Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2020 Sep 16.
Published in final edited form as: Evolution. 2020 May 17;74(6):1098–1111. doi: 10.1111/evo.13990

Co-evolving wing spots and mating displays are genetically separable traits in Drosophila

Jonathan H Massey 1,2, Gavin R Rice 3, Anggun Firdaus 4, Chi-Yang Chen 4, Shu-Dan Yeh 4, David L Stern 2,*, Patricia J Wittkopp 1,5,*
PMCID: PMC7494213  NIHMSID: NIHMS1624996  PMID: 32363590

Abstract

The evolution of sexual traits often involves correlated changes in morphology and behavior. For example, in Drosophila, divergent mating displays are often accompanied by divergent pigment patterns. To better understand how such traits co-evolve, we investigated the genetic basis of correlated divergence in wing pigmentation and mating display between the sibling species Drosophila elegans and D. gunungcola. Drosophila elegans males have an area of black pigment on their wings known as a wing spot and appear to display this spot to females by extending their wings laterally during courtship. By contrast, D. gunungcola lost both of these traits. Using Multiplexed Shotgun Genotyping (MSG), we identified a ~440 kb region on the X chromosome that behaves like a genetic switch controlling the presence or absence of male-specific wing spots. This region includes the candidate gene optomotor-blind (omb), which plays a critical role in patterning the Drosophila wing. The genetic basis of divergent wing display is more complex, with at least two loci on the X chromosome and two loci on autosomes contributing to its evolution. Introgressing the X-linked region affecting wing spot development from D. gunungcola into D. elegans reduced pigmentation in the wing spots but did not affect the wing display, indicating that these are genetically separable traits. Consistent with this observation, broader sampling of wild D. gunungcola populations confirmed the wing spot and wing display are evolving independently: some D. gunungcola males performed wing displays similar to D. elegans despite lacking wing spots. These data suggest that correlated selection pressures rather than physical linkage or pleiotropy are responsible for the coevolution of these morphological and behavioral traits. They also suggest that the change in morphology evolved prior to the change in behavior.

Keywords: Drosophila, correlated traits, pigmentation, courtship behavior, optomotor-blind

Introduction

Animals often use colorful morphological structures to communicate with prospective mates during courtship (McKinnon and Pierotti, 2010). In vertebrates and invertebrates, pigmented bodies or wings often evolve together with specific components of courtship behavior that animals use to display their colorful anatomy (Loxton, 1979; Endler, 1991; Sinervo et al., 2000; White et al., 2015). These correlated differences evolve both within and between populations, frequently distinguishing males from females or closely related species (Gray and McKinnon, 2007; McKinnon and Pierotti, 2010). In the handful of case studies examining the genetic basis of such co-evolving traits, linkage mapping and genome-wide association studies (GWAS) have shown that loci affecting pigmentation patterning tend to map to the same region of the genome as loci affecting variation in mating behaviors (Lindholm and Breden, 2002; Kronforst et al., 2006; Yeh et al., 2006; Thomas et al., 2008; Küpper et al., 2016; Lamichhaney et al., 2016; Merrill et al., 2019; reviewed in McKinnon and Pierotti, 2010). That is, physical linkage of genetic variants often appears to underlie phenotypic correlations between mating behavior and pigmentation. Although linked loci tend to explain much of the variation observed for both traits, it remains unclear in all of these cases which genes underlie the observed correlations in phenotypic differences. A key challenge, therefore, is determining how frequently these patterns of genomic architecture underlie correlated evolution and whether a single pleiotropic locus or separate linked loci are involved.

Disentangling whether pleiotropic or physically linked loci underlie patterns of correlated evolution between pigmentation and mating behavior is important for understanding how natural selection generates differences between sexes and species. If two beneficial traits are genetically correlated due to separate, physically linked loci, theory predicts that natural or sexual selection (e.g., through predation or female choice) will act to minimize recombination between the causal loci (Charlesworth and Charlesworth, 1976). It has been hypothesized that one solution to this problem might involve the evolution of chromosomal inversions that suppress recombination between two or more linked loci (Kirkpatrick and Barton, 2006), and in multiple cases linked loci contributing to correlated evolution have been shown to lie within inversions (Thomas et al., 2008; Kupper et al., 2016; Lamichhaney et al., 2016). Alternatively, mutations at a single pleiotropic gene could cause correlated components of pigmentation and mating behavior to evolve simultaneously. Although such mutations have not yet been identified for pigmentation and behavior, they have been found for other co-evolving traits (e.g., Chang et al., 2009; Kent et al., 2009; McLean et al., 2011; Duveau et al., 2012; Endler et al. 2018; Nagy et al., 2018). Distinguishing between these genetic modes of phenotypic evolution requires, in part, high-resolution mapping of correlated traits.

In the Oriental Drosophila melanogaster species group, male-specific wing spots are phylogenetically correlated with mating displays (Kopp and True, 2002; Figure 1A). Species with wing spots perform elaborate wing display dances during courtship, extending their wings laterally, turning their dorsal wing surfaces toward the female, and waving them up and down; species without wing spots lack display behavior (Kopp and True, 2002, Figure 1A,B). Correlated gains and losses of both traits have evolved repeatedly (Kopp and True, 2002, Figure 1A). For example, in D. elegans and D. gunungcola, sibling species from this group that are estimated to have diverged 2–2.8 million years ago (Prud’homme et al., 2006), D. elegans (Bock and Wheeler, 1972) males possess wing spots and perform wing displays, whereas D. gunungcola (Sultana et al., 1999) males lost both traits (Kopp and True, 2002; Prud’homme et al., 2006; Yeh et al., 2006; Figure 1B; Video 1; Video 2). Previously, Yeh et al., (2006) and Yeh and True (2014) discovered that D. elegans and D. gunungcola can generate fertile F1 hybrid female offspring in the lab and they performed interspecific crosses to study the genetic basis of wing spot and wing display divergence. Through quantitative trait locus (QTL) mapping, they showed that evolution of linked loci on the X chromosome contributed to divergence in both traits (Yeh et al., 2006; Yeh and True, 2014). One QTL explaining wing spot size variation was linked to the pigmentation gene yellow, supporting the hypothesis that yellow cis-regulatory divergence contributes to wing pigmentation evolution (Wittkopp et al., 2002a; Gompel et al., 2005; Prud’homme et al., 2006). It remained unclear, however, whether the same or different loci on the X chromosome underlie correlated differences in wing spot and wing display between these species.

Figure 1. Wing pigmentation and wing display behavior in D. elegans, D. gunungcola, and F1 hybrids.

Figure 1

Open in a new tab

(A) Phylogeny of the “Oriental” Drosophila melanogaster species group adapted from Kopp and True (2002) and Prud’homme et al. (2006). Plus (+) signs indicate species possess wing spots and/or wing displays, and minus (−) signs indicates wing spots and/or wing displays are absent. (B) Males in D. elegans (left) possess wing spots and perform bilateral wing display behaviors in front of females during courtship (Video 1). Wing spots and wing displays are absent in D. gunungcola males (right) (Video 2). (C) F1 hybrid males inheriting their X chromosome from D. elegans mothers (F1E, left) possess wing spots and perform wing display behavior like D. elegans (Video 3). F1 hybrid males inheriting their X chromosome from D. gunungcola mothers (F1G, right) are spotless and perform wing displays with low bilateral wing angles (Video 4). (D) Quantification of wing spot size (see Methods) in male D. elegans and F1E. Wing spots are larger in D. elegans than F1E (Student’s t-test; t = −2.8057; df = 11.43; P = 0.017; two-tailed). (E) Quantification of maximum bilateral wing display angles during courtship (see Methods) in male D. elegans and F1 hybrids. F1G hybrids showed lower maximum wing display angles than D. elegans and F1E hybrids (One-way ANOVA: F2,71 = 20.92; P < 7.18 × 10−8; post-hoc Tukey HSD was significant between D. elegans and F1G: P < 2.0 × 10−7 and between F1E and F1G: P < 7.1 × 10−5). Gray triangles represent individual replicates.

To distinguish between these possibilities, we re-examined the genetic basis of wing spots and wing display divergence between D. elegans and D. gunungcola. Specifically, we (1) generated recombinant backcross progeny segregating for both traits, (2) assembled chromosome-length scaffolds of D. elegans, (3) used Multiplexed Shotgun Genotyping (MSG) (Andolfatto et al., 2011) to estimate recombination crossover positions across the genome, (4) generated quantitative measures of both wing spots and wing display behavior to estimate the effect size of loci contributing to divergence, and (5) generated advanced, recombinant introgressions on the X chromosome in an attempt to separate quantitative trait loci (QTL) underlying wing spots and wing display behavior. These experiments showed that a single locus on the X chromosome behaves like a genetic switch for wing spot divergence. Males with their wing spots turned off via introgression, however, performed wing displays like D. elegans males, indicating that the two traits are genetically separable. These findings suggest that wing spot and wing display behavior might have originally diverged independently. Consistent with this hypothesis, newly collected D. gunungcola strains from Indonesia appear to completely lack wing spots but retain the ability to perform wing displays. This observation suggests that the loss of wing spots occurred prior to the loss of wing display in the reference strain of D. gunungcola used in this study and in prior work.

