Proc. Natl. Acad. Sci. USAVol. 92, pp. 5184-5188, May 1995Genetics

The su(Hw) protein bound to gypsy sequences in onechromosome can repress enhancer-promoter interactions in thepaired gene located in the other homologPAVEL G. GEORGIEV*t# AND VIcrOR G. CORCESt*Institute of Gene Biology, Russian Academy of Sciences, Moscow 117334, Russia; and tDepartment of Biology, The Johns Hopkins University,Baltimore, MD 21218

Communicated by Joseph G. Gall, Carnegie Institution of Washington, Baltimore, MD, January 17, 1995 (received for review July 21, 1994)

ABSTRACT The suppressor of Hairy-wing [su(Hw)] pro-tein exerts a polar effect on gene expression by repressing thefunction of transcriptional enhancers located distally fromthe promoter with respect to the location of su(Hw) bindingsequences. The directionality of this effect suggests that thesu(Hw) protein specifically interferes with the basic mecha-nism of enhancer action. Moreover, mutations in modifier ofmdg4 [mod(mdg4)] result in the repression of expression of agene when the su(Hw) protein is bound to sequences in thecopy of this gene located in the homologous chromosome. Thiseffect is dependent on the presence of the su(Hw) bindingregion from the gypsy retrotransposon in at least one of thechromosomes and is enhanced by the presence of additionalgypsy sequences in the other homology. This phenomenon isinhibited by chromosomal rearrangements that disrupt pair-ing, suggesting that close apposition between the twQ copies ofthe affected gene is important for trans repression of tran-scription. These results indicate that, in the absence of mod-(mdg4) product, the su(Hw) protein present in one chromo-some can act in trans and inactivate enhancers located in theother homolog.

Mutations in the suppressor of Hairy-wing [su(Hw)] genereverse mutant phenotypes resulting from the insertion of thegypsy retrotransposon at a variety of loci in Drosophila mela-nogaster (1). This effect has been extensively studied by usingthe yellow (y) gene (2). The gypsy-induced y2 allele displays atissue-specific mutant phenotype characterized by abnormalpigmentation of the wings and body cuticle, whereas all othertissues of the larva and adult show wild-type coloration (3). Inthis mutation, gypsy was inserted at bp -700 from the tran-scription start site of the y gene, such that the enhancers thatcontrol'y expression in the wings and body cuticle are locatedupstream of the gypsy insertion site (4-6). The resultingmutant phenotype is caused by the inability of these enhancersto act on the y promoter due to the presence of specificsequences of gypsy present in the transcribed untranslatedregion of this element (2). The region of gypsy responsible forits mutagenic effect contains 12 copies of a sequence similarto the octamer repeat found in enhancer elements of variousvertebrate genes (7-9). This octamer motif constitutes thebinding site for the su(Hw) protein (10), which acts as apositive regulator of the spatial and temporal expression ofgypsy (11). At the same time, binding of su(Hw) to the octamersequences of gypsy results in the repression of enhancerslocated distally with respect to the promoter from the gypsyinsertion point, thus causing the characteristic phenotype ofthey2 allele (12). The same type of effect is responsible for thegypsy-induced mutant phenotype in other Drosophila genes(13, 14).

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement" inaccordance with 18 U.S.C. §1734 solely to indicate this fact.

The gypsy element not only mediates interactions betweenenhancers and the promoter within a gene but also seems to playa role in allowing enhancers of a gene to act on the promoterlocated in the homologous chromosome (15). These trans inter-actions between enhancers and promoters explain the phenom-enon of transvection, the pairing-dependent interallelic comple-mentation between mutations located in homologous chromo-somes (16, 17). Transvection or interallelic complementation atthe y locus is one of several phenomena that involve pairingbetween chromosomes and have been termed trans-sensing ef-fects (18). One well-studied case of a trans-sensing effect is thatof dominant position-effect variegation at the brown (bw) locusof Di;osophila (19). This effect is caused by heterochromaticsequences located in the vicinity of the bw locus that not only caninactivate the gene located in cis but also can cause transinactivation of the homologous paired gene. The possibility thatthe su(Hw) binding region has similar properties could not bestudied in the past, since trans effects could not be distinguishedfrom cis interactions in this system. To approach this problem, wehave used a second gene, modifier of mdg4 [mod(mdg4)], thatalso affectsgypsy-induced phenotypes (20). Mutations in this genecause an enhancement in the gypsy-induced phenotype of y2,resulting in adult flies in which all cuticular structures have nully coloration. mod(mdg4)U1, the best characterized allele of mod-(mdg4), is caused by the insertion of the Stalker transposableelement. Molecular analysis of the locus indicates that thismutation accumulates "10-fold less mod(mdg4) RNA than wildtype, suggesting that mod(mdg4)"1 is a hypomorph that produceslower than normal amounts of the wild-type protein (T. Gerasi-mova and V.G.C., unpublished observations). The effect of themod(mdg4) mutation on the expression of they gene iny2 is notobserved in the presence of null mutations in su(Hw), suggestingthat the mod(mdg4) gene encodes a protein that interacts withsu(Hw) and modulates the interaction between this protein andy enhancers and promoter. Here we describe results that suggestthat a mutation in the mod(mdg4) gene allows the su(Hw) proteinto act in trans and inhibit the action ofy enhancers located in thehomologous chromosome on the promoter of their own gene.

