DISCUSSION AND CONCLUSIONS (continued)

7. Do Heliconius possess unique developmental genetic mechanisms for generating wing pattern?

It has long been assumed that the diversity of lepidopteran wing patterns must arise from very different developmental mechanisms which in some cases can be seen competing for space on the same wing (e.g., see Figure 7 lower right). The most elegant summary of the early literature in this regard is that of Monroe (1983), beautifully expanded for butterfly patterns by Nijhout (1991). Viewed in the broader context of lepidopteran wing patterns, Heliconius only seem exceptional in the degree to which important aspects of wing pattern, such as key pattern boundaries, appear to be determined early in development by putting scale type specification under the control of regulatory genes which, in turn, respond to compartments or other lineage restriction boundaries.

Koch et al. (1998) and ffrench-Constant and Koch (this volume), working on sexually dimorphic Papilio, suggest one alternative for how genes might control qualitative changes in wing pattern. A single gene, by controlling production of hormones which regulate rate of scale maturation relative to circulating pigment precursors, can change scale pigmentation from yellow to melanic over most of the wing in females. However, sexual mosaics and gynandromorphs in Papilio indicate that if scale precursor cells are genetically male, threshold responses to the same signals are below that required for melanization. I have proposed in this chapter that in Heliconius, such scale cell lineage autonomy can also arise more generally through somatically or germ line heritable changes in the settings of regulatory genes (realizators) in early development, providing another way that a few genes can govern entire pattern field of the wing. This possibility is supported by recent molecular studies of the expression of Drosophila regulatory genes during butterfly wing development as reviewed by Brakefield and Monteiro (this volume). They point out that mutations in such a gene might alter for example, whether the eyespots they study are expressed dorsally and ventrally. In spite of the elegance of this scenario and its recognition of regulatory switches that govern major fields of pattern as described for Heliconius, it cannot fully contend with the wild and discrete variety one must consider in this genus.

Window boundaries in Heliconius (Figure 6) are apparently determined by a primary, non-plastic pattern program similar to that which determines position of eyespot foci that are sources of morphogens. Second-order patterns are created as scale cells mature differently according to local concentrations of such morphogens on pupal wings of Precis (Nijhout 1980 a b, Bard and French 1984). Primary patterns of elements such as eyespot foci are established by early acting regulatory genes (see review in Brakefield and Monteiro this volume). The possibility of homology between Heliconius window boundaries and certain eyespot foci in Precis is supported by the expression pattern of Hedgehog signaling protein on the HW of Precis (Keys et al 1999). In addition to helping define a major eyespot center in a standard nymphaline ground plan species, Hedgehog defines a boundary across the pupal wing which, in relation to wing venation, corresponds closely to the anterior boundary of the HW window of H. cydno. It should now be possible to establish whether window boundaries in Heliconius are essentially like eyespot foci stretched into lines, rather than lineage restriction boundaries established in earlier development as I have proposed.

In Precis, and other nymphaline ground plan species, scales developing in zones around eyespot foci on the pupal wing are developmentally plastic with respect to thresholds of response to morphogen concentrations and thus plastic with respect to abiotic factors such as cold shock (Nijhout 1984). The idea that genes changing the threshold-sensing properties of scale precursor cells towards some diffusing morphogen (ffrench-Constant and Koch this volume) could change a pattern boundary agrees with my hypothesis for how the proximal boundary of the FW shutter (Figure 6 b) is determined in the MCS clade of Heliconius. In contrast to the FW window-wall boundary nearby (Figure 6 e),which apparently is determined in early embryogenesis, the FW shutter's distal boundary shows plasticity with respect to its precise position and sharpness (personal observation). This and other elements such as the white marginal spots (previously mentioned) suggest that in many aspects of pattern, Heliconius differs in degree, but not in kind, from other nymphaline butterflies. Indeed, expression patterns for the Drosophila distal-less gene in Heliconius wing discs are virtually identical to those seen in Precis (S. Carroll, J. Gates, pers. comm., and see review in French and Monteiro 1994).