Materials and Methods

Fly stocks

The D. elegans HK (Hong Kong) and D. gunungcola SK (Sukarami) lines used in this study were a gift from John True (Stony Brook University). Species stocks were kept on a 12 h light-dark cycle at 23°C on a University of Michigan “R food” diet containing molasses (http://lab-express.com/flyfoodsupplies.htm#rfood) (Wirtz and Semey, 1982). Maintaining these species on R food at high densities (50–100 flies per vial) allowed for the parental population to build up to thousands of flies to collect hundreds of virgins for interspecific crosses (see below). Neither D. elegans nor D. gunungcola pupate on the sides of the vial, so adults were flipped out when 3rd instar L3 larvae developed and Fisherbrand filter paper (cat# 09-790-2A) was added to the food to create pupation space.

Generating hybrid progeny

Virgin males and females of D. elegans HK and D. gunungcola SK (the same lines used previously in Yeh et al., 2006; Yeh and True, 2014) were isolated upon eclosion and stored in groups of ten for one week on University of Michigan “M food”, which is the standard cornmeal diet from the Bloomington Drosophila Stock Center (https://bdsc.indiana.edu/information/recipes/bloomfood.html) with 20% higher agar content. Virgin males from D. elegans were crossed to virgin females from D. gunungcola, and virgin males from D. gunungcola were crossed to virgin females from D. elegans in groups of ten males and ten females to generate fertile F1 female and sterile F1 male hybrids. These crosses took ~3–4 weeks to produce hybrid progeny. The switch from R food to M food for interspecific crosses was necessary, because R food tended to accumulate condensation and bacterial growth much faster than M food when few flies occupied a vial. Since crossing D. elegans and D. gunungcola to generate F1 hybrids tends to take several more weeks than within species crosses, the switch to M food diet allowed for maximum breeding time and the development of dozens of hybrid progeny. Once hybrid females eclosed from both interspecific cross directions, they were pooled into the same vial and aged for ten days. We did not keep track of F1 hybrid female maternity, because previous work (Yeh and True, 2014) found no effect of F1 hybrid maternity on trait means for wing spots and wing display in backcross populations. Multiple high-density groups of ~60 F1 hybrid females were then backcrossed to ~60 virgin male D. elegans flies in individual vials on M food diet to create the D. elegans backcross recombinant population (724 individuals). To create the D. gunungcola backcross recombinant population (241 individuals), groups of ~60 F1 hybrid females were backcrossed to ~60 virgin male D. gunungcola flies in individual vials on M food diet; this backcross was less successful at producing recombinant progeny than the D. elegans backcross direction.

Behavioral assays

Virgin D. elegans females were isolated upon eclosion, aged 10–20 days, and stored in groups of 30–40 for courtship assays. F1 hybrid and recombinant backcross males were isolated individually in M food vials using CO2 upon eclosion for at least 5 days before each courtship assay. For each assay, a single individual male was gently aspirated into a custom built 70 mm diameter bowl arena that matches the specifications in Simon and Dickinson (2010). Next, a single virgin D. elegans female was aspirated into the chamber and videotaped for the next 20 min, using a Canon VIXIA HF R500 camcorder mounted to Manfrotto (MKCOMPACTACN-BK) aluminum tripods. Videos were recorded between 09:00 and 16:00 at 23°C. D. elegans virgin females were used in all courtship assays in case any D. elegans female cues were necessary to elicit male wing display behavior. After each assay, both the male and female were aspirated back into an M food vial and left for up to 5 days, after which each male was frozen in individual 1.5 mL Eppendorf tubes for wing spot quantification (see Methods-Quantification of wing spots), genomic DNA (gDNA) extraction, and sequencing (see Methods-Library preparation and sequencing). All courtship videos (~900 total) are available here: https://deepblue.lib.umich.edu/data/concern/data_sets/j098zb17n?locale=en.

Quantification of wing display behavior

F1 hybrid and recombinant males from both backcross directions performed variable wing display behaviors during courtship as described previously (Yeh et al., 2006; Yeh and True, 2014). To generate quantitative measurements of wing display variation between individuals, each courtship video was played using QuickTime (version 10.4) (Apple Inc., Cupertino, CA) software in a MacOS environment, and digital screenshots were manually taken for each wing display bout, defined as a bilateral wing extension performed near the female (Supplementary Figure S1). Next, for each individual fly, wing display screenshots were compared to each other to identify the maximum wing display bout per fly, defined by comparing the distance between the tips of each wing relative to the center of the fly. These maximum wing display screenshots were then imported into ImageJ software (version 1.50i) (Wayne Rasband, National Institutes of Health, USA; http://rsbweb.nih.gov/ij/) to manually measure the “Maximum wing display angle” for F1 hybrid and recombinant males. In ImageJ, each screenshot image was inverted using the “Find Edges” function to enhance the contrast between the arena background and the edges of the fly wings (Supplementary Figure S1). Next, the “Polygon Selections” tool was used to fit an ellipse around the fly body using the “Fit Ellipse” function (Supplementary Figure S1). A Macros function (Supplementary File S1) was then used to generate major and minor axes inside the ellipse to identify the center of the fly body (Supplementary Figure S1). Finally, the “Angle Tool” was used to measure the “Maximum wing display angle” centering the vertex at the intersection of the major and minor axes and extended from wing tip to wing tip (Supplementary Figure S1). “Maximum wing display angle” varied between ~50° and ~220° between backcross recombinant individuals. Raw data for Figure 1E is available in Supplementary File S2, and raw data for Figure 3 are deposited on Dryad (doi:10.5061/dryad.gb5mkkwm5).

Figure 3. QTL analysis and effect plots for wing display divergence.

Figure 3

Open in a new tab

(A) Maximum wing display angles varied in D. elegans and D. gunungcola backcross recombinants. Maximum wing display angles were quantified by measuring the angle between each wing tip using ImageJ software (see Methods). (B) Maximum wing display QTL map for the D. elegans (red) and D. gunungcola (blue) backcross. LOD is indicated on the y-axis. Individual SNP markers are indicated with black tick marks along the x-axis. Horizontal red and blue lines mark P = 0.01 for the D. elegans and D. gunungcola backcross, respectively. (C) Effect plots for the X chromosome and Muller Element B QTL peaks from the D. elegans backcross (left) and for the X, Muller Element B, and E QTL peaks from the D. gunungcola backcross (right). No epistatic interactions were detected between QTLs (see Methods) (Two-way ANOVA: F1,402 = 0.146; P = 0.70 for the D. elegans backcross; Three-way ANOVA: F1,137 = 0.050 (X:B), 0.034 (X:E), 1.75 (B:E), 0.799 (X:B:E); P = 0.82 (X:B), 0.86 (X:E), 0.19 (B:E), 0.37 (X:B:E) for the D. gunungcola backcross). Gray triangles represent individual replicates.

Quantification of wing spots

Since wing spots fully form ~24 h after eclosion in D. elegans, all parental male D. elegans, D. gunungcola, F1 hybrids, and backcross recombinants were aged at least 7 days before being frozen at −20C in 1.5 mL Eppendorf tubes. Next, using a 20 Gauge stainless steel syringe tip (Techcon) (cat# TE720100PK) the right wing of each fly was cut away from the thorax and placed on a glass microscope slide (Fisherbrand) (cat# 12-550-15) to image using either a Leica MZFLIII stereoscope equipped with a Leica DC480 microscope camera or a Canon EOS Rebel T6 camera equipped with a Canon MP-E 65 mm macro lens. Each camera was calibrated using an OMAX 0.1 mm slide micrometer to define pixel density in ImageJ software. JPEG images of wings were imported into ImageJ to measure wing spot size relative to total wing area (wing spot size / total wing area). We quantified wing spot size, rather than wing spot intensity, because we aimed to map previously identified wing spot size QTL (Yeh et al., 2006; Yeh and True 2014) and their relationship with wing display behavior. Total wing area (wing length × wing width) was approximated using length and width proxies following methods described in Yeh and True (2014). Using the “Polygon Selections” tool, the margins of black pigmentation defining each “Wing spot size” was traced and the polygon area quantified in mm2 using the “Measure” function. “Wing spot size” varied between 0 mm2 (spotless) and 0.15 mm2 between recombinant individuals. Raw data for Figure 1D is available in Supplementary File S3, and raw data for Figure 2 are deposited on Dryad (doi:10.5061/dryad.gb5mkkwm5).

Figure 2. QTL analysis, effect plots, and in situ hybridization for wing pigmentation divergence.

