Evolution of Developmental Control Mechanisms

Sex-specific repression of dachshund is required for Drosophila sexcomb development

Joel Atallah n, Gerard Vurens, Setong Mavong, Alexa Mutti, Don Hoang, Artyom KoppDepartment of Evolution & Ecology, University of California, One Shields Avenue, Davis, CA 95616, USA

a r t i c l e i n f o

Article history:Received 11 August 2013Received in revised form7 December 2013Accepted 11 December 2013Available online 19 December 2013

Keywords:DrosophilaSex combdachshundDistallessSex determinationNovelty

a b s t r a c t

The origin of new morphological structures requires the establishment of new genetic regulatory circuits tocontrol their development, from initial specification to terminal differentiation. The upstream regulatorygenes are usually the first to be identified, while the mechanisms that translate novel regulatoryinformation into phenotypic diversity often remain obscure. In particular, elaborate sex-specific structuresthat have evolved in many animal lineages are inevitably controlled by sex-determining genes, but thegenetic basis of sexually dimorphic cell differentiation is rarely understood. In this report, we examine therole of dachshund (dac), a gene with a deeply conserved function in sensory organ and appendagedevelopment, in the sex comb, a recently evolved male-specific structure found in some Drosophila species.We show that dac acts during metamorphosis to restrict sex comb development to the appropriate legregion. Localized repression of dac by the sex determination pathway is necessary for male-specificmorphogenesis of sex comb bristles. This pupal function of dac is separate from its earlier role in legpatterning, and Dac at this stage is not dependent on the pupal expression of Distalless (Dll), the mainregulator of dac during the larval period. Dll acts in the epithelial cells surrounding the sex comb duringpupal development to promote sex comb rotation, a complex cellular process driven by coordinated cellrearrangement. Our results show that genes with well-conserved developmental functions can be re-usedat later stages in development to regulate more recently evolved traits. This mode of gene co-option maybe an important driver of evolutionary innovations.

& 2013 Elsevier Inc. All rights reserved.

Introduction

New morphological structures often evolve through the spatialand temporal redeployment of pre-existing genes (Carroll, 2008).For example, male-specific wing spots found in some Drosophilaspecies are associated with novel expression patterns of wingless(Werner et al., 2010) and Dll (Arnoult et al., 2013) in pupal wings.In addition to this new and lineage-specific function, both genesplay deeply conserved roles in wing patterning during earlierdevelopmental stages (Held, 2002). One of the questions raised bythese findings is how a gene can be co-opted to regulate novelphenotypes without pleiotropic effects on its ancestral functions.A related question is how complex, integrated phenotypes arise inthe course of evolution (Monteiro and Podlaha, 2009; Salazar-Ciudad and Marín-Riera, 2013). Novel characters frequently consistof a suite of simpler traits that work together to perform a specificfunction (Wagner et al., 2008). While it is often relatively easy toidentify a co-opted “master control gene” (Gehring, 1996) at a highlevel in the genetic hierarchy, determining how this gene regulateseach of the many component traits can be more difficult.

The sex comb, a male-specific array of modified bristles thatdevelops on the front legs of Drosophila melanogaster and many of itsclose relatives, is a powerful system for addressing questions aboutthe origin and diversification of complex novel traits (Kopp, 2011;Seher et al., 2012). The sex comb evolved in the Sophophora subgenuswithin the genus Drosophila (Barmina and Kopp, 2007). The “teeth”(bristle shafts) that make up the sex comb show a number ofstructural modifications compared to the ancestral bristle morphol-ogy (Fig. 1). In D. melanogaster and some other species, sex combsalso rotate by up to 901 during pupal development, from a transverseto an oblique or longitudinal orientation (Atallah et al., 2012, 2009a,2009b; Tanaka et al., 2009; Tokunaga, 1962) (Fig. 1E).

Sex comb evolution involved the deployment of doublesex (dsx),a transcription factor that mediates sex-specific differentiation inDrosophila and other insects, in a novel spatial pattern in theprothoracic leg (Tanaka et al., 2011). Subsequent changes in dsxexpression are associated with the evolution of the remarkablydiverse sex comb structures seen in different Drosophila species.dsx acts in concert with Sex combs reduced (Scr), the HOX generesponsible for the development of the prothoracic segment (Held,2010), to induce sex comb formation. Sex comb development isrestricted to a specific region within the foreleg by the action ofgenes that control proximo-distal, anterior–posterior, and dorso-ventral leg patterning (Kopp, 2011).

Contents lists available at ScienceDirect

journal homepage: www.elsevier.com/locate/developmentalbiology

Developmental Biology

0012-1606/$ - see front matter & 2013 Elsevier Inc. All rights reserved.http://dx.doi.org/10.1016/j.ydbio.2013.12.017

n Corresponding author.E-mail address: jatallah@ucdavis.edu (J. Atallah).

