Emergence and Diversification of Fly Pigmentation Through Evolution of a Gene Regulatory Module

See allHide authors and affiliations

Science  22 Mar 2013:
Vol. 339, Issue 6126, pp. 1423-1426
DOI: 10.1126/science.1233749

Seeing Spots

Some flies in the Drosophila melanogaster lineage exhibit wing spots, which vary among species. Examining the underlying genetics of spot determination, Arnoult et al. (p. 1423) provide evidence for a two-step scenario for the origin and diversification of patterning novelty in these fly wings. The findings suggest that the two-step model may generally apply to the emergence and diversification of traits in plants and animals.


The typical pattern of morphological evolution associated with the radiation of a group of related species is the emergence of a novel trait and its subsequent diversification. Yet the genetic mechanisms associated with these two evolutionary steps are poorly characterized. Here, we show that a spot of dark pigment on fly wings emerged from the assembly of a novel gene regulatory module in which a set of pigmentation genes evolved to respond to a common transcriptional regulator determining their spatial distribution. The primitive wing spot pattern subsequently diversified through changes in the expression pattern of this regulator. These results suggest that the genetic changes underlying the emergence and diversification of wing pigmentation patterns are partitioned within genetic networks.

Whether emergence and diversification of morphological traits both occur through changes at the same tier of a gene regulatory network is currently an open question in evolutionary biology. We addressed this question by studying a wing pigmentation pattern that evolved recently in the Drosophila melanogaster group. Although this male-specific wing spot assumes different shapes, colors, and intensities among species (Fig. 1 and fig. S1), phylogenetic reconstruction indicates that it emerged only once in this lineage and subsequently diversified (1).

Fig. 1

A spot of black pigmentation (left) on male Drosophila wings assumes various shapes among species. These variations are precisely prefigured by the distribution of the Yellow protein (right) in the wing of young adults. The phylogenetic relationships of these species are indicated to the left. Arrowheads highlight subtle variations between closely related species.

We examined the expression pattern of yellow (y), a gene required for the production of black pigment patterns, including the wing spot (2). Yellow distribution in the wing of pupae or young adults precisely foreshadows the adult pigmentation pattern in the various species we studied (Fig. 1), suggesting that understanding the evolution of y expression patterns would illuminate how the wing spot appeared and diversified. The wing spot expression pattern of y in D. biarmipes, a wing-spotted species, is encoded by an evolutionarily derived cis-regulatory element (CRE) of the yellow locus (2), the spot CRE [hereafter referred to as y spotbia675 (3)]. Therefore, to decipher how the wing spot expression of y is spatially regulated, we sought to identify the transcription factors controlling the localized activity of the y spotbia196 CRE, a minimal version of y spotbia675 CRE with a similar activity (2).

The y spotbia196 CRE is functional in D. melanogaster wing, and we used its transcriptional activity as a readout in an RNA interference (RNAi) screen targeting selected genes to identify its transcriptional activators. With this functional screen of the ~350 transcription factors expressed in a D. melanogaster late pupal wing (table S1), we identified a handful of candidates (fig. S2). Among five candidate activators, Distal-less (Dll) first caught our attention because of its well-defined role in wing patterning. When Dll is down-regulated, the activity of the y spotbia196 CRE is severely affected (Fig. 2, A and B, and fig. S2C). Reciprocally, we overexpressed Dll throughout the wing, which resulted in an increase and an expansion of the y spotbia196 activity (Fig. 2C). These genetic manipulations reveal that Dll is both necessary and sufficient to control the activity of the y spotbia196 CRE in a D. melanogaster context.

Fig. 2

In a D. melanogaster pupal wing, the activity of a spotbia196-reporter construct (A) is severely impaired when the transcription factor Dll is depleted (B) and enhanced when Dll is overexpressed (C). The effect mediated by Dll down-regulation (B) is mirrored by the mutation of Dll binding sites in the spotbia196 CRE (D). The diagrams represent the spotbia196 CRE and the localization of the Dll binding sites (gray boxes) and the mutated sites (crossed boxes).

By scanning the spotbia196 CRE, we identified several putative Dll binding sites. We tested the ability of Dll to bind these sites in vitro (fig. S3, A and B) and, thus, discovered that Dll can physically interact with four of them. Mutations of these sites preventing Dll binding in vitro also impaired the activity of the y spotbia196 CRE in vivo (Fig. 2D and fig. S3C). Of note, the same mutations also abolish the y spotbia196 response to Dll overexpression (fig. S3D). Together, these results indicate that the y spotbia196 CRE contains multiple binding sites for Dll and suggest that the evolution of a regulatory link between Dll and yellow was essential for the emergence of the wing spot.