Figure 2

Open in a new tab

(A) Wing spots vary in size and shape in D. elegans and D. gunungcola backcross recombinants. Wing spots were traced (pink) and quantified relative to proxies for total wing area (length × width) using ImageJ software (see Methods). (B) Wing spot QTL map for the D. elegans (red) and D. gunungcola (blue) backcross. LOD (logarithm of the odds) is indicated on the y-axis. The x-axis represents the physical map of Muller Elements X, B, C, D, E, and F based on the D. elegans assembled genome (see Methods). While D. elegans and D. gunungcola have six separate chromosomes (Yeh et al., 2006; Yeh and True, 2014), they are each syntenic with the D. melanogaster genome accordingly: X = X, B = 2L, C = 2R, D = 3L, E = 3R, F = 4. Individual SNP markers are indicated with black tick marks along the x-axis. Horizontal red and blue lines mark P = 0.01 for the D. elegans and D. gunungcola backcross, respectively. (C) Effect plots for the X chromosome QTL peak from the D. elegans backcross (left) and D. gunungcola backcross (right). Gray triangles represent individual replicates. (D) D. elegans and D. gunungcola backcross recombinants containing X chromosome breakpoints immediately flanking the wing spot QTL peak were aligned to compare the effects of each on wing pigmentation. Regions in red represent D. elegans linked loci, and regions in blue represent D. gunungcola linked loci. Recombinants possessing D. elegans loci to the left of ~10.32 Mbp are spotless, while recombinants possessing D. elegans loci to the right of ~10.74 Mbp possess dark wing spots. (E) Two recombinants define the wing spot locus to a ~440 Kbp region containing 15 candidate genes. omb is the strongest wing pigmentation candidate gene given evidence from prior work (see Results and Discussion). (F) In situ hybridization of D. elegans and D. gunungcola pupal wings probed for omb mRNA (purple) at 30 h after pupal formation (APF) (see Supplementary Figure S8 for additional replicates).

Library preparation and sequencing

We estimated chromosome ancestry “genotypes” for 724 D. elegans backcross progeny and 241 D. gunungcola backcross progeny with a single Multiplexed Shotgun Genotyping (MSG) (Andolfatto et al., 2011) library using 965 barcoded adaptors following methods described in Cande et al., (2012). In brief, to extract gDNA from all male backcross individuals, single flies were placed into individual wells of 96-well (Corning, cat# 3879) plates containing a single steel grinding bead in each well (Qiagen, cat# 69989). Eleven plates in total were prepared for 965 individual gDNA extractions. gDNA was isolated and purified using the solid tissue extraction procedure from a Quick-DNA 96 Kit (Zymo, cat# D3012) and a paint shaker to homogenize tissue. gDNA was tagmented using a hyperactive version of Tn5 transposase charged with annealed adaptor oligos following the methods described in Picelli et al. (2014). Unique barcoded adaptor sequences were ligated to each sample of tagmented gDNA with 14 cycles of PCR using OneTaq 2x Master Mix (NEB, cat# M0482S), and all samples were pooled into a single multiplexed sequencing library. Agencourt AMPure XP beads (Beckman Coulter, cat# A63881) were used to size select ~150–800 bp fragments and eluted in 35 uL of molecular grade water (Corning, cat# MT46000CI). The library was quantified by qPCR and sequenced in a single lane of Illumina HiSeq by the Janelia Quantitative Genomics Team.

In addition to generating the backcross sequencing library, both D. elegans HK and D. gunungcola SK parental species were sequenced at 20x coverage using an Illumina MiSeq Reagent Kit (v.3, 600 cycle PE) to facilitate genome assembly. In brief, gDNA was extracted using a Quick-DNA Microprep Kit (Zymo, cat# D4074) from 10 pooled females for each species and quantified on a Qubit 2.0 (Invitrogen). These samples were sent to the University of Michigan DNA Sequencing Core to prepare 300 bp paired-end libraries, which were quantified by qPCR and sequenced in a single lane of Illumina MiSeq.

Genome assembly

In brief, Illumina reads from all 965 backcross recombinants were used to perform MSG on the Baylor College of Medicine D. elegans genome assembly (accession number: GCA_000224195.2). Using custom scripts in R and Python (https://github.com/masseyj/elegans), the recombination fraction between the Baylor and MSG contigs was calculated and plotted to manually tabulate joins and splits between newly assembled contigs. These new contigs were then used to assemble approximately chromosome length scaffolds in D. elegans (accession number: WVIB00000000) and partially assembled scaffolds in D. gunungcola (accession number: WTSR00000000).

Marker generation with Multiplexed Shotgun Genotyping

Following methods described previously (Andolfatto et al., 2011; Cande et al., 2012), we used the MSG software pipeline (https://github.com/JaneliaSciComp/msg/tree/master/instructions) to perform data parsing and chromosome ancestry estimation to generate markers for quantitative trait locus (QTL) analysis. In brief, using data from the Illumina backcross sequencing library (see Supplementary File S4 for the number of reads per individual), we mapped reads to the assembled D. elegans and D. gunungcola parental genomes to estimate chromosome ancestry for each backcross individual. We generated 3,425 and 3,121 markers for the D. elegans and D. gunungcola backcrosses, respectively, for QTL analysis [markers, phenotypes, and procedures for QTL mapping are deposited on Dryad (doi:10.5061/dryad.gb5mkkwm5)]. PDFs of chromosomal breakpoints for each recombinant are available here: https://deepblue.lib.umich.edu/data/concern/data_sets/j098zb17n?locale=en.

QTL analysis

QTL analysis was performed using R/qtl (Broman et al., 2003) in R for Mac version 3.3.3 (R Core Team 2018) in a MacOS environment. Ancestry data for both backcross directions were imported into R/qtl using a custom script (https://github.com/dstern/read_cross_msg), which directly imports the conditional probability estimates produced by the Hidden Markov Model (HMM) of MSG (Andolfatto et al., 2011). We performed genome scans with a single QTL model using the “scanone” function of R/qtl and Haley-Knott regression (Haley and Knott, 1992) for “Wing spot size” and “Maximum wing display angle”. Note, for “Wing spot size”, 68 and 42 recombinants from the D. elegans and D. gunungcola backcross populations, respectively, were excluded from the QTL mapping because their wings were too damaged to quantify spot variation. Similarly, for “Maximum wing display angle”, 314 and 94 recombinants from the D. elegans and D. gunungcola backcross populations, respectively, were excluded from the QTL mapping because these males did not perform any courtship behavior during the assay. Significance of QTL peaks at α = 0.01 was determined by performing 1000 permutations of the data. Effect sizes for each QTL peak were individually estimated by comparing the mean “Wing spot size” or “Maximum wing display angle” between individuals that inherited either D. elegans or D. gunungcola alleles at each QTL peak position [markers, phenotypes, and procedures for QTL mapping are deposited on Dryad (doi:10.5061/dryad.gb5mkkwm5)].

Since we detected multiple QTL peaks on separate chromosomes for “Maximum wing display angle”, we tested for the presence of epistatic interactions using two methods: First, we performed two- and three-way ANOVAs comparing the effect of each QTL peak in multiple QTL peak genetic backgrounds and found no evidence of an interaction. For two-way ANOVAs, we tested for any statistically significant interactions for max wing display angles between two different QTL peaks in the D. elegans backcross. For three-way ANOVAs, we tested for any statistically significant interactions for max wing display angles between three different QTL peaks in the D. gunungcola backcross. Second, we performed genome-wide pairwise tests using the “scantwo” function of R/qtl and Haley-Knott regression to test for non-additive interactions across all markers; LOD significance thresholds at α = 0.05, 0.01, and 0.001 were determined by performing 1000 permutations of the data for each model (Supplementary Figure S2, Supplementary Tables S1,S2).

Annotating the wing spot QTL interval

To annotate genes within the ~440 Kbp fine-mapped wing spot locus, we performed nucleotide BLAST (BLASTn) (Johnson et al., 2008) searches against the D. melanogaster genome (taxid: 7227) using ~10 Kbp windows of assembled D. elegans chromosomal regions spanning the wing spot QTL interval. Using the “GBrowse” tool on Flybase (Thurmond et al., 2018), we mapped regions of microsynteny to identify the orientation of each gene and exported the respective D. melanogaster coding region (CDS) FASTA sequences to align with the D. elegans X chromosome.

In situ hybridization

Fly genomic DNA (gDNA) was extracted from ten homogenized D. elegans and D. gunungcola females using a Quick-DNA Microprep Kit (Zymo, cat# D3021). The following forward and reverse primers were designed and synthesized by Integrated DNA Technologies (IDT) to PCR amplify 321 bp DNA templates targeting exon 5 of the omb locus in D. elegans: 5’-GCTGAGGATCCATTCGCTAGATTTG-3’ and 5’-GTTGTTGGAACTAGAGTTGTTGGTG-3’, and D. gunungcola: 5’- GCTGAGGATCCATTCGCTAGATTTG-3’ and 5’-GTTGTTGGAACTGGAGTTGTTGGTG-3’. Reverse primers were designed beginning with a T7 RNA polymerase binding sequence (TAATACGACTCACTATAG) to facilitate in vitro transcription. Raw PCR products were then used to generate digoxigenin-labeled RNA probes using a T7 RNA in vitro transcription kit (Promega / Life Technologies). RNA was ethanol precipitated and resuspended in water to analyze on a Nanodrop. Each probe was stored at −20°C in 50% formamide before in situ hybridization.