Developmental Biology 386 (2014) 440–447

In D. melanogaster and its relatives, sex combs are homologousto the distal transverse bristle rows (TBRs) and to the morphologi-cally similar bristles just anterior to these rows (longitudinal rows6 and 7); the same precursors are found in conspecific females andin males of species that lack sex combs. In contrast to the othertransverse row bristles, which are straight, pencil-shaped, andlightly pigmented, sex comb teeth are thicker, curved, blunt andheavily melanized (Fig. 1). The distinctive shape of sex comb teethdepends on dsx (Tanaka et al., 2011), but the downstream mechan-isms that mediate its effects have remained unknown. In this study,we investigate the role of dachshund (dac), a gene involved inproximo-distal leg patterning and sensory organ development, inestablishing bristle diversity in the Drosophila foreleg. Our resultsshow that dac acts downstream of dsx in the developmentalpathway that controls the morphology of sex comb teeth, explain-ing one of the many aspects of sex-specific morphogenesis.

Materials and methods

Ectopic expression and RNAi

The GAL4/UAS system (reviewed in Phelps and Brand, 1998)was used to drive ectopic gene expression or RNAi. In cases where

this proved lethal, or where an earlier phenotype masked a laterone (e.g., by ablating leg segments), the TARGET system (McGuireet al., 2003) was used to limit the activity of the GAL4 to specificperiods in development. Flies carrying tub-GAL80ts and the desiredGAL4 and UAS constructs were shifted from 18 1C to 30 1C at thedesired timepoint. GAL80 binds to GAL4 at the colder temperature,preventing activation of UAS transgenes, but it is inactive at thewarmer temperature, allowing GAL4 to bind to UAS and activategene expression. The following constructs were used: UAS-dac(Shen and Mardon, 1997); UAS-Dll (Gorfinkiel et al., 1997); UAS-DsxM (Lee et al., 2002); UAS-dac-RNAi (Dietzl et al., 2007); UAS-Dll-RNAi (Dietzl et al., 2007); UAS-TraF (Ferveur et al., 1995); sca-GAL4;rn-GAL4 (St. Pierre et al., 2002); Dll-GAL4.

Mitotic clones

The FLP/FRT system (Theodosiou and Xu, 1998) was used togenerate dac mutant clones through mitotic recombination. Wecrossed dacE462/CyO (Bloomington Drosophila Stock Center) anddac4/CyO (Mardon et al., 1994) males to hs-FLP; FRT 40A::GFPfemales to generate hypomorphic and null mitotic clones, respec-tively. Larvae were heat-shocked for 1 h at 37 1C at 72 h afterhatching to induce mitotic recombination. Null or hypomorphicdac clones were marked by the absence of GFP. In addition to being

Fig. 1. The sex comb. (A) Scanning electron micrograph of a female D. melanogaster tarsus, showing all five segments (ts1-ts5). (B) Close-up view of the distal ts1 of female D.nikananu. The female bristle pattern is widely conserved in Drosophila and related genera. Sensory organs in both sexes include needle-shaped mechanosensory longitudinalrow bristles (large arrow), curvilinear chemosensory bristles (small arrow) and pencil-shaped transverse row bristles (arrowhead). Note that transverse bristle rows (TBRs), andthe bristles just anterior to them (right) continue to the distal tip of the segment. (C) In male D. melanogaster, the most distal TBR on ts1 is replaced by a rotated sex comb(arrow), consisting of enlarged bristles with curved tips. Arrowhead indicates the TBR just proximal to the sex comb. (D) Bristles just anterior and dorsal to the sex comb(arrows) are displaced proximally from the distal tip of the segment. (E) The sex comb is initially transverse and rotates as a consequence of male-specific tissue extension in thedistal ts1 (black arrows). This extension also causes the bristles just anterior and dorsal to the comb to migrate proximally, displacing them from the tip of the segment. Sexcomb bristles are depicted by black circles, longitudinal row bristles by large gray circles, transverse row bristles by small gray circles, chemosensory bristles by open circles, andthe central bristle by an asterisk. Distal is down and anterior is to the right in all panels. Scale bars: 10 μm.

J. Atallah et al. / Developmental Biology 386 (2014) 440–447 441

reduced in these clones, dac was also reduced in sex comb teeth thatwere generated non-autonomously through secondary processes (e.g. enlargement of the sex comb region through cell proliferation), butthese were identifiable because they contained GFP.

Pupal dissection, tissue staining, and imaging

White prepupae were collected and aged in a 25 1C incubator.Dissections were carried out in phosphate buffered saline (PBS),and pupae were fixed in 4% formaldehyde or 4% paraformalde-hyde. Antibody and Phalloidin staining followed standard proto-cols with previously described modifications for pupal legs(Atallah et al., 2009a; Tanaka et al., 2009). The following primaryantibodies were used: mouse anti-Dac (Mardon et al., 1994) (1:2);guinea pig anti-Dll (McKay et al., 2009) (1:800); rat anti-DsxM

(Hempel and Oliver, 2007) (1:50) and rat anti-E-cadherin (Odaet al., 1994) (1:2).