RESULTSA Mutation in the mod(mdg4) Gene Allows Interchromosomal

Repressive Effects on y Transcription. The repressive effect ofsu(Hw) on y expression in y2 is limited to the chromosome inwhich the su(Hw) binding sites are present in the gypsy element,sincey2/+ flies show normal coloration of all cuticular structures.To test whether this negative effect of the su(Hw) protein on yexpression could be transmitted to a gene present in the homol-ogous paired chromosome, we analyzed the effect of themod(mdg4)ul mutation on the phenotype of y2/y+ flies. Theamount of pigmentation in different cuticular structures was

Abbreviation: LTR, long terminal repeat.*Permanent address: Engelhardt Institute of Molecular Biology, Rus-sian Academy of Sciences, 32 Vavilov Street, Moscow, Russia.

5184

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 29

, 202

0

Proc. Natl. Acad. Sci. USA 92 (1995) 5185

visually determined and assigned a value between 0 (completelack of pigmentation characteristic of the y' null allele) and 5(wild-type pigmentation). Table 1 shows a summary of the resultsof this analysis. The nwd(nzdg4)U1 mutation has no effect ony+/y+. However,y2/y+ flies with the mod(mdg4)uJ mutation showa decrease of pigmentation in the leg bristles from values of 5 to3 and an even more dramatic reduction in the coloration of theantenna from values of 5 to 1. This effect can only be explainedif expression ofboth copies of they gene is impaired in the bristlesand the antenna, suggesting that the mutation of mod(mdg4)allows thegypsy element present in one chromosome to affect thetranscription of they gene located in the other homolog. To testwhether this inhibition might also be caused by the su(Hw)protein, we first analyzed its dependence on the presence of thegypsy element by studying the effect of the mod(mdg4)ul mutationon the phenotype ofheterozygousy alleles, such asy4P, not causedby gysy insertion. Results from this experiment show that thecoloration of the leg bristles and the antenna ofy4P/y+ flies is thesame in the presence or absence of mod(mdg4) protein, suggest-ing that the presence of thegsy element in at least one copy ofthe y gene is important in mediating the interchromosomalrepression ofy transcription.The su(Hw) Binding Region Mediates Interchromosomal

Trans Repression of Transcription of the y Gene. To deter-mine the nature of the gypsy sequences involved in interchro-mosomal repression ofy expression, we analyzed the effect ofthe mod(mdg4)ul mutation on various derivatives ofy2 causedby the deletion of different regions of the gypsy element (Fig.1). The y59b allele lacks gypsy nt 1-4230, including the 5' longterminal repeat (LTR), the su(Hw) binding sites, and a largepart of the coding region (13). The coloration ofyS9b/y+ fliesis wild-type and is not affected by the mod(mdg4)ul mutation(Table 1), suggesting that the region responsible for transrepression of transcription is deleted in thegypsy element in they locus of the y59b allele. Further insight into the nature of thisregion was obtained by analyzing the effect ofmod(mdg4)ul onthe phenotype of heterozygousy69/y+ flies. The gypsy elementresponsible for the y69 allele lacks nt 1-653, including the 5'LTR and part of the transcribed untranslated region, butcontains all 12 copies of the su(Hw) binding site (13). Although

Table 1. Trans repression of transcription in heterozygouscombinations of y2 and its derivatives

