from 12A10 to 13A2 and the distal tip,1A1 to 1 B remain. This region is appended to the centric heterochromatin of the Xchromosome. The general structure and origin of the Dp(1;f)LJ9 mini-chromosome isdiagrammed in Figure 30. Although not mentioned by Hardy eta!, (1984) it may beassumed that the heterochromatin was broken by the X-ray treatment as the mostproximal euchromatic gene (su(f)) is missing from the mini-chromosome andvariegation for a number of genes adjacent to the heterochromatin is observed (seebelow).Mitotic and meiotic stability.As the centric heterochromatin (or a region within) is involved in the normalsegregation of chromosomes, I tested the stability of the mini-chromosome in bothmeiotic and mitotic cell divisions (Figure 31). The mini-chromosome shows a verylow rate of non-disjunction from males (0.1-0.3%, n=1 546). This rate is comparable tothat observed for a standard X chromosome (Bridges 1916). The rate of non220Figure 30. Diagram of the structure and origin of the Dp(1,i)LJ9 mini-chromosome.The top figure diagrams the wild type chromosome. The middle figure shows thestructure of the In(1)sc29chromosome and the lowest figure shows the structure of theDp(1;f)LJ9 mini-chromosome. The relative positions of the narrow abdomen, tiny andgarnet genes are shown.221Structure and origin of the Dp(1:f)LJ9 mini-chromosomeX RAYSn&tygTi1B113A2 12A10THE MINI-CHROMSOSOMEDp(1 ;f)LJ9\F/WILD TYPE X CHROMOSOME1 B1 B/i 3A2 tAO ic/i 3B29In(1)sc2b I222Figure 31. Meiotic stability of the Dp(1;f)LJ9 mini-chromosome.The top portion of the figure diagrams the production of regular and non-disjunctiongametes from the attached-X and duplication bearing females (X”X/Dp(1;f)LJ9). Thefrequency with which the various genotypes arise is shown in the Punnit square. Thetotal number of flies scored for the maternal cross was 745.Cross: X’X/Dp(1;f)LJ9 0 y zag53d/y0&yzag53d/Dp(1,.f)LJ9 a’The lower portion of the figure diagrams the production of regular and non-disjunctiongametes from the attached XY duplication bearing male (XY/Dp(1;f)LJ9). Thefrequency with which the various genotypes arise is shown in the Punnit square. Thefrequency of the y za g53d/Q genotype arising from the paternal cross is given as arange because all of the five individuals of this genotype arose from one vialcontaining only three females. Thus this genotype may represent a pre-meiotic loss ofthe mini-chromosome. The total number of flies scored for the paternal cross was1546.Cross: X’Y/Dp(1;f)LJ9 a’ ® y zag53d/yza g53d y zag53d/Dp(1,’f)LJ9c/223DISJUNCTION IN FEMALESregulargametesDpnon-disjunctiongametesYô(+Op 053dy53drn 0.47 tRIP 0.10Q37 RIP 006 RIPCROSS: XIDp(1;f)LJ9 y4 53di yDISJUNCTION IN MALESregular non-disjunctiongametes gametes)t’ Dp fv+Dp 00.56 0.44 0 0.001-0.003CROSS: )&/Dp(1 ;f)LJ9 x yz 53diyz224disjunction from females is 16% (n=745) which presumably reflects the inability of themini-chromosome to pair and recombine with the full length attached X-chromosomesin the female. Mitotic non-disjunction was monitored by looking for mosaic patches ofmutant yellow tissue in a fly bearing a mutant yellow gene on its normal Xchromosome (y zag53d/Dp(1;f)L.j9, y or y zag53d/yza53d/Dp(1;f)LJ9, y) as themini-chromosome carries the wild-type allele of yellow. Only 2 instances of mosaicismwere seen in over 10,000 flies suggesting that mitotic non-disjunction does nothappen at an appreciable rate (interesting but probably not relevant is the fact thatthese two mosaics were perfect bilateral gynandromorphs and came from the sameparent). Thus it would appear that the mini-chromosome is transmitted faithfullythrough both meiosis and mitosis suggesting that the bulk of the centricheterochromatin, telomere and any other functions necessary for normal disjunctionare not compromised.Imprinted variegation of the mini-chromosome.The variegation is dependent on the sex of the transmitting parent:When females carrying the mini-chromosome (XX/Dp(1;f)LJ9) are crossed to yzag53d/ymales, the male progeny (y zag53d/Dp(1;f)LJ9) appeared wild type. This is theexpected phenotype as the mini-chromosome carries the wild type genes for yellowand garnet (the za allele has no independent phenotype and serves only to lighten thegarnet mutant phenotype in the background). In contrast, when the males carrying themini-chromosome were crossed to females of the same y za g53d strain, the wild-typegarnet gene on the mini-chromosome showed variegated expression in thegenotypically identical y zag53d/Dp(1,i)LJ9) sons (Figure 32). In most cases thevariegation was expressed as no, one, two or three large wild type spots (garnet+) on225Figure 32. The garnet phenotype in flies with paternally or maternally inherited Dp(1;f)LJ9 mini-chromosomes.A. The top figure shows two male flies of identical genotype; yzag53d/Dp(1;f)LJ9. Thefly on the left, with variegated eyes, has a paternally transmitted mini-chromosomewhereas the fly on the right bears a maternally transmitted mini-chromosome.Maternal cross: X”X/Dp(1;f)LJ9 ® y za g53d/y01yzag53d/Dp(1;f)LJ9Paternal cross: X’Y/Dp(1;f)LJ9 ‘ ® y zag53d/y za g53d>y zag53d/Dp(1;f)LJ9fB. The lower figure shows two female flies of identical genotype; y zag53d/yzag53d/Dp(1;f)LJ9. The fly on the left, with variegated eyes, has a paternally transmittedmini-chromosome whereas the fly on the right bears a maternally transmitted mini-chromosome.Maternal cross: y zag53d/yzag53d/Dp(1;f)LJ9 a y zag53d/yzag53dj‘I,y zag53d/yzag53d/Dp(1,.f)LJ9Paternal cross: y zag53d/yza g53d ® y za53d/y/Dp(1;f)LJ91y zag53d/yzag53d/Dp(1;f)LJ9In all cases the genotypes being compared differ only in the parental origin of the minichromosome. They are otherwise genotypically identical and isogenic. The yellowmutation was used to monitor the presence of the mini-chromosome without bias as toeye phenotype. The zestea allele was used to lighten the background garnet eye226colour of the g53d allele. In the absence of wild type garnet expression from the mini-chromosome the background eye colour is a pale orange due to the za and g53dmutations. In cells in which the garnet+ gene on the mini-chromosome is expressedthe eye colour is wild type. In most of the following experiments males of yzag53d/Dp(1;f)LJ9 were used to monitor the imprint as they were the genotype that arosefrom both standard maternal and paternal crosses. These phenotypes persist for onegeneration only. For example, both the females shown in part B will transmit nonvariegating mini-chromosomes regardless of whether they themselves showvariegation for the garnet gene.227Nthe pale orange background colour. These spots appeared to correspond to clonallydetermined regions of the eye (Jannings 1970). This mosaic expression was probablynot due to mitotic loss of the mini-chromosome in the eye since the mini-chromosomeis mitotically stable in the rest of the fly. This imprinted expression is seen in both maleand female progeny (Figure 32) and persists for only one generation (Table 35).To summarize, genotypically identical progeny, produced by reciprocal crosses andthus differing only in the parental origin of the sex chromosomes and the mini-chromosome, show very different phenotypes. When the mini-chromosome is derivedfrom the male parent the garnet gene variegates extensively. In contrast, geneticallyidentical offspring resulting from the reciprocal cross in which the mini-chromosome isderived from the female parent, showed no or very weak variegation (Figure 32).Thus the parental origin determines the extent of variegation observed in thegenotypically identical progeny. Therefore the expression of the wild type garnet geneis dependent on the sex of the parent transmitting the mini-chromosome. This situationconstitutes a classical example of genomic imprinting. The many parallels betweenthis example of imprinting and that seen in coccids and mammals will be discussedbelow.Tissue specificity of the imprint:This parent-dependent expression is not limited to the eye. Examination of malpighiantubules in individuals bearing maternally versus paternally derived mini-chromosomesdemonstrate that the garnet gene expression in this tissue also variegated extensivelywhen the mini-chromosome is transmitted by the father (Figure 33). Malpighiantubules from individuals in which the mini-chromosome is maternally derived showonly occasional unpigmented spots. Testes sheaths from individuals of these samegenotypes showed a uniform wild type pigmentation as expected since the g53d allele229Figure 33. garnet phenotype in malpighian tubules bearing maternally andpaternally derived Dp(1;f)LJ9 mini-chromosome.The top figure shows a typical malpighian tubule from a male of the genotype yzag53d/Dp(1;f)LJgf. where the mini-chromosome was inherited from the father. Thedark areas are individual pigmented cells. The clear areas are regions in which nopigment is produced.The lower figure shows a typical malpighian tubule from a male of the identicalgenotype y za53d/Dp(1;f)LJ9. In this instance the mini-chromosome was inheritedfrom the mother. Only one unpigmented region is seen. Frequently there are nonealthough up to three unpigmented regions have been found within one tubule fromindividuals of this genotype.Wild type malpighian tubules are always uniformly pigmented, unless damaged inremoval (data not shown).Crosses:Maternal X’X/Dp(1;f)LJ9 ® y za g53d/Y0&...). y zag53d/Dp(1;f)LJ9fPaternal: XY/Dp(1;f)LJ9 ‘ ® y zag53d/y za g53d—y zag53d/Dp(1,.f)LJ9f2304•1N)()I—.4-Idoes not affect testes sheath pigmentation.Imprinted expression of narrow abdomen and tinyTo determine if the parental effect is peculiar to the garnet gene or is due to somegeneral feature of the mini-chromosome, I tested two other, closely linked, cellautonomous genes narrow abdomen (na) and tiny (ty) to determine if the expressionof these genes is also dependent on parental origin. When the mini-chromosome wasinherited from the father, the wild type gene on the mini-chromosome showed variablehypomorphic expression (Figure 34 and 35) in males mutant for narrow abdomen ortiny on the standard X chromosome (ty/Dp(1;f)LJ9 or na/Dp(1;f)LJ9). In contrast, whenthe mini-chromosome was maternally inherited, the genotypically identical progenywere generally wild type. Thus the imprinting effect seems to encompass at least threegenes in the small euchromatic portion of the mini-chromosome. The “imprinting”effect for all these genes was paternal. Interestingly, the severity of the imprint, asassessed by the difference between gene expression when either maternally orpaternally inherited, decreased with distance from the centromere and the centricheterochromatin.Imprinting and position effect variegation:The clonal phenotype of the garnet variegated expression is distinctly reminiscent ofposition effect variegation. The variable hypomorphic expression of the genes tiny andnarrow abdomen is also indicative of position-effect variegation. As it is probable thatthe centric heterochromatin was broken when the mini-chromosome was generated Itested several different types of modifiers of position effect variegation to determine ifthey modified the mosaic expression of the garnet gene. They did, (see below)suggesting that the mosaic expression of the garnet, narrow abdomen and tiny genes232Figure 34. Phenotype of narrow abdomen and tiny in flies bearing maternally orpaternally derived Dp(1;f)LJ9 mini-chromosomes.The top portion of the figure shows two genetically identical flies bearing a mutation forthe narrow abdomen (na) gene on the standard X-chromosome and the wild typeallele of narrow abdomen on the mini-chromosome (na/Dp(1;f)LJ9). The individual onthe left bears a maternally derived mini-chromosome whereas the genotypicallyidentical fly on the right has a paternally derived mini-chromosome. Arrows point to theabdomen which is considerably more extended (more mutant) in the flies bearing apaternally derived mini-chromosome.Crosses:Maternal X’X/Dp(1;f)LJ9 ® na/Ye’—> na/Dp(1;f)LJ9cIPaternal: XY/Dp(1;f)LJ9 010 na/In(1)d149 —na/Dp(1;f)LJ9o’The lower figure shows thoracic bristles from flies with a mutation for the tiny (ty) geneon the regular X-chromosome and the wild type tiny gene on the mini-chromosome(ty/Dp(1;f)LJ9). The individual on the left bears a maternally derived mini-chromosome.The genotypically identical fly on the right bears a paternally derived minichromosome. Arrows point to the thoracic bristles which are smaller and finer (moremutant) in the flies bearing a paternally derived mini-chromosome.Crosses:Maternal X’X/Dp(1;f)LJ9 0 g2 ty/Yo’—> na/Dp(1;f,)LJ9o’Paternal: X’Y/Dp(1;f)LJ9 010 g2 ty/In(1)d149 .-÷na/Dp(1;f)LJ9o’233nalDp(1;f)LJ9MATMATtyiDp(1,f) LJ9PATPAT234Figure 35. Expression of garnet, narrow abdomen and tiny in genotypically identicalflies bearing either paternally or maternally derived Dp(1;f)LJ9 mini-chromosome.This figure shows the level of variegation for three closely linked wild type genes,garnet (position 44.4), narrow abdomen (position 44.5) and tiny (position 45.2), on theDp(1;f)LJ9 mini-chromosome.The level of expression of these genes was assessed visually as follows: The level ofexpression of each of these genes was visually estimated. Each fly, or eye in the caseof garnet expression, was assigned a score of 0, 1/2 or 1 for extreme mutant,hypomorphic and wild type expression respectively. These values were averaged andthis value is presented with the calculated standard error of the mean.A fly was scored as tiny if its bristles appeared severely minute. In practice this meantapproximately half the length of regular bristles on the sibs. A fly was scored as narrowabdomen if the abdomen appeared exceptionally long and thin. A fly was scored as amutant for garnet if its eyes were completely or extensively variegated. As the scoringis somewhat subjective the crosses were scored by using flies grown on the same setof media at the same time and the crosses were scored blind.235Parent-dependent expression ofnarrow abdomen. tiny jjsgarnetEXPRESSION OF GENES ON THEMINI-CHROMOSOMEmini-chromosomederived frommothermini-chromosomederived fromfathernarrow abdomen tiny garnet(na) (ty) (g)0.67±0.02 0.52±0.02 0.86±0.020.51 ±0.02 0.24±0.02 0.06 ±0.02236transmitting parent, the immediate cause of the imprinting of the garnet, narrowabdomen and tiny genes is imprinted position effect variegation. The possible causesof this imprinted variegation are explored below.Possible causes of the imprinting effect.There are a number of possible causes of imprinting. Obviously the mechanism ofimprinting must be established on the haploid genome and thus rely on some featureof meiosis or gametogenesis that differs between females and males or on thesubsequent fates of the gametes until fusion of the pronuclei. I directly tested theinvolvement of a number of factors in the establishment and maintenance of theimprint. The following possible causes of the imprinting effect were tested:1. The Y chromosome2. Allele specific interactions3. Maternal effect of the imprinted gene4. Maternally transmitted modifiers of variegation5. Unusual euchromatic sequence or DNA structure near the garnet gene6. Special heterochromatic featuresDetails of experiments designed to test these possibilities are provided below.1. The Y chromosome.Imprinting is clearly associated with the sex of the parent from which the imprintedchromosome is derived. Thus, imprinting must be a consequence of eitherphysiological or genotypic differences