All tissues underwent primary dissection in PBS, fixed for 30 mins in 4% PFA, washed 3X in PBT and underwent secondary dissection in PBT, were then washed 2X in MeOH, and 2X in EtOH before being stored at −20°C. Male D. elegans and D. gunungcola L3 wing discs were dissected first to validate that our omb probes detected an mRNA expression pattern similar to D. melanogaster (Grimm and Pflugfelder, 1996; Supplementary Figure S3). Next, pupal wings were dissected at 30 and 48 h after pupal formation (APF) to probe for omb mRNA. To prepare pupal wings, appropriately staged pupae underwent a primary dissection: were cut in half along the anterior-posterior axis using Astra Platinum Double Edge Razor Blades, and fat body was washed out of the pupal casing using a pipette and PBS prior to fixation. After fixation, pupal wings underwent a secondary dissection to pull off the cuticle surrounding each wing and then washed using the procedure described above. Finally, in situ hybridization was carried out as previously described (Vincent et al., 2019). Briefly, we used an InsituPro VSi robot to rehydrate in PBT, fix in PBT with 4% PFA, and prehybridize in hybridization buffer for 1 hr at 65°C. Samples were then incubated with probe for 16 h at 65°C before washing with hybridization buffer and PBT. Samples were blocked in PBT with 1% bovine serum albumin (PBT+BSA) for 2 hours. Samples were then incubated with anti-digoxigenin Fab fragments conjugated to alkaline phosphatase (Roche) diluted 1:6000 in PBT+BSA. After additional washes, color reactions were performed by incubating samples with NBT and BCIP (Promega) until purple stain could be detected under a dissecting microscope. Samples were mounted in glycerol on microscope slides coated with poly-L-lysine and imaged at 10X magnification on a Leica DFC450C camera.

Generating advanced recombinant introgressions on the X chromosome

To try to isolate the QTL effects for “Wing spot size” and “Maximum wing display angle” localized to the X chromosome according to the D. elegans backcross experiment, F1 hybrid females were generated using the procedures described above. F1 hybrid females were then backcrossed to D. elegans males in ~12 replicate vials, and backcross males lacking wing spots were isolated to measure “Maximum wing display angles” during courtship as described above. This procedure was repeated for seven generations to generate BC3-BC9 backcross individuals: backcross females were backcrossed en masse (across ~12 vials) to D. elegans males, and BC3 backcross males lacking wing spots were isolated to measure “Maximum wing display angles” during courtship with D. elegans virgins (and so on to BC9). At each generation, an attempt was made to create stable introgression lines of advanced recombinant males lacking wing spots, but all failed to produce offspring, suggesting that D. gunungcola X-linked loci might also contain hybrid sterility factors. After seven generations of backcrossing, gDNA from all backcross males lacking wing spots was extracted and sequenced for MSG as described above. Backcross males lacking wing spots from BC4-BC9 were homozygous for D. elegans genomic regions across all autosomes but varied for the amount of D. gunungcola genome regions on the X chromosome.

Introgression of black body color alleles from D. gunungcola into D. elegans

In the D. gunungcola backcross, QTL mapping for wing spot size revealed QTL peaks linked to Muller Element C and E when spotless recombinants were excluded from the analysis (Supplementary Figure S4; Supplementary Table S3). The Muller Element E QTL peak is located near the ebony gene, which appears to contribute to variation in body color between D. elegans and D. gunungcola (unpublished data). We therefore reasoned that introgressing dark body color from D. gunungcola into D. elegans would introgress the Muller Element E QTL peak underlying wing spot size differences. After six generations of backcrossing dark brown female recombinants with D. elegans males, we crossed dark brown male and female recombinants together to create black offspring homozygous for the introgressed region. We then performed MSG on a single, dark black introgression line and found that it was homozygous for ~1.5 Mb of D. gunungcola alleles linked near the Muller Element E QTL peak (Supplementary Figure S4A,C).

Observing and collecting wild D. gunungcola and D. elegans in Indonesia

Throughout early July 2018, D. elegans and D. gunungcola were recorded performing courtship in East Java, Indonesia on Ipomoea sp. and Brugmansia sp. flowers using Canon VIXIA HF R500 camcorders mounted to Manfrotto (MKCOMPACTACN-BK) aluminum tripods. Both species were observed in sympatry on flowers near Coban Rondo Waterfall in Batu, Batu City, East Java, Indonesia (−7.884985, 112.477311). After video recording courtship, males and females were captured using a mouth pipette and gently aspirated into glass vials containing standard fly media (glucose, corn meal, yeast extract, and agar). Isofemale lines of D. gunungcola from Bumiaji District (Batu City, East Java Province, Indonesia) were established in the laboratory on standard fly media at 24°C temperature. We quantified (see Methods- Quantification of wing display behavior), to the best of our knowledge, the first recorded observations of D. gunungcola wing displays on flowers in the field and in the laboratory (Supplementary Figure S11; Videos 8, 9, and 11). Across ~10 independent observations of courting D. gunungcola pairs at Bumiaji District and Coban Rondo, all males performed a type of wing display despite completely lacking wing spots. To confirm species identification of D. gunungcola and D. elegans from the field sites mentioned above, we dissected and imaged male genitalia and compared with the laboratory strains (D. gunungcola SK and D. elegans HK) used in this study and described previously (Sultana et al., 1999; Kopp and True, 2002) (Supplementary Figure S9). The distal paramere [also called the pregonite (Rice et al., 2019)] was especially diagnostic of species identity (Supplementary Figure S9). We also performed low coverage sequencing of the new D. gunungcola strains’ genomes from Coban Rondo (see Methods-Library preparation and sequencing) and aligned coding sequences from the omb locus with the D. gunungcola SK lab strain (Supplementary Figure S10). A nonsynonymous coding change that distinguished the laboratory D. gunungcola SK strain from D. elegans HK also distinguished the new D. gunungcola Coban Rondo strain from D. elegans HK, matching the D. gunungcola SK sequence (Supplementary Figure S10).

Statistics

Statistical tests were performed in R for Mac version 3.3.3 (R Core Team 2018) using Student’s t-test (two-tailed) to test for statistically significant effects of pairwise comparisons of continuous data with normally distributed error terms. For tests comparing more than two groups, ANOVAs were performed with post hoc Tukey HSD for pairwise comparisons adjusted for multiple comparisons. See “QTL analysis” methods for statistical tests used during QTL mapping.

Results and Discussion

X-linked sequence divergence contributed to wing spot and wing display divergence

D. elegans males perform elaborate wing display dances (Video 1) in front of females during courtship, displaying the presence of darkly pigmented wing spots (Figure 1B), whereas its sibling species, D. gunungcola, lost wing spots (Yeh et al., 2006; Prud’homme et al., 2006) and wing displays (Figure 1B; Video 2). Despite these differences in sexual traits, D. elegans and D. gunungcola can mate and form viable F1 hybrids in the lab (Yeh et al., 2006; Yeh and True, 2014). Sequence divergence on the X chromosome has previously been implicated in the divergence of wing spots and wing display behavior (Yeh et al., 2006; Yeh and True, 2014). To confirm this effect of the X-chromosome, we quantified variation in wing spot size and wing display behavior in F1 hybrid males from reciprocal crosses between D. elegans and D. gunungcola. These F1 hybrids inherited their X chromosome from either D. elegans or D. gunungcola (whichever species was their mother) and autosomes from both species. Consistent with prior work, F1 hybrid males inheriting the X chromosome from D. elegans mothers (F1E) possessed wing spots, whereas F1 hybrid males inheriting the X chromosome from D. gunungcola mothers (F1G) did not (Figure 1C,D). These wing spots of F1E males were smaller, however, than the wing spots seen in D. elegans (Figure 1D, Student’s t-test, P = 0.02). Differences in wing display behavior were also apparent between F1E (Video 3) and F1G hybrids (Video 4), which is also consistent with prior work (Yeh et al., 2006; Yeh and True, 2014). More specifically, we found that although both F1 hybrids performed wing displays during courtship, F1E hybrids tended to open their wings more widely than F1G hybrids during display performance (Figure 1C). We quantified variation in this wing display trait between F1 hybrids by measuring the maximum bilateral wing display angles (Figure 1C) during courtship (see Methods). We found that F1E hybrids performed wing displays comparable to D. elegans males (Figure 1E, post-hoc Tukey HSD, P = 0.6), whereas F1G males showed, on average, lower display angles (Figure 1E, post-hoc Tukey HSD, P = 7.1 × 10−5). Together these data confirm that divergence of one or more loci on the X chromosome contribute to divergence in wing spot size and wing display behavior between D. elegans and D. gunungcola.