Pigmentation

Bristles An-Genotype Th Ab W L Hair tenna

Y+/Y+ 5 5 5 5 5 5y+/y+; mod(mdg4)/mod(mdg4) 5 5 5 5 5 5y2/y2 5 5 5 5 5 1y2/y2; mod(mdg4)/mod(mdg4) 0 0 0 0 0 0y2ly+ 5 5 5 5 5 5y2/y+; mod(mdg4)/mod(mdg4) 5 5 5 3 5 1y4P/y+ 5 5 5 5 5 5y4P/y+; mod(mdg4)/mod(mdg4) 5 5 5 5 5 5yS9b/y+ 5 5 5 5 5 5yS9b/y+; mod(mdg4)/mod(mdg4) 5 5 5 5 5 5y69/y+ . 5 5 5 5 5 5y69/y+; mod(mdg4)/mod(mdg4) 5 5 5 2 4 0y2PRI/y+ 5 5 5 5 5 5y2PRl/y+; mod(mdg4)/mod(mdg4) 5 5 5 4 5 1y88d/y+ 5 5 5 5 5 5y88d/y+; mod(mdg4)/mod(mdg4) 5 5 5 5 5 4

Numbers indicate levels of pigmentation determined by visualobservation of adult flies under a dissecting microscope. Pigmentationlevels corresponding to those of a null y allele were assigned a valueof 0, whereas normal pigmentation is indicated by a value of 5. Bristleslocated in various parts of the fly are abbreviated as follows: Th,thorax; Ab, abdomen; W, wing; L, legs.

Ci) wing blade (a) mouth partsdenticle befts

(ZZ) body cuticle tarsal claws

bristlesQ denticle beltsaristae

* su(Hw)

FIG. 1. Structure of the y locus in wild-type and mutant alleles.Exons of the y gene are represented by thick lines and ovals withdifferent markings represent various tissue-specific transcriptionalenhancers that control y gene expression in the respective tissues.Insertions responsible for different alleles are represented by triangles.The gypsy element is inserted in the 5' region of they gene in the y2allele and its derivativesy2PRl, y2PR2, y",yb,9, andylul. Solid boxesflanking gypsy represent LTRs, and arrows indicate the direction oftranscription of this element. The ylul allele also contains a copy of theStalker element inserted into the second exon of they gene in additionto the gypsy element in the 5' region. Deletions ofDNA sequences arerepresented by horizontal lines spanning the deleted region. Thesu(Hw) binding site is represented by a solid sphere located in the 5'transcribed untranslated region of gypsy. Insertions of the hobo andjockey transposable elements into the su(Hw) binding region resultingin partial reversion of the y phenotype are represented by triangles.The y-700 and y-800 mutations were made in vitro by inserting thesu(Hw) binding region at bp -700 and -800, respectively, from thetranscription start site of they gene; the resulting plasmids were theninserted into y- flies by germ-line transformation (12).

y69/y+ flies display normal coloration, the presence of a muta-tion in mod(mdg4) results in a -marked decrease in the pig-mentation of the hairs, leg bristles, and the antenna (Table 1).These results suggest that a region between gypsy nt 653 and4230 is involved in mediating interchromosomal repression ofy transcription. This region contains the 12 copies of theoctamer motif that interacts with the su(Hw) protein.To test whether the su(Hw) binding region is indeed re-

sponsible for the observed interchromosomal repression of yexpression, we analyzed the effect of the mod(mdg4)ul muta-tion on gypsy-induced y alleles that carry alterations in thesu(Hw) binding region. Theyw" mutation is a partial revertantof y2 in which the gypsy element lacks 5 of the 12 su(Hw)binding sites (Fig. 1) (13). The mod(mdg4)u1 mutation failed todecrease pigmentation of cuticular structures in yUd/y+ flies(Table 1). This result indicates that gypsy regions that interactwith su(Hw) mediate the negative effect on transcriptiontransmitted between homologous chromosomes. This conclu-sion was confirmed by analyzing the effect of the mod(mdg4)ulmutation on a complete revertant of y2. This revertant, des-ignated y+2MC, is caused by a deletion of all 12 copies of thesu(Hw) binding site and the insertion in this region of a copyof the jockey element (Fig. 1) (T. Gerasimova, D. Gerasimov,D. Gdula, and V.G.C., unpublished data). The mod(mdg4)ulmutation does not affect the pigmentation ofy+2Mc/y+2Mc ory+2MC/y+ flies (Table 2), in agreement with the absence ofsu(Hw) binding sites in the gypsy element present in they gene

Genetics: Georgiev 'and Corces

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 29

, 202

0

5186 Genetics: Georgiev and Corces

Table 2. Trans repression of transcription in trans heterozygotesof y+2MC and derivatives of y2