Stained pupal legs were imaged using confocal microscopy. 3-Dprojections of multiple confocal slices were generated using theZeiss LSM Image Browser and ImageJ. Adults legs were imagedusing either light microscopy or scanning electron microscopy(Atallah et al., 2009a).

Results

dac is specifically repressed in sex comb teeth

In larval leg imaginal discs, before the sex comb and most otherbristles are specified, dac expression correlates with its function inpatterning medial leg segments, from the femur to proximal tarsus(Kojima, 2004; Mardon et al., 1994). Flies homozygous for loss-of-function dac alleles lack the corresponding leg region (Mardonet al., 1994). At the pupal stage, dac is expressed at a high level inthe first tarsal segment (ts1) and weakly in the second segment(ts2) (Randsholt and Santamaria, 2008). In D. melanogaster, incontrast to many other Drosophila species, the sex comb developsin ts1 but not ts2 (Barmina and Kopp, 2007). Gain-of-function dacalleles cause ectopic sex combs in ts2, suggesting that strong dacexpression promotes sex comb development (Devi et al., 2013;Docquier et al., 1997). We found, however, that dac expression isspecifically reduced in the sex comb teeth in D. melanogaster(particularly the shaft and socket cells; Fig. 2) and other species(Figs. 3 and S2; Atallah et al., 2009b). This paradoxical resultmotivated us to examine the role of dac in sex comb developmentmore closely.

Dac is still present in ts1 as late as 30 h after pupariation (AP),when the bristle arrangement bears a strong resemblance to theadult pattern (Fig. 2). The Dac protein is near-ubiqutious in theepithelium, and is strongly expressed in the chemosensory andtransverse row bristles. However, it is strikingly absent from theshaft and socket cells of the sex comb teeth (Fig. 2 A and B). Incontrast, Dac is present in the most distal transverse bristle row(TBR) in females, which is homologous to the male sex comb(Fig. 2C, D). These observations suggest that dac repression in bristlecells may be necessary for male-specific sex comb differentiation.

The repression of dac occurs during a critical time in sex combdevelopment. At the prepupal stage, when many of the leg sensoryorgan precursor (SOP) cells are first specified (Orenic et al., 1993),we see no evidence of reduced Dac expression in the presumptivesex comb region (Fig. S1 A and B). At 16 h AP, on the other hand,when the SOPs that make up the sex comb are beginning to divideand migrate to their adult positions (Atallah et al., 2009a), there isclear evidence of dac repression in these precursors (Fig. S1 E andF). Legs that are slightly younger than 16 h AP still show some dacexpression in the sex comb region (Fig. S1 C and D), suggestingthat the loss of Dac occurs rapidly at around 16 h AP. Bristle shaftoutgrowth begins at around 20 h AP, indicating that dac repressionprecedes bristle morphogenesis by several hours.

dac repression correlates with male-specific bristle modifications

While sexually monomorphic aspects of the tarsal bristle patternshow strong conservation, sex comb morphology has diversifiedremarkably (Kopp, 2011; Figs. 3 and S2). In Drosophila ananassae(Fig. S2A) and several of its relatives, transverse (non-rotating) sexcombs are present on distal ts1 and throughout ts2. UnlikeD. melanogaster, the sex comb teeth of D. ananassae are straightand not heavily melanized. In male D. ananassae pupae, Dac isrepressed in the developing sex comb teeth, correlating with strongexpression of DsxM in these bristles (Fig. S2 A–E). In contrast, Dac isstrong and DsxM is low or absent in the unmodified transverse rowbristles that occupy the proximal and anterior regions of ts1. On ts2,all TBRs show reduced Dac, consistent with their differentiation intosex comb teeth. Similarly, Dac is strongly repressed in the sex combof Drosophila bipectinata, a species that is closely related toD. ananassae but has a rotating sex comb only on ts1, composed ofenlarged, curved, melanized teeth (Fig. 3 A–D). These findings,together with previous results from species in the montium subgroupand Oriental lineage of the melanogaster group (Atallah et al., 2009b),

Fig. 2. Sexually dimorphic dac expression in developing pupal legs. The distal ts1 of 30 h AP sca-GAL4 UASmCD8::GFP legs is shown. (A and B) In male forelegs labeled withE-cadherin (green) and Dac (red), dac is strongly expressed in the transverse row (arrowheads) and chemosensory bristles (small arrows), but is reduced in the sex comb(large arrow). (C and D) In a female leg, dac (red) is expressed in the distal TBR (arrow), which is homologous to the male sex comb. Scale bars: 10 μm. (For interpretation ofthe references to color in this figure legend, the reader is referred to the web version of this article.)

J. Atallah et al. / Developmental Biology 386 (2014) 440–447442

suggest that Dac is specifically repressed in sex comb bristles regard-less of their exact morphology or position in the leg.