Evolution of at least three loci contribute to wing spot divergence

To identify the location of X-linked (as well as autosomal) loci contributing to divergence in wing spot size, we quantified wing spot size variation in 656 recombinant males produced by backcrossing F1 hybrid females to D. elegans males and 199 recombinant males produced by backcrossing F1 hybrid females to D. gunungcola males. These backcross males showed a range of wing spot sizes (Figure 2A). Using Multiplexed Shotgun Genotyping (MSG) (Andolfatto et al., 2011), we inferred the allele most likely inherited from the F1 mother (D. elegans or D. gunungcola) for each genomic position in each recombinant. We then performed quantitative trait locus (QTL) mapping for wing spot size and identified a single, highly significant QTL peak on the X chromosome (Figure 2B and Table 1). In both backcross directions, variation linked to this wing spot QTL peak explained almost all of the difference in wing spot size between D. elegans and D. gunungcola (Figure 2C). Repeating the QTL mapping after excluding recombinant individuals lacking wing spots, however, allowed us to identify additional QTLs of smaller effect on Muller Elements C (chromosome 2R in D. melanogaster) and E (chromosome 3R in D. melanogaster) in the D. gunungcola (but not D. elegans) backcross population (Supplementary Figure S4A; Supplementary Table S3). Observing these QTL only in the D. gunungcola backcross populations suggests that they are caused by recessive D. gunungcola alleles, which are never homozygous in the D. elegans backcross population. Introgressing the QTL region on Muller Element E from D. gunungcola into D. elegans through 5 generations of backcrossing (Supplementary Figure S4C) reduced the size of wing spots (Supplementary Figure S4D,E). This region includes the ebony gene, which has previously been shown to be able to inhibit the development of dark pigments in D. melanogaster (Wittkopp et al., 2002b). Crossing this introgression line to D. elegans masked most of the reduction in spot size (Supplementary Figure S4D,E), consistent with the D. gunungcola QTL allele being recessive to the D. elegans allele. Taken together, these data indicate that the majority of wing spot divergence between D. elegans and D. gunungcola maps to a single, large-effect QTL on the X chromosome, but that wing spot size is also influenced by loci on Muller Elements C and E.

Table 1.

QTLs detected for wing spot size and maximum wing display angle divergence

Trait Backcross Chromosome QTL interval (bp)a QTL peak (bp) LOD
Wing spot size D. elegans X 10,297,836–10,744,020 10,304,581 220
Max wing display angle D. elegans X 8,729,737–15,691,924 9,006,035 18.9
Max wing display angle D. elegans B 5,773,911–13,325,000 9,001,485 4.66
Wing spot size D. gunungcola X 10,474,499–11,584,862 11,223,359 38.9
Max wing display angle D. gunungcola X 16,885,658–25,539,528 24,196,217 4.23
Max wing display angle D. gunungcola B 7,078,659–12,180,268 10,093,006 6.28
Max wing display angle D. gunungcola E 3,813,413–11,535,144 9,604,970 7.59
Open in a new tab
a

LOD drop 1.5 support interval

A 440 kb locus behaves like a genetic switch for wing spots

To further refine the X-linked QTL, we more closely examined the genotypes and phenotypes of recombinants with inferred crossover positions immediately flanking the wing spot QTL peak (Figure 2D, Supplementary Figure S5). Doing so allowed us to identify a ~440 kb region containing a QTL that acts like a genetic switch controlling the presence or absence of the wing spot (Figure 2D, Supplementary Figure S5). This region includes 15 genes (Figure 2E) and notably excludes the X-linked pigmentation gene, yellow, which has previously been suggested to contribute to wing spot development and evolution through changes in a spot-specific cis-regulatory element (Wittkopp et al., 2002a; Gompel et al., 2005; Prud’homme et al., 2006; Yeh et al., 2006; Arnoult et al., 2013; Yeh and True, 2014; Supplementary Figure S6). One of these 15 genes is optomotor-blind (omb) (Figure 2E), which encodes a T-box-containing transcription factor (Pflugfelder et al., 1992a; Pflugfelder et al., 1992b) that has previously been implicated in pigmentation patterning (Thompson, 1959; Kopp and Duncan, 1997), pigmentation evolution (Brisson et al., 2004), and distal wing patterning (Grim and Pflugfelder, 1996). In D. melanogaster, gain- and loss-of-function omb alleles cause expansion and contraction of abdominal pigmentation bands, respectively (Kopp and Duncan, 1997), and variation in abdominal pigmentation patterning in D. polymorpha is strongly associated with polymorphisms at the omb locus (Brisson et al., 2004).

Although we identified two nonsynonymous protein coding changes between D. elegans and D. gunungcola (Supplementary File S5), omb is required for the development of many structures throughout the body (Pflugfelder, 2009); we, therefore, reasoned that genetic divergence in omb would be more likely to affect its expression than its protein function (Stern and Orgogozo, 2008). To look for differences in omb expression between D. elegans and D. gunungcola that might affect wing spot development, we used in situ hybridization to detect omb mRNA in the developing wing of both species (Figure 2F). In D. melanogaster, omb is expressed in a broad stripe that overlaps the wing pouch region in larval L3 wing discs (Grimm and Pflugfelder, 1996). omb expression in the wing pouch is required for distal wing development, as demonstrated by D. melanogaster omb hypomorphs that show disrupted distal wing tip development in adults (Grimm and Pflugfelder, 1996). We hypothesized, therefore, that differences in D. elegans and D. gunungcola omb expression patterning during pupal wing development might prefigure changes in wing spot pigmentation observed in adult males, similar to the changes in wingless expression shown to prefigure wing spots in D. guttifera (Werner et al., 2010). Consistent with the expression of omb-lacZ in pupal wings of D. melanogaster (Álamo Rodríguez et al., 2004), we detected omb mRNA in the wing hinge and distal wing tip 30 h after puparium formation (APF) in D. elegans and D. gunungcola (Figure 2F). We were unable to identify any consistent differences in the omb expression patterns between D. elegans and D. gunungcola, although it is possible that we may not have detected subtle differences in expression patterns. In addition, it is possible that the changes in omb protein sequence contribute to differences in wing spot patterning, or that other genes in the minimal mapped interval are the true cause of the difference in wing spot patterning.

Evolution at multiple loci contributed to wing display divergence

To identify loci contributing to divergence in wing display behavior, we quantified variation in maximum wing display angles (see Methods) in 410 D. elegans and 147 D. gunungcola backcross recombinant males, again observing a range of phenotypes (Figure 3A). We identified multiple significant QTL contributing to variation in wing display (Figure 3B; Table 1). In the D. elegans backcross, we mapped a QTL on the X chromosome that overlaps with the wing spot QTL (Figure 3B; Table 1). We also mapped a QTL on Muller Element B (chromosome 2L in D. melanogaster) (Figure 3B; Table 1). In the D. gunungcola backcross, we mapped QTLs on the X chromosome as well as Muller Elements B and E (Figure 3B; Table 1). The observation of a peak on Muller Element E only in the D. gunungcola backcross suggests the presence of alleles affecting wing display behavior that are recessive and/or interact epistatically with divergent sites elsewhere in the genome.

To test for epistatic interactions contributing to wing display divergence, we performed a two-dimensional genome scan to search for non-additive interactions across all markers in both backcross directions and found no significant interactions (Supplementary Figure S2; Supplementary Tables S1,S2). We also tested for evidence of non-additive interactions among the wing display QTL peaks themselves by performing two- and three-way ANOVAs in the D. elegans and D. gunungcola backcrosses, respectively, and found no evidence of significant interactions between loci (Figure 3C). Instead, each wing display QTL peak appears to behave approximately additively, with D. gunungcola alleles contributing to lower maximum wing display angles (Figure 3C). Surprisingly, the effect of the X-linked QTL on wing display angle in the D. gunungcola backcross in multiple genetic backgrounds was similar to the estimated effect size of the X-linked QTL in the D. elegans backcross (compare panels in Figure 3C) despite the much lower LOD score of the X-linked QTL in the D. gunungcola backcross population (Figure 3B; Table 1). We suggest that while the detected QTL in the D. gunungcola backcross appear to interact additively with each other, undetected QTL elsewhere in the genome are likely masking the X-effect in the D. gunungcola backcross map. While the purpose of the two-dimensional genome scan (Supplementary Figure S2; Supplementary Tables S1,S2) was to detect these effects, our sample size is likely too small to identify small-effect epistatic interactions.

Males lacking wing spots perform normal wing displays

While it remains unclear which gene evolved to cause the majority of wing spot divergence, fine-mapping the locus controlling the presence or absence of the wing spot allowed us to test whether the locus that turns off wing spots in D. gunungcola also affects wing display behavior. To perform this test, we introgressed D. gunungcola alleles causing a loss of the wing spot into D. elegans by repeated backcrossing (see Methods). We recovered three introgression lines lacking wing spots and found that all three lines had inherited the ~440 kb region observed in mapping experiments to act like a genetic switch controlling wing spot development (Figure 4A,B), independently confirming the causal role of the switch region in wing spot divergence. We noticed, however, that several advanced recombinants developed a wing spot “shadow” (Figure 4B), possibly due to the effects of other D. elegans alleles affecting wing spot development. We next asked whether the spotless advanced recombinants performed wing displays with lower wing display angles than D. elegans males. Surprisingly, we found that all advanced recombinants inheriting the D. gunungcola allele eliminating the wing spot performed wing displays indistinguishable from D. elegans males during courtship (Figure 4B,C; Videos 5–7). Thus, the loci controlling the wing spot and courtship behavior are genetically separable.