The results obtained in D. melanogaster stimulated the investigation of the role of Distal-less in wing pigmentation pattern formation in D. biarmipes. We established a loss-of-function assay based on the expression of artificial short hairpin microRNA (shRNA) (4) and validated the system in D. biarmipes by silencing yellow in the pupal wing (fig. S5, A and C), which phenocopies a y mutant. The down-regulation of Dll, using the same shRNA approach, resulted in a dramatic reduction of the pigmentation spot (Fig. 3, A and B, and fig. S4, A to D); however, the uniform gray shading of the wing was not affected. Reciprocally, the overexpression of Dll throughout the wing yielded ectopic pigmentation across the wing (Fig. 3, A and C, and fig. S4, E to G and J). Together, these results establish that Dll plays an essential role in wing spot formation in D. biarmipes.

Fig. 3

Compared with the wild type (A), Dll down-regulation reduces the intensity of the pigmentation spot in D. biarmipes (B), whereas, conversely, the overexpression of Dll results in ectopic pigmentation across the wing (C). These effects are reflected in the activity of the yellow spotbia675-reporter construct (D to F), which is reduced when Dll is knocked down and expanded when Dll is overexpressed. They are also reflected in the expression of ebony, in a reciprocal manner (G to I). Note that the activity of spotbia675 in a WT D. biarmipes context (D) precisely prefigures the native Yellow distribution and the shape of the pigmentation spot. Arrowheads in (G) to (I) indicate the boundaries of ebony distribution in the presumptive spot area in the wild type.

To test the specificity of the effect of Dll on wing spot pattern formation in D. biarmipes, we first examined the other transcription factors identified in our RNAi screen. Contrary to Dll, none of the candidates that we overexpressed across the wing were sufficient to induce ectopic pigmentation (fig. S4, H to M). These results highlight the specific role of Dll in the wing pigmentation spot formation. Second, we ectopically expressed Dll in the wing of the nonspotted species D. ananassae, a species with an ancestral pattern predating the wing spot evolution (Fig. 1) (1). In this species, Dll overexpression had no effect on wing pigmentation (fig. S6G), suggesting that the regulatory interactions between Dll and downstream pigmentation genes are absent. Consistent with this, y spotbia675 CRE activity (Fig. 3, D to F), y mRNA level (fig. S6), and Yellow distribution (fig. S5, B, E, and F) all respond to Dll expression changes in a D. biarmipes wing, whereas, in contrast, all remain unaffected when Dll is overexpressed in D. ananassae (fig. S6). Together, these results indicate that the regulatory link between Dll and yellow in the wing probably evolved in the common ancestor of wing-spotted species, after the divergence with the D. ananassae lineage.

The evolutionary co-option of Dll by yellow in the wing was presumably an essential step in the emergence of the wing spot, but it was certainly not solely sufficient for the pigmentation pattern to appear. The wing spot of the D. biarmipes yellow mutant shifts from black to light beige, but its shape remains unaffected (fig. S5C), indicating that at least one additional pigmentation gene is involved. This is consistent with the view that black pigment deposits result from multiple enzymatic activities (2, 5). We therefore hypothesized that during evolution, Dll became a key regulator controlling the shape of the wing pigmentation spot by coordinating the spatial deployment of several enzymatic activities. Overexpression of Dll throughout the wing of a D. biarmipes yellow mutant resulted in the expansion of the light beige color (fig. S5D), suggesting that Dll is governing the spatial distribution of multiple pigmentation genes. Hence, we tested whether Dll regulates ebony, a pigmentation gene whose expression pattern correlates with the wing spot shape in D. biarmipes (2). In wild-type (WT) males, ebony expression is repressed in the presumptive wing spot area (Fig. 3G). When Dll expression is down-regulated, ebony is no longer repressed in the spot area (Fig. 3H), whereas reciprocally, Dll overexpression reduces the ebony expression level across the wing (Fig. 3I). These results show that Dll controls the spatial distribution of ebony in the wing. More generally, our findings suggest that the wing spot emergence was brought on by the evolution of regulatory links between Dll and multiple pigmentation genes.

The wing spot diversified in shape and intensity among species of the D. melanogaster lineage (fig. S1), and Yellow distribution mirrors this diversification (Fig. 1). We therefore asked how yellow expression has changed among wing-spotted species. We noticed that the y spotbia675 CRE precisely recapitulates the native y distribution in D. biarmipes (Fig. 3D), but not in D. melanogaster (Fig. 2A) (2). This difference indicates that some trans-acting factors have likely diverged between the two species. As Dll is an activator of the y spotbia675 CRE, we compared its expression pattern between D. melanogaster and D. biarmipes. Dll is expressed in a gradient pattern in the wing blade around the margin during early pupal development in D. biarmipes (fig. S7, A and B), suggesting a conserved role in early wing development (6). However, ~30 hours after puparium formation, when the wing patterning is complete, Dll expression diverges between the two species. Although in D. melanogaster Dll remains expressed in a gradient from the wing blade (fig. S7, K and L), in D. biarmipes a distinct expression pattern element prefiguring the shape of the pigmentation spot becomes visible (Fig. 4 and fig. S7, C to G). Therefore, y spotbia675 CRE activity follows the spatial distribution of Dll in both D. melanogaster and D. biarmipes, reinforcing that Dll dictates the spatial distribution of yellow in the wing. Importantly, the spatially restricted expression pattern of Dll in D. biarmipes contrasts with the uniform distribution of other candidate transcription factors identified in our screen (fig. S7, I and J). These observations support the notion that Dll governs the spatial regulation of pigmentation genes involved in the wing spot formation.