Pigmentation

Bristles

Genotype Th Ab W L Hair Antennay+2MC/y+2MC 5 5 5 5 5 1y+2MC/y+2MC;

mod(mdg4)/mod(mdg4) 5 5 5 5 5 1y+2MC/y+ 5 5 5 5 5 5

y+2MC/y+;

mod(mdg4)/mod(mdg4) 5 5 5 5 5 5y2/y+2MC 5 5 5 5 5 1y2/y+2MC;

mod(mdg4)/mod(mdg4) 3 4 3 1 1 0y4P/y+2MC 5 5 5 5 5 1y4P/y+2MC;

mod(mdg4)/mod(mdg4) 5 5 5 5 5 1yS9b/y+2MC 5 5 5 5 5 1yS9b/y+2MC;

mod(mdg4)/mod(mdg4) 5 5 5 5 5 1y2PRI/y+2MC 5 5 5 5 5 1y2PR1/y+2MC;

mod(mdg4)/mod(mdg4) 3 4 3 2 2 0y2PR2/y+2MC 5 5 5 5 5 1y2PR2/y+2MC;

mod(mdg4)/mod(mdg4) 3 4 3 1 2 0y88d/y+2MC 5 5 5 5 5 1y88d/y+2MC;

mod(mdg4)/mod(mdg4) 5 5 5 2 5 1y69/y+2MC 5 5 5 5 5 1y69/y+2MC;

mod(mdg4)/mod(mdg4) 0-1 1-2 1 0 0 0ylulsZ3/y+2MC 5 5 5 5 5 1ylu1Z3/y+2MC;

mod(mdg4)/mod(mdg4) 3 4 3 1 1 0

Abbreviations and numbers indicating levels of pigmentation are asdescribed in Table 1.

of these flies. This result confirms the essential role of thesu(Hw) binding region in mediating interchromosomal repres-sion of transcription. Since the su(Hw) protein binds to thesesequences and interferes with the effect of cis enhancers on thepromoter located in the same gene, it follows that the effectson y expression observed in the background of themod(mdg4)ul mutation could be due to the ability of su(Hw)protein to also repress enhancer action in trans.Trans Repression of Transcription at the y Locus Is En-

hanced by Sequences Other than the su(Hw) Binding Site. Totest whether other sequences present in the gypsy element arealso required to mediate trans inactivation of enhancer func-tion by the su(Hw) protein, we analyzed the effect of mod-(mdg4)ul on trans heterozygotic combinations of the y2 alleleand they+2MC revertant. Results presented in Table 2 indicatethat the mod(mdg4)ul mutation dramatically decreases bristlepigmentation in y2/y+2MC trans heterozygotes. This effect isstronger than in y2/y+ flies (see Table 1). This enhancementmight be due to the formation of closer contacts betweenpaired chromosomes mediated by sequences of the gypsyelement. Further insight into the nature of the sequencesresponsible for the enhancement of trans repression of tran-scription was obtained by analyzing the effect of mod(mdg4)ulon the phenotype of trans heterozygotes between y+2MC andvarious y mutants (Table 2). Interchromosomal repression ofy transcription follows the same rules in combinations of they+2MC revertant with other gypsy-inducedy mutations as in thecase of y+ combined with the same mutants, except that thephenotypes are more extreme wheny+2MC is present. No effectwas observed in yS9b/y+2MC flies, again supporting the conclu-

sion that the su(Hw) binding region, absent in these twomutants, is essential for trans repression of transcription(Table 2). To test whether the phenotypic enhancement ob-served when copies of the gypsy element are present in bothhomologs is only due to the additional presence of su(Hw)binding sites, we tested the phenotype of combinations ofy+2MC with y alleles in which these sequences are deleted orpartially nonfunctional due to the insertion of transposableelements. We first analyzed this effect by using two partialrevertants of , y2PRI, and y2P1m, caused by the insertion of thejockey and hobo transposable elements into the su(Hw) bindingregion, respectively (7) (see Fig. 1). The mod(mdg4)ul muta-tion has a moderate inhibitory effect on y transcription inbristles and hairs in flies heterozygous for these two partial yrevertants in combination with y+2MC (Table 2). This effect isstronger than in y2PRl/y+ (see Table 1), suggesting that addi-tional gypsy sequences present in y+2MC and distinct from thesu(Hw) binding region are responsible for the observed en-hancement. The same conclusion can be obtained by examin-ing the effect of mod(mdg4)ul on y88d/y+2MC flies, which isstronger than that observed withy8m/y+ heterozygotes (Tables1 and 2). The stronger effect observed in the presence ofadditional gypsy sequences might be simply due to a moreintimate pairing between homologous chromosomes in theregion of the y locus.

