Phenotypic variation in Heliconius erato crosses shows that iridescent structural colour is sex-linked and controlled by multiple genes

Abstract

Bright, highly reflective iridescent colours can be seen across nature and are produced by the scattering of light from nanostructures. Heliconius butterflies have been widely studied for their diversity and mimicry of wing colour patterns. Despite iridescence evolving multiple times in this genus, little is known about the genetic basis of the colour and the development of the structures which produce it. Heliconius erato can be found across Central and South America, but only races found in western Ecuador and Colombia have developed blue iridescent colour. Here, we use crosses between iridescent and non-iridescent races of H. erato to study phenotypic variation in the resulting F₂ generation. Using measurements of blue colour from photographs, we find that iridescent structural colour is a quantitative trait controlled by multiple genes, with strong evidence for loci on the Z sex chromosome. Iridescence is not linked to the Mendelian colour pattern locus that also segregates in these crosses (controlled by the gene cortex). Small-angle X-ray scattering data show that spacing between longitudinal ridges on the scales, which affects the intensity of the blue reflectance, also varies quantitatively in F₂ crosses.

Keywords: Heliconius; butterflies; evolution; iridescence; quantitative genetics; structural colour.

Figure 1.

Scanning electron microscope image showing the structures on a Heliconius wing scale. Longitudinal ridges, composed of overlapping lamellae, are connected by cross-ribs.

Figure 2.

Cross-design and examples of colour pattern variation in H. erato F₁, F₂ and backcross generations. Examples of the Cr genotypes are shown in the F₂ generation.

Figure 3.

RGB values were measured in the hatched areas highlighted on the right wings and averaged for each butterfly. Left wings were used when the right side were too damaged. SAXS measurements were taken along the dotted line shown on the left forewing. (Online version in colour.)

Figure 4.

Four measurements of forewing band width were taken (bold arrows) along with three further measurements to standardize wing size (dotted arrows), using wing veins as points of reference.

Figure 5.

Representative SAXS patterns for a single frame of a male H. e. cyrbia parent. (a) The 2D pattern reveals approximately perpendicular scattering intensity from scale features. From their orientation, length scales of the scattered intensity and previous interpretations, we infer that they correspond to the spacing between ridges and cross-ribs. (b) Full azimuthal integration of the scattered intensity as a function of the magnitude of the momentum transfer vector q. The peaks corresponding to ridge and cross-rib spacing are indicated together with the measurements in real space.

Figure 6.

Mean BR values across H. erato generations. F₁ and F₂ individuals largely fall between the parental demophoon and cyrbia races. The backcross generation (BC) are highly skewed towards cyrbia, which is the race they were crossed with.

Figure 7.

If there are loci of interest on the Z chromosome, F₁ females with an iridescent cyrbia father will be bluer than those with a non-iridescent demophoon father because they inherit a ‘cyrbia’ Z chromosome. In the F₂, males always inherit a complete, non-recombined Z chromosome from their maternal grandfather, so if he is iridescent they will be bluer than offspring from the reciprocal cross.

Figure 8.

F₁ females with an iridescent cyrbia father were significantly bluer than those with a demophoon father. There were no differences in males.

Figure 9.

Mean BR values for F₂ males with an iridescent maternal grandfather (MGF) were higher than those with an iridescent maternal grandmother, although not significantly. Females in both groups had similar BR values.

Figure 10.

In the F₂ generation, BR values did not differ with the different Cr phenotypes. Cr^dCr^d represents the demophoon genotype with the yellow bar present on the hindwing, and Cr^cCr^c is the cyrbia genotype with the white margin. Cr^dCr^c is heterozygous and has neither of these elements.

Figure 11.

An increase in longitudinal ridge spacing correlated with a decrease in BR values. Blue colour slightly decreased with cross-rib spacing, but ridge spacing and cross-rib spacing were also highly correlated. The cross-hairs show the standard error from the 33 to 133 SAXS point measurements for each individual. Blue lines indicate the fitted linear regression, with the dotted lines showing the 95% confidence interval. (Online version in colour.)

Figure 12.

Variation in longitudinal ridge spacing in the F₂ suggests that it is controlled by multiple genes. In the F₁, those with an iridescent father had lower ridge spacing, reflecting the higher BR values seen in this cross.

Figure 13.

Cross-rib spacing also shows continuous variation in the F₂ generation and extremes extended beyond the values of the few parental individuals which were measured.

Figure 14.

Males have narrower longitudinal ridge spacing than females in the F₂. This difference is less pronounced and not significant in the parental races. Significant differences in cross-rib spacing were seen in cyrbia and in the F₂, with males again having narrower spacing. These results are consistent with the finding that males have higher measures of blue colour.