Fig. 4

(A) Differences in Dll distribution among species correlate with the shape of the spot (arrowheads indicate subtle spatial differences in the pigmentation patterns between D. biarmipes, D. elegans, and D. pulchrella prefigured by Dll distribution divergence). (B) A regulatory model of the emergence and diversification of the wing spot. The evolutionary transition from unspotted [(1), similar to D. ananassae] to spotted species [(2), similar to D. rhopaloa] required the assembly of a new genetic module. Subsequently, the spatial distribution of this module diversified through changes in Dll regulation, thereby creating different wing pigmentation patterns (3). The blue color indicates the evolution of new regulatory links in (2) and changes in Dll regulation in (3). Wings in (A) are at ~33% of pupal development. y, yellow; e, ebony.

As Dll pupal wing expression has changed between D. melanogaster and D. biarmipes and because Dll controls the spatial distribution of the pigmentation genes necessary to produce black pigment, we asked whether the diversification of Dll expression could account for the various wing pigmentation patterns among wing-spotted species in the D. melanogaster lineage. We found that, in all cases, the expression pattern of Dll predicts Yellow distribution and the adult pigmentation pattern at the wing tip (Fig. 4A and fig. S7H). These results suggest that the divergence of Dll wing expression pattern drives the diversification of wing pigmentation patterns in the D. melanogaster lineage.

On the basis of these results, we propose that the wing spot emerged through the independent and likely progressive recruitment by several pigmentation genes of multiple transcriptional regulators (2, 7), including Dll as a common spatial coordinator (Fig. 4B), probably through the evolution of binding sites for Dll in cis-regulatory sequences of these genes. These new regulatory links between Dll and the pigmentation genes created a novel genetic module sufficient for the formation of a primitive wing spot pattern. Subsequently, this module was deployed in various spatial patterns in the wings of different species through changes in Dll distribution, thereby creating diverse wing pigmentation patterns.

This work illustrates how a transcription factor governing several effector genes to form a morphological trait has become a privileged genetic target to modify the spatial pattern of this trait during evolution (8). A similar logic may apply generally to the numerous morphological patterns that vary spatially between related species (812): The emergence of these morphological novelties may result from the evolutionary modifications of effector genes, whereas their spatial diversification involves the redeployment of upstream patterning genes.

Supplementary Materials

Materials and Methods

Figs. S1 to S8

Tables S1 and S2

References (1329)

References and Notes

  1. See the materials and methods and other supplementary materials on Science Online.
  2. Acknowledgments: We thank A. Kopp, S. B. Carroll, J. David, R. Mann, A. Mazo, M. Yamaguchi, and the Vienna Drosophila RNAi Center (VDRC, Vienna), Bloomington (Indiana), National Center of Genetics (Kyoto), and San Diego Drosophila Stock Centers for flies and antibodies; A. Polette and L. Mille for developing a custom quantification program for wing pigmentation; J.-L. Mari for advice on programming; G. Storelli for help with cloning; A. Defaye for qPCR guidance; J.-Y. Sgro for microarray analysis; A. Kopp, F. Leulier, and J. Cande for helpful comments on the manuscript; Flybase for information support; and Le Calendal for logistical support. L.A. and K.F.Y.S. were supported by fellowships from the French government (Ministère de la Recherche de l’Enseignement Supérieur), D.M. by the Portuguese Foundation for Science and Technology (Fundação para a Ciência e a Tecnologia) and the GABBA graduate program, and K.F.Y.S. by the Fondation ARC pour la Recherche sur le Cancer. This work was funded by a EURYI award, a Human Frontier Science Program Career Development Award, the Agence Nationale de la Recherche, the France-Berkeley Fund, the Fondation Schlumberger pour l’Enseignement et la Recherche, and the CNRS. RNAi lines from the VDRC are covered by a Materiels Transfer Agreement. The microarray data set is accessible in the National Center for Biotechnology Information Gene Expression Omnibus database (accession no. GSE43673). Author contributions: B.P. and N.G. conceived the project; all authors designed the experiments; K.F.Y.S. and D.M. performed the RNAi screen; K.F.Y.S., D.M., and L.A. did the antibody stainings; K.F.Y.S., L.A., D.M., C.M., and B.P. cloned the molecular constructs; and D.M. and B.P. ran the biochemistry experiments. L.A. designed and carried out the shRNA experiments in D. biarmipes, did the qPCR analysis, optimized transgenesis, and developed the wing-peeling technique. L.A. and N.G. quantified the effects on pigmentation; K.F.Y.S., L.A., D.M., J.M., N.G., and B.P. made transgenic flies; L.A., N.G., and K.F.Y.S. took the images; all authors analyzed the data; and B.P. and N.G. wrote the paper, with feedback from L.A. and the other authors.
View Abstract

Navigate This Article