between the parents. The only geneotypicdifference between the parents is the presence of a Y chromosome in males in place237of one of the X chromosomes. The Y chromosome is a large heterochromatic body,known to affect many processes such as chromosome pairing, expression ofheterochromatic genes and position effect variegation. Therefore the parental effectmight be due to a direct interaction between the Y chromosome and the mini-chromosome. If this were the case, the mini-chromosome should be imprinted whentransmitted by XXY females. This hypothesis was tested by comparing variegation ofgenotypically identical XIDp progeny from XXY and standard XX females. Thephenotype of the progeny of these crosses was identical (Figure 36). Thus the mini-chromosome responds to the sex of the parent rather than the presence or absence ofa Y chromosome.2. Allele specific interactions.Imprinting of the mini-chromosome could be a consequence of an interaction betweenthe wild type allele of garnet on the mini-chromosome and the g53dallele on theregular X chromosome. If this were the case, a parental imprinting effect would not beseen with other garnet alleles. To determine if the imprinting effect was allele specific,the variegation of the mini-chromosome was examined in strains heterozygous for thewild type garnet gene on the mini-chromosome and eight different garnet alleles onthe regular X chromosome. In every case variegation was evident when the mini-chromosome was introduced paternally but not when introduced maternally (Table27). This suggests that the imprinting effect is not due to some combination of thegarnet alleles. This result and the imprinting of the nearby narrow abdomen and tinygenes also argue against specific allelic interaction as the cause of the imprint.3. Maternal effect of the garnet gene.It is possible that the differential expression of the garnet gene (and the other twogenes) results from maternal rescue of the mutant phenotype rather than paternal238Figure 36. The Y chromosome does not cause the imprint.The possibility that the imprint was caused directly by the Y chromosome, as opposedto the sex of the parent, was tested by examining the variegation of the standard testgenotype, yzag53d/Dp(1;f)LJ9, when the mini-chromosome was derived fromfemales with or without a Y chromosome. The first experimental column shows theresults of transmission of the mini-chromosome from males (which necessarilypossess a Y chromosome). The next two experimental columns show the level ofgarnet expression in genotypically identical yzag53d/Dp(1;f)LJ9 progeny in whichthe mini-chromosome was derived from females with and without a Y chromosome,respectively. The values shown in the middle of the figure give an indication of thelevel of expression of the garnet gene as an indicator of the imprint. The expression ofthe garnet gene was monitored both visually and quantitatively by microflourimetricassay. The eye phenotype is indicated schematically at the bottom of the figure.Values are taken from Table 34.Crosses:Paternal: X’Y/Dp(1;f)LJ9 ‘ ® y zag53d/y za g53d—>y zag53d/Dp(1;f)LJ9cJMaternal : WITH Y CHROMOSOMEX’X/Dp(1;f)LJ9 0 any male y zag53d/yf‘I,X’X/Y/Dp(1;f)LJ9 ®y zag53d/y‘Iy zag53d/Dp(1;f)LJ9fMaternal: WITHOUT Y CHROMOSOMEX’X/Dp(1;f)LJ9 0 y zag53d/Y0&yzag53d/Dp(1;f)LJ9239The imprint is not caused by the Y chromosomeF’xYDp(1 ;f)U901y zag53d,Dpu9)CkYDp(1 ;f)LJ9Izag53d,DpgDp(1 ;f)LJ9zag53d,p9visualestimatepigmentassay0. 10 ± 0.02 1.0±0 0.92 ± 0.0141 ±2 94±4 100± 3SVARIEGATED WILD TYPE WILD TYPE(IMPRINTED) (NOT IMPRINTED) (NOT IMPRINTED)240Table 27. The effect of different garnet alleles on the imprint.The first column lists the garnet allele. The second and third columns indicatesvariegation when the Dp(1;f)LJ9 mini-chromosome is transmitted maternally orpaternally, respectively.Cross:Maternal cross: X’X/Dp(1;f)LJ9 ® g*/yl > g*/Dp(1;f)LJgfPaternal cross: X’Y/Dp(1;f)LJ9cJ ® g*/g*> g*/Dp(’1;f)LJ9c/where g* is the given allele of garnet.241The effect of different garnet alleles on the imprintgarnet allele on the phenotype of phenotype ofregular X g/Dp g/Dpchromosome Maternal Paternalg1 wild type variegatedg2 wild type variegatedg3 ND variegatedg4 ND variegatedg5Oe wild type variegatedg53d wild type variegatedg61 wild type variegatedgEMS wild type NDgP wild type NDgS3 wild type variegated242regardless of parental origin, but if the mini-chromosome were the only source of wildtype garnet+ product, and if the mother were able to deposit wild type garnet productin the egg, a conventional maternal effect could be misinterpreted as paternalimprinting. This explanation is unlikely. Firstly, more than fifty different garnet alleleshave been examined and none show a maternal effect (Lindsley and Zimm 1992).Secondly, the tissue profile of the garnet gene failed to reveal high levels of garnetmRNA in ovaries and eggs (Figure 27) as would be expected for a maternallydeposited product. Finally this argument would have to be extended to the otherimprinted genes, narrow abdomen and tiny. Nevertheless, to completely exclude thisargument I genetically tested the garnet gene for a maternal effect on eyepigmentation. No maternal effect was detected (Table 28). I also directly tested themini-chromosome for maternal effect rescue by a wild-type copy of the garnet gene. Ifthe imprinting effect was due to maternal rescue of the garnet mutant phenotype,providing a wild type garnet allele (elsewhere than on the mini-chromosome) shouldeliminate the imprinting effect. Specifically, female parents with a wild type garnetallele should deposit enough wild-type product in the egg cytoplasm to rescue thevariegation of the garnet allele on the Dp(1;f)L9 mini-chromosome derived from thepaternal parent. This was not the case (Table 28). Thus a maternal effect for thegarnet gene can be excluded as a cause of the imprint.4. Physiological compensation.The physiological compensation model is a variant of the maternal effect hypothesis.This model posits that the female parent produces, and transmits to the egg, factorscapable of suppressing variegation, but only in response to the variegating Dp(1;f)LJ9mini-chromosome in her cells. This scenario is formally equivalent to that suggested toexplain the imprinting effect in the imprinted diseases Huntington’s Chorea (Bird, Caroand Pillins 1974) and fragile X syndrome (Van Dyke and Weiss 1986). Adapted to this243Table 28. Maternal effect of the garnet gene.The potential maternal effect of the garnet gene was examined in two ways. The firsttest involved generating genotypically identical zag53d/Dp males (where the Dpchromosome was paternally derived) produced from females heterozygous (column 2)or homozygous (column 3) for the garnet gene. The amount of variegation for garnetwas assayed visually (first row of data) as described in the legend to Figure 34 andby microflourometric measurement of pteridine pigments (second data row, the valuesare expressed as percent wild type pigmentation).Cross: X’Y/Dp(1;f)LJ9cf 0 ÷÷+/y za g53d or y zag53d/yza g53d1-y zag53d/Dp(1;f)LJ9cfThe lower portion of the table shows a direct test for a maternal effect of the garnetgene. The amount of pigment was quantitated in homozygous daughters (first row ofdata) and hemizygous sons (second row of data) derived from females heterozygous(column 2) or homozygous for garnet (column 3). All values are again given aspercent wild type pigment levels.Cross: y za g53d/y1f® yzag53d/yza g53d or +/y za g53d1y zag53d/Dp(1;f)LJ9244Maternal effect of the garnet gene1. Effect on imprint heterozygous homozygousmother mothery zag53d/Dp(1;f)LJ9 + /y zag53d yzag53d/yzag5dvisual estimate 0.14 ± 0.03 0.10 ± 0.02pigment assay 73±8 41 ±22. Effect on garnet heterozygous homozygouspigment mother mother+/yzag53d yzag53d/yzag5dyzag53d/yza g53d 4±1 10±1yzag53d/Y 6±1 12±2245example of imprinting, this model posits that flies bearing the variegating mini-chromosome experience a physiological stress. They compensate for this stress byproducing substances that reduce the level or frequency of variegation. (Thesehypothetical substances would have to be germ line specific as the female somashows variegation when bearing a paternally transmitted mini-chromosome, Figure37b.) It these substances were additionally transmitted cytoplasmically then femalesthat possessed a mini-chromosome would produce and transmit substances capableof rescuing the variegation, and thus their progeny would appear wild type. Malesbearing a mini-chromosome would obviously be unable to transmit such substances totheir progeny.Initially, this is seemed a plausible explanation for the imprinting effect. Variegationcan be deleterious (unpublished observations) and many modifiers of position effectvariegation are early acting and are maternally deposited (Sinclair et al 1992, Dorn etal. 1993). To examine this model, crosses were performed in such a way thatgenotypically identical progeny were generated, all of which receive the mini-chromosome from their paternal parent. The crosses differ in whether or not thematernal parent bore a mini-chromosome and thus might possess and transmit thehypothetical compensatory molecules. y zag53d/Dp progeny from normal motherswith a paternally derived mini-chromosome had an arbitrary visual score of variegationof 0.05 ± 0.01 corresponding to pigment levels of 48±2% (see figure legend).Genotypically identical progeny from mothers possessing a variegator had a score of0.07 ± 0.04, corresponding to pigment levels of 46 ± 9% (Figure 37). These valuesare not satirically different. Thus there is no evidence that the imprint is due to maternaltransmission of substances capable of suppressing variegation.246Figure 37. Test of the physiological compensation model.A. The top portion of this figure diagrams the physiological compensation model whichhas been invoked to the explain reduced severity of Huntington’s chorea when thedisease gene is maternally inherited. The model essentially postulates that maternaltransmission compensatory substances (shown as smiling faces in the leftmost figure)produced in response to the abnormal condition, mitigates the impact of the disease inthe offspring of these “conditioned” females. In genetic terms this is essentially modifiergene products with a maternal effect. This models has been proposed implicitly byBird, Caro and Pillings (1975) and Van Dyke and Weiss (1980) and explicitly by DavidBaillie (personal communication).B. The lower figure shows a test of the Drosophila equivalent of this model. Essentiallythe model postulates that mini-chromosome-bearing mothers will transmit cytoplasmicfactors to the embryo which are capable of mitigating the extent of variegation. The firsttwo rows show the maternal and paternal genotypes. The third row indicates theparent contributing the mini-chromosome to the embryo. The fourth row shows thepotential for contribution of maternal modifiers which might modify the imprint. The lastrows show the phenotype and amount of variegation of the diagnostic y zag53d/Dp(1;f,)LJ9 progeny as a measure of the imprint, assessed both bymicroflourimeter and visual assay. Pigment values for the microflourimeter assay areexpressed as percent wild type levels.247Crosses.Experimental 1. X’X/Dp(1;f)LJ9 ® y zag53d/y0&....> y zag53d/Dp(1,.f)LJ9fExperimental 2*. y zag53d/y zag53d/Dp(1;f)LJ9 ® XY/Dp(1;f)LJ9cf—>y zag53d/Dp(1;f,)LJ9cfExperimental 3. X’Y/Dp(1;f)LJ9 cf ® y zag53d/yza g53d>y zag53d/Dp(1;f)LJ9* Two mini-chromosomes are lethal due to presence of the male diplo-lethal region. Asa result the y zag53d/Dp(1;f,)LJ9 must have a paternally derived mini-chromosome asthe y za g53d homolog must come from the mother.248A. The physiological compensation modelB.LESSMATERNALMOREPATERNALMINI-CHROMOSOMEMATERNAL CYTOPLASMPATERNALPATERNAL GENOTYPEMATERNAL GENOTYPESOURCE OFMINI-CHROMOSOMEvisualestimatey zag5sq,YXXIDpLJ9motherpotentialmodifiersWILD TYPEyz53d,yIDpXX/DpfatherpotentialmodifiersXYIDpa53d 53dyzg iyzgfatherVARIEGATEDMOTHER+1+(later onset) (earlier onset)MATERNALCONTRIBUTIONpercentwild typepigmentnone98±1 46±9 48±20.82 ± 0.02PHENOTYPE0.07 ± 0.04VARIEGATED0.05 ± 0.012495. Unusual structure of the euchromatic region encompassing the garnet gene.It is possible that the garnet gene possesses some unusual feature which wouldpermit the ingress of heterochromatin when the mini-chromosome was received fromthe male. The DNA sequence of the garnet gene itself is generally unremarkable(Figure 26). Nevertheless, it is possible that there is some gross structuralabnormality in the region surrounding the garnet gene which could allow theaggressive invasion of heterochromatin, although in a parental specific manner.Localized changes in somatic copy number of genes showing position effectvariegation have been observed in polytene tissue (Karpen and Spradling 1990). Ifaltered copy number were the cause of the differential expression of the garnet gene,one would predict that individuals with maternally, versus paternally, derived mini-chromosomes would differ in DNA content in the vicinity of the garnet gene. To testthis, DNA was extracted from flies which had either a paternally or a maternallyderived mini-chromosome and an X chromosome marked with the g2 allele, which hasa restriction polymorphism 2 kb from the 3’ end of the garnet gene. These individualswere tested for alteration in copy number of the garnet gene by Southern blot analysis.Figure 38 shows that while there is considerable under-representation of DNA at thegarnet locus, the degree of under-representation does not correlate with the amount ofvariegation at the phenotypic level, nor with the parental origin of the mini-chromosome. Thus the under-representation seems unrelated to the imprinting andmay simply reflect the limited amount of euchromatin present in the mini-chromosome.The control of under representation at the garnet locus of this mini-chromosome iscomplex. The results shown in Figure 38 are derived from males bearing an Xchromosome marked with y e(g) cv g2 and the mini-chromosome. Very few males ofthis genotype are obtained because the mini-chromosome appears to variegate for adiplo-male-lethal locus identified in this region (Stewart and Merriam 1973, Belote and250Figure 38. Under-representation of the garnet gene in the Dp(1;f)LJ9 mini-chromosome.This figure shows Southern analysis of whole, newly eclosed, y e(g) cvg2/Dp(1;f)LJ9males where the mini-chromosome was inherited maternally (first lane - MAT.) orpaternally (second lane - PAT.). DNA was extracted from (0-1 day) males of theappropriate genotype, restricted with Eco RI, separated on a 1% agarose gel,transferred to nylon membrane and probed with the garnet c-DNA derived from theimaginal disc library. Whole flies were used as a source of DNA as the garnet geneseems to be expressed in a variety of tissues (Figure 27) and males of this genotypewere sufficiently rare to exclude analysis of isolated tissues. The g2 allele which marksthe normal X chromosome is associated with a Eco RI restriction polymorphism 2 kb 3’to the garnet gene and thus serves as an internal control. This generates the lowermobility, 8.5 kb band. The wild type garnet gene on the mini-chromosome producesthe higher mobility, 6.6 kb, band. The lowest band is a 0.2 kb Eco RI band generatedby both the normal X chromosome and the mini-chromosome.251I 0,Lucchesi 1980). The lethality seems to occur principally at the pupal stage, structurallynormal pharate adults