Evolving spots, again and again

a–c, The wing spots on male flies of the Drosophila genus. Drosophila tristis (a) and D. elegans (b) have wing spots that have arisen during convergent evolution. Drosophila gunungcola (c) instead evolved from a spotted ancestor. d, Males wave their wings to display the spots during elaborate courtship dances.

It's all about style. When you're out and about looking for mates, what tends to draw the eye first are general signals—health and vigor, symmetry, absence of blemishes or injuries, that sort of thing—but then we also look for that special something, that je ne sais quoi, that dash of character and fashionable uniqueness. In humans, we see the pursuit of that elusive element in shifting fashions: hairstyles, clothing, and makeup change season by season in our efforts to stand out and catch the eye in subtle ways that do not distract from the more important signals of beauty and health.

Flies do the same thing, exhibiting genetic traits that draw the attention of the opposite sex, and while nowhere near as flighty as the foibles of human fashion, they do exhibit considerable variability. Changes in body pigmentation, courtship rituals, and pheromones are all affected by sexual selection, but one odd feature in particular is the presence of spots on the wing. Flies flash and vibrate their wings at prospective mates, so the presence or absence of wing spots can be a distinctive species-specific element in their evolution. One curious thing is that wing spots seem to be easy to lose and gain in a fly lineage, and species independently generate very similar pigment spots. What is it about these patterns that makes them simultaneously labile and frequently re-expressed?

The distribution of that wing spot is complicated, as the cladogram below shows. The data suggest that the last common ancestor of all those flies in the genus Drosophila was spotless, and that spots evolved independently in the melanogaster and obscura groups, and that spots were independently lost at least five times in the melanogaster group. Off, on, on, off—it's like they've been doing a slow-motion irregular blinking over the last few million years. What predisposes these spots to appear and reappear?

For each species, the presence or absence of a wing spot is indicated with a solid black or a solid white circle, respectively. For key nodes, the black portion of each pie chart and the percentages shown next to them indicate the posterior probability that the ancestor was spotted. Dotted branches indicate inferred losses of the wing spot. Branch lengths correspond to absolute time estimates. Numbers in parenthesis identify nodes discussed in the text. Species examined in this study are underlined.

I've summarized the general principles of regulatory logic in this system before—you can get a refresher course in the basics at that link. The short, short summary is that there is a matrix of positional information in the wing, invisible to our eyes, that consists of an array of transcription factors: proteins that bind to genes in the nucleus and switch other genes off and on. Each gene has things called cis regulatory elements, or CREs, which are sites associate with a gene that can be bound by transcription factors, and which regulate whether the gene is off or on.

Think of a CRE as something like a lock for a gene, and transcription factors as the keys. Different keys are found in different regions of the animal, and a gene will be active in that region if one of the keys happens to fit its CRE lock. There are two ways evolution can change the animal so that a particular gene is switched on: 1) the key (transcription factor) can change so that it now fits a particular lock (CRE). This is difficult, though, because a transcription factors affect many more than just one gene: modifying the key changes its effect on many different locks. 2) the lock can change so that a key available in a particular region opens it. This is much more discrete, in that the change only affects one gene, rather than many.

Prud'homme et al. have identified the gene responsible for the pigment spot in the fly wing—it's called yellow—and have also examined in detail the CREs of the yellow gene in multiple species.

For instance, D. melanogaster has a wing spot, while D. gunungcola does not. There are quite a few nucleotide differences in the yellow CRE between these two species, but by taking them apart piece by piece, they've narrowed the critical differences to between 2 and 7 nucleotides in a region that's a bit more than 700 base pairs long. What that means is that they've taken apart the lock and found the few tumblers that are different, that make all the difference in whether the key will open the lock, or not. They can take the D. melanogaster CRE, and put those few D. gunungcola nucleotides in it, and poof, the wing spot goes away. Similarly, they can take the D. melanogaster nucleotides and put them in D. gunungcola CRE, and presto, they get a D. gunungcola with wing spots. It takes very few changes to make the spot switch off or on—the point being that evolution of this feature is easy.

As evolution would predict, though, species that diverged more recently have greater similarities in their regulatory structure than more distantly related species. A comparison between spotted D. melanogaster and spotted D. tristis (look in the cladogram above; they're in different groups, diverged over 30 million years ago, and acquired their spots independently) reveals that they have completely different CREs regulating yellow expression. The CRE controlling wing spot expression in D. melanogaster is 5' to the yellow gene, but in D. tristis it's actually imbedded in an intron inside the yellow gene. In this case, comparisons with another species in the obscura group, D. guanche, shows that this intronic CRE is associated with expression of pigment in the wing veins; D. tristis seems to have co-opted a regulatory element used in one context and re-used it to generate the wing spot.

The evolutionary history of the gain and loss of CREs in these flies is diagrammed below.

a, b, In the D. tristis male pupal wing (a), Yellow expression prefigures the pigmentation pattern (b). c, d, The D. tristis y wing large orthologue (c) drives a uniform low level of reporter expression in the pupal wing, whereas the D. tristis y intron (d) drives the novel expression pattern in the spot region as well as the wing veins. e, y spot expression patterns evolved twice by the co-option and modification of two different pre-existing yellow CREs (symbolized by yellow → orange circles and pink → green circles) and were lost by the repeated inactivation of one CRE.

What it all suggests is that one of the major sources of evolutionary innovation may be changes in cis regulatory regions. Especially in genes that are pleiotropic (that have multiple effects, and as suggested by this work, have multiple CREs that are active under different conditions), tweaking a few nucleotides in a regulatory switch can generate expression of genes in novel domains, without causing major disruptions in the global organization of the organism. It also suggests that the existing framework of transcriptional factors represents a constraint on evolution and development—that it is easier to modify gene expression within an existing domain than to generate an entirely new domain.

Prud'homme B, Gompel N, Roka A, Kassner VA, Williams TM, Yeh S-D, True JR, Carroll SB (2006) Repeated morphological evolution through cis-regulatory changes in a pleiotropic gene. Nature 440:1050-1053.

Wray GA (2006) Spot on (and off). Nature 440:1001-1002.