Additional evidence for a role of pairing in the enhancementof interchromosomal repression of transcription was obtainedby analyzing the effect of the mod(mdg4)"1 mutation on thephenotype of combinations of y2 and some of its derivativeswith the y2+#9 revertant allele (Table 3). In this mutation, thegypsy element lacks half of the 5' LTR, and the completecentral region including the su(Hw) binding sites is replaced bya copy of thejockey element (21). The mod(mdg4)ul mutationhas a stronger negative effect ony expression iny2/y2+#9 thanin y2/y+ flies (Table 3) but is much weaker than in y2/y+2MC.The effect of mod(mdg4)u1 on other combinations of y2+#9withy2 derivatives (Table 3) follows the same pattern observedin combinations ofy+ ory+2MC with the same mutants (Tables1 and 2), but the enhancement is intermediate between that

Table 3. Trans repression of transcription in trans heterozygouscombinations ofy2+#9 withy2 and its derivatives

Pigmentation

Bristles

Genotype Th Ab W L Hair Antenna

y2+#9ly2+#9 5 5 5 5 5 1y2+#9/y2+#9.

mod(mdg4)/mod(mdg4) 5 5 5 5 5 1y2ly2+#9 5 5 5 5 5 1Y;I/; +#9;-

mod(mdg4)/mod(mdg4) 4-5 5 5 2 3-4 0y4Ply2+#9 5 5 5 5 5 1y4Ply+#9;mod(mdg4)/mod(mdg4) 5 5 5 5 5 1

yS9b/y2+#9 5 5 5 5 5 1,V9bly2+#9;

mod(mdg4)/mod(mdg4) 5 5 5 5 5 1y2PRJ/y2+#9 5 5 5 5 5 1y2PRI/y2+#9;

mod(mdg4)/mod(mdg4) 5 5 5 3 4 0yWly2f+#9 5 5 5 5 5 1yWly2f+#9;mod(mdg4)/mod(mdg4) 5 5 5 5 5 1

y69/y2+#9 5 5 5 5 5y9ly2+#9;mod(mdg4)/mod(mdg4) 4 5 4 1 3 0Abbreviations and numbers indicating levels of pigmentation are as

described in Table 1.

Proc. Natl. Acad. Sci. USA 92 (1995)

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 29

, 202

0

Proc. Natl. Acad. Sci. USA 92 (1995) 5187

observed withy+ andy+2Mc. For example, mod(mdg4)ul has noeffect on combinations of y2+#9 with alleles not induced bygypsy (y4P) or alleles in which the 5' half of this element ismissing (y59b). The enhancement of the y phenotype is inter-mediate on alleles in which the su(Hw) binding region ispartially deleted or truncated by insertion of transposableelements (y2PRJ andy8m") and is stronger on alleles that containan intact su(Hw) binding region (2 and y69) (Table 3). In allcases, the extent of enhancement is weaker than that observedin combinations of the same mutations withy+2mC but strongerthan in combinations with y+, suggesting that gypsy regionsother than the su(Hw) binding region favor interchromosomalrepression of enhancer function, probably by mediating tighterpairing between homologous chromosomes at the location ofthey gene. The importance of additional gypsy sequences thatfavor pairing for interchromosomal repression of gene expres-sion is clearly observed in the dramatic effect of mod(mdg4)ulon the phenotype ofy6I9/y+2MC flies, which show an almost nullphenotype of all bristles and hairs of the adult (Table 2). Thislarge effect cannot be explained only on the basis of thepresence of an intact su(Hw) binding region in y69, since it ismore pronounced that in y2/y+2MC (Table 2). One possibleexplanation is that they69 allele also contains a deletion of thepromoter and coding region of the y gene. The lack of the cispromoter in the gene containing su(Hw) binding sites mightthen favor interactions with the promoter located in trans, thusenhancing the mutant phenotype. The importance of theabsence of the promoter is underscored by results obtainedfrom the analysis of other gypsy-induced nully mutations. Theylul allele is ay2 derivative caused by the insertion of the Stalkerretrotransposon in the second exon of they gene, whereasyu2and ylu3 resulted from the deletion of the same exon; thesethree mutations still contain a copy of the gypsy elementinserted in the 5' region of they gene. Although, likey69, theseare null mutations carrying deletions within the y gene, themod(mdg4)ul mutation fails to enhance the phenotype ofylul/y+2MC ylu2/y+2MC, and ylu3/y+2MC to the same extent as

