Desiccation resistance and pigmentation variation reflects bioclimatic differences in the Drosophila americana species complex

Abstract

Background: Disentangling the selective factors shaping adaptive trait variation is an important but challenging task. Many studies-especially in Drosophila-have documented trait variation along latitudinal or altitudinal clines, but frequently lack resolution about specific environmental gradients that could be causal selective agents, and often do not investigate covariation between traits simultaneously. Here we examined variation in multiple macroecological factors across geographic space and their associations with variation in three physiological traits (desiccation resistance, UV resistance, and pigmentation) at both population and species scales, to address the role of abiotic environment in shaping trait variation.

Results: Using environmental data from collection locations of three North American Drosophila species-D. americana americana, D. americana texana and D. novamexicana-we identified two primary axes of macroecological variation; these differentiated species habitats and were strongly loaded for precipitation and moisture variables. In nine focal populations (three per species) assayed for each trait, we detected significant species-level variation for both desiccation resistance and pigmentation, but not for UV resistance. Species-level trait variation was consistent with differential natural selection imposed by variation in habitat water availability, although patterns of variation differed between desiccation resistance and pigmentation, and we found little evidence for pleiotropy between traits.

Conclusions: Our multi-faceted approach enabled us to identify potential agents of natural selection and examine how they might influence the evolution of multiple traits at different evolutionary scales. Our findings highlight that environmental factors influence functional trait variation in ways that can be complex, and point to the importance of studies that examine these relationships at both population- and species-levels.

Fig. 1

Distribution map of collection locations and major environmental variables across North America for our three focal species. Panel a shows collection records for each of D. novamexicana (green), D. a. americana (blue), and D. a. texana (red) as obtained from the TaxoDros database (taxodros.uzh.ch, see methods). Closed circles indicate the nine sample locations for populations used in this study. Panel b-d show heatmaps of spatial variation in annual mean temperature (b), annual mean radiation (c), and annual mean moisture index (d), as obtained from the Worldclim and Climond databases. Cuticle dissection images inset into Panel A are representative male cuticles from (left to right) Grand Junction, Colorado, White River, Arkansas, and New Orleans, Louisiana

Fig. 2

PC1 and PC2 values for all collection records (N = 149). PC axes are inverted to mirror spatial orientation of populations in geographic space. Filled circles indicate the 9 focal populations used to assess trait variation. Across all collection records, species differ along PC1 (F(2, 82.2, P < 0.001) and PC2 (F(2, 59.76), P < 0.001). Post-hoc Tukey tests indicated that all 3 species differed from one another for both PC1 and PC2 (all P values < 0.001)

Fig. 3

Distribution of individual values for desiccation resistance (top, circles) and pigmentation (bottom, diamonds), according to species (X axis) and population within species (shade variation among points). The y-axis values for individual cuticles on the pigmentation graph correspond to a computed greyscale value represented by the gradient bar. The thick bars indicate species means; thin bars indicate 95% confidence intervals around the mean. Desiccation resistance differs between species (F(2, 29.76); P < 0.0001) and populations within species (F(6, 4.48); P = 0.0004). Post-hoc tests indicate significant differences in all pairwise contrasts (D.nov – D. am: P < 0.0001, D.nov – D. tex: P = 0.00039; D.am – D. tex: P = 0.0023). Pigmentation differs between both species (F(2, 11.86), P < 0.0001) and populations within species (F(6, 3.13), P = 0.0083). Post-hoc contrasts indicate D. novamexicana is significantly lighter than D. a. americana (P < 0.0001) and D. a. texana (P < 0.0001); D. a. americana and D. a. texana do not differ (P = 0.96)

Fig. 4

Relationship between mean desiccation resistance (circles) or pigmentation (diamonds) in each population and PC1 (top panel) or PC2 (bottom panel). The left-side y-axis corresponds to desiccation resistance in minutes survived, while the right-side y-axis show pigmentation values on that correspond to a greyscale value represented by the gradient bar. PC1 is not significantly associated with desiccation resistance (r(7) = 0.15; P = 0.71) or pigmentation (r(7) = 0.53; P = 0.15) was across all 9 populations (top, grey trend lines). PC2 is associated with both desiccation resistance (r(7) = 0.74; P = 0.022) and pigmentation (r(7) = 0.68; P = 0.044) (bottom, black trend lines), prior to multiple testing correction