form but do not eclose. This diplo-lethality appears sensitive tothe zeste gene as it is alleviated by za and z1 alleles. When this experiment wasrepeated using a za g2 strain, to increase the number ofg2/Dp progeny for molecularanalysis, the under representation was not observed (data not shown). Thus theunder-representation appears to be zeste sensitive. Whether there is a causalrelationship between zeste mediated pairing, transcriptional repression, the malediplo-lethality and sequence under-representation is unclear. The question of under-representation is further complicated by recent results which suggest that the observed“under-representation” is an artifact of DNA modification which inhibits transfer to solidsupport during Southern blotting (Glasser and Spradling 1994). Whether this is alsotrue for the under-representation around the garnet gene in the Dp(7;f)LJ9 mini-chromosome is currently being investigated by the Glasser lab.6. The role of heterochromatin.There are a number of features shared by imprinting and position effect variegationwhich might suggest a mechanistic link between the two phenomena. Asheterochromatin formation has been posited to cause position effect variegation andas heterochromatin seems widely involved in imprinting, I decided to test factorsknown to modify position-effect variegation for both their effect on the variegation,which is the mode of the expression of the imprint, and also on the transmission of thevariegation, that is, the imprint itself.There are three general groups of factors that affect position effect variegation. Theyare: environmental factors such as temperature, chemical factors, such as sodiumbutyrate and genetic factors such as the presence of extra heterochromatin in the cell,usually in the form of an additional Y chromosome. I tested all these factors for their253Table 29. Variegation of the Dp(1,i)LJ9 mini-chromosome: The effect ofdevelopmental temperature.Cultures generating g/Dp(1;f)LJ9 males, where the mini-chromosome was transmittedeither maternally or paternally, were raised at 18, 22, 25 and 290. The amount ofpigment in the eyes of these progeny was determined by microflourimeter assay. Nonspecific effects of temperature on pigment (such as effects on fly size or number) werecontrolled by assaying siblings. No such effects were noted (data not shown). The toprow indicates the temperature of the culture. MAT and PAT indicate the parental originof the mini-chromosome. Values derived from the paternal cross are shown in italicsfor contrast. The garnet allele of the non-Dp bearing parent, either g53d, g500or g2, isindicated in the first column. ND = not done. All values are given as percent wild typepteridine pigments.Crosses:Maternal cross: X”X/Dp(1;f)LJ9 ® y zg53d/y or y zag5Oe/yf ory e(g) cvg2/Yo’yza g”/Dp(1,’f)LJ9o’Paternal cross: X”Y/Dp(1;f)LJ9o’ ® y zag53d/yza g53d ory zag5Oe/y zag5Oe ory e(g) cvg2/ye(g) cvg2‘Iyza254The effect of culture temperature on variegation of Dp(1;f)LJ9.culture temperature18° 22° 25° 29°g53dMAT 72±8 73+8 64±6 39±5PAT 44± 14 37± 13 36±7 28±4g5OeMAT 70±8 80±9 71±8 44±7PAT 47±17 50±6 52±9 35±5g2MAT ND 93±10 105±8 90±11PAT 83±9 89±9 84±15 84±7255Table 30. Variegation of the Dp(1,i)LJ9 mini-chromosome: The effect of sodiumbutyrate.Cultures which generatedg53d/Dp(1;f)LJ9 males, where the mini-chromosome wastransmitted either maternally or paternally, were supplemented with no, 100, 150, 200,250 or 300 mM sodium butyrate (top row). The amount of pigment in the eyes of theseprogeny was determined by microflourimeter assay (row five and six, all valuesexpressed as percent wild type pteridine levels). Non specific effects of temperature onpigment (such as effects on fly size or number) were controlled by assayingphenotypically wild type XX//Y siblings. These results are shown in the second row.Flies variegating for In(1)wm4 were grown alongside the Dp(1;f)LJ9 crosses onidentical sets of supplemented media as a control for the effectiveness of butyratetreatment (row three and four). ND not done.Crosses:Maternal cross: X’X/Dp(1;f)LJ9 ® y za g53d/y1f1-y zag53d/Dp(1;f)LJ9cfPaternal cross: X’Y/Dp(1;f)LJ9o’ ® y zag53d/y za g53d‘Iy zag53d/Dp(1;f)LJ9m4 cross: wm4/w ®wm4/Ycf> m4 and ci’.256The effect of butyrate on variegation of Dp(1;f)LJ9.butyrate concentration0 mM 100 mM 150 mM 200 mM 250 mM 300 mMX’XhY 98±4 98±4 91±3 87±3 ND NDwm4 14±3 18±4 34±4 23±3 40±11 65±10wm40’ 1±1 19±6 9±1 18±6 18±13 13±10yzag53d/DpLJ9(maternal) 87±6 81 ±4 71 ±3 64±6 ND NDyzag53d/DpLJ9(paternal) 40±2 44±5 38±3 31±2 34±2 34±3257Table 31. Variegation of the Dp(1;f)LJ9 mini-chromosome: The effect of Ychromosome dosage:Effect of Y chromosome dosage on garnet variegation of Dp(1;f)LJ9 is shown forindividuals in which the mini-chromosome was derived either maternally or paternally.The first and second columns show the amount of variegation assayed both visually(row 1) and by microflourimeter assay (row 2, all values expressed as percent wildtype pteridine levels) for genotypes which are X/Dp and X/Dp + Ywhere the mini-chromosome is paternally derived. The third and fourth columns show the equivalentgenotypes with a maternally derived mini-chromosome.Genotype of the X/Dp flies is yzag53d/Dp(1;f)LJ9. That of the X/Dp + Yflies isy zag53d/Dp(1;f)LJ9/y.Crosses:Maternal cross: X’XJY/Dp ® y zag53d/y‘Iy zag53d/Dp(1,.f)LJ9 and y zag53d/Dp(1;f)LJ9/ycJThese progeny were separated by progeny testing.Paternal cross: y zag53d/y/Dp(1;f)LJ9 ‘® y zag53d/yza g53d‘I,yzag53d/Dp(1;f)LJ9cJ and yzag53d/Dp(1;f)LJ9/ycfThese progeny were separated by progeny testing.258The effect of Y chromosome dosage on variegation of Dp(1;f)LJ9.Paternal MaternalX/Dp X/Dp + Y X!Dp X/Dp + Yvisual estimate 0.07 ± 0.03 0.97 ± 0.03 1.0 ± 0 1.0±0pigment assay 37 ± 9 99 ±4 79±5 94±4259ability to affect both the extent and the transmission of variegation; that is themaintenance and establishment of the imprint. All these factors which modify classicalposition effect variegation, also modified the variegation associated with this instanceof imprinting. Tables 29, 30, and 31 show the effect of temperature, sodiumbutyrate, and the presence of an extra Y chromosome on variegation, respectively.The effect of developmental temperature and butyrate was the reverse of the canonicalresponse. While a Hreverseu response to temperature is hardly unprecedented(Spofford 1976) the biological significance is unclear. Growth on butyratesupplemented media also enhanced the variegation of the garnet gene on the mini-chromosome, again, an unconventional response. The effect of these geneticmodifiers of position effect variegation was, with the exception of the Y chromosome,not particularly dramatic. However, in general, the modifiers of position-effectvariegation did modify the variegation of the garnet gene on the mini-chromosome.I next tested these same modifiers to determine if they could affect the imprint, that isthe meiotic transmission of the variegation. Table 32, 33, and 34 show the results ofthese tests. In no case did any of these conditions, which alter the variegation, alter theparental imprint. This conclusion is naturally subject to the criticism thatheterochromatin is a complex, and largely uncharacterized structure. Those factorsinvolved in the establishment of the imprint may in fact be components ofheterochromatin but simply different from those tested here.The effect of the Y chromosome is particularly striking. Although the strongestmodifiers of variegation, capable of completely obliterating variegation of a paternallyderived mini-chromosome, it had no effect on the transmission of the variegation - theimprint. The effect of the Y chromosome on imprinting was examined in four ways: A260Table 32. Imprinting of the Dp(1;f)LJ9 mini-chromosome: The effect of developmentaltemperature.Cultures generating g*/Dp(1;f)LJ9 males, where the mini-chromosome was transmittedboth maternally and paternally, were raised at 18, 22, 25 and 29°. In addition theparents were raised throughout their lives (from egg onwards) at 18, 22, 25 and 29°The amount of pigment in the eyes of these progeny was determined bymicroflourimeter assay, all values are expressed as percent wild type pteridine levels.Non specific effects of temperature on pigment (such as effects on fly size or number)were controlled by assaying siblings. No such effects were noted (data not shown).Data from experiments with three different garnet allele in the non-Dp bearing parent,either g53d,g5Oe or g2 is shown separately in each table. The temperature at whichthe parents were raised is shown on the left. The culture temperature of the crosswhich generated the diagnostic g/Dp(1;f)LJ9 progeny is shown at the top. Valuesderived from the paternal cross are shown in italics for contrast. ND = not done. Thereis limited information for the 290 series of paternal crosses as the XY/Dp males weregenerally sterile when raised at this temperature.Crosses:Maternal cross: X’X/Dp(1;f)LJ9 ® y zag53d/yf ory za g53d/y01or y e(g) cvg2/YQ’Paternal cross: X’Y/Dp(1;f)LJ9o’ ® y zag53d/yza g53d ory zag5Oe y zag5Oe or y e(g) cvg2/y e(g) cvg2261Imprinting of the Dp(1;f)LJ9 mini-chromosome: The effect of developmentaltemperature.g53dculture temperature18 22° 25° 29°180 MAT 94±3 89±3 63±5 33± 10parental PAT 42±4 37±2 50±3 39±222° MAT 85±4 87±3 61±2 49±2temperature PAT 27±3 36±3 54±3 30±325° MAT 80±4 96±4 58±3 54±1PAT 30±2 34±4 27±2 41±229° MAT 91±5 92±30 85±4 59±4PAT ND 16±4 ND NDg5Oeculture temperature18 22 25° 29°18° MAT 96±1 87±2 69±4 61±2parental PAT 69±4 37±2 58±2 41±322 MAT 89±4 89±3 75±4 60±2temperature PAT 81±5 56±3 53±3 33±425° MAT 95±3 88±4 60±3 58±2PAT 73±3 58±7 52±2 52±129° MAT 108±8 89±3 83±3 59±5PAT ND ND ND 45±6culture temperature18° 22° 25° 29°18° MAT 98±3 112±6 99±9 96±4parental PAT 96±3 67± 10 ND 99±522 MAT 103±6 102±3 108±3 99±3temperature PAT 80±5 79±4 87±4 92±325 MAT 112± 10 107±5 100±3 104±2PAT ND ND 114± 30 100± 329° MAT ND 105±5 113±6 133±30PAT ND ND ND ND262Table 33. Imprinting of the Dp(1,’f)LJ9 mini-chromosome: The effect of sodiumbutryrate.Cultures which generated g/Dp(1;f)LJ9 males, where the mini-chromosome wastransmitted either maternally or paternally, were raised on media supplemented with 0or 200 mM sodium butyrate. In addition the parents were also raised on either 0 or 200mM sodium butyrate. The amount of pigment in the eyes of these progeny wasdetermined by microflourimeter assay. All values are expressed as percent wild typepteridine levels. The first set of numbers reflects the amount of pigment in yzag53d/Dp(1;f)LJ9 progeny with a maternally inherited mini-chromosome. The secondset shows pigment levels of this genotype when the mini-chromosome is paternallyinherited. ND = not done.Crosses:Maternal cross: X’X/Dp(1;f)LJ9 ® y zag53d/y1f—* y zag53d/Dp(1;f)LJ9cJPaternal cross: X’Y/Dp(1;f)LJ9o’ ® y zag53d/yza g53d— yzag53d/Dp(1;f)LJ9cJ263Imprinting of the Dp(1;f)LJ9 mini-chromosome: The effect of sodium butyrate.parent concentration 0 mM 0 mM 200 mM 200 mMprogeny concentration 0 mM 200 mM 200 mM 0 mMMaternal mini- 87±6 64±6 67±3 77±3chromosomePaternal mini- 40±2 31 ±2 ND 71 ±5chromosome264Table 34. Imprinting of the Dp(1;f)LJ9 mini-chromosome: The effect of Y chromosomedosage.Four tests of the effect of the Y chromosome on imprinting of Dp(1;f) LJ9 wereperformed.1. The direct paternal test was intended to determine if an additional Y chromosome inthe Dp bearing father affected imprinting. Genotypically identical y zag53d/Dp(1,’f)LJ9progeny, both with a paternally transmitted mini-chromosome, where generated froman XY + Dp and an XYY + Dp fathers (first and second data columns). The results wereassessed by visual inspection and by microflourimeter pigment assay (first and seconddata row, values are expressed as percent wild type pteridine levels). y zag53d/Dp(1,.f)LJ9 progeny were distinguished from their y zag53d/Dp(7;f)LJ9/y siblingsby progeny testing.Cross: XX/Dp ® )Q’Y (any male)X’)QY/Dp 0 X’Y/Dpo’‘I,XY/Dp or X”Y/Y/Dpo’ ®y zag53d/y za g53d yzag53d/Dp(1;f)LJ9f2. A direct maternal effect was tested by examining progeny derived from XX vs. XXYduplication bearing mothers. Genotypically identical y za53d/Dp(1;f)LJ9 progeny,both with a maternally transmitted mini-chromosome, where generated from an XX +Dp and an XXY + Dp mothers (first and second data columns). The results wereassessed by visual inspection and by microflourimeter pigment assay, the latter valuesare expressed as percent wild type pteridine levels. y za53d/Dp(1;f)LJ9 progenywere distinguished from their yzag53d/Dp(7,.f)LJ9/y siblings by progeny testing.265Cross: XX/Dp ® any maleX”X/Dp or X”XJY/Dp ®y zag53d/yçj& y zag53d/Dp(1;f)LJ9cf3. The third test was of a paternal effect on the mini-chromosome variegationunrelated to the mini-chromosome. Genotypically identical y zag53d/Dp(1;f)LJ9progeny, both with a maternally transmitted mini-chromosome, where generated froman XY and an XYY (no mini-chromosome) fathers (first and second data columns). Theresults were assessed by visual inspection and by microflourimeter pigment assay, thelatter values are expressed as percent wild type pteridine levels. y za53d/Dp(1;f)LJ9progeny were distinguished from their yzag53d/Dp(1,.f)LJ9/y siblings by progenytesting.Cross: X’Y/O XY/Y‘IX’Y/O vs. X’Y/Yd ® y zag53d/yzag53d/Dp _> y zag53d/Dp(1;f)LJ9c4. The final test was of a maternal effect on the mini-chromosome variegationunrelated to the mini-chromosome. Genotypically identical y zag53d/Dp(7;f)LJ9progeny, both with a paternally transmitted mini-chromosome, where generated froman XX and an XXY (no mini-chromosome) mothers (first and second data columns).The results were assessed by visual inspection and by microflourimeter pigmentassay, the latter values are expressed as percent wild type pteridine levels. y zag53d/Dp(1;f)LJg progeny were distinguished from their y zag53d/Dp(7;f)LJ9/y siblingsby progeny testing.cross: X’Y/Dp ® y zag53d/y za g53d1y za g53d/y za g53d or y za g53d/xy ® X’Y/Dpo’— y zag53d/Dp(7;f)LJ9266Imprinting of the Dp(1;f)LJ9 mini-chromosome: The effect of Y chromosome dosage.1. Direct paternal effect 2. Direct maternal effectXJY/Dp vs. X/Y/Y/Dp X’X’/Dp vs. X’X//X’X/XY+Dp XY+Y+Dp XX+Dp XX+Y+Dpfather father mother mothervisual 0.10±0.02 0.12±0.02 visual 0.92±0.01 1.0±0estimate estimatepigment 41 ±2 48±3 pigment 100 ±3 94±4assay assay3. Paternal Y effect 4. Maternal V effectXYvs. XYYx Dp females XXvs. XXY x Dp malesXY ® Dp XYY 0 Dp XX 0 Dp XXY 0 Dpmother mother father fathervisual 0.73 ± 0.03 1.0±0 visual 0.05 ± 0.02 0.03 ± 0.03estimate estimatepigment 75 ± 10 81 ± 10 pigment 25 ±5 34 ± 11assay assay267direct effect in the male, and female (both the Y chromosome and mini-chromosomepresent in the same individual); and as an indirect maternal, and paternal effect (Ychromosome and mini-chromosome not present in the same parent). The presence ofan extra Y chromosome did not alter the imprint regardless of whether it was presentwith or without the mini-chromosome, or in males or females. There was no evidenceof a maternal or paternal Y chromosome effect on variegation as reported by Khesinand Bashirov (1978). If we accept for the moment, that these modifiers of position effectvariegation act via heterochromatin, then it seems that heterochromatin is involved inthe somatic propagation of the imprint. Clearly an additional Y chromosome caneliminate the garnet variegation, which is the signature of the imprint of the Dp(1;f)LJ9mini-chromosome.268DISCUSSION-imprintingIn