Figure 4. D. elegans males possessing the D. gunungcola wing spot locus perform normal wing displays.

Figure 4

Open in a new tab

(A) Multiplexed Shotgun Genotyping (MSG) (Andolfatto et al., 2011) was used to estimate genome-wide ancestry assignments for three introgression lines generated by repeatedly backcrossing the D. gunungcola wing spot QTL region into a D. elegans genetic background (see Methods). The posterior probability that a region is homozygous for D. elegans (red) or D. gunungcola (blue) ancestry is plotted along the y-axis. The dotted line marks the location of the fine-mapped wing spot region (Figure 2D,E; Table 1). (B). None of the introgressions possessed dark wing spots (although a light wing spot “shadow” is visible). (B,C) Every introgression performed max wing display angles indistinguishable from D. elegans males (One-way ANOVA: F3,42 = 0.449; P = 0.72). Gray triangles represent individual replicates.

The repeated co-evolution of male-specific wing spots and wing display behavior in multiple species (Kopp and True, 2002) combined with the presence of overlapping QTL for these traits on the X chromosome (Yeh et al., 2006; Yeh and True, 2014; and this study) suggested that a single pleiotropic gene might be contributing to the evolution of both traits. The finding that D. elegans introgression lines lacking a wing spot performed a normal wing display argues against this hypothesis and indicates instead that these two traits arose independently between this species pair. To further investigate how these divergent traits might have evolved, we recorded courtship behavior in a wild population of D. gunungcola in Indonesia; to the best of our knowledge, all prior studies of D. gunungcola pigmentation and courtship used the one previously available lab strain (Sultana et al., 1999). Surprisingly, we found that all D. gunungcola males observed in the wild population lacked wing spots (Supplementary Figure S7) but performed wing displays (Supplementary Figure S11; Videos 8, 9), confirming that these are genetically distinct traits. The wing displays performed by these flies appeared to show a lower maximum wing extension angle than D. elegans (Supplementary Figure S11), similar to the wing display behavior seen in F1 hybrids between D. elegans and D. gunungcola with D. gunungcola mothers (Figure 1C; Video 4). Analysis of new lab strains founded by flies captured from this D. gunungcola population showed similar male courtship behavior in the lab as observed on flowers (Supplementary Figure S11; Video 11).

It remains unknown whether the absence of wing display behavior in the D. gunungcola SK lab strain seen since Yeh et al., (2006) (collected originally from Sumatra in 1999) is either (1) segregating within and/or among wild populations of D. gunungcola or (2) limited to the SK strain and might have been lost by chance or adaptation to the lab environment. Observing that flies from the newly isolated strain of D. gunungcola displayed a similar wing display behavior in the field and after being reared in the lab (Supplementary Figure S11) does, however, argue that the absence of wing display behavior in the SK line is unlikely due to phenotypic plasticity caused by the lab setting. Additional sampling of natural D. gunungcola populations from throughout its species range (or at least resampling of the population from which the SK lab line was derived) is needed to distinguish between these possibilities. What we can say at this time, however, is that the new strain of D. gunungcola performed a similar wing display in the field and in the lab, with the angle of the wing display appearing to be consistently less than the wing display angle measured for D. elegans (Supplementary Figure S11). Because we mapped QTL explaining variation in the wing display angle (Figure 3) rather than the presence or absence of wing display, some QTL identified here might also contribute to variation in wing display angle segregating within wild populations. We therefore conclude that although the absence of wing spots appears fixed in D. gunungcola, the absence of wing display behavior does not. These observations suggest that the loss of male-specific wing spots predates the loss of male wing display behavior in this species.

Conclusions

Male-specific wing spots and wing display behavior have co-evolved in Drosophila multiple times (Kopp and True, 2002). By studying the genetic basis of these divergent traits between D. elegans and D. gunungcola, we showed that the changes in wing spot and wing display were not caused by changes in a single, pleiotropic gene despite overlapping QTL (Yeh et al., 2006; Yeh and True, 2014). Rather, we found that distinct loci contribute to divergence in each of these traits, with the genetic architecture of divergent wing behavior being more complex than that of the divergent wing spot pigmentation. Both traits were affected by divergent gene(s) located on the X chromosome that are in physical linkage, however, causing alleles of these distinct loci to be co-inherited. This linkage might have facilitated the coordinated evolution of these traits.

The specific genes contributing to divergence in wing spot and wing display remain unknown, but optomotor-blind is a strong candidate for the X-linked gene contributing to the loss of the wing spot. Introgression lines and additional sampling of D. gunungcola from a wild population also showed that the loss of wing spots and wing display are not inexorably linked: in both cases, males lacking wing spots still performed a wing display behavior. Coordinated evolution of morphological and behavioral traits such as these is often observed in animal species, but it is often unclear which change evolved first. In this case at least, it seems that the divergence of morphology preceded the divergence of behavior.

Supplementary Material

Video 1

Video 1 D. elegans HK wing display behavior

Download video file (245.9KB, mp4)
Video 2

Video 2 D. gunungcola SK courtship and copulation

Download video file (1.1MB, mp4)
Video 5

Video 5 Introgression 1 wing display behavior

Download video file (197.4KB, mp4)
Video 6

Video 6 Introgression 2 wing display behavior

Download video file (81.9KB, mp4)
Video 3

Video 3 F1E wing display behavior

Download video file (2.1MB, mp4)
Video 4

Video 4 F1G wing display behavior

Download video file (2.3MB, mp4)
Video 7

Video 7 Introgression 3 wing display behavior

Download video file (64.7KB, mp4)
Video 8

Video 8 D. gunungcola wing display behavior at Coban Rondo Waterfall in East Java, Indonesia (Version 1)

Download video file (1MB, mp4)
Video 9

Video 9 D. gunungcola wing display behavior at Coban Rondo Waterfall in East Java, Indonesia (Version 2)

Download video file (493.1KB, mp4)
Video 11

Video 11 D. gunungcola (Batu City, Indonesia) wing display behavior in the laboratory

Download video file (244.5KB, mp4)
Video 10

Video 10 D. elegans wing display behavior in Tumpang, Indonesia

Download video file (2.1MB, mp4)
Supplementary Figure 1

Supplementary Figure S1 ImageJ procedure for measuring maximum wing display angles Screenshots of each wing display were captured for every recombinant courtship video. The maximum wing display bout was identified for each fly by quickly comparing screenshots that varied in wing display angles (from wing tip to wing tip) and picking by eye the display with the largest angle. Next, for each fly, the maximum wing display angle was quantified in ImageJ by using 1) Find Edges function, 2) polygon tool to Fit Ellipse around the fly body, 3) Ellipse Macros (Supplementary File S1) to fit the major and minor axes of the ellipse, and 4) draw Angle tool, fitting the angle vertex at the major and minor axes intersection to calculate the wing display angle from wing tip to wing tip.

Supplementary Figure 2

Supplementary Figure S2 LOD scores estimated from a two-dimensional, two QTL scan of maximum wing display angles (A) For the D. elegans backcross, the Interaction LODi, which estimates the likelihood that the effect of genotypes at one marker depend on genotypes at another, is displayed in the upper left triangle; the Full LODf, which estimates the effect of both additive and non-additive interactions between genotypes (see Supplementary Table S1 for LOD thresholds), is displayed in the lower right triangle (Broman and Sen, 2009). The color scale on the right indicates LOD values for LODi (left) and LODf (right). (B) For the D. gunungcola backcross, the Interaction LODi is displayed in the upper left triangle; the Full LODf (see Supplementary Table S2 for LOD thresholds) is displayed in the lower right triangle. The color scale on the right indicates LOD values for LODi (left) and LODf (right).

Supplementary Figure 5

Supplementary Figure S5 Fine-mapping the wing spot locus D. elegans and D. gunungcola backcross recombinants containing X chromosome breakpoints immediately flanking the wing spot QTL peak were aligned to compare the effects of each on wing pigmentation. Regions in red represent D. elegans linked loci, and regions in blue represent D. gunungcola linked loci. All recombinants possessing D. gunungcola loci to the right of ~10.95 Mbp are spotless.

Supplementary Figure 6

Supplementary Figure S6 Effects of the yellow gene on wing spot size and wing display behavior in D. elegans (A) Loss-of-function D. elegans HK yellow mutants develop smaller wing spots than D. elegans HK wild-type males (Student’s t-test; t = 4.7759; df = 15.28; P = 0.0002; two-tailed) and (B) show lower maximum wing display angles (Student’s t-test; t = 3.0294; df = 50.82; P = 0.004; two-tailed).

Supplementary Figure 3

Supplementary Figure S3 In situ hybridization of D. elegans and D. gunungcola L3 wing discs Male D. elegans (left) and D. gunungcola (right) L3 wing discs were dissected and stained with probes targeting omb mRNA.

