bric à brac controls sex pheromone choice by male European corn borer moths

Abstract

The sex pheromone system of ~160,000 moth species acts as a powerful form of assortative mating whereby females attract conspecific males with a species-specific blend of volatile compounds. Understanding how female pheromone production and male preference coevolve to produce this diversity requires knowledge of the genes underlying change in both traits. In the European corn borer moth, pheromone blend variation is controlled by two alleles of an autosomal fatty-acyl reductase gene expressed in the female pheromone gland (pgFAR). Here we show that asymmetric male preference is controlled by cis-acting variation in a sex-linked transcription factor expressed in the developing male antenna, bric à brac (bab). A genome-wide association study of preference using pheromone-trapped males implicates variation in the 293 kb bab intron 1, rather than the coding sequence. Linkage disequilibrium between bab intron 1 and pgFAR further validates bab as the preference locus, and demonstrates that the two genes interact to contribute to assortative mating. Thus, lack of physical linkage is not a constraint for coevolutionary divergence of female pheromone production and male behavioral response genes, in contrast to what is often predicted by evolutionary theory.

Fig. 1. Tissue-specific and cell-specific gene expression.

a Heat map showing tissue-specific, sex-specific, and strain-specific mean expression of Bap18, LIM, BgiB ago, and bab as determined by RT-qPCR. Red signifies higher expression; pp prepupal stage; d days. ANOVAs per candidate gene, and separately for each strain, yielded significant variation among all tissue by life-stage combinations, after a Benjamini–Hochberg multiple-test correction. Letters indicate significant differences in two-sided Tukey’s HSD post-hoc tests (P < 0.05). Z strain: Bap18 F = 44.83, df = 20, P = 4 × 10⁻¹⁶; BgiB F = 3.81, df = 20, P = 9.01 × 10⁻⁵; LIM F = 37.97, df = 20, P = 4 × 10⁻¹⁶; ago F = 52.11, df = 20, P = 4 × 10⁻¹⁶; bab F = 155.2, df = 20, P = 4 × 10⁻¹⁶; E strain: Bap18 F = 7.1, df = 18, P = 2.41 × 10⁻⁰⁷; BgiB F = 7.13, df = 18, P = 1.26 × 10⁻⁰⁷; LIM F = 181.8, df = 18, P = 4 × 10⁻¹⁶; ago F = 18.31, df = 18, P = 2.61 × 10⁻¹³; bab F = 26.29, df = 18, P = 7.35 × 10⁻¹⁶. b Double whole mount in situ hybridization of differentially labeled transcripts of bab in red (digoxigenin) and odorant receptors and co-receptor in green (biotin). Top images show cells co-expressing both bab and Orco in 4-day-old male pupal antenna. Bottom images show separate but neighboring cells expressing bab and OR7, OR4, or Orco (co-localization in the same sensillum) in 2-day-old adult male antenna. In merged images on the right, yellow indicates overlapping of the two signals. Successful demonstrations in pupal antennae were obtained 4 times for bab/Orco combinations. Successful demonstrations in adult antennae were obtained 22 times for bab/OR4 combinations, 8 times for bab/OR7 combinations and 7 times for bab/Orco combinations. The scale bar corresponds to 20 µm. Source data are provided as a Source Data file.

Fig. 2. Electrophysiological and behavioral analysis of bab-recombinant lines.

a Crossover points relative to exons of bab within lines L44-Z and L44-E. Boxes represent exons 1–5, orange gene regions originated from the Z-strain and blue from the E-strain. Locations of introns 1A and 1B are noted. Flanking genes ago and not are each represented by a single box. b Electroantennogram (EAG) response ratio of pure strain and bab-recombinant males. Data are presented as mean ± SEM of EAG response (in mV) to Z11-14:OAc divided by response to E11-14:OAc. Sample sizes of measured animals are Z-strain n = 10, E-strain n = 10, L44-Z n = 20, L44-E n = 20. Z-strain and Z-like responses are shown in orange, E-strain and E-like responses are shown in blue. P values report results of two-sided Tukey’s HSD post-hoc tests after an ANOVA (F = 38.67, df = 3, P = 1.13 × 10⁻¹³). c Wind tunnel responses of L44-E and L44-Z males to the Z-strain pheromone lure (97% Z-isomer, 3% E-isomer) are shown in orange and to the E-strain lure (1% Z-isomer, 99% E-isomer) in blue. P values report results of chi-square tests for L44-Z (χ² (2, n = 38 animals) = 64.89) and L44-E (χ² (2, n = 35 animals) = 57.76). Resting, no response to pheromone; WF wing-fanning response to pheromone, HP + WF wing-fanning and hair-pencil extrusion response to pheromone. Source data are provided as a Source Data file.

Fig. 3. Polymorphisms associated with pheromone response in field populations.

Bayes factor (BF) for polymorphism (SNP, indel, SV) association with pheromone trap after accounting for population structure plotted along the Z chromosome (in Mb). a Male Resp QTL region with candidate genes (gray) and bab (green). b The first intron of bab. (a, b) Bayes factor (BF) > 20 dB (triangle) and eBP_is > 2 (purple) indicate the strongest evidence for an association. Vertical green lines indicate bab exon boundaries, including rare splice variant exon 1.5. Note orientation is opposite of Fig. 2.

Fig. 4. Linkage disequilibrium between the autosomal gene controlling female pheromone blend (pgFAR) and the Z chromosome.

a Of amino-acid changing mutations detected at pgFAR, the one showing maximum r² with each Z-linked polymorphism (281,385 SNPs) is plotted for the 18–19 Mb associated interval on the Z chromosome (22,689 SNPs) (n = 62 males). Purple points depict r² values falling above the 99.99th percentile (dashed line; r² = 0.66, 26 SNPs) for the ~21-Mb Z chromosome. bab gene structure (green) and neighboring genes (gray) are shown. The solid line depicts a sliding window average of r² (1 kb with 100-bp step). b A pgFAR SNP underlying a leucine (Z allele) to isoleucine (E allele) substitution shows a maximum chromosome-wide r² of 0.71 with a SNP within intron 1B of bab (position 18.71 Mb). All males homozygous for the E allele (A/A) at pgFAR were homozygous for the C SNP at bab (23 individuals). Most males homozygous for the Z allele at pgFAR (C/C) were homozygous for the T SNP at bab (27 of 31 individuals, 87%). pgFAR heterozygotes were more evenly associated with bab genotypes. 71% of males caught in the E trap were pgFAR^A bab1^C multilocus homozygotes (22 of 31) and 84% of males caught in the Z trap were pgFAR^C bab1^T multilocus homozygotes (26 of 31). c A representative pgFAR SNP and Z-linked SNP on scaffold 325 having an r² value close to the Z chromosome median value of 0.06. Data (b, c) are presented as an estimated linear regression line (black) with a 95% confidence interval (gray band), and point color indicates the pheromone blend to which each field-caught male was attracted. Males caught using Z and E pheromone lures are shown in orange and blue, respectively. Source data are provided as a Source Data file.