To reconstruct the evolutionary origin of dac repression, weexamined more distantly related Drosophila species. In Drosophilaguanche, a member of the obscura species group within the OldWorld Sophophora, we see reduced Dac expression in the sex combslocated on ts1 and ts2 (Fig. S2 F–K). On the other hand, in Drosophilawillistoni, a more basal Sophophora species that lacks sex combs, Dacis present in the distal TBR of ts1, which is homologous to the sexcomb of D. melanogaster and other species (Fig. S3 A–C). It is alsoexpressed in the bristles just anterior to the TBRs (longitudinal rows6 and 7), which are homologous to the sex combs of Drosophilaficusphila and the montium species subgroup (Atallah et al., 2009b;Kopp, 2011). A similar pattern is seen in Drosophila virilis, a distantly

related species that primitively lacks sex combs (Fig. 3 E–G). Theseresults suggest that male-specific repression of dac coincides withthe origin of the sex comb (Fig. 3H).

dac represses sex comb tooth development

The previous finding that ectopic dac induces sex combs on ts2(Docquier et al., 1997), together with the repression of dac in sexcomb teeth, suggests that its functions in sex comb developmentare stage-specific. We used the TARGET system (McGuire et al.,2003) to ectopically express dac during the D. melanogaster earlypupal stage, when the leg has extended significantly (Ward et al.,2003) and the bristle pattern is being set up (Atallah et al., 2009a).When dac is ectopically expressed only in SOPs using the sca-GAL4

Fig. 3. dac expression in distantly related species. (A–D) In D. bipectinata, a fly sporting two sex combs on ts1, Dac is absent from both sex combs (arrows) but present in theTBRs (arrowheads). (E–G) In males of D. virilis, a fly that primitively lacks sex combs, dac is expressed in the distal TBR of ts1 (arrow) and the adjacent bristles of longitudinalrows 6 and 7 (arrowheads). (H) A phylogeny of the species referred to in this figure and the Data Supplement, showing the typical bristle pattern on the first two tarsalsegments (sex comb teeth are represented by black circles, and other bristles by open circles). Scale bars: 10 μm.

J. Atallah et al. / Developmental Biology 386 (2014) 440–447 443

driver, the sex comb is partly transformed toward TBR morphol-ogy: the teeth become thinner, lighter, and less curved, and losetheir characteristic rounded tips (Fig. 4B). In some individuals, thesex comb is split into two or more separate groups of bristles,suggesting a loss of adhesion between sex comb SOPs.

We also expressed dac more broadly throughout the distal tarsalepithelium using the rn-GAL4 driver. The effect varied considerablydepending on the time in development when expression wasactivated. Temperature shift shortly before pupariam formation(effectively activating ectopic dac during the early pupal stage,given the lag in full GAL4 activation of approximately 16 h aftertransfer to the warmer temperature (Pavlopoulos and Akam, 2011))resulted in only a modest disorganization and transformation of sexcomb teeth (Fig. 4C). Earlier activation of dac with a temperatureshift approximately 24 h before puparium formation (i.e. drivingexpression in larval imaginal discs), blocked sex comb rotation,reduced the number of sex comb teeth on ts1, induced ectopic sexcomb teeth on ts2, and reduced the number of leg segments (Fig. 4D and E). Sex comb rotation is driven by coordinated rearrangementof the surrounding epithelial cells (Atallah et al., 2009a), suggestingthat the early role of dac in this process is unrelated to its laterfunction in controlling bristle morphology.

Since ectopic dac represses sex comb formation in the earlypupa, we tested whether reduced dac expression is sufficient totransform other bristles into sex comb teeth. Down-regulating dac

in males by RNAi using the rn-GAL4 driver transforms a few distaltransverse row bristles toward a sex comb-like phenotype (Fig. 5A).However, no other bristles in the rn-expressing region show anytransformation. Since RNAi produces only a partial knock-down (Dacprotein can still be detected by antibody staining), we generatedhypomorphic and null dac clones by mitotic recombination during thethird larval instar (Fig. 5 B–D). We found many males with ectopic sexcomb teeth, but all of them occurred in the anterior region of the legwhere the sex comb, TBRs, and longitudinal rows 6 and 7 are normallyfound. Bristles within null dac clones that were generated in otherparts of the leg were not transformed into sex comb teeth. Takentogether, these results suggest that male-specific dac repression cannottransform bristles into sex comb teeth outside of a limited anteriorregion of ts1. This is the same region that has high expression of Scr,which promotes TBR formation and is necessary for sex combdevelopment (Shroff et al., 2007), and where wingless (wg) regulatesthe development of ventral leg structures (reviewed in Held, 2002),suggesting that the ability of dac tomodify bristle shaft morphogenesismay be dependent on Scr and/or wg.

dac is repressed by the sex determination pathway

Sex comb development is promoted by the sex determinationpathway through the male-specific isoform of the doublesextranscription factor (dsxM) (Tanaka et al., 2011). Strong dsxM