Supplementary Figure 4

Supplementary Figure S4 Effects of Muller Element E on wing spot divergence (A) Wing spot QTL map for D. elegans (red) and D. gunungcola (blue) backcross recombinants. Note, all recombinant individuals that lacked wing spots were removed from this QTL analysis to identify loci contributing to wing spot size variation independent of wing spot presence or absence. LOD (logarithm of the odds) is indicated on the y-axis. The x-axis represents the physical map of Muller Elements X, B, C, D, E, and F based on the D. elegans assembled genome (see Methods). Individual SNP markers are indicated with black tick marks along the x-axis. Horizontal red and blue lines mark P = 0.01 for the D. elegans and D. gunungcola backcross, respectively. (B) Images illustrating D. elegans and D. gunungcola body color differences. (C) Multiplexed Shotgun Genotyping (MSG) (Andolfatto et al., 2011) was used to estimate genome-wide ancestry assignments for a single introgression line generated by repeatedly backcrossing D. gunungcola into a D. elegans genetic background (see Methods). The posterior probability that a region is homozygous for D. elegans (red) or D. gunungcola (blue) ancestry is plotted along the y-axis. (D) Representative wing spot images of D. elegans and D. gunungcola species parents, the introgression line genotyped in (B), and an F1 heterozygote generated by crossing D. elegans females to introgression males. (E) Quantification of wing spot size differences between each genotype. Results of Tukey HSD post hoc tests following one-way ANOVA are shown (One-way ANOVA F2,88 = 78.6; P < 2.0 × 10−16; post-hoc Tukey HSD was significant between D. elegans and Introgression: P < 1.0 × 10−7, D. elegans and F1 Inrogression/D. elegans heterozygote: P = 0.02, and Introgression and Inrogression/D. elegans heterozygote: P < 1.0 × 10−7. Gray triangles represent individual replicates.

Supplementary Figure 9

Supplementary Figure S9 Comparison of genital morphology between D. elegans and D. gunungcola strains Male genitalia were dissected and imaged for comparison between D. elegans and D. gunungcola strains. The medial region of D. gunungcola clasper exhibits two groups of black teeth, which consist of three claw-like teeth and two short, pointy teeth. D. elegans has two groups of 2–3 short black rounded teeth at the medial margin of its clasper. The anal plate of both species has a concave medial margin. One difference between both species is visible on the anal plate’s ventromedial tooth: The corner of the D. gunungcola ventromedial tooth exhibits a sharp angle (>90°), while D. elegans exhibits a smoother curve. The parameres of both species are elongated, however D. gunungcola’s parameres are round at the posterior end and finely serrated, while the posterior end of D. elegans’ parameres are curved and pointy. The aedeagus of both species is expanded at the apical end, narrowed at the mid region, and broadened at the basal end.

Supplementary Figure 10

Supplementary Figure S10 omb protein alignment between D. elegans and D. gunungcola strains Genomic DNA isolated from field-collected D. gunungcola from Coban Rondo, Indonesia was sequenced to compare with the D. gunungcola SK and D. elegans HK laboratory strains (see Methods - Library preparation and sequencing). We were only able to recover part of the omb coding sequence from our raw reads. A nonsynonymous coding change (asterisk) that distinguished the laboratory D. gunungcola SK strain from D. elegans HK also distinguished the new D. gunungcola Coban Rondo strain from D. elegans HK, matching the D. gunungcola SK sequence.

Supplementary Figure 11

Supplementary Figure S11 Comparison of D. elegans and D. gunungcola wing display behavior in the field and laboratory On flowers in the field, D. elegans males from Taiwan and Indonesia perform wing displays qualitatively and quantitatively similar to the D. elegans HK laboratory strain (Figure 1B,E). We recorded D. gunungcola performing wing displays for the first time on flowers in Indonesia, and these displays appear to be innate given their performance in a laboratory environment without any previous social experience. Measurements of wing display angles, however, suggest that newly collected D. gunungcola from Coban Rondo cannot perform wing displays as wide as D. elegans display angles. Instead, D. gunungcola wing displays look qualitatively similar to F1G hybrid displays (Figure 1C), although these flies showed no other signs of being hybrids with D. gunungcola.

Supplementary Figure 8

Supplementary Figure S8 In situ hybridization of D. elegans and D. gunungcola pupal wings probed for omb mRNA omb mRNA (purple) was probed at 30 and 48 h after pupal formation (APF) in both males and females.

Supplementary Figure 7

Supplementary Figure S7 Male wings from new D. gunungcola isolates Male D. gunungcola from five newly collected isofemale lines from Coban Rondo, Indonesia do not develop wing spots.

Supplementary File 1
Supplementary File 3
Supplementary File 2
Supplementary File 4
Supplementary File 5
Supplementary Table 2
Supplementary Table 1
Supplementary Table 3

Acknowledgements

We thank members of the Wittkopp, Stern, Yeh, and Rebeiz labs for helpful discussions. For fly strains, we thank John True (Stony Brook University). For guidance throughout the in situ hybridization work, we thank Mark Rebeiz (University of Pittsburgh). For arranging the Material Transfer Agreement for D. gunungcola and D. elegans, we thank Nia Kurniawan (University of Brawijaya, Indonesia); for hosting us in Indonesia, we thank Karuniawan Wicaksono (University of Brawijaya, Indonesia); for assistance with field collections, we thank Hagus Tarno (University of Brawijaya, Indonesia). Funding: University of Michigan, Department of Ecology and Evolutionary Biology, Peter Olaus Okkelberg Research Award, National Institutes of Health (NIH) training grant T32GM007544, and Howard Hughes Medical Institute Janelia Graduate Research Fellowship to J.H.M.; NIH R01 GM089736 and 1R35GM118073 to PJW.