this chapter I have described a mini-chromosome in Drosophila melanogasterwhich exhibits genomic, parent-specific or gamete-specific imprinting. The imprint isindependent of the sex of the progeny and is completely reversible. This imprint ismanifest as parent-dependent expression of the garnet eye colour gene, and of atleast two other genes on the mini-chromosome, tiny and narrow abdomen. Imprintingof the mini-chromosome was first noted as the parent-specific expression of the garnetgene. While other genes on the mini-chromosome also appear to be imprinted, theimprint at the garnet gene is the most dramatic expression of the imprint because of itsstriking mosaic phenotype in the eye. This mosaic phenotype arises because thegarnet gene is cell autonomous and the inactivation of the paternally derived garnetallele is not complete. Imprinting of a non-cell autonomous gene would be manifest asa quantitative difference in the level of gene expression which would be less readilyapparent.The Dp(1;f)LJ9 mini-chromosome shows extensive variegation (inactivation) for thegarnet+ gene when it is transmitted by a male. When this mini-chromosome istransmitted by a female the garnel gene is fully expressed. Thus according to theterminology of Reik (1992) the mini-chromosome is paternally imprinted. It should benoted that this terminology is arbitrary. There is no data to indicate whether imprintingacts to inactivate otherwise active genes or visa versa or whether the ground state ofthe garnet+ gene on the mini-chromosome is variegated or fully expressed.The immediate cause of disruption of garnet expression is position effect variegation.That the mosaic imprinted phenotype is due to conventional position effect variegation,is shown by the response of this variegator to the standard factors which modify269position effect variegation. The variegation is unusual in that is dependent of the sex ofthe transmitting parent thus the variegation itself is imprinted. As this example ofgenomic imprinting involves position effect variegation, the mechanism wherebyposition effect variegation causes gene inactivation is of considerable importance.While number of models of the mechanism of position effect variegation have beenproposed (Frankham 1988, Karpen 1994), there is experimental evidence to supportthree general classes of models; the somatic elimination model, the nuclearcompartmentalization model and chromatin or heterochromatin formation models. Ihave shown that neither the expression or the imprint of the garnet gene correlateswith under-representation of this gene (Figure 38), arguing against a role for somaticelimination. Of the two other models, experimental evidence slightly favours thechromatin formation model although a role for functional nuclear localization isintriguing and can not be ruled out. These two models are discussed in more detailbelow, in context of the mechanism of imprinting.If we accept, for the moment, that position effect variegation is due to the illicit spreadof heterochromatin than this striking phenotype could arise from either, a parentdependent ability to induce variegation, or from a parent-dependent difference in thedistance of spread or aggressiveness of the invading heterochromatin. The latter ismore likely for the following reasons. There is unlikely to be a parent specificdifference in the ability of the mini-chromosome to form centric heterochromatin. If thiswere the case, increased non-disjunction of paternally derived mini-chromosomewould be expected. Frequent non-disjunction has been observed with another minichromosome which was broken near the centromere and may have variegated forcentromere function (Wines and Henikoff 1992). However, such mitotic non-disjunctionwas not observed with the Dp(1;f)LJ9 mini-chromosome used in this study. Nor is theimprint likely to result from parent specific recognition of heterochromatic boundary270sequences (if in fact such sequences exist) as the same sequence would be presentor absent on the mini-chromosomes whether it was derived from males or females.Finally, the effect of temperature on the imprint suggests that the maternally derivedmini-chromosome can variegate under some circumstances, If individuals with amaternally derived mini-chromosome are raised at high temperature a few individualsshow limited inactivation of the garnet gene (Table 29). Thus I propose thatvariegation occurs on the mini-chromosome regardless of its parental origin, but thedistance over which heterochromatin spreads, and thus the likelihood of inactivatingthe garnet reporter gene is dependent on the parental origin. This situation isdiagrammed in Figure 39. Implicit in this model is that the imprint is evident onlywithin a limited range along the mini-chromosome. Also implicit in this model is theproposal that the imprint is nucleated with in the centric heterochromatin and spreadsdistally. The resulting parent-dependent differences in heterochromatin formationbecome less pronounced with distance from the centromere. This is in accordancewith the extent of the imprinting of the three neighboring genes, garnet, narrowabdomen and tiny (Figure 35).I tested a number of factors to determine the cause of the imprinting. The imprinting isnot allele specific. Nor is it a trivial artifact of maternal action of the marker gene,garnet. More interestingly, the imprint does not seem to be due to maternally-contributed compensatory substances, induced by physiological stress, as has beenproposed for the imprinting effects associated with Huntington’s Chorea and fragile Xmental retardation (Bird, Caro and Pilling 1974, VanDyke and Weiss 1986,respectively). Investigation of the role of heterochromatin in imprinting initially seemedpromising given the widely observed involvement of heterochromatin in imprinting andposition effect variegation. Tests of the stability and generation of the imprint in the271Figure 39. Model of the parent-dependent spread of heterochromatin responsible forimprinting of the Dp(1;f)LJ9 mini-chromosome.This figure diagrams the percent of cells showing phenotypic inactivation due to thespread of heterochromatin as a function of distance from the heterochromaticboundary. The effect of culture temperature is shown as altering the distance of thisspread rather than the occurrence of variegation.272Model of the mechanism of parent-dependent variegation ofDp(1;f)LJ9100%percentinactivecells0%variegation ofmaternally derivedmini-chromosomevariegation ofpaternally derivedmini-chromosomeDistance from heterchromatic boundarygarnet tiny narrowabdomen273presence of environmental, chemical and genetic factors which modify position effectvariegation, and are proposed to modify heterochromatin formation and integrity,suggest that at least in this instance of imprinting, the role of heterochromatin isrestricted to the somatic expression of the imprint but is not involved in theestablishment of the imprint.While these data might challenge the role of chromatin structure in initiation theimprint, heterochromatin formation is clearly necessary for the somatic manifestation ormaintenance of the imprint. This was shown by the response of the variegation of themini-chromosome to conventional modifiers of position effect variegation. Threefactors which generally modify position effect variegation (temperature, butyrate, and Ychromosome dosage) altered the variegation of this mini-chromosome. The effects ofthese modifiers deserve additional comment. The effect of temperature is the reverseof the conventional effect. Usually high temperature suppresses position effectvariegation. A “reverse” response to high temperature is not unprecedented by anymeans (Spofford 1967), however, the other examples of reverse temperaturesensitivity have been attributed to temperature sensitive hypomorphic alleles on thenon-variegating homolog. This cannot be the case for the garnet variegation on themini-chromosome as the temperature effect is seen with several alleles, none of whichare temperature sensitive. While the mechanism of position effect variegation isunknown, the canonical response to temperature suggests that the process ofheterochromatin formation is limiting. The simplest interpretation of the reverseresponse of the Dp(1;f)LJ9 variegation to temperature might suggest that it iseuchromatin formation which is limited in this chromosome. Why or how this mightoccur is not clear. The unconventional response to butyrate supplemented media isalso interesting. Sodium butyrate has been proposed to suppress position effectvariegation by altering the de-acetylation of histones and so disrupting chromatin274condensation. If this is in fact occurring, it is not clear why this would enhanceheterochromatin formation on the mini-chromosome. Finally, the moderation ofvariegation by temperature and sodium butyrate was quite modest. This difference ineffectiveness between modifiers of position effect variegation is in contrast to theireffect on conventional (non-parental) variegators (Clegg et al. 1992).In contrast, the effect of an extra Y chromosome (X/Y/Dp versus X/Dp) was much moredramatic, essentially eliminating variegation even for paternally derived mini-chromosomes. It may be that the Y chromosome effects variegation by mechanismsdistinct from that of the other modifiers. In this regard, Talbert and Henikoff (1994) haveargued that the suppressing effect of the Y chromosome on position-effect variegationis due to occlusion of the heterochromatin-forming compartment of the nucleus ratherthan directly on heterochromatin formation. The Y chromosome might also alter thesomatic pairing of the mini-chromosome. In contrast to these models, Zuckerkandel(1974) proposed that the Y chromosome suppresses position effect variegation byacting as a binding site for heterochromatic-specific proteins. It is possible that the Ychromosome is a particularly potent suppresser of Dp(1;f)LJ9 variegation because it isa powerful competitor for the specific heterochromatic proteins which bind to the mini-chromosome. One possible candidate might be the proteins which bind to the Stellategene cluster or proteins which bind to repetitive sequences near to the garnet gene(see below). In the absence of data on the mechanism of position effect variegation,any number of possibilities can be envisioned.Why do only some variegators show parental effects?All published parental effects in Drosophila involve position effect variegation (Table26). This may indicate an obligatory requirement for the large scale chromosome275rearrangements associated with position effect variegation. This scenario is hard tojustify conceptually. Absence of published reports of parental imprinting of transgenesin Drosophila, as is seen in mice, may reflect nothing more than investigators’ aversionto non-Mendelian expression which defeats the purpose for which transgenes areusually generated. There have been sporadic unpublished reports of parentdependent transgene expression as seen in mice (C. Bazinet, personalcommunication, C. Berg, personal communication).Although all published parental effects in Drosophila (excluding conventionalmaternal effects) involve variegating rearrangements, by no means is every exampleof position effect variegation associated with parental effects (Spofford 1976). Anotheroddity evident from Table 26 is that all the examples of parental effects in Drosophilainvolve the X chromosome. They do not however share any other commonchromosomal element nor is the heterochromatin of the X chromosome alwaysinvolved. Thus this may be simply a spurious coincidence.As imprinting amongst variegating rearrangements is decidedly a rarity, this raises thequestion of what features distinguish those variegating rearrangements which doshow parent-specific variegation. The existence of cloned and characterized DNAsequences from the vicinity of the garnet gene made this an obvious candidate toinvestigate any sequence or larger scale peculiarities which might explain theimprinting.Specific sequence motifs within the garnet gene are unlikely to be the cause of theparental imprinting. Although the 3’ region does contain a polyglutamine repeatsequence which has been associated with some imprinted genes in humans (Green1993) the repeat seems too short and inconsequential to explain the imprinted spread276of heterochromatin through at least three genes. The trinucleotide repeats found in thefragile X syndrome, spinal and bulbomuscular atrophy, Huntington’s disease,spinocereballar ataxia type I and myotonic dystrophy are present as between 5-50copies in the normal alleles and 40-4000 copies in the imprinted disease causingalleles. Large scale structural features of the mini-chromosome might also beassociated with the imprint. The Steilate genes are highly repetitive gene clusterspresent on the X and Y chromosomes and degenerate members are present in theproximal heterochromatin of the X chromosome (Shevelyov 1992, E. V.Benevolenskaya, personal communication). The Stellate genes are highly transcribedin spermatogenesis but are of unknown function (Livak 1990, Palumbo et al. 1994).The proximity of Stellate sequences to the garnet gene, as well as possibledegenerate Steilate sequences in the X heterochromatin, may meant that the garnetgene is flanked by multiple Stellate repeats. Given the propensity of heterochromatinto engage in non-homologous pairing (Eberl, Duyf and Hilliker 1993, Talbert, LeCieland Henikoff 1994), the ability of repeated sequence to nucleate heterochromatinformation (Dorer and Henikoff, 1994), the fact that many imprinted transgenes inmammals are present as multiple repeats (Surani, Reik and Allen 1988) and thepresence of repeat sequence within a region of an imprinted transgene defined as thecis-acting imprinting signal (Chaillet at al. 1995), the Stellate sequences might induceaberrant heterochromatin structure or nuclear localization of the mini-chromosomewhich could lead, somehow, to imprinting. Intriguingly, the euchromatic andheterochromatic Stellate sequences can be seen to form ectopic fibres, which indicateillicit pairing, in polyene chromosome preparations (Palumbo et al. 1994). The manyStellate repeats on the Y chromosome might explain the remarkable effectiveness ofthe Y chromosome in suppressing variegation of the mini-chromosome it thesesequences compete for binding proteins, Of course the presence, number and functionof the degenerate heterochromatic Stellate sequences in the mini-chromosome277remains to be confirmed. A similar argument could be made for any other middle orhighly repetitive sequences present near garnet and on the Y chromosome. In thisregard it is interesting that the 12DE region adjacent to garnet is peppered withrepetitive sequences (Leung et al. 1987).Although the consequences of imprinting are dramatically manifest in development,neither the evolutionary forces which lead to imprinting or the mechanisms are known.The former issue has occupied the attention of many biologists, but in the absence ofdata on the mechanism, all evolutionary scenarios remain completely speculative. Themechanism of imprinting is of considerable interest in itself, and is also of medicalimportance given the number of human diseases in which imprinting has beenimplicated.Mechanism of imprinting:While it is evident that differences in packaging genetic material in eggs and spermmay facilitate differential gene expression in early