Fig. 4. Effects of ectopic dac expression. (A) Wild-type D. melanogastermale distal ts1. Sex comb teeth (arrow) are enlarged and have round tips. (B) Ectopic expression of dac inthe proneural cells that make up the bristles (tub-GAL80ts/scaGAL4 UAS-mCD8::GFP; UAS-dac/þ , transferred from 18 1C to 30 1C as wandering larvae) partially reverts the sexcomb to a TBR-like morphology by making bristle shafts straighter, thinner, and less heavily pigmented (arrow). (C–E) When dac is ectopically expressed in a broad region of theepithelium (tub-GAL80ts/þ ; rn-GAL4/UAS-dac), from the distal half of TS1 through TS4, the effect is dependent on the timing of induction. When flies are transferred from 18 1Cto 30 1C around the time of pupariation, the effect is minimal. (C) When transferred approximately 24 h before pupariation, distal segments are lost, the number of sex combteeth on TS1 is reduced and the sex comb fails to rotate (D). Additional teeth appear on ts2 (arrow). (E) Close-up of the distal ts1 in panel D. Scale bars: 10 μm.

Fig. 5. Effects of reduced dac expression. (A) Reduced dac expression (rn-GAL4/UAS-dacRNAi) induces ectopic sex comb teeth in the distal TBRs (arrow). (B–D) Using the FLP/FRT system, dac null mitotic clones were generated during the mid-third larval instar. Ectopic sex comb teeth form in the anterior leg region (arrow), which are seen here at30 h AP. In other parts of the leg, clones that lack Dac (marked by the absence of GFP, arrowheads) do not induce sex combs. Scale bars: 10 μm.

J. Atallah et al. / Developmental Biology 386 (2014) 440–447444

expression, and reduced Dac, appear to be hallmarks of the sexcomb in all species we have examined (Figs. 2, 3 and S2; Atallahet al., 2009b; Tanaka et al., 2011). This reciprocal pattern suggeststhat dsxM may control sex comb tooth morphogenesis, at least inpart, by down-regulating dac. To test this hypothesis, we expressedthe female isoform of transformer (traF) in the distal tarsus using

the rn-GAL4 driver. TraF switches dsx splicing from the male to thefemale isoform (dsxF) (reviewed in Pomiankowski et al., 2004),transforming the sex comb into a female-like TBR (Fig. 6A).As predicted, dac is derepressed in the transformed sex combteeth (Fig. 6 B and C). Conversely, ectopic dsxM expression infemales transforms the homologous female TBR into a sex comb

Fig. 6. The sex determination pathway represses dac in the sex comb. (A–C) In UAS-traF/UAS-GAL4; rn-GAL4/þ males (A), Dac is present (B and C) in the feminized sex comb(arrow) at levels comparable to more proximal TBRs (arrowhead). (D and E) In females expressing dsxM (UAS-dsxM/þ; rn-GAL4/þ), Dac is reduced in the ectopic sex combteeth (arrow). Scale bars: 10 μm.

Fig. 7. Effects of reduced Dll expression in the sex comb region. (A) Ectopic Dll expression (sca::GAL4/UAS-Dll) in bristles can reduce their size. Sex comb teeth may becomeshorter and lose their pigmentation (arrows). (B) When Dll is broadly reduced in the distal tarsus (UAS-DllRNAi/tub-GAL80ts; rn-GAL4/þ), sex comb rotation is impeded (thickarrow), and bristles anterior to the comb do not migrate proximally (arrowhead). Some TBR bristles become darker and slightly thicker (thin arrow), appearing more sex-comb-like. The strong phenotype shown in the image is typically seen when Dll is reduced early in development, with the transfer to 30 1C occurring approximately 20 hprior to pupariation. (C) The effect of reduced Dll on ts1 length in males and females. (D–G) Reduced Dll expression does not visibly affect the pattern of Dac (red in panels Dand E), which is still strong in the TBRs (arrowheads) and reduced in the sex comb (arrow), or DsxM (green in panel G), which remains elevated in the sex comb andsurrounding region. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

J. Atallah et al. / Developmental Biology 386 (2014) 440–447 445

(Tanaka et al., 2011). We find that dac expression is reduced inthese ectopic sex combs, mimicking the male pattern (Fig. 6D andE). In contrast, ectopic dac has no detectable effect on dsxexpression (data not shown). Together, these results indicate thatdsx is responsible for the sexually dimorphic expression of dac.

Distalless is required for sex comb rotation

dac expression in the larval leg disc is largely regulated by Dll(Giorgianni and Mann, 2011). Similar to Dac, Dll is reduced in sexcomb teeth at the pupal stage, but unlike Dac it is not upregulatedin the TBRs or chemosensory bristles (Fig. S4). Ectopic Dll expres-sion in bristle cells, including the sex comb, leads to slightlyreduced bristle size (Fig. 7A). On the other hand, reduction of Dllexpression in males by RNAi during late larval and early pupalstages partially transforms the distal TBRs towards sex comb-likemorphology by increasing their thickness and pigmentation(Fig. 7B). Although Dll-RNAi reduces nuclear expression of Dll toundetectable levels, it does not visibly affect Dac expression at thisstage (Fig. 7D, E).