y69/y+2MC (Table 2). This result suggests that the extreme effectof mod(mdg4)u1 on the phenotype of y69/y+2MC is not simplydue to the deletion of y sequence but rather to the specificabsence of a functional promoter in the y69 gene.Trans Repression of Transcription Is Pairing-Dependent

but Is Not Affected by Mutations in the zeste (z) Gene. In theabsence of the mod(mdg4) protein, the repressive effect ofsu(Hw) is stronger on enhancers located in the same chromo-some than on those located in trans. This suggests thatrepressive interchromosomal effects are weaker than the pos-itive effects observed in the transvection phenomenon (15) orthe intrachromosomal effects observed in the repression ofenhancer-promoter interactions in the y gene by gypsy se-quences located in the same chromosome (12). The weaknessof the interchromosomal effects induced by the presence ofsu(Hw) binding sequences correlates with an increased sensi-tivity of these effects to chromosomal pairing under conditionsin which standard transvection effects are not affected. Whenthe tip of the X chromosome, containing the y2 locus, istranslocated to the Y chromosome, pairing between the y

genes located in the X and Y chromosomes is still possible, andyS9b/y2.y males display wild-type coloration (15). Therefore,interchromosomal activation of the y promoter located in theY chromosome by enhancers located in the X chromosome can

take place due to the pairing that can still occur between thesetwo chromosomes. Nevertheless, the mod(mdg4)ul mutationfailed to enhance the y phenotype of y+2MC/y2-Y males, sug-gesting that interchromosomal repression of y expression ispairing-dependent and more sensitive to pairing effects thanstandard transvection (Table 4). To confirm this result, we

used several strains that carry a deletion of the y and acaetegenes and, in addition, contain a copy of the y2 locus that hasbeen inserted in different places of the X chromosome by

Table 4. Effects of chromosomal pairing on trans repressionof transcription

PigmentationBristles

Genotype Th Ab W L Hair Antennay+2MC/ y 5 5 5 5 5 1y+2MC/y y;

mod(mdg4)/mod(mdg4) 5 5 5 5 5 1y-700/y+2MC 5 5 5 5 5 1y-700/y+2MC;

mod(mdg4)/mod(mdg4) 5 5 5 5 5 1y-800/y+2MC 5 5 5 5 5 1y-8001y+2MC;

mod(mdg4)/mod(mdg4) 5 5 5 5 5 1yD-CTRL/y+2MC 5 5 5 5 5 1yD-CTRL/y+2MC;

mod(mdg4)/mod(mdg4) 5 5 5 5 5 1Abbreviations and numbers indicating levels of pigmentation are as

described in Table 1.

P-element-mediated germ-line transformation (10, 22). Thesestrains, named y-700y-800, and yD-CTRL, display the samephenotype as y2 and fail to mediate transvection in combina-tions with y59b due to the inability of the two copies of the ylocus to pair (15). Combinations of these mutations withy+2MCshow normal pigmentation of all cuticular structures, but themod(mdg4)ul mutation fails to enhance the y phenotype oftrans heterozygous combinations between y+2MC and thesevarious strains. These results suggest again that the inability ofthe two copies of the y gene to pair interferes with the inter-chromosomal repression of y expression, supporting the con-clusion that chromosomal pairing is a prerequisite for thiseffect. Surprisingly, trans repression ofy transcription does notdepend on the presence of z protein. Neither the null Zv77h northe Z)p6 and z1 alleles alter the effect of the mod(mdg4)u1mutation on the phenotype of y2/y+ flies (data not shown).The su(Hw) Protein Is Responsible for Trans Repression of

Transcription at they Locus. The fact that interchromosomalrepression of y transcription is completely dependent on thepresence of su(Hw) binding sites suggests that this phenome-non, as is the cis inactivation ofy expression, is caused by thesu(Hw) protein. In agreement with this assumption, strongalleles that lack detectable levels of su(Hw) protein, such assu(Hw)2 or SU(HW)69k, cancel the effect of mod(mdg4)ul ontrans repression of gene expression (Table 5). The sensitivityof this phenomenon to levels of the su(Hw) protein can betested with su(Hw) mutations that affect levels rather than thestructure of the protein. In the presence of half the normallevel of su(Hw) protein, as is the case in heterozygous com-binations of su(Hw)2 or su(Hw)69k, the mod(mdg4)ul mutationenhances the y phenotype of y+2MC/y2, but to a lesser extentthan in flies with wild-type levels of su(Hw) protein (Table 5).Interestingly, trans inactivation effects are specially sensitive tomutations in su(Hw) affecting the C-terminal acidic domain ofthis protein such as su(Hw)J (23) (Table 5). Even in a het-erozygote, the su(Hw)J allele completely reverses the effect ofmod(mdg4)ul on the y phenotype of y2/y+2MC flies, whereasnull alleles in a heterozygote do not affect trans repression oftranscription. This result suggests that the acidic domains ofthe su(Hw) protein play an important role in the interchro-mosomal inactivation of y expression.