embryogenesis, a simple responseto the different physiology and morphology of the germ cell formation and structuremay not suffice as a mechanism for imprinting. In any case, this explanation fails toaddress the question of why and how some genes show parent specific expressionwhereas others do not. Based on observations that pairing, recombination and geneconversion events coincide with condensation at pachytene stage of meiosis, Monkand Grant (1990) and Hall (1990) have proposed that imprinting is a side effect ofthese events. Details of how or why this might occur are unclear and as such thismodel is difficult to test empirically. Functional differences in nuclear localization, invarious guises, has been proposed as a potential basis for imprinting.278Modulation of gene expression by nuclear compartmentalization has been inferredfrom both cytological and genetic evidence. The existence of nuclear compartments isindisputable but ascribing specific functions to them remains problematic. Forexample, the nucleolus is an obvious example of a specialized nuclear compartment.It has been suggested that other structures found in the nucleus, such as the coiledbodies, perichromatin fibrils and interchromatin granules, correspond to otherfunctional compartments associated with splicing. Recent evidence, however,suggests that this is not the case (Mattaj, 1994). The first intimation that nuclearcompartments might have specific functions in gene regulation stemmed fromobservation of the specific orientation of chromosomes in the nucleus, with telomeresand centromeres apposed to the nuclear membrane (RabI 1885). A similarchromosome organization in yeast is correlated with variable gene inactivationresembling position effect variegation (Laurenson and Rhine 1992), however, a causalrelationship between either telomere-nuclear membrane apposition in S. cerevisiae orcentromere-nuclear membrane apposition in S. pombe remains to be shown. HeslopHarrison (1990) has proposed that imprinting results from differential location of thetwo parental or species haploid sets of chromosomes in the zygotic nuclei. There aremany well documented and striking examples of differential location of the twochromosome sets in plant inter-specific hybrids (summarized by Heslop-Harrison1990) however, these studies did not correlate the location of the two genome setswith gene expression. Several investigators have argued for a similar effect of nuclearlocalization on gene expression in position effect variegation (Hessler 1957, Wakimotoand Hearn 1990, Talbert, LeCiel and Henikoff 1994) however cytological correlationsare generally lacking. If a causal relationship between nuclear localization and geneexpression could be shown in position effect variegation, it might serve to supporthypotheses suggesting a connection between nuclear location and gene expressionin imprinting.279In mammals and plants, DNA methylation has long been known to be linked to geneinactivation. There are correlations between methylation and the gene inactivationresulting from imprinting in mammals. Unfortunately the convenience of determiningthe methylation status of a gene has led to methylation being treated synonomouslywith imprinting without other evidence of genetic activity. But, the correlation betweenmethylation and gene activity is by no means absolute. Such a correlation is seen inonly one of six examples of imprinted transgenes (Surani, Reik and Allen 1988) and ithas been shown in a number of cases that the decision to inactivate a gene occursbefore methylation is evident. The Hprt gene on the imprinted X chromosome isinactivated several days before differential methylation is established (Lock, Takagiand Martin 1987). Likewise, differential methylation of the imprinted genes H19, Igf2and Igf2r does not generally persist through meiosis, thus can not constitute the initialimprinting signal. For example, the parent-specific methylation associated withimprinted expression of H19 occurs late in embryogenesis and is not propagatedthrough the male meiosis (Ferguson-Smith et al 1993). Most of the parent-specificmethylation pattern in the promoter and coding region of Igf2r/Mpr is established latein embryogeneis and is erased in meiosis. There is, however, in this case, one site inan intron which maintains its methylation status though the female meiosis (Stoger etal 1993). In contrast, the imprinted gene Igf2 seems devoid of parent specificmethylation in the immediate vicinity of the gene, and the gene is identicallymethylated in eggs and sperm (Sasaki et al 1992, Brandeis et al 1993). Thusmethylation seems not to constitute the initial decision or even an early event inimprinting. The role of methylation in imprinting may be relegated to maintenance ofthe decision. Finally, methylation is not causally involved paramutation in maize(Patterson, Thorpe and Chandler 1993), a phenomenon that resembles imprinting, nordoes it correlate with either imprinting or heterochromatin formation in coccids280(Scarbrough, Hattman and Nur 1984) nor is it found in Dipteran insects where thereare striking examples of imprinting. Thus methylation is unlikely to play a role in thiscase, or others, of imprinting in Drosophila.The role of chromatin structure in imprinting is less experimentally tractable than that ofmethylation, but, its near universal involvement in imprinting and the involvement ofposition effect variegation in imprinting in Drosophila, makes it a logical candidate forinvestigation. The role of heterochromatin formation in position effect variegation hasbeen extensively investigated. Evidence for a chromatin-based model of position effectvariegation stems from cytological, genetic and molecular work. A correlation betweengene inactivation and the cytological appearance of heterochromatin in thecorresponding region of the chromosome (Hartmann-Goldstein 1967, Spofford 1976)has been observed for many genes. A role for chromatin formation in position effectvariegation is further suggested by the sensitively of position effect variegation inDrosophila (and the seemingly related telomere mediated position effect in S.cerevisiae) to alterations in histone levels. A number of other genetic modifiers ofposition effect variegation have also proven to encode chromatin associated proteins(reviewed by Orlando and Paro, 1995). Finally, alterations in nucleosome positioning,phasing and chromatin accessibility have been found associated with position effectvariegation (Wallrath and Elgin, 1995). Thus there is reasonable evidence to supportthe assertion that gene inactivation associated with position effect variegation isrelated to changes in chromatin structure. Tartof and Bremer (1990) have extendedthese observations to imprinting and have made the rather stringent prediction thatimprinted regions will correspond to regions of intercalary heterochromatin. No ectopicpairing sites, a sign of intercalary heterochromatin have been reported around thegarnet gene. Nevertheless, the role of chromatin structure in imprinting can not bedisregarded. The results presented in chapter 3 of this thesis imply that281heterochromatin in Drosophila may perform a maintenance function, similar to thatplayed by methylation in mammals. There is evidence that the maintenance stage inmammals can be resolved into early and late stages. Intriguingly, preliminary evidencesuggests that the early maintenance stage may be mediated by chromatin bindingproteins (Monk 1988, Barlow 1994, Ohlsson, Barlow and Surani, 1994). This maysuggest a parallel between the maintenance mechanisms of the imprint betweenmammals and Drosophila. Different mechanisms of late somatic memory might simplyreflect differences in the life cycles of mammals versus Drosophila. Most cells of amammal continue to undergo cell division throughout the adult portion of life. Thecontinuation of cell division requires a high fidelity memory mechanism. Covalentmodification of DNA based by methylation, clearly, would fulfill this function. Incontrast, adult flies are mitotically quiescent, thus there might be less rigorousrequirement for the memory mechanism and a “chaotic” system such a chromatinpartitioning might suffice.A major impediment to resolution of the precise role of heterochromatin in imprintingis that the structure of chromatin and molecular architecture of heterochromatinremains undefined, leaving any models about the role of heterochromatin necessarilyvague. One approach to resolving the mechanism of imprinting is to isolate geneticmodifiers of the imprinting process by screening for mutations which enhance,eliminate or switch the parental specificity of imprinted genes. Genes which modifyimprinting would be expected to encode products which are components of themachinery which recognizes the sexual context of the imprinted region and thus mightact in one sex only (which depends on whether gene activity or inactivity is the defaultstate). Some of these gene products would also have to recognize and isolate theregion to be imprinted, make the initial decision to imprint and then, at least initially,propagate and enforce the imprinted state. Some progress has been made in282Table 35. Effect of attached versus free sex chromosomes on imprinting ofDp(1;f)LJ9.Genotypically identical y zag53d/Dp(1;f)LJ9 progeny, with either a paternallytransmitted mini-chromosome (columns 2 and 3) or a maternally transmitted mini-chromosome (columns 4 and 5), were generated from stocks with either attached sexchromosomes (column 2 and 4) or freely segregating sex chromosomes (column 3and 5). The results were assessed by visual inspection (first data row) and bymicroflourimeter pigment assay (second data row, values are expressed as percentwild type pteridine levels). yzag53d/Dp(1;f)LJ9 progeny were distinguished from theiry zag53d/Dp(1,.f)LJ9/y siblings by progeny testing.Crosses: paternal: X’Y/Dp or y zag53d/y/Dpo’ ®y zag53d/yza g53dmaternal: X’X/Dp or yzag53d/yzag53d/Dpyzag53d/yor X’Y/OO’283Effect of attached versus free sex chromosomes on imprinting of Dp(1;f)LJ9.Paternal cross Maternal crossattached free attached freechromosomes chromosomes chromosomes chromosomesvisual 0.10 ± 0.02 0.071 ± 0.03 0.92 ± 0.01 0.73 ± 0.03estimatepigment 41±2 37±9 100±3 85±3assay284identifying the location of imprinting genes (imprinting genes are those genes whichcontrol the imprinting process as opposed to imprinted genes) in mice, (Allen, Norrisand Surani 1990, Babinet et al 1990, Cattanach and Beechey 1990, DeLoia andSofter 1990, Reik, Howlett and Surani 1990, Surani et al 1990, Engler et al 1991,Foreijt and Gregorova 1992 and Sapeinza et al 1992, Chaillet et al. 1995) andhumans (Sapienza -personal communication). However these genes have not beenotherwise characterized. As imprinting in mammals is early acting and an importantdevelopmental process, the genes which effect it are likely to be both pleiotropic andlethal when mutant. Such genes cannot easily be detected by crossing different micestrains each homozygous for potential modifier genes. Human genetics poses similarbut more extreme problems. For the purpose of identifying, cloning and characterizingsuch genes, the sophisticated genetic and molecular tools available in Drosophilawould make this organism an excellent system for examining and dissecting imprintingphenomenon. The existence of genetic modifiers of imprinting in Drosophila issuggested by the data shown in Table 35. This table shows that the extent of theimprint differs slightly, but reproducibly, when the mini-chromosome is transmitted byfemales with free versus attached X chromosomes. This difference in the extent ofvariegation may represent segregation of sex linked imprinting genes with minoreffects in the genetic background of these two strains.In summary, the garnet gene has been used as a tool to examine various biologicalprocesses. This chapter describes a mini-chromosome which is subject to parentspecific imprinting. This imprint is manifest as a parent-dependent variegation of thegarnet gene. The result is imprinted expression of the garnet gene. The genetic andmolecular information on the garnet gene allowed me to test several factors whichhave been postulated to cause imprinting. The primary finding is that heterochromatin,which is implicitly involved in imprinting in a wide variety of organisms, may act as a285memory mechanism, not a primary determinant of the imprint. If heterochromatin isrelegated to a “memory” mechanism, as is methylation in mammals, then there mustbe an independent system which is involved in establishing the imprint. This processis undoubtedly complex but isolation of modifiers of imprinting should help toilluminate the mechanics and possibly the evolutionary and developmental rational ofthe phenomenon of genomic imprinting.286BibliographyAllen, N. D., D. G. Cran, S. C. Barton, S Hettle, W. Reik and M. A. Surani. 1988.Transgenes as probes for active chromosomal domains in mouse development.Nature 333: 852-855.Allen, N. D., M. L. Norris and M. A. Surani. 1990. Epigenetic control of transgeneexpression and imprinting by genotype-specific modifiers. Cell 61: 853-861.Altenburg, L. S. and E. Alternberg. 1959. The mutagenicity of 2,5-bis-ethylenehydroquinone in Drosophila. Genetics 44: 498.Babinet, C., V. Richoux, J.-L. Guénet and J.-P. Renard. 1990. The DDk inbredstrain as a model for the study of interactions between parental genomes andegg cytoplasm in mouse pre implantation development. p. 81-87 In: Genomicimprinting (development 1990 supplement). Ed: M. Monk and A. Surani.Company of Biologists Ltd., 1990, Cambridge.Baker, W. K.. 1963. Genetic control of pigment differentiation in somatic cells.Am. Zool. 3: 57-69.Baker, W. K. and J. B. Spofford. 1959. Heterochromatic control of position-effectvariegation in Drosophila. Biological Contributions. Univ. Texas PuIb.5914:135-154.Barlow, D. P. 1994. Imprinting: a gamete’s point of view. Trends Genet. 10: 194-199.Bartolomei, M. S., A. C. Webber, M. E. Brunkow and S. M. Tilghman. 1993.Epigenetic mechanisms underlying the imprinting of the mouse H19 gene.Genes and Devel. 7: 1663-1673.Beadle, G. W. 1937a. Development of eye colors in Drosophila: Fat bodies andmalpighian tubes in relation to substances. Genetics 22:587-611.Beadle, G. W. 1 937b. The inheritance of the color of malpighian tubes inDrosophila melanogaster. Amer. Nat. 71: 277-279.Beadle, G. W. and B. Ephrussi. 1935a. Différenciation de Ia couleur de l’oeilcinnabar chez a Drosophile (Drosophila melanogaste,). Comptes RendusAcad. Sci. Paris 201: 642-646.Beadle, G. W. and B. Ephrussi. 1935b. Transplantation in Drosophila. Proc. Nat.Acad. Sci. (USA) 21:642-646.Beadle, G. W. and B. Ephrussi. 1936. The differentiation of eye pigments inDrosophila as studied by transplantation. Genetics 21: 225-247.287Beadle, G. W. and B. Ephrussi. 1937. Development of eye colors in Drosophila:Diffusible substances and their interrelations. Genetics 22: 76-86.Bel, Y. and J. Ferré. 1986. Biosynthesis of pteridines and metabolism ofaromatic amino acids in Drosophila melanogaster. p. 335-338 In: Chemistryand biology of pteridines. Walter de Greuyer and Co. Berlin, N.Y.Belote, J. M. and J. C. Lucchesi. 1980. Male-specific lethal mutations ofDrosophila melanogaster. Genetics 96: 165-186.Benzer, S. 1955. Fine structure of a genetic region in bacteriophage. Proc. NatI.Acad. Sci. USA. 41 :344-354.Bingham, P. M., and B. H. Judd. 1981. A copy of the copia transposable elementis very tightly linked to the wa allele of the white locus of D. melanogaster. Cell25: 705-711.Birchier, J. A., U. Bhadra, L. Rabinow, R. Linsk and A. T. Nguyen-Huynh. 