Reduction of Dll expression using the rn-GAL4 driver also leadsto shorter leg segments, reducing ts1 length by approximately 38%in males and 22% in females (Fig. 7C). Interestingly, it alsosuppresses sex comb rotation without affecting the number ormorphology of the teeth (Fig. 7B). Furthermore, bristles anterior tothe sex comb that migrate proximally in wild-type males (Fig. 1Dand E) remain at the distal tip of TS1 in Dll RNAi flies (Fig. 7B). Allof these phenotypes depend on the time in development when Dllis reduced, with earlier reduction leading to more pronouncedphenotypes (unpublished data). These results provide furtherevidence that both sex comb rotation and the proximal migrationof anterior bristles are caused by local male-specific tissue exten-sion, and suggest that Dll contributes to this process.

Since dsxM is expressed in the epithelium around the develop-ing sex comb and is necessary for sex comb rotation, we testedwhether suppressing the rotation using Dll RNAi affected DsxMexpression. We found, however, that DsxM was strongly expressedin these cells even when the sex comb does not rotate (Fig. 7 F–G),showing that the function of Dll in tissue extension is notmediated by dsx.

Discussion

Starting from the ancestral leg ground-plan, which is conservedin all Drosophila females, conspicuously different phenotypes aregenerated during metamorphosis in sex-comb bearing males.These complex traits require a large number of changes in cellbehavior, but the genetic pathways responsible for these changesare largely unknown (Kopp 2011). Here, we show that localizedrepression of dac, a critical leg patterning gene, by the sexdetermination pathway is necessary for the male-specific devel-opment of the bristles that make up the sex comb.

The number of tarsal segments that bear sex combs varies acrossspecies (Barmina and Kopp, 2007). When viewed from this per-spective, dac plays a positive role in sex comb development (Deviet al., 2013; Docquier et al., 1997; Randsholt and Santamaria, 2008).The overall plan of the developing leg, including segmental identity,is established during the larval stage. Since Dac expression isweaker in the second tarsal segment than in the first (Mardonet al., 1994), it is possible that increasing Dac levels during this earlyperiod transforms ts2 toward ts1 identity, in much the same waythat reduction of bab causes homeotic transformations of ts2, ts3and ts4 to a ts1 state (Couderc et al., 2002; Godt et al., 1993). Thismay occur through dac-mediated repression of bab (Galindo et al.,2002) or through another mechanism.

If the same gene is to act as both an activator and a repressor ofa trait, these functions must be separated in time or space. Wehave found that the downregulation of dac in the sex comb doesnot occur until the sex comb SOPs are specified, long after theproximodistal pattern of the leg has been established. The role ofdac in the fine-scale repression of sex comb teeth in specific cellsin the pupal tarsus is therefore distinct from its earlier larvalfunction in establishing segmental identity throughout the leg.Consistent with this, dac expression in pupal leg bristles does notrequire Dll at this stage, the main regulator of dac in the larval legdisc (Giorgianni and Mann, 2011). While it has long been knownthat both dac and Dll play conserved roles in sensory organdevelopment (reviewed in Angelini and Kaufman, 2005), ourfindings show that dac also has a more specific function indifferentiating between alternative bristle morphologies. In thiscapacity, its function is similar to that of Pox neuro, a gene thatdifferentiates between mechanosensory and chemosensory bristlefates (Nottebohm et al., 1992).

In addition to bristle modification, sex comb development inmany species involves a complex and poorly-understood processof tissue rotation. We have shown that two sexually dimorphictraits, sex comb rotation and the male-specific proximal migrationof the adjacent bristles, both require Dll-dependent extension ofts1, providing further evidence that these traits are coupled at thecellular level. This observation also indicates that Dll, like dac,contributes to sex comb development during the pupal stage inaddition to its earlier and more general role in proximo-distal legpatterning (reviewed in Kojima, 2004).

Changes in dac and Dll expression and function in restricted cellpopulations may not have any negative pleiotropic consequencesif they occur late in development, after the deeply conserved earlyfunctions of these genes in leg patterning and growth have beenfulfilled. This opens the door for evolutionary changes in geneexpression underlying the dazzling diversity of sex comb morphol-ogies observed in nature. More generally, we suggest that co-option of early-acting “master regulators” for new functions atlater developmental stages may be an important mechanismcontributing to evolutionary innovations.

Acknowledgments

We thank Graeme Mardon for a dac4 FRT40A stock, Douglas Allenfor a UAS-dac line and Isabel Guerrero and Konrad Basler for UAS-Dllstocks. Nana Liu, Peter Dennis, Andy Hon, Nancy Napan, SethThompson and Michael Garland provided valuable assistance withthe experiments. This work was funded by NIH Grant R01 GM082843to AK. Part of the research was carried out in the lab of Ellen Larsen atthe University of Toronto, whose help is greatly appreciated. Fly stockswere obtained from the Bloomington Drosophila Stock Center and theUC San Diego Drosophila Species Stock Center, and the DevelopmentalStudies Hybridoma Bank provided antibodies.