DISCUSSIONThe presence of su(Hw) protein bound togypsy sequences inthe 5' region of the y gene interferes with the effect ofenhancers located distally from the gypsy insertion site on they promoter (5, 12). The mechanism of this repressive effect

Genetics: Georgiev and Corces

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 29

, 202

0

5188 Genetics: Georgiev and Corces

Table 5. Effects of su(Hw) mutations on trans repressionof transcription

Pigmentation

Bristles

Genotype Th Ab W L Hair Antennay+2MC/y2 5 5 5 5 5 1y+2MC/y2;

mod(mdg4)/mod(mdg4) 3 4 3 1 1 0y+2MC/y2; su(Hw)2

mod(mdg4)/su(Hw)2mod(mdg4) 5 5 5 5 5 1

y+2MC/y2; su(Hw)2mod(mdg4)/+ mod(mdg4) 3 4 3 1 1 0

y+2MC/y2; Su(HW)69kmod(mdg4)/su(Hw)69kmod(mdg4) 5 5 5 5 5 1

y+2MC/y2; su(HW)69kmod(mdg4)/+ mod(mdg4) 3 4 3 1 1 0

y+2MC/y2; su(Hw)Jmod(mdg4)/su(Hw)Jmod(mdg4) 5 5 5 5 5 1

y+2MC/y2; su(Hw)J

mod(mdg4)/+ mod(mdg4) 5 5 5 5 5 1

Abbreviations and numbers indicating levels of pigmentation are asdescribed in Table 1.

on enhancer function is not known, but several hypotheseshave been put forward (10, 24, 25). One possibility we favoris that binding of su(Hw) to the octamer motifs present ingypsy might cause directional changes in chromatin structureby establishing boundaries between higher-order domains ofgene activity. This view is supported by the ability of su(Hw)to buffer the expression of the white (w) gene from chro-mosomal position effects in a manner independent of thelocation of this w gene in the genome (25). We haveexamined a series of genetic interactions between gypsy-induced mutation at the y locus. These interactions are notobserved in wild-type flies but can be studied in the presenceof mutations in the mod(mdg4) gene. Mutations in mod-(mdg4) cause an enhancement of they2 phenotype and resultin the absence of pigmentation in all tissues of the fly (20).The mod(mdg4) gene encodes a protein that interacts withsu(Hw). In the absence of mod(mdg4) protein, heterozygousy2/y+ flies show a mutant y phenotype. This effect isameliorated by mutations that affect the integrity of thesu(Hw) binding site present in the gypsy-induced allele,suggesting that the presence of su(Hw) protein bound tothese sequences is responsible for the observed phenome-non. Considering the mechanism by which the mod(mdg4)ulmutation enhances the y phenotype iny2/y2 flies and the factthat the y+ chromosome in y2/y+ flies contains an intact ygene, an explanation of these results must involve the abilityof the su(Hw) present in they gene in the y2 chromosome toalso affect the transcription of the gene located in the y+homolog. This phenomenon of interchromosomal repressionis enhanced by the presence of gypsy sequences in the y+chromosome and is specially strong when this chromosomecontains a revertant ofy2 that lacks the su(Hw) binding sitebut otherwise contains an intact gypsy element. This resultsuggests that close pairing between homologs at the y locus,mediated by gypsy sequences other than the su(Hw) bindingregion, is important for the trans inhibitory effect of su(Hw)on enhancer-promoter interactions. The importance ofchromosomal pairing is underscored by the lack of inter-

chromosomal repression when one of the copies of the ylocus is translocated to a different location in the genome.The phenomenon of interchromosomal repression of tran-

