In moth pheromone communication signals, both quantitative and qualitative intraspecific differences have been found across geographic regions. Such variation has generally been hypothesized to be due to selection, but evidence of genetic control of these differences is largely lacking. To explore the patterns of variation in pheromone signals, we quantified variation in the female sex pheromone blend and male responses of two closely related noctuid moth species in five different geographic regions for 2-3 consecutive years. We found significant variation in the ratios of sex pheromone blend components as well as in male response, not only between geographic regions but also within a region between consecutive years. The temporal variation was of a similar magnitude as the geographic variation. As far as we know, this is the first study reporting such temporal variation in moth chemical communication systems. The geographic variation seems to at least partly be controlled by genetic factors, and to be correlated with the quality of the local chemical environment. However, the pattern of temporal variation within populations suggests that optimization of the pheromonal signal also may be driven by within-generation physiological adjustments by the moths in response to their experience of the local chemical environment. Kirkpatrick and Ravigne 2002). However, the evolution of behavioral isolation is poorly understood (Coyne and Orr 2004). To gain insight into the evolution of prezygotic behavioral isolation it is essential to quantify intraspecific variation in the premating signals, on which selection may operate, and biotic and abiotic factors that contribute to this variation.Premating signals are generally hypothesized to be under stabilizing selection

Evolutionary diversification of sexual communication systems in moths is perplexing because signal and response are under stabilizing selection in many species, and this is expected to constrain evolutionary change. In the moth Heliothis virescens, we consistently found high phenotypic variability in the female sex pheromone blend within each of four geographically distant populations. Here, we assess the heritability, genetic basis and behavioural consequences of this variation. Artificial selection with field-collected moths dramatically increased the relative amount of the saturated compound 16:Ald and decreased its unsaturated counterpart Z11-16:Ald, the major sex pheromone component (high line). In a cross between the high-and low-selected lines, one quantitative trait locus (QTL) explained 11-21% of the phenotypic variance in the 16:Ald/Z11-16:Ald ratio. Because changes in activity of desaturase enzymes could affect this ratio, we measured their expression levels in pheromone glands and mapped desaturase genes onto our linkage map. A delta-11-desaturase had lower expression in females producing less Z11-16:Ald; however, this gene mapped to a different chromosome than the QTL. A model in which the QTL is a trans-acting repressor of delta-11 desaturase expression explains many features of the data. Selection favouring heterozygotes which produce more unsaturated components could maintain a polymorphism at this locus.

The two moth species Heliothis virescens (Hv) and H. subflexa (Hs) are closely related, but have vastly different feeding habits. Hv is a generalist and an important pest in many crops in the USA, while Hs is a specialist feeding only on plants in the genus Physalis. In this study, we conducted a comparative population genetic analysis to assess whether and how generalist and specialist life styles are reflected in differences in population structures. In Hv 98% of the total variation occurred within populations. The overall differentiation (F(ST) ) between regions was 0.006 and even lower between years (0.0039) and hosts (0.0028). Analyses of population structure suggest that all individuals form one genetically homogeneous population, except for at most 12 individuals (6%) that diverged from this cluster. Population homogeneity likely results from the high mobility of Hv and its generalist feeding behaviour. Hs exhibited substantially more population structure. Even though 96% of the total variation was attributable to within-population variability, F(ST) -values between Hs populations were 10 times higher than between Hv populations. Hs populations showed significant isolation by distance. Analyses of Hs population structure suggest at least two subpopulations and thus some degree of metapopulation structure. We speculate that the patchy distribution of Physalis- the exclusive food source of Hs - contributes to differences in population structure between these closely related species. The finding that the specialist shows more population differentiation than the generalist corroborates the notion that host specialization is not an evolutionary dead end but a dynamic trait.

Even though premating isolation is hypothesized to be a major driving force in speciation, its genetic basis is poorly known. In the noctuid moth Heliothis subflexa, one group of sex pheromone components, the acetates, emitted by the female, plays a crucial isolating role in preventing interspecific matings to males of the closely related Heliothis virescens, in which females do not produce acetates and males are repelled by them. We previously found intraspecific variation in acetates in H. subflexa: females in eastern North America contain significantly more acetates than females in Western Mexico. Here we describe the persistence of this intraspecific variation in laboratory-reared strains and the identification of one major quantitative trait locus (QTL), explaining 40% of the variance in acetate amounts. We homologized this intraspecific QTL to our previously identified interspecific QTL using restriction-associated DNA (RAD) tags. We found that a major intraspecific QTL overlaps with one of the two major interspecific QTL. To identify candidate genes underlying the acetate variation, we investigated a number of gene families with known or suspected acetyl- or acyltransferase activity. The most likely candidate genes did not map to our QTL, so that we currently hypothesize that a transcription factor underlies this QTL. Finding a single, large QTL that impacts variation in pheromone blends between and within species is, to our knowledge, the first such example for traits that have been demonstrated to affect premating isolation.

No abstract