Appendix A. Supplementary material

Supplementary data associated with this article can be found inthe online version at http://dx.doi.org/10.1016/j.ydbio.2013.12.017.

References

Angelini, D.R., Kaufman, T.C., 2005. Insect appendages and comparative ontoge-netics. Dev. Biol. 286, 57–77.

Arnoult, L., Su, K.F.Y., Manoel, D., Minervino, C., Magriña, J., Gompel, N., Prud’-homme, B., 2013. Emergence and diversification of fly pigmentation throughevolution of a gene regulatory module. Science 339, 1423–1426.

J. Atallah et al. / Developmental Biology 386 (2014) 440–447446

Atallah, J., Liu, N.H., Dennis, P., Hon, A., Godt, D., Larsen, E.W., 2009a. Cell dynamicsand developmental bias in the ontogeny of a complex sexually dimorphic traitin Drosophila melanogaster. Evol. Dev. 11, 191–204.

Atallah, J., Liu, N.H., Dennis, P., Hon, A., Larsen, E.W., 2009b. Developmentalconstraints and convergent evolution in Drosophila sex comb formation. Evol.Dev. 11, 205–218.

Atallah, J., Watabe, H., Kopp, A., 2012. Many ways to make a novel structure: a newmode of sex comb development in Drosophilidae. Evol. Dev. 14, 476–483.

Barmina, O., Kopp, A., 2007. Sex-specific expression of a HOX gene associated withrapid morphological evolution. Dev. Biol 311, 277–286.

Carroll, S.B., 2008. Evo-devo and an expanding evolutionary synthesis: a genetictheory of morphological evolution. Cell 134, 25–36.

Couderc, J.L., Godt, D., Zollman, S., Chen, J., Li, M., Tiong, S., Cramton, S.E., Sahut-Barnola, I., Laski, F.A., 2002. The bric à brac locus consists of two paralogousgenes encoding BTB/POZ domain proteins and acts as a homeotic andmorphogenetic regulator of imaginal development in Drosophila. Development129, 2419–2433.

Devi, T.R., Amruthavalli, C., Shyamala, B. v., 2013. Evolution of sex comb from theprimitive bristle pattern in Drosophila is associated with modification in thedevelopmental regulatory protein dachshund. Genesis 51, 97–109.

Dietzl, G., Chen, D., Schnorrer, F., Su, K.C., Barinova, Y., Fellner, M., Gasser, B., Kinsey,K., Oppel, S., Scheiblauer, S., Couto, A., Marra, V., Keleman, K., Dickson, B.J.,2007. A genome-wide transgenic RNAi library for conditional gene inactivationin Drosophila. Nature 448, 151–156.

Docquier, F., Randsholt, N.B., Santamaria, P., 1997. Gain of function mutations of thedachshund (dac) gene induce proximo-distal leg transformations in Drosophilamelanogaster. Int. J. Dev. Biol. 41, 13S.

Ferveur, J.F., Stortkuhl, K.F., Stocker, R.F., Greenspan, R.J., 1995. Genetic feminizationof brain structures and changed sexual orientation in male Drosophila. Science267, 902–905.

Galindo, M.I., Bishop, S.A., Greig, S., Couso, J.P., 2002. Leg patterning driven byproximal–distal interactions and EGFR signaling. Science 297, 256–259.

Gehring, W.J., 1996. The master control gene for morphogenesis and evolution ofthe eye. Genes Cells 1, 11–15.

Giorgianni, M.W., Mann, R.S., 2011. Establishment of medial fates along theproximodistal axis of the Drosophila leg through direct activation of dachshundby Distalless. Dev. Cell 20, 455–468.

Godt, D., Couderc, J.L., Cramton, S.E., Laski, F.A., 1993. Pattern formation in the limbsof Drosophila: bric a brac is expressed in both a gradient and a wave-like patternand is required for specification and proper segmentation of the tarsus.Development 119, 799–812.

Gorfinkiel, N., Morata, G., Guerrero, I., 1997. The homeobox gene Distal-less inducesventral appendage development in Drosophila. Genes Dev. 11, 2259–2271.

Held Jr., L.I., 2002. Imaginal discs: the genetic and cellular logic of patternformation, 1st ed. Cambridge University Press, Cambridge, UK.

Held Jr., L.I., 2010. How does Scr cause first legs to deviate from second legs?Drosoph. Inf. Serv. 93, 132–146.

Hempel, L.U., Oliver, B., 2007. Sex-specific DoublesexM expression in subsets ofDrosophila somatic gonad cells. BMC Dev. Biol. 7, 113.

Kojima, T., 2004. The mechanism of Drosophila leg development along theproximodistal axis. Dev. Growth Differ. 46, 115–129.

Kopp, A., 2011. Drosophila sex combs as a model of evolutionary innovations. Evol.Dev. 13, 504–522.