References

  1. Andolfatto P, Davison D, Erezyilmaz D, Hu TT, Mast J, Sunayama-Morita T, & Stern DL (2011). Multiplexed shotgun genotyping for rapid and efficient genetic mapping. Genome Research, 21(4), 610–617. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Álamo Rodríguez D, Felix JT, & Díaz-Benjumea FJ (2004). The role of the T-box gene optomotor-blind in patterning the Drosophila wing. Developmental Biology, 268(2), 481–492. [DOI] [PubMed] [Google Scholar]
  3. Arnoult L, Su KF, Manoel D, Minervino C, Magriña J, Gompel N, & Prud’homme B (2013). Emergence and diversification of fly pigmentation through evolution of a gene regulatory module. Science, 339(6126), 1423–1426. [DOI] [PubMed] [Google Scholar]
  4. Bock IR, and Wheeler MR (1972) The Drosophila melanogaster species group. University of Texas Publication 7213, 1–102. [Google Scholar]
  5. Brisson JA, Templeton AR, & Duncan I (2004). Population genetics of the developmental gene optomotor-blind (omb) in Drosophila polymorpha: evidence for a role in abdominal pigmentation variation. Genetics, 168(4), 1999–2010. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Broman KW, Wu H, Sen S, Churchill GA (2003) R/qtl: QTL mapping in experimental crosses. Bioinformatics, 19: 889–890. [DOI] [PubMed] [Google Scholar]
  7. Broman KW, & Sen S (2009). A Guide to QTL Mapping with R/qtl (Vol. 46). New York: Springer. [Google Scholar]
  8. Cande J, Andolfatto P, Prud’homme B, Stern DL, & Gompel N (2012). Evolution of multiple additive loci caused divergence between Drosophila yakuba and D. santomea in wing rowing during male courtship. PLoS One, 7(8), e43888. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Chang SH, Jobling S, Brennan K, & Headon DJ (2009). Enhanced Edar signalling has pleiotropic effects on craniofacial and cutaneous glands. PloS one, 4(10). [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Charlesworth D, & Charlesworth B (1976). Theoretical genetics of Batesian mimicry II. Evolution of supergenes. Journal of Theoretical Biology, 55(2), 305–324. [DOI] [PubMed] [Google Scholar]
  11. Duveau F, & Félix MA (2012). Role of pleiotropy in the evolution of a cryptic developmental variation in Caenorhabditis elegans. PLoS biology, 10(1). [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Endler JA (1991). Variation in the appearance of guppy color patterns to guppies and their predators under different visual conditions. Vision Research, 31(3), 587–608. [DOI] [PubMed] [Google Scholar]
  13. Endler L, Gibert JM, Nolte V, & Schlötterer C (2018). Pleiotropic effects of regulatory variation in tan result in correlation of two pigmentation traits in Drosophila melanogaster. Molecular ecology, 27(16), 3207–3218. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Gompel N, Prud’homme B, Wittkopp PJ, Kassner VA, & Carroll SB (2005). Chance caught on the wing: cis-regulatory evolution and the origin of pigment patterns in Drosophila. Nature, 433(7025), 481. [DOI] [PubMed] [Google Scholar]
  15. Gray SM, & McKinnon JS (2007). Linking color polymorphism maintenance and speciation. Trends in Ecology & Evolution, 22(2), 71–79. [DOI] [PubMed] [Google Scholar]
  16. Grimm S and Pflugfelder GO (1996). Control of the gene optomotor-blind in Drosophila wing development by decapentaplegic and wingless. Science, 271, 1601–04. [DOI] [PubMed] [Google Scholar]
  17. Haley CS, Knott SA (1992). A simple regression method for mapping quantitative trait loci in line crosses using flanking markers. Heredity, 69: 315–324. [DOI] [PubMed] [Google Scholar]
  18. Johnson M, Zaretskaya I, Raytselis Y, Merezhuk Y, McGinnis S, & Madden TL (2008). NCBI BLAST: a better web interface. Nucleic Acids Research, 36(suppl_2), W5–W9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Kent CF, Daskalchuk T, Cook L, Sokolowski MB, & Greenspan RJ (2009). The Drosophila foraging gene mediates adult plasticity and gene–environment interactions in behaviour, metabolites, and gene expression in response to food deprivation. PLoS Genetics, 5(8). [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Kirkpatrick M, & Barton N (2006). Chromosome inversions, local adaptation and speciation. Genetics, 173(1), 419–434. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Kopp A, & Duncan I (1997). Control of cell fate and polarity in the adult abdominal segments of Drosophila by optomotor-blind. Development, 124(19), 3715–3726. [DOI] [PubMed] [Google Scholar]
  22. Kopp A, & True JR (2002). Evolution of male sexual characters in the oriental Drosophila melanogaster species group. Evolution & Development, 4(4), 278–291. [DOI] [PubMed] [Google Scholar]
  23. Kronforst MR, Young LG, Kapan DD, McNeely C, O’Neill RJ, & Gilbert LE (2006). Linkage of butterfly mate preference and wing color preference cue at the genomic location of wingless. Proceedings of the National Academy of Sciences, 103(17), 6575–6580. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Küpper C, Stocks M, Risse JE, dos Remedios N, Farrell LL, McRae SB, … & Kitaysky AS (2016). A supergene determines highly divergent male reproductive morphs in the ruff. Nature Genetics, 48(1), 79. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Lamichhaney S, Fan G, Widemo F, Gunnarsson U, Thalmann DS, Hoeppner MP, … & Chen W (2016). Structural genomic changes underlie alternative reproductive strategies in the ruff (Philomachus pugnax). Nature Genetics, 48(1), 84. [DOI] [PubMed] [Google Scholar]
  26. Lindholm A, & Breden F (2002). Sex chromosomes and sexual selection in poeciliid fishes. The American Naturalist, 160(S6), S214–S224. [DOI] [PubMed] [Google Scholar]
  27. Loxton RG (1979). On display behaviour and courtship in the praying mantis Ephestiasula amoena (Bolivar). Zoological Journal of the Linnean Society, 65(1), 103–110. [Google Scholar]
  28. McKinnon JS, & Pierotti ME (2010). Colour polymorphism and correlated characters: genetic mechanisms and evolution. Molecular Ecology, 19(23), 5101–5125. [DOI] [PubMed] [Google Scholar]
  29. McLean CY, Reno PL, Pollen AA, Bassan AI, Capellini TD, Guenther C, … & Wenger AM (2011). Human-specific loss of regulatory DNA and the evolution of human-specific traits. Nature, 471(7337), 216–219. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Merrill RM, Rastas P, Martin SH, Melo MC, Barker S, Davey J, … & Jiggins CD (2019). Genetic dissection of assortative mating behavior. PLoS Biology, 17(2), e2005902. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Nagy O, Nuez I, Savisaar R, Peluffo AE, Yassin A, Lang M, … & Courtier-Orgogozo V (2018). Correlated evolution of two copulatory organs via a single cis-regulatory nucleotide change. Current Biology, 28(21), 3450–3457. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Pflugfelder GO, Roth H and Poeck B (1992a). A homology domain shared between Drosophila optomotor-blind and mouse Brachyury is involved in DNA binding. Biochem. Biophys. Res. Comm, 186, 918–25. [DOI] [PubMed] [Google Scholar]
  33. Pflugfelder GO, Roth H, Poeck B, Kersher S, Schwarz H, Jonschker B and Heisenberg M (1992b). The lethal (1) optomotor-blind gene of Drosophila melanogaster is a major organizer of optic lobe development: Isolation and characterization of the gene. Proc. Nat. Acad. Sci, 89, 1199–1203. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Pflugfelder GO (2009). omb and circumstance. Journal of neurogenetics, 23(1–2), 15–33. [DOI] [PubMed] [Google Scholar]
  35. Picelli S, Björklund ÅK, Reinius B, Sagasser S, Winberg G, & Sandberg R (2014). Tn5 transposase and tagmentation procedures for massively scaled sequencing projects. Genome Research, 24(12), 2033–2040. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Prud’homme B, Gompel N, Rokas A, Kassner VA, Williams TM, Yeh SD, … & Carroll SB (2006). Repeated morphological evolution through cis-regulatory changes in a pleiotropic gene. Nature, 440(7087), 1050–1053. [DOI] [PubMed] [Google Scholar]
  37. R Core Team. 2013. R: A Language and Environment for Statistical Computing. Available from: http://www.r-project.org/.
  38. Rice G, David JR, Kamimura Y, Masly JP, Mcgregor AP, Nagy O, … & Siegal ML (2019). A standardized nomenclature and atlas of the male terminalia of Drosophila melanogaster. Fly, 13(1–4), 51–64. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Simon JC, & Dickinson MH (2010). A new chamber for studying the behavior of Drosophila. Plos One, 5(1), e8793. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Sinervo B, Miles DB, Frankino WA, Klukowski M, & DeNardo DF (2000). Testosterone, endurance, and Darwinian fitness: natural and sexual selection on the physiological bases of alternative male behaviors in side-blotched lizards. Hormones and Behavior, 38(4), 222–233. [DOI] [PubMed] [Google Scholar]
  41. Stern DL, & Orgogozo V (2008). The loci of evolution: how predictable is genetic evolution?. Evolution, 62(9), 2155–2177. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Sultana F, Kimura MT, & Toda MJ (1999). Anthophilic Drosophila of the elegans species-subgroup from Indonesia, with description of a new species (Diptera: Drosophilidae). Entomological Science, 2, 121–126. [Google Scholar]
  43. Thomas JW, Cáceres M, Lowman JJ, Morehouse CB, Short ME, Baldwin EL, … & Martin CL (2008). The chromosomal polymorphism linked to variation in social behavior in the white-throated sparrow (Zonotrichia albicollis) is a complex rearrangement and suppressor of recombination. Genetics, 179(3), 1455–1468. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Thompson PE (1959). Drosophila. Info. Service 33, 99 [Google Scholar]
  45. Thurmond J, Goodman JL, Strelets VB, Attrill H, Gramates LS, Marygold SJ, … & Kaufman TC (2018). FlyBase 2.0: the next generation. Nucleic Acids Research, 47(D1), D759–D765. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Vincent BJ, Rice GR, Wong GM, Glassford WJ, Downs KI, Shastay JL, … & Rebeiz M (2019). An atlas of transcription factors expressed in male pupal terminalia of Drosophila melanogaster. G3: Genes, Genomes, Genetics, g3–400788. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Werner T, Koshikawa S, Williams TM, & Carroll SB (2010). Generation of a novel wing colour pattern by the Wingless morphogen. Nature, 464(7292), 1143. [DOI] [PubMed] [Google Scholar]
  48. Wirtz RA and Semey HG (1982). The Drosophila kitchen - equipment, media preparation, and supplies. Drosophila Information Service. 58, 176–180. [Google Scholar]
  49. White TE, Zeil J, & Kemp DJ (2015). Signal design and courtship presentation coincide for highly biased delivery of an iridescent butterfly mating signal. Evolution, 69(1), 14–25. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Wittkopp PJ, Vaccaro K, & Carroll SB (2002a). Evolution of yellow gene regulation and pigmentation in Drosophila. Current Biology, 12(18), 1547–1556. [DOI] [PubMed] [Google Scholar]
  51. Wittkopp PJ, True JR, & Carroll SB (2002b). Reciprocal functions of the Drosophila yellow and ebony proteins in the development and evolution of pigment patterns. Development, 129(8), 1849–1858. [DOI] [PubMed] [Google Scholar]
  52. Yeh SD, Liou SR, & True JR (2006). Genetics of divergence in male wing pigmentation and courtship behavior between Drosophila elegans and D. gunungcola. Heredity, 96(5), 383. [DOI] [PubMed] [Google Scholar]
  53. Yeh SD, & True JR (2014). The genetic architecture of coordinately evolving male wing pigmentation and courtship behavior in Drosophila elegans and Drosophila gunungcola. G3: Genes, Genomes, Genetics, 4(11), 2079–2093. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Video 1

Video 1 D. elegans HK wing display behavior

Download video file (245.9KB, mp4)
Video 2

Video 2 D. gunungcola SK courtship and copulation

Download video file (1.1MB, mp4)
Video 5

Video 5 Introgression 1 wing display behavior

Download video file (197.4KB, mp4)
Video 6

Video 6 Introgression 2 wing display behavior

Download video file (81.9KB, mp4)
Video 3

Video 3 F1E wing display behavior

Download video file (2.1MB, mp4)
Video 4

Video 4 F1G wing display behavior

Download video file (2.3MB, mp4)