1994.Weakener of white (Wow), a gene that modifies the expression of the white eyecolor locus and that suppresses position effect variegation in Drosophilamelanogaster. Genetics 137: 1057-1070.Bird, E. D., A. J. Caro and J. B. Pilling. 1974. A sex related factor in theinheritance of Huntington’s chorea. Ann. Human Genet. 37: 255-260.Blakeslee, R. W. 1938. Report of the Director. Carnagie Institute of WashingtonYearbook. 42: p156-i 61.Bonse, A.. 1967. Untersuchungen uber die chemische Natatur und die Bildungder Harnkong lomerate in die Malpighischen Gefassen der Mutante rosy vonDrosophila melanogaster. Z. Naturforsh. 22: 1027-1029.Brandeis, M., T. Kafri, M. Ariel, J. R. Chaillet, J. McCarrey, A. Razin and H.Cedar. 1993. The ontogeny of allele-specific methylation associated withimprinted genes in the mouse. EMBO J. 12: 3669-3677.Brehme, K. S. and M. Demerec. 1942. A survey of malpighian tube color in theeye color mutants of Drosophila melanogaster. Growth 6: 351-355.Bridges, C. B. 1916. Non-disjunction as proof of the chromosome theory ofheredity. Genetics 1:1-52 and 107-163.Bridges, C. B. and K. A. Breme. 1944. The mutants of Drosophila melanogaster.Carnagie Inst. Wash. PubI. 552.Brown, G. M. 1989. Biosynthesis of H4Biopterin and related compounds. InBiology and Chemistry of Pteridines. Ed: H.-Ch. Curtius, S. Ghisla, N. Blau.Walter de Gruyter and Co. Berlin, N. York. p. 199-21 2.288Brown, G. M. and C. L. Fan. 1975. The synthesis of pterins catalyzed byenzymes from Drosophila melanogaster. In: Chemistry and Biology ofPteridines. Ed: W. Pfleiderer. Walter de Gruyter, Berlin, N. York.Brown, G. M., G. G. Krivi, C. L. Fan, and T. R. Unnasch. 1978. The biosynthesisof pteridines in Drosophila melanogaster. In Biology and Chemistry ofPteridines. Ed: Kisliuk and Brown. Elsevier North Holland, Inc. p. 81-86.Brun, L. 0., P. Borsa, V. Gaudichon, J. J. Stuart, K. Aronstein, C. Coustau and R.H. ffrench Constant. 1995. Functional haplodiploidy. Nature 374:506.Calatayud, M. T., D. A. Jacobson and J. Ferré. 1989. New eye colour mutantsaffecting the biosynthesis of pteridine and ommochromes in Drosophilamelanogaster. In Chemistry and Biology of Pteridines. Ed: H.-Ch. Curtius, S.Ghisla and N. Blau. Walter de Gruyter and Co. N. York, Berlin. p. 591 -594.Campos-Ortega, J. A. 1988. Cellular interactions during early neurogenesis ofDrosophila melanogaster. Trends Neurosci. 11: 400-405.Carlson, E. A. 1959. Comparative genetics of complex loci. Quart. Rev. Biol.34:33-67.Casteel, D. B. 1929. Histology of the eyes of X-rayed Drosophila. J. Exp. Zool.53: 373-385.Castro-Sierra, E. and S. Ohno. 1968. Allelic inhibition at the autosomallyinherited gene locus for liver alcohol dehydrogenase in chicken-quail hybrids.Biochem. Genet. 1: 323-335.Cattanach, B. M. and C. V. Beechey. 1990. Autosomal and X-chromosomeimprinting. p. 53-72. In: Genomic imprinting (development 1990 supplement).Ed: M. Monk and A. Surani. Company of Biologists Ltd., 1990, Cambridge.Cattanach, B. M. and J. N. Perez. 1970. Parental influence on X-autosometranslocation-induced variegation in the mouse. Genet. Res. Camb. 15:43-53.Chaillet, J. R., D. S. Badar and P. Leder. 1995. Regulation of genomic imprintingby gametic and embryonic processes. Genes Dev. 9: 1177-1187.Chandra, H. S. and S. W. Brown. 1975. Chromosome imprinting and themammalian X chromosome. Nature 253: 165-168.Chandra, H. S. and V. Nanjundiah. 1990. The evolution of genomic imprinting.p. 47-53. In: Genomic imprinting (development 1990 supplement). Ed: M. Monkand A. Surani. Company of Biologists Ltd., 1990, Cambridge.Chovnick, A. 1957. Pseudoallelism at the garnet locus in Drosophilamelanogaster. Genetics 42:365.289Chovnick, A. 1958. Aberrant segregation and pseudoalleleism at the garnetlocus in Drosophila melanogaster. Proc. Nati. Acad. Sd. USA. 44:333-337.Chovnick, A. 1958. Structural and functional aspects of pseudoallelism inDrosophila melanogaster. Proc. 10th mt. Cong. Genetics. 2:49-50.Chovnick, A. 1961. The garnet locus in Drosophila melanogaster. 1.Pseudoallelism. Genetics 46:493-507.Chovnick, A. 1989. Intragenic recombination in Drosophila: The rosy locus.Genetics 123:621-624.Chovnick, A., A. Schalet, R. P. Kemaghan and M. Krauss. 1964. The rosy cistronin Drosophila meIanogaster genetic fine structure analysis. Genetics 50:1245-1259.Chovnick, A., R. J. Lefkowitz and D. R. McQuinn. 1956. Complexity at the garnetlocus in Drosophila melanogaster. Genetics 41:637.Christensen, H. N. 1973. On the development of amino acid transport systems.Fed. Proc. 32: 19-37.Clarke, A. 1990. Genetic imprinting in clinical genetics. p. 131-139. In: Genomicimprinting (development 1990 supplement). Ed: M. Monk and A. Surani.Company of Biologists Ltd., 1990, Cambridge.Cohen, J.. 1962. Position-effect variegation at several closely linked loci inDrosophila melanogaster. Genetics 47:647-659.Cremer-Bartels, G. 1975. Pteridines in the mammalian retina and light effects. p.861-870. In: The chemistry and Biology of Pteridines. Ed: W. Pfleiderer. Walterde Gruyter, Berlin, N. York.Crouse, H. V.. 1960. The controlling element in sex chromosome behavior insciara. Genetics 45: 1429-1443.DeLoia, J. A. and D. Solter. 1990. A transgene insertional mutation at animprinted locus in the mouse genome. p. 73-79. In: Genomic imprinting(development 1990 supplement). Ed: M. Monk and A. Surani. Company ofBiologists Ltd., 1990, Cambridge.Demakova, 0. V. and E. S. Belyaeva. 1988. Effect of mating direction on theposition effect variegation of T(1;2)dorvar7in Drosophila melanogaster. Dros.Info. Service. 67:19-20.Dobzhansky, T. 1946. Genetics of natural populations. XIII. Recombination andvariability in populations of Drosophila pseudoobscura. Genetics 31: 269-290.290Dorer, D. R. and S. Henikoff. 1994. Expansions of transgene repeats causeheterochromatin formation and gene silencing in Drosophila. Cell 77:993-1002.Dorn, R., J. Szidonya, G. Korgo, M. Schnert, H. Taubert, E. Archoukieh, B.Tschiersch, H. Morwietz, G. Wustmann, G. Hoffman and G. Reuter. 1993. Ptransposon-induced dominant enhancer mutations of position-effect variegationin Drosophila melanogaster. Genetics 133: 279-290.Dorn, R., V. Krauss, G. Reuter and H. Saumweber. 1993. The enhancer ofposition-effect variegation of Drosophila, E(var)3-93D, codes for a chromatinprotein containing a conserved domain common to several transcriptionalregulators. Proc. Nat. Acad. Sci. USA 90:11376-11380.Dorsett, D. L., J. J. Yim and K. B. Jacobson. 1978. Biosynthesis of drosopterinsin the head of Drosophila melanogaster. In: The chemistry and Biology ofPteridines. Ed: R. L. Kisliuk and G. M. Grown. Elsevier, North Holland. p. 99-104.Dressen, T. D., D. H. Johnson and S. Henikoff. 1988. The brown protein ofDrosophila melanogaster is similar to the white protein and to components ofactive transport complexes. Mol. Cell Biol. 8:5206-5215.Dryja, T. P., S. Mukai, R. Petersen, J. M. Rapaport, D. Walton and D. W. Yandell.1989. Parental origin of mutations of the retinoblasoma gene. Nature 339: 556-558.Eberl, D. F., B. J. Duyf and A. J. Hilliker. 1993. The role of heterochromatin in theexpression of a heterochromatic gene, the rolled locus of Drosophilamelanogaster. Genetics 134: 277-292.Engler, P., D. Haasch, C. A. Pinkert, L. Doglio, M. Glymour, R. Brinster and U.Storb. 1991. A strain-specific modifier on mouse chromosome 4 controls themethylation of independent transgene loci. Cell 65: 939-947.Ephrussi, B. and G. W. Beadle. 1935a. La transplantation des disquesimaginaus chez Ia Drosophile. Comptes Rendus Acad. Sci. Paris. 201: 98-99.Ephrussi, B. and G. W. Beadle. 1935b. La transplantation des ovaires chez IaDrosophile. Bull. Biol. 69: 492-502.Ephrussi, B. and G. W. Beadle. 1935c. Sur les conditions del’autodifferenciation des carateres mendeliens. Comptes Rendus Acad. Sci.Paris. 201: 1148-1150.Ephrussi, B. and G. W. Beadle. 1937a. Development of eye colors inDrosophila: Production and release of cn+ substance by the eyes of differenteye color mutants. Genetics 22: 479-483.291Ephrussi, B. and G. W. Beadle. 1 937b. Development of eye colors inDrosophila: Transplantation experiments on the interaction of vermilion withother eye colors. Genetics 22: 65-75.Ephrussi, B. and J. L. Herold. 1944. Studies of eye pigments of Drosophila. I.Methods of extraction and quantitative estimation of the pigment components.Genetics 29: 148-175.Ferguson-Smith, A. C., H. Sasaki, B. M. Cattanach and M. A. Surani. 1993.Parental-origin-specific epigenetic modification of the mouse H19 gene. Nature362: 751-755.Ferré, J., F. Silva, M. D. Real and J. L. Mensua. 1983. Comparative study of theeye colour mutants of Drosophila meIanogaster Quantitation of the eye-pigments and related metabolites. In Chemistry and Biology of Pteridines. Ed: J.A. Blair. Walter de Gruyter and Co. N. York, Berlin. p. 669-673.Ferré, J., F. Silva, M. D. Real and J. L. Mensua. 1986. Pigment patterns inmutants affecting the biosynthesis of pteridines and xanthomatin in Drosophilamelanogaster. Biochem. Genet. 24: 545-569.Flavell, R. B. and M. O’Dell. 1990. Variation in inheritance cytosine methylationpatterns in wheat at the high molecular weight glutinin and ribosomal RNA geneloci. p. 15-20. In: Genomic imprinting (development 1990 supplement). Ed: M.Monk and A. Surani. Company of Biologists Ltd., 1990, Cambridge.Forejt, J. and S. Gregurova. 1992. Genetic analysis of genomic imprinting: Animprintor- 1 gene controls inactivation of the paternal copy of the mouse Tmelocus. Cell 70: 443-450.Frankham, R. 1988. Molecular hypotheses for position-effect variegation: anti-sense transcription and promotor occlusion. J. Theor. Biol. 135: 85-1 07.Fuyama, Y. 1984. Gynogenesis in Drosophila melanogaster. Jpn. J. Genetics59: 91-96.Gilchrist E. J. and D. G. Moerman. 1992. Mutations in the sup-38 gene ofCaenorhabditis elegans suppress muscle-attachment defects in unc-52mutants. Genetics 132: 431-442.Glaser, R. L, and A. C. Spradling. 1994. Unusual properties of genomic DNAmolecules spanning the euchromatic - heterochromatic junction of a Drosophilaminichromosome. Nuc. Acids. Res. 22: 5068-5075.Glass, H. B. 1934. A study of dominant mosaic eye-color mutants on Drosophila.I. Phenotypes and loci involved. Amer. Nat. 68: 107-114.Glassman, E. 1956. Kynurenine formamidase in mutants of Drosophila.Genetics 41: 5666-574.292Goldberg, A., A. Schalet and A. Chovnick. 1962. On the lethality of doublemutants of Hnr3and various ry mutant alleles. Dros. Info. Serv. 36: 67-68.Green, H. 1993. Human genetic diseases due to codon reiteration: Relationshipto an evolutionary mechanism. Cell 74: 955-956.Green, M. M. 1959. Spatial and functional properties of pseudoalleles at thewhite locus in Drosophila melanogaster. Heredity 13:303-315.Gumaratne, P. H., A. Mansukhani, S. E. Lipari, H. C. Loiu, P. W. Matindale andM. L. Goldberg. 1986. Molecular cloning, germ-line transformation andtranscriptional analysis of the zeste locus of Drosophila melanogaster. Proc.Nat. Acad. Sd. USA 83: 701-705.Gunaratne, P. H., M. Nakao, D. H. Ledbetter, J. S. Sutcliffe and A. C Chinault.1995. Tissue-specific and allele-specific replication timing controls in theimprinted human Prader-Willi syndrome region. Genes Dev. 9: 808-820.Haas, 0. A., A. Argyriou-Tirita and T. Lion. 1991. Parental origin of thechromosomes involved in the translocation t(9;22). Nature 359: 414-41 6.Hadchouel, M., H. Farza, D. Simon, P. Tiollais and C. Pourcel. 1987. Maternalinhibition of hepatitis B surface antigen gene expression in transgenic micecorrelates with de novo methylation. Nature 329: 454-456Hadorn E. 1962. Fractionating the fruitfly. Sci. Am (April): 101-110.Hadorn, E. and H. K. Mitchell. 1951. Properties of mutants of Drosophilamelanogaster and changes during development as revealed by paperchromatography. Proc. Nat. Acad. Sci. 37: 650-665.Hall, J. G. 1990. How imprinting is relevant to human disease. p. 141-148. In:Genomic imprinting (development 1990 supplement). Ed: M. Monk and A.Surani. Company of Biologists Ltd., 1990, Cambridge.Hardy, R. W., D. L. Lindsley, K. J. Livak, B. Lewis, A. L. Siversten, G. L. Joslyn, J.Edwards and S. Bonaccorsi. 1984. Cytogenetic analysis of a segment of the Ychromosome of Drosophila melanogaster. Genetics 107: 591-610.Harnly, M. H. and B. Ephrussi. 1937. Development of eye colors in Drosophila:Time of action of body fluid on cinnabar. Genetics 22:393-401.Hartmann-Goldstein, I. J.. 1967. On the relationship betweenheterochromatinization and variegation in Drosophila, with special reference totemperature sensitive periods. Genet. Res. 10: 143-1 59.Hayman, D. L. and R. H. Maddern. 1969. Report-D. melanogaster-new mutants.Dros. Info. Service 44:69.293Hearl, W. G., D. Dorsett and K. B. Jacobson. 1983. The common precursor ofsepiapterin and drosopterin in Drosophila: enzymatic and chemical synthesis.In Chemistry and Biology of Pteridines. Ed: J. A. Blair. Walter de Gruyter and Co.N. York, Berlin. p. 397-401.Hearl, W. G., K. B. Jacobson. 1984. Eye pigment granules of Drosophilamelanogaster. Isolation and characterization for synthesis of sepiapterin andprecursors of drosopterin. Insect Biochem. 14:329-335.Henikoff, S., D. Nash, R. Hards, J. Bleskan, J. F. Wodford, F. Naguib and D.Patterson. 1986. Two Drosophila melanogaster mutations block successivesteps of de novo purine synthesis. Proc. Nat. Acad. Sci. USA 83: 3919-3923.Heslop-Harrison. 1990. Gene Expression and parental dominance in hybridplants. p. 21-28. In: Genomic imprinting (development 1990 supplement). Ed: M.Monk and A. Surani. Company of Biologists Ltd., 1990, Cambridge.Hess 0. 1970. Independence between modification of genetic position effectsand formation of Iampbrush loops by the Y chromosome of Drosophila hydel.Mol. Gen. Genet. 107: 224-242.Hessler, A. 1958. V-type position effects at the light locus in Drosophilamelanogaster. Genetics 43: 395-403.Hessler, A. Y.. 1961. A study of parental modification of variegated positioneffects. Genetics 46: 463-484.Hexter, W. H. 1956. Pseudoallelism at the g locus. Dros. Info. Serv. 30:121.Hexter, W. M. 1958a Probable gene conversion in Drosophila. Proc. 10th Int.Cong. Genetics. 2:120.Hexter, W. M. 1 958b. On the nature of the garnet locus in Drosophilamelanogaster. Proc. NatI. Acad. Sci. USA. 44:768-771.Hilliker, A. J., B. Duyf, D. Evans and J. P. Phillips. 1992. Urate-null rosy mutantsof Drosophila melanogaster are hypersensitive to oxygen stress. Proc. Nat.Acad. Sci. 89: 4343-4347Holliday R. 1990. Genomic imprinting and allelic exclusion. p. 125-129. In:Genomic imprinting (development 1990 supplement). Ed: M. Monk and A.Surani. Company of Biologists Ltd., 1990, Cambridge.Holliday, R. 1987. The inheritance of epigenetic defects. Science 238: 163-170.Hopkins, F. G. 1889. Note on yellow pigments in butterflies. Nature 40: 335.294Howland, R., E. A. Glancy and B. P. Sonnenblick. 1937. Transplantation of wildtype and eye disks among four species of Drosophila. Genetics 22:434-442.Janning, W. 1970. Bestimmung des heterochromatisierungsstodiums beimwhite-positionseffekt mittles röntgen induzierter mitotischer rekombination in deraugenanlage von Drosophila melanogaster. Mol. Gen. Genet. 107: 128-149.Johannsen, A. 0. 1924. Eye structure in normal and eye-mutant Drosophila. J.Morph. 39: 337-350.Jones, W.K. and J. M. Rawls. 