Lee, G., Hall, J.C., Park, J.H., 2002. Doublesex gene expression in the central nervoussystem of Drosophila melanogaster. J. Neurogenet. 16, 229–248.

Mardon, G., Solomon, N.M., Rubin, G.M., 1994. dachshund encodes a nuclear proteinrequired for normal eye and leg development in Drosophila. Development 120,3473–3486.

McGuire, S.E., Le, P.T., Osborn, A.J., Matsumoto, K., Davis, R.L., 2003. Spatiotemporalrescue of memory dysfunction in Drosophila. Science 302, 1765–1768.

McKay, D.J., Estella, C., Mann, R.S., 2009. The origins of the Drosophila leg revealedby the cis-regulatory architecture of the Distalless gene. Development 136,61–71.

Monteiro, A., Podlaha, O., 2009. Wings, horns, and butterfly eyespots: how docomplex traits evolve? PLoS Biol. 7, e1000037.

Nottebohm, E., Dambly-Chaudière, C., Ghysen, A., 1992. Connectivity of chemosen-sory neurons is controlled by the gene poxn in Drosophila. Nature 359, 829–832.

Oda, H., Uemura, T., Harada, Y., Iwai, Y., Takeichi, M., 1994. A Drosophila homolog ofcadherin associated with Armadillo and essential for embryonic cell-celladhesion. Dev. Biol. 165, 716–726.

Orenic, T.V., Held, L.I., Paddock, S.W., Carroll, S.B., 1993. The spatial organization ofepidermal structures: hairy establishes the geometrical pattern of Drosophilaleg bristles by delimiting the domains of achaete expression. Development 118,9–20.

Pavlopoulos, A., Akam, M., 2011. Hox gene Ultrabithorax regulates distinct sets oftarget genes at successive stages of Drosophila haltere morphogenesis. Proc.Natl. Acad. Sci. 108, 2855–2860.

Phelps, C.B., Brand, A.H., 1998. Ectopic gene expression in Drosophila using GAL4system. Methods 14, 367–379.

Pomiankowski, A., Nöthiger, R., Wilkins, A., 2004. The evolution of the Drosophilasex determination pathway. Genetics 166, 1761–1773.

Randsholt, N.B., Santamaria, P., 2008. How Drosophila change their combs: the Hoxgene Sex combs reduced and sex comb variation among Sophophora species.Evol. Dev. 10, 121–133.

Salazar-Ciudad, I., Marín-Riera, M., 2013. Adaptive dynamics under development-based genotype-phenotype maps. Nature 497, 361–364.

Seher, T.D., Ng, C.S., Signor, S.A., Podlaha, O., Barmina, O., Kopp, A., 2012. Geneticbasis of a violation of Dollo’s law: re-evolution of rotating sex combs inDrosophila bipectinata. Genetics 192, 1465–1475.

Shen, W., Mardon, G., 1997. Ectopic eye development in Drosophila induced bydirected dachshund expression. Development 124, 45–52.

Shroff, S., Joshi, M., Orenic, T.V., 2007. Differential Delta expression underlies thediversity of sensory organ patterns among the legs of the Drosophila adult.Mech. Dev. 124, 43–58.

St. Pierre, S.E., Galindo, M.I., Couso, J.P., Thor, S., 2002. Control of Drosophilaimaginal disc development by rotund and roughened eye: differentiallyexpressed transcripts of the same gene encoding functionally distinct zincfinger proteins. Development 129, 1273–1281.

Tanaka, K., Barmina, O., Kopp, A., 2009. Distinct developmental mechanismsunderlie the evolutionary diversification of Drosophila sex combs. Proc. Natl.Acad. Sci. USA 106, 4764–4769.

Tanaka, K., Barmina, O., Sanders, L.E., Arbeitman, M.N., Kopp, A., 2011. Evolution ofsex-specific traits through changes in HOX-dependent doublesex expression.PLoS Biol. 9, e1001131.

Theodosiou, N.A., Xu, T., 1998. Use of FLP/FRT system to study Drosophila develop-ment. Methods 14, 355–365.

Tokunaga, C., 1962. Cell lineage and differentiation on the male foreleg ofDrosophila melanogaster. Dev. Biol. 4, 489–516.

Wagner, G.P., Kenney-Hunt, J.P., Pavlicev, M., Peck, J.R., Waxman, D., Cheverud, J.M.,2008. Pleiotropic scaling of gene effects and the “cost of complexity”. Nature452, 470–472.

Ward, R.E., Reid, P., Bashirullah, A., D’Avino, P.P., Thummel, C.S., 2003. GFP in livinganimals reveals dynamic developmental responses to ecdysone during Droso-phila metamorphosis. Dev. Biol. 256, 389–402.

Werner, T., Koshikawa, S., Williams, T.M., Carroll, S.B., 2010. Generation of a novelwing colour pattern by the Wingless morphogen. Nature 464, 1143–1148.

J. Atallah et al. / Developmental Biology 386 (2014) 440–447 447