L.E. Gilbert - Heliconius Wing Pattern Evolution

Adaptive novelty in Heliconius.....by L.E. Gilbert (continued)

METHODS, MODELS, METAPHORICAL AIDS

Scale vs. Pattern-level control in development

Heliconius wing as a computer screen

To explain my working hypothesis for the genetic control of wing pattern in the MCS clade of Heliconius, I find it convenient to envision the wing as a computer screen. The following scheme of how genes determine scale phenotypes is based on Gilbert et al. (1988). If each scale is analogous to a pixel, then in the case of Heliconius, there are three basic pixel types

The placement of these three pixels on the Heliconius "wing monitor screen" would be approximated most simply (and correspond best to the B&W plates in this chapter) if we imagine turning off color on the monitor and simply viewing the image formed by white (Type 1), black (type II), and gray (type III) pixels. These three scale types are the essential subunits at the disposal of the pattern-forming mechanism in Heliconius: i.e., those genes that determine the two-dimensional array of three pixel types on the Heliconius screen. Whether wing pixel type I is white or yellow, pixel type II dull black or reflective, or pixel type III brown or red is genetically trivial as far as the mechanism which forms spatial pattern is concerned; just as turning a monitor's color control to black and white mode does not affect the essential pattern viewed on the monitor screen. In the case of Heliconius, the three major scale types are morphologically distinct at the ultrastructural level (Gilbert et al. 1988) and resemble photoreceptor subunits of Drosophila ommatidia in that structure correlates with associated pigments (Tomlinson and Ready 1986).

Unlike pixels on a computer screen, scales of a Heliconius wing develop from undifferentiated trichogen cells (scale precursors). If one selects a coordinate point on the wing that bears either type I, II, or III scales and observes the fate of scale phenotypes in that site over a series of crosses, it appears that "realizator genes" (sensu Garcia-Bellido 1977) for scale type generally interact in the following order of dominance or epistasis at the scale-level: III>II>I. Type I scales do not require pigments to mature and stiffen and are thus considered a default condition. If realizator gene M switches on, then the scale may develop into melanic (II) or xanthommatin (III). Type III scales occur if realizator gene X switches on in a trichogen cell with M also in the "on" position. (Gilbert et al. 1988, used the term "selector gene" which is here changed to "realizator" for compatibility with Garcia-Bellido's model). These binary rules account for most of what we observe in crosses within and between species in the MCS clade. There are exceptions, with some cases of melanic (type II) regions being dominant/epistatic to xanthommatin (type III) regions (e.g., see "forceps" shutter below). Furthermore, on the ventral hind wing only, regions in which scales should be I/II heterozygotes exhibit a strange structural modification that increases reflectance and allows the position of the recessive white/yellow pattern to be seen as a reflectance shift in otherwise melanic scales, a useful diagnostic for heterozygous genotypes.

If one assumes 1) that the realizator genes specifying the three scale types in the Heliconius model act by altering the rate of scale development relative to the timing of pigment deposition and 2) that pigment type determines scale morphology, then the epistatic relationships mentioned above are consistent with developmental data on similarly pigmented scales in other butterflies (e.g., Koch 1998). Nevertheless, realizator genes are relatively uninteresting unless a class of regulatory genes that Garcia-Bellido (1977) termed "selector genes" specify the regions of realizator gene activities across the wing in order to create a pattern. (Metaphorically I think of selector genes as the computer program that chooses among folders containing bitmap files of pre-scanned images to be viewed in a particular window on the screen as array of pixels [ = scales on the wings of Heliconius]). Initially I thought that these metaphorical windows on Heliconius wings might correspond to developmental "compartments" as described or hypothesized in Drosophila (e.g., Lawrence and Morata 1976, Kaufmann et al. 1978, Kaufmann 1981). Compartment theory suggested that cell surface properties, determined by the "on" or "off" states of a hierarchy of selector genes, help organize polyclones of cells that come to occupy discrete, sharply-bounded zones or compartments.

Although aspects of compartment theory are now understood (e.g,. Blair and Ralston, 1997), work over the last 15 years has demonstrated other mechanisms that may establish regions of the Drosophila wing differing in regulatory gene expression (e.g., O'Brochta and Bryant 1985), For simplicity, I will occasionally refer to such regions of the Heliconius wing as "compartments" with the reader's understanding that the term is not used in its strict drosophilian sense even though the strict sense definitions may actually apply to some parts of Heliconius wings. Obviously, extending the simplistic computer screen metaphor to spatial patterns of scales becomes more complex if a variety of lineage restriction mechanisms are at work and if zones of the wing surface are reserved for more ancestral nymphaline developmental schemes. However, simplistic models can be useful for guiding research even when we suspect they are largely wrong.

BACK TO TABLE OF CONTENTS