1988. Genetic and molecular mapping ofchromosome region 85A in Drosophila melanogaster. Genetics 120: 733-742.Jowett, T. 1986 Preparation of nucleic acids p.275-286. In: Drosophila - apractical approach. Ed. D. E. Roberts. IRL Press, Oxford, Washington.Kajii, T. and K. Ohama. 1977. Androgenic origin of hydatidiform mole. Nature268: 633-634.Kamdor, K. P., M. E. Shelton and V. Finnerty. 1994. The Drosophilamolybdenum cofactor gene cinnamon is homologous to three EscherichIa colicofactor proteins and to the rat protein gephyrin. Genetics 137: 791 -801.Karpen, G. H.. 1994. Position-effect variegation and the new biology ofheterochromatin. Curr. Op. Genet. and Dev. 4: 281-291.Karpen, G. H. and A. C. Spradling. 1990. Reduced DNA polytenization of aminichromosome region undergoing position-effect variegation in Drosophila.Cell 63: 97-107.Keith, T.P., M. A, Riley, M. Kreitman, R. C. Lewantin, D. Curtis and G. Chambers.1987. Sequence of the structural gene for xanthine dehydrogenase (rosy locus)in Drosophila melanogaster. Genetics 116: 67-73.Kermicle, J. L. and M. Alleman. 1990. Gametic imprinting in maize in relation tothe angiosperm life cycle. p. 9-14. In: Genomic imprinting (development 1990supplement). Ed: M. Monk and A. Surani. Company of Biologists Ltd., 1990,Cambridge.Khesin, R. B., and V. N. Bashkirov. 1978. Maternal influence upon the V-typegene position effect in Drosophila melanogaster. Molec. Gen. Genet. 163: 327-334.KIar, A. J. 1987. Differentiatied parental DNA strands confer developmentalasymmetry on daughter cells in fission yeast. Nature 326: 466-470.KIar, A. J. 1990. Regulation of fission yeast mating-type interconversion bychromosome imprinting. p3-8. In: Genomic imprinting (development 1990295supplement). Ed: M. Monk and A. Surani. Company of Biologists Ltd., 1990,Cambridge.Kuhn, D. T. 1972. Report-D. melanogaster-new mutants. Dros. Info. Service49:38.Laird, C. D. 1987. Proposed mechanism of inheritance and expression of thehuman fragile-X syndrome of mental retardation. Genetics. 117: 587-599.Laird, C. D. 1990. Huntington’s disease: proposed mechanism of mutation,inheritance and expression. Trends Genet. 6: 242-247.Laurenson, P. and J. Rine. 1992. Silencers, silencing and heritabletranscriptional states. Microbiol. Rev. 56: 543-560.Lederer, E. 1940. Les pigments des invertebres. Biol. Rev. Camb. Phil. Soc. 15:273-306.Lewis, E. B. 1951. Pseudoallelism and gene evolution Cold Spring HarborSymp. Quant. Biol. 16:159-174.Linder, D., B. McCaw, X. Kaiser and F. Hecht. 1975. Parthenogenetic origin ofbenign ovarian teratoma. New EngI. J. Med. 292: 63-66.Lindsley, D. L. and G. G. Zimm. 1992. The genome of Drosophila melanogaster.Academic Press. N.Y.Livak, K. J.. 1990. Detailed structure of the Drosophila melanogaster Stellategenes and their transcripts. Genetics 124: 303-31 6.Lock, L. F., N. Takugi and G. R. Martin. 1987. Methylation of the Hprt gene in theinactive X occurs after chromosome inactivation. Cell 48: 39-46.Lucchesi, J. C. 1968. Synthetic lethality and semi-lethality among functionallyrelated mutants of Drosophila melanogaster. Genetics 59: 37-44.Lyon, M. F. 1993. Epigenetic inheritance in mammals. Trends Genet. 9: 123-128.Mackay, W. J. and J. M. O’Donnell. 1983. A genetic analysis of the pteridinebiosynthitic enzyme, guanosine triphosphate cyclohydrolase, in Drosophilamelanogaster. Genetics 105: 35-53.Mainx, F. 1938. Analyse der genwirkung durch faktoren kombination. Versuchemit den augenfarbenfaktoren von Drosophila melanogaster. Z. lnductiv.Abstammungs-Vererbungslehre (Mol. Gen. Genet.) 75: 256-276.Mattaj, I, W.. 1994. Splicing in space. Nature 372:727-728.296Matzke, M. A. and A. J. M. Matzke. 1995. Homology-dependent gene silencingin transgenic plants: what does it really tell us? Trends Genet. 11: 1-3.McCarthy, A. and H. Nickla. 1980. Morphology of the carnation-light syntheticlethal focus in Drosophila melanogaster. Experimentia 36:1361-1362.McClintock, B. 1944. The relation of homozygous deficiencies to mutations andallelic series in maize. Genetics 29: 478-502.McLean, J. R., R. Boswell and J. O’Donnell. 1990. Cloning and molecularcharacterization of a metabolic gene with developmental functions inDrosophila. Genetics 126: 1007-1019.Metz, C. W.. 1938. Chromosome behavior, inheritance and sex determination inSciara. Am. Naturalist 72:485-520.Meyerowitz, E. M. and D. R. Kankel. 1978. A genetic analysis of visual systemdevelopment in Drosophila melanogaster. Dev. Biol. 62: 112-142Modelell, J., W. Bender and M. S. Messelson. 1983. Drosophila melanogastermutations suppressible by the suppressor-of-Hairy-wing are insertions of a 7.3kb mobile element. Proc. Nat. Acad. Sci. USA 80: 1678-1 682.Monk, M. 1988. Genomic imprinting. Genes and Dev. 2: 921-925.Monk, M. and M. Grant. 1990. Preferential X-chromosome inactivation, DNAmethylation and imprinting. In: Genomic imprinting (development 1990supplement). Ed: M. Monk and A. Surani. Company of Biologists Ltd., 1990,Cambridge.Moore, T., and D. Haig. 1991. Genomic imprinting in mammalian development:A parental tug-of-war. Trends Genet. 7: 45-49.Mori, K. 1937. A study on the development of pigments in various eye colormutants of Drosophila. Japan. J. Genet. 13: 81-99.Moses, K., M. C. Ellis and G. M. Rubin. 1989. The glass gene encodes a zinc-finger protein required by Drosophila photoreceptor cells. Nature 340: 531-536.Mottus, R., R. Reeves and T. A. Grigliatti. 1980. Butyrate suppression of positioneffect variegation in Drosophila melanogaster. Mol. Gen. Genet. 178: 465-469.Muller, H. J. 1958. An androgenic homozygous male. Dros. Info. Serv. 32: 140.Nájera, C. 1985. Report-D. melanogaster-new mutants. Dros. Info. Service61 :21 5.Narayanan, Y. and J. A. Weir. 1964. Paper chromatography of pteridines ofprune and clot stocks of Drosophila melanogaster. Genetics 50: 387-392.297Nash, W. G. 1971. deep orange and carnation: Another lethal gene combinationin D. m. Dros. Info. Serv. 47:73.Nicholls, R. B., J. H. M. Knoll, M. G. Butler, S. Karam and M. Lalande. 1989.Genetic imprinting suggested by maternal heterodisomy in non-deletion PraderWilli syndrome. Nature 342: 281 -285.Nickla, H. 1977. Maternal effects determine effect lethal phase of carnation-lightsynthetic lethal in Drosophila melanogaster. Nature 268:638-639.Nickla, H., T. Lilly and A. McCarthy. 1980. Gene activity in the carnation-lightsynthetic lethal in Drosophila melanogaster. Experimentia 36:402-403.Nolte, D. J. 1943. Appearance of unexpected eye colors. Dros. Info. Serv. 17:63.Nolte, D. J. 1944. White of y wan interaction-product. Dros. Info. Serv. 18: 54.Nolte, D. J. 1950. The eye pigmentary system of Drosophila: The pigment cells.J. Genetics 50:79-99.Nolte, D. J. 1 952a. The eye pigmentary system of Drosophila: II. Phenotypiceffects of gene combinations. J. Genetics 51:130-141.Nolte, D. J. 1952b. The eye pigmentary system of Drosophila: III. The action ofeye-colour genes. J. Genetics 51:142-186.Nolte, D. J. 1 954a. The eye pigmentary system of Drosophila: IV. The pigmentsof the vermilion group of mutants. J. Genetics 52:111-126.Nolte, D. J. 1954b. The eye pigmentary system of Drosophila: V. The pigmentsof the light and dark groups of mutants. J. Genetics 52:127-135.Nolte, D. J. 1955. The eye pigmentary system of Drosophila: VI. The pigments ofthe ruby and red groups of mutants. J. Genetics 53:1-10.Nolte, D. J. 1959a. The eye pigmentary system of Drosophila: VII. The whitelocus. Heredity 13:219-231.Nolte, D. J. 1959b. The eye pigmentary system of Drosophila: VIII. Series ofmultiple alleles. Heredity 13:233-241.Nolte, D. J. 1 959c. The eye pigmentary system of Drosophila: IX. Heterozygouseffects of eye-colour genes. Heredity 13:219-231.Nur, U. 1970. Translocations between eu- and heterochromatic chromosomes,and spermatocytes lacking a heterochromatic set in male mealy bugs.Chromosoma 29: 42-61.298Nur, U. 1990. Heterochromatinization and euchromatinization of wholegenomes in scale insects (Coccoidea: Homoptera). P. 29-34. In: Genomicimprinting (development 1990 supplement). Ed: M. Monk and A. Surani.Company of Biologists Ltd., 1990, Cambridge.Nur, U., J. H. Werren, D. G. Eickbush, W. D. Burke and T. H. Eichbush. 1988. A“selfish” B chromosome that enhances its transmission by eliminating thepaternal genome. Science 240: 512-514.Ohlsson, R., D. Barlow and A. Surani. 1994. Impressions of imprints. TrendsGenet. 10: 41 5-41 7.Orlando, V. and R. Paro. 1995. Chromatin multiprotein complexes involved inthe maintenance of transcription patterns. Curr. Op. Genet. and Dev. 5: 174-179.Pak, W. L., Grabowski, S. R. 1980. Physiology of the visual and flight systems.In: Genetics and Biology of Drosophila, Ed. M. Ashburner and E. Noviski.Academic Press, London, New York, Sidney, Toronto and San Francisco.Palumbo, G., S. Bonaccorsi, L. G. Robbins and S. Pimpinelli. 1994. Geneticanalysis of Stellate elements of Drosophila melanogaster. Genetics 138: 1181-1197.Parkhurst, S. M. and V. G. Corces. 1985. forked, gypsys and suppressors inDrosophila. Cell 41: 429-437.Paro, R. 1990. Imprinting a determined state into the chromatin of Drosophila.Trends in Genetics 6: 116-121.Patterson, G. I., C. J. Thorpe and V. L. Chandler. 1993. Paramutation, an allelicinteraction, is associated with a stable and heritable reduction of transcription ofthe maize b regulatory gene. Genetics 135: 881 -894.Payne, F. and M. Denny. 1921. The heredity of orange eye color in Drosophilamelanogaster. Am. Nat. 55: 377-381.Pfleiderer, W. 1993. Natural pteridines - A chemical hobby. In: Chemistry andbiology of pteridines and folates. p. 1-16. Eds: J. E. Ayling, M. G. Nair and C. M.Baugh. Plenum Press, N. York.Phillips, J. P and H. S. Forrest. 1976. Ommochromes and Pteridines. p. 541-617. In: Genetics and Biology of Drosophila. Ed. Ashburner and Noviski,Academic Press, N.Y.Phillips, J. P, H. S. Forrest and Kulkarni. 1973. Terminal synthesis ofxanthommatin in Drosophila melanogaster. Ill. Mutational pleiotropy andpigment granule association of phenoxazinone synthesis. Genetics 73: 45-56.299Pirrota, V., H. Stellar and M. P. Bozzetti. 1985. Multiple upstream regulatoryelements control the expression of the Drosophila white gene. EMBO J. 4:3501-3508.Prokofyeva-Belgovskaya, A. A.. 1947. Heterochromatinization as a change ofchromosome cycle. J. of Genetics 48:80-98.Rabinow, L., A. T. Nguyen-Huyah and J. A. Birchler. 1991. A transactingregulatory gene that inversely affects the expression of the white, brown andscarlet loci in Drosophila Genetics 129: 463-480.Rabi, C. 1885. Uber Zelltheilung. Morphologisches Jahrbuch 10: 214-330.Rangunathang, R., W. A. Harris and C. S. Zuker. 1991. The molecular geneticsof invertebrate phototransduction. Trends Neurosci. 14: 486-493.Ready, D. P. 1989. A multifaceted approach to neural development. TrendsNeurosci. 12: 102-110.Real, M. D., J. Ferré and J. L. Mensua. 1985. Methods for the quantitativeestimation of the red and brown pigments of Drosophila melanogaster. Dros.Info. Serv. 61:198-1 99.Reame, A. G., D. A. Knecht and A. Chovnick. 1991. The rosy locus in Drosophilamelanogaster Xanthine dehydrogenase and eye pigments. Genetics 129:1099-1109.Reedy, J. J. and F. P. Cavalier. 1971. Epistasis in eye colors of Drosophilamelanogaster. J. Heredity 62:131-134.Reik, W. 1992. Genomic imprinting in mammals. p. 203-229. In: Results andproblems in cell differentiation 18. Ed. W. Hennig. Springer-Verlag. Berlin,Heidelberg.Reik, W., S. K. Howlett and M. A. Surani. 1990. Imprinting by DNA methylation:From transgenes to endogenous gene sequences. p. 99-106. In: Genomicimprinting (development 1990 supplement). Ed: M. Monk and A. Surani.Company of Biologists Ltd., 1990, Cambridge.Renfranz, D. J. and S. Benzer. 1989. Monoclonal antibody probes discriminateearly and late mutant defects in development of the Drosophila retina. Dev.Biol.. 136: 411-429.Reuter, G., I. Wolff and B. Friede. 1985. Functional properties of theheterochromatic sequences inducing m4 position-effect variegation inDrosophila melanogaster. Chromosoma 93: 132-139.Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecular cloning: Alaboratory manual. Ed: C. Nolan, Cold Spring Harbor Laboratory Press, N.Y.300Sapienza, C. 1990. Sex-linked dosage-sensitive modifiers as imprinting genes.p. 107-113. In: Genomic imprinting (development 1990 supplement). Ed: M.Monk and A. Surani. Company of Biologists Ltd., 1990, Cambridge.Sapienza, C., J. Paquette, P. Pannunzio, S. Albrechtson and K. Morgan. 1992.The polar-lethal Ovum Mutant gene maps to the distal portion of the mousechromosome 11. Genetics 132: 241-246.Sasaki, H., B. A. Jones, R. J. Chaillet, A. C. Ferguron-Smith, S. C. Barton, W.Reik and M. A. Surani. 1992. Paternal imprinting: potentially active chromatin ofthe repressed maternal allele of the mouse insulin-like growth factor II (IGF2)gene. Genes Dev. 6: 1843-1 856.Scarbrough, K., S. Hattman and U. Nur. 1984. Relationship of DNA methylationlevel to the presence of heterochromatin in mealy bugs. Mol. Cell Biol. 4: 599-603.Schalet, A. 1957. Spontaneous mutations at specific X chromosome loci inDrosophila melanogaster. Genetics 42:393.Schalet, A. P. 1986. The distribution of, and complementation relationshipsbetween spontaneous X-linked recessive lethal mutations recovered fromcrossing long-term laboratory stocks of Drosophila melanogaster. Mutation Res.163: 115-144.Schmidtke, J. P. KuhI and W. Engel. 1976. Transitory hemizygosity of paternallyderived alleles in hybrid trout embryos. Nature 260: 31 9-320.Schott, D. R., M. C. Baldwin and V. Finnerty. 1986. Molybdenum hydroxylases inDrosophila. Ill. Further characterization of the low xanthine dehydrogenasegene. Biochem. Genet.. 24:509-527.Schultz, J. 1935. Aspects of the relation between genes and development inDrosophila. Am. Nat. 69: 30-54.Schwinck, I and L. Schwinck. 1972. Report-D. melanogaster-new mutants.Dros. Info. Service 49:38.Schwinck, I. 1975. Aurodrospterins in eye colour mutants of Drosophilamelanogaster. In: The chemistry and Biology of Pteridines. Ed: W. Pfleiderer.Walter de Gruyter, Berlin, N. York. p. 919-929.Schwink, I. 1978. Drosopterins and the in vivo modulation of their synthesis byimplantation of metabolites in eye colour mutants of Drosophila melanogaster.In: The chemistry and Biology of Pteridines. Ed: R. L. Kisliuk and G. M. Grown.Elsevier, North Holland. p. 141-146.301Searles, L. L. and R. A. Voelker. 1986. Molecular characterization of theDrosophila vermillion locus and its suppressible alleles. Proc. Nat. Acad. Sci.USA 83: 404-408.Shepherd, J. C. W., U. Walldorf, P. Hug and W. J. Gehring. 1989. Fruit flies withadditional expression of elongation factor EF-1 oc live longer. Proc. Nat. Acad.Sci. USA 86: 7520-7521.Shevelyov, Y. Y.. 1992. Copies of a Stellate gene variant are located in the Xheterochromatin of Drosophila melanogaster and are probably expressed.Genetics 132: 1033-1037.Shoup, J. R. 1966. The development of pigment granules in the eye of wildtypeand mutant Drosophila melanogaster. J. Cell Biol. 29: 223-249.Sinclair, D. A., A. A. Ruddell, J. K. Brock, N. J. Clegg, V. K. Lloyd and T. A.Grigliatti. 1992. A cytogenetic and genetic characterization of a group of closelylinked second chromosome mutations that suppress position-effect variegationin Drosophila melanogaster. Genetics 130: 333-344.Spofford, J. B. 1967. Single-locus modification of position-effect variegation inDrosophila melanogaster. I. white variegation. Genetics 57: 751-766.Spofford, J. B.. 1959. Parental control of position-effect variegation: I. Parentalheterochromatin and expression of the white locus in compound-X Drosophilamelanogaster. Proc. Nat. Acad. Sci. 45: 1003-1007.Spofford, J. B.. 1961. Parental control of position-effect variegation. II. Effect ofsex of parent contributing white-mottled rearrangement in Drosophilamelanogaster. Genetics 46:1151-1167.Spofford, J. B.. 1976. Position-effect variegation in Drosophila. p. 955-1018. InGenetics and Biology of Drosophila. Ed. Ashburner and Novitski. AcademicPress. N.Y.Stewart, B. and J. R. Merriam. 1973. Segmental aneuploidy of the Xchromosome. Dros. Info. Service 50: 167-170.Stoger, R., P. Kubicka, C.-G. Lui, T. Katri, A. Razin, H. Cedar and D. P. Barlow.1993. Maternal-specific methylation of the imprinted mouse lgf2r locus identifiesthe expressed locus as carrying the imprinting signal. Cell 73:61 -71.Sturtevant, A. H. 1932. The use of mosaics in the study of the developmentaleffects of genes. Proc. 6th Int. Cong. Genet., Vol. 1: 304-307.Sturtevant, A. H. 1956. A highly specific complementary lethal system inDrosophila melanogaster. Genetics 41: 118-123.302Sturtevant, A. H. and E. Novitski. 1941. The homologies of the chromosomeelements in the genus Drosophila. Genetics 26: 517-541.Sturtevant, A. H., C. B. Bridges, T. H. Morgan, L. V. Morgan and J. U. Chili. 1929.Garnet. p. 24 In: Contributions to the genetics of Drosophila simulans andDrosophila melanogaster. Carnagie Inst. Wash. PubI. 399. Carnagie Inst.Washington.Sullivan, D. T. and M. C. Sullivan. 1975. Transport defects as the physiologicalbasis for eye color mutants of Drosophila melanogaster. Biochem. Genet. 13:603-613.Sullivan, D. T., R. J. Kitos and M. C. Sullivan. 1973. Developmental andgenetics studies on kynurenine hydroxylase from Drosophila melanogaster.Genetics 75: 651-661.Sullivan, D. T., S. L. Grillo, and R. J. Kitos. 1974. Subcellular localization of thefirst three enzymes of the ommochrome synthetic pathway in Drosophilamelanogaster. J. Exp. Zool. 188: 225-234.Surani, M. A., R. Kothary, N. D. Allen, P. B. Singh, R. Fundel, A. C. RegusonSmith and S. C. Barton. 1990. Genomic imprinting and development in themouse. p. 89-98. In: Genomic imprinting (development 1990 supplement). Ed:M. Monk and A. Surani. Company of Biologists Ltd., 1990, CambridgeSurani, M. A., W. Reik and N. D. Allen. 1988. Transgenes as molecular probesfor genomic imprinting. Trends. Genet. 4: 59-62.Taira, T. 1960. Is the double recessive Hflr3 ry homozygote a synthetic lethal?Dros. Info. Service 34: 107.Talbert, P. B., C. D. S. LeCiel and S. Henikoff. 1994. Modification of theDrosophila heterochromatic mutation brownDomlflaflt by linkage alterations.Genetics 136: 559-571.Tartof, K. D. and M. Bremer. 1990. Mechanisms for the construction anddevelopmental control of heterochromatin formation and imprinted chromosomedomains. p35-45. In: Genomic imprinting (development 1990 supplement). Ed:M. Monk and A. Surani. Company of Biologists Ltd., 1990, Cambridge.Tautz, D. and C. Pfeifle. 1989. A non-radioactive in situ hybridization method forthe localization of specific RNAs in Drosophila embryos reveals transitionalcontrol of the segmentation gene hunchback. Chromosoma 98:81-85.Tearle, R. G., J. M. Belote, M. Mckewan, B. S. Baker and A. J. Howells. 1989.Cloning and characterization of the scarlet gene of Drosophila melanogaster.Genetics 122: 595-606.303Tearle, R.. 1991. Tissue specific effects of ommochrome pathway mutations inDrosophila melanogaster. Genet. Res. Camb. 57:257-266.Teng, D. H. F., C. M. Engele and T. ft Venkatesh. 1991. A product of the prunelocus of the Drosophila is similar to mammalian GTPase-activating proteins.Nature 353: 437-440.Thaker, H. M., Kankel, D. R. 1992. Mosaic analysis gives an estimate of theextent of genomic involvement in the development of the visual system inDrosophila melanogaster. Genetics 131:883-894.Toguchida, J., K. Ishizaki, M. S. Sasaki, Y. Nakumara, M. Ikenaga, M. Kato, M.Sugimot, Y. Kotoura and T. Yamamuro. 1989. Preferential mutation of thepaternally derived RB gene as the initial event in sporadic osteosarcoma.Nature 338: 156-158.Tomlinson, A. 1985. Cellular interactions in the developing Drosophila eye.Development 104: 183-193.Tower, J., G. H. Karpen, N. Craig and A. C. Spradling. 1993. Preferentialtransposition of Drosophila P-elements to nearby chromsomal sites. Genetics133: 347-359.Valencia, R. M. 1966. Report-D. melanogaster-new mutants. Dros. Info. Service.41:58.Van Dyke, D. L., and L. Weiss. 1986. Maternal effect on intelligence in fragile Xmales and females. Am. J. Med. Genet. 23: 723-737.Varmuza, S. and M. Mann. 1994. Genomic imprinting - defusing the ovariantime bomb. Trends. Genet. 10: 118-123.Venkatesh, T. R., Zipursky, S. L., Benzer, S., 1985. Molecular analysis of thedevelopment of the compound eye in Drosophila. Trends in nuerosci. 8:251-257.Wakimoto, B. and M. Hearn. 1990. The effects of chromosome rearrangementson the expression of heterochromatic genes in chromosome 2L of Drosophilamelanogaster. Genetics 125: 141-151.Walirath, L. L. and S. C. R. Elgin. 1995. Position effect variegation in Drosophilais associated with an altered chromatin structure. Genes and Dev. 9:1263-1277.Warner, C. K., D. T. Watts and V. Finnerty. 1980. Molybdenum hydroxylases inDrosophila. I. Preliminary studies of pyridoxal oxidase. Mol. Gen. Genet. 180:449-453.Wennberg, R. A. 1988. A novel requirement for the X-chromosome in P-Mhybrid dysgenesis and the interaction of the garnet and enhancer of garnet loci304in Drosophila melanogaster. MSc. Thesis. University of British Columbia.Canada.Wiederrecht, G. J. and G. M. Brown. 1984. Purification and properties of theenzymes from Drosophila melanogaster that catalyze the conversion ofdihydroneopterin triphosphate to the pyrimidodiazepine precursor of thedrosopterins. J. Biol. Chem. 259: 14121-14127.Wiederrecht, G. J., D. R. Paton and G. M. Brown. 1984. Enzymatic conversion ofdihydroneopterin triphosphate to the pyrimido diazepine intermediate involvedin the biosynthesis of the drosopterins of Drosophila melanogaster. J. Biol.Chem. 259:2 1 95-2200.Wiley, K. and H. S. Forrest. 1981. Terminal Synthesis of xanthommatin inDrosophila melanogaster. IV. Enzymatic and non enzymatic catalysis. Biochem.Genet. 19:1211-1221.Wines, D. R. and S. Henikoff. 1992. Somatic instablity of a Drosophilachromosome. Genetics 131: 683-691.Wright, 5. 1932. Complementary factors for eye color in Drosophila. Am. Nat.66: 282-283.Yamamoto, A. H., D. J. Komma, C. D. Shaffer, V. Pirrotta and S. A. Endow. 1989.the claret locus in Drosophila encodes products required for eye color and formeiotic chromosome segregation. EMBO J. 8: 3543-3552.Zipursky, S. L. 1989. Molecular and genetic analysis of Drosophila eyedevelopment: sevenless, bride of sevenless and rough. Trends in Neurosci.12:183-1 89.Zipursky, S. L., T. R. Vankatesh, D. B. Teplow and S. Benzer. 1984. Neuronaldevelopment in the Drosophila retina: monoclonal antibodies as molecularprobes. Cell 36: 15-26.Zuckerkandl, E. 1974. Recherches sur les proprietes et l’activité biologique deIa chromatine. Biochimie 56: 937-954.305Appendix 1.Determination of eye pigment levels.As the principal phenotype of the garnet and other eye colour genes describedin this thesis is alterations in eye pigments, it was necessary to accurately andprecisely quantify the amount of eye pigments.A. Efficiency and Selectivity of pigment extraction.The first step in quantitation of eye pigments is to extract these pigmentsefficiently and specifically from the eye. The different biochemical properties ofthe pteridine (red) and ommochrome (brown) pigments have been exploited todifferentially extract those pigments from fly eyes. A number of such procedureshave been published, that of Real, Ferré and Mensua (1985) was adapted forthis work. It was necessary to show that these procedures efficiently extractedthe intended pigments, either the pteridines or xanthommatin, withoutcontamination by the other class of pigment. Figure A shows the pigmentsextracted from wild type (Oregon R), brown and vermilion mutants as a test ofthe efficiency and selectivity of the extraction method. The brown mutant shouldhave no pteridine pigments whereas the vermilion should not have theommochrome pigment. The top series of data show that the procedure forpteridine pigment extraction is both efficient and highly selective. Essentially noommochrome pigment is extracted. The second data series show that theprocedure for ommochrome pigment extraction is less selective, some pteridinepigments are extracted. It also appears less efficient as the levels of the306ommochrome pigment, xanthommatin, extracted from the brown mutant areconsistently lower than from wild type.The extraction method for pteridine pigments was not only more specific andreproducible but considerably less time consuming than the method for theommochrome pigment. Consequently, only the pteridine pigments wereassayed in most experiments.A.Efficiency of red pigment extractionQ%O.R 100 <1% 98±5Efficiency of brown pigment extractionO.R.averagereading 106±8 73±3 5±2%O.R. 100±8 69±8 5±2B. Contribution of the different pteridine pigments to total pteridine pigmentlevel.Unlike the single ommochrome pigment, there are numerous (28+) pteridinepigments, each of which may contribute to the reading obtained for the totalpteridine pigments. To determine the degree to which each of these pigmentscontributed to the total fluorescence recorded for the pteridines,307chromatography plates on which the pteridine pigments had been partiallyseparated was scanned under UV illumination using the same conditions usedto quatitate total pteridine pigment levels. The percent of total fluorescenceattributed to each pigment is shown in Figure B. The major contribution is fromthe drosopterin pigments (70%) and the second largest is from the unseparatedresidue (15%). Based on the distinctive orange colour, florescence of theresidue is probably due largely to drosopterins. Thus approximately 85% of totalfluorescence is due to drosopterin pigments. While the system could probablybe optimized to preferentially detect other pigments by changing the extractiontechnique and the UV illumination wavelength, the drosopterins are stablemolecules and a convenient indicator of pteridine pigment levels.B.Fluorescence of separated pteridine pigments.pigment percent total pigmentresidue 15drosopterin 70mystery spot 1 2mystery spot 2 1isoxanthopterin 1mystery spot 3 2xanthopterin 2sepiapterin 12-amino-4-hydroxypteridine 1biopterin 1isosepiapterin 5C. Comparison between different methods of pteridine pigment quantification.Most published methods of pigment quantification rely on a change in UVabsorbance as measured by spectrophotometer. I found this method to be toocumbersome and slow to be suited for large numbers of pigment level308determinations. A method involving quantification of fluorescence by amicroflourimeter allowed rapid scanning of multiple spots of extracted pigments.The top portion of Figure C shows a comparison between pigment levelsdetermined by spectrophotometric assay (Y axis) and by microflourometricassay (X axis). Each reading represents the results of pigment quantificationfrom one group of frozen fly heads. The graph incorporates data from twoseparate experiments. One set of data was produced by Kevin Swanson, andundergraduate research assistant.The relationship between the two methods seems to be slightly sinusoidal. Thisindicates that at high pigment levels the spectrophotometric method providesmore discrimination. However, the lower portion of the graph shows that for lowpigment levels (<20% by the microflourometric method) that themicroflourometric method is the more accurate. These results are notunexpected as the two methods exploit different properties of the pigments.Nevertheless, the relationship between the two methods is roughly linear. Thegreater speed afforded by the microflourometric assay and the lower variabilitymade the microflourometric assay the method of choice for pigmentquantitation.Although the microflourometric assay was more sensitive than thespectrophotometric assay, it still involved combining five heads to extractsufficient pigment. Combining heads posses no problem for genotypes whereall individuals should have the same amount of pigment. For mosaicphenotypes, such as the flies with variegating genotypes studied in chapter 3,each individual has different amounts of pigmentation. As combining headswould obscure these differences, I adapted the microflourometer assay to309measure pigment from individual heads. The lower figure shows a comparisonof microflourometer readings using one head (Y axis) and 5 heads (X axis).Each assay was again performed on the same set of frozen heads andincorporate two independent sets of experiments. One set of data was providedby David Dyment, an undergraduate student working on a directed studiesproject. The results of the two assays are roughly comparable. There ishowever, discrepancy between the two assays systems at higher pigmentlevels. The deviation appears as a random scatter about the mean suggestingthat it represents genuine differences between the pigment of individual headsand the average reading of five heads. Data derived from uniformly pigmentedheads shows far less variation (data not shown).310Comparison between microflou rimeter and spectrophotometerassays for pteridine pigment levelsU,cV(1$a,IL.a,a,Ea,0€00.(I)I0Ea)-c1microflourimeter readings (5 head assay)5 head microflourimeter assay311.100-90n— — —40’)— — — —70 — ——n_ — —I- I— — —_I -50— — — — — . — ‘— — — —).440— . —30--—-.-2”‘.1 I10— -0— — — — — — —+0 10 20 30 40 50 60 70 80 90 100Comparison between five and one head + + ÷microflourimeter assay +1009080706050403020100*++,- ——*4 —::41:- --k--- —-÷-- L ± — — — — — —+++I’ll0102030405060708090100IIAppendix 2.Cloning of the garnet gene.The cloning of the garnet gene forms the basis of the molecular analysisrecorded in chapter 2 of this thesis. This appendix is included to record thecloning of the gene. The garnet gene was cloned by Donald A. Sinclair using aP-element induced garnet gene isolated by Richard Wennberg (Wennberg1988).A P-element induced allele of garnet, and four revertants thereof, was isolatedfrom a naturally occurring P-element strain, S6-1 (see Figure 15). Afterreplacement of the autosomes and most of the X-chromosome byrecombination, two P-elements remained, one at cytological position 1 2B, thelocation of the garnet gene, and one at 1 2E. Genomic DNA isolated from the gPmutation, restricted with Eco RI and probed with the P-element containing HindIll fragment of the p1125.1 plasmid revealed an approximately 8.5 kb fragment.A library made from size fractionated DNA from the gP strain yielded eight Pelement containing clones when probed with the same P-element containingprobe. Hybridization of these clones to polytene chromosmes showed thatseven of these clones represented the 1 2E P-element and the last one, atcytological position 12B, potentially identified the garnet gene. Using the non Pelement containing DNA from this clone, a series of lambda phage cloneswhere isolated from a wild type genomic (EMBL 3) library (Figure 17 and 30).A 6.6 Eco RI fragment common to all the phage clones was subcloned intopuclg. This fragment was used to probe two c-DNA libraries (Figure 27) and anumber of spontaneous garnet mutations (Figure 19 and 22).312


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