scription described here is very similar to the dominant positioneffects described for the bw locus (for a review, see ref. 26). In thelatter case, the effect of sequences that have an altered chromatinstructure and repress expression of the adjacent gene can betransmitted in trans to the gene located in the other homolog. Inboth cases this effect is pairing-dependent but is not affected bymutations in the z gene. In the case of dominant brown variega-tion, sequences responsible for this effect are located in a regionof heterochromatin translocated next to the bw locus. In the caseof su(Hw), the structure of the chromatin created by binding ofthis protein to its target sequence is not known, but its ability tobuffer gene expression from any position effects has been inter-preted on the basis of the creation of boundaries betweendomains of higher-order chromatin structure (25); these bound-aries could be structurally similar to heterochromatin, furthersupporting the parallels between dominant position effects andthe trans repression of transcription described here. Interchro-mosomal repression of gene expression can be explained as aconsequence of chromatin changes mediated by su(Hw) and thesubsequent interaction of this protein, or other proteins recruitedto the area, with enhancer-bound transcription factors present inthe paired gene in the homologous chromosome. This transinteraction explains the sensitivity of this phenomenon to disrup-tions of chromosomal pairing. Similarly, dominant position ef-fects in the bw locus have been explained by the inhibitory effectof heterochromatic proteins located in one homolog on se-quences responsible for the regulation of bw expression presentin the paired gene (27). Thus, the su(Hw) and mod(mdg4)proteins might represent entry points into the in vivo study of therelative roles of chromatin structure in enhancer function ineukaryotes.

This research was supported by the Russian State Program Frontiersin Genetics, by an Award from the Fogarty International Center toP.G.G., and by U.S. Public Health Service Grant GM-35463 to V.G.C.

1.

2.3.4.5.6.7.

8.9.

10.11.12.13.14.

15.16.17.18.19.20.21.

22.23.

24.25.26.27.

Rutledge, B. J., Mortin, M. A., Schwarz, E., Thiery-Mieg, D. & Meselson,M. (1988) Genetics 119, 391-397.Corces, V. G. & Geyer, P. K. (1991) Trends Genet. 7, 86-90.Parkhurst, S. & Corces, V. (1986) Mol. Cell. Biol. 6, 47-53.Geyer, P. K., Spana, C. & Corces, V. G. (1986) EMBO J. 5, 2657-2662.Geyer, P. K. & Corces, V. G. (1987) Genes Dev. 1, 996-1004.Martin, M., Meng, Y. B. & Chia, W. (1989) MoL Gen. Genet. 218, 118-126.Geyer, P. K, Green, M. M. & Corces, V. G. (1988) Proc. Natl. Acad. Sci.USA 85, 8593-8597.Peifer, M. & Bender, W. (1988) Proc. Natl. Acad. Sci. USA 85,9650-9654.Mazo, A. M., Mizrokhi, L. J., Karavanov, A. A., Sedkov, Y. A., Krichev-skaya, A. A. & Ilyn, Y. V. (1989) EMBO J. 8, 903-911.Spana, C. & Corces, V. G. (1990) Genes Dev. 4, 1505-1515.Smith, P. M. & Corces, V. G. (1995) Genetics, in press.Geyer, P. K. & Corces, V. G. (1992) Genes Dev. 6, 1865-1873.Holdridge, C. & Dorsett, D. (1991) Mol. Cell. Biol. 11, 1894-1900.Jack, J., Dorsett, D., Delotto, Y. & Liu, S. (1991) Development (Cambridge,U.K) 113, 735-748.Geyer, P. K, Green, M. M. & Corces, V. G. (1990)EMBO J. 9, 2247-2256.Lewis, E. B. (1954) Am. Nat. 88, 225-239.Judd, B. H. (1988) Cell 53, 841-843.Tartof, K. D. & Henikoff, S. (1991) Cell 65, 201-203.Henikoff, S. & Dreesen, T. D. (1989)Proc. Natl. Acad. Sci. USA 86,6704-6708.Georgiev, P. G. & Gerasimova, T. I. (1989) Mol. Gen. Genet. 220, 121-126.Geyer, P. K., Green, M. M. & Corces, V. G. (1988) Proc. Natl. Acad. Sci.USA 85, 3938-3942.Smith, P. A. & Corces, V. G. (1992) Mol. Gen. Genet. 233, 65-70.Harrison, D. A., Gdula, D. A., Coyne, R. S. & Corces, V. G. (1993) GenesDev. 7, 1966-1978.Wolffe, A. (1994) Curr. Biol. 4, 85-87.Roseman, R. R., Pirrotta, V. & Geyer, P. K. (1993) EMBO J. 12, 435-442.Henikoff, S. (1990) Trends Genet. 6, 422-426.Dreesen, T. D., Henikoff, S. & Loughney, K. (1991) Genes Dev. 5,331-340.

Proc. Natl. Acad. Sci. USA 92 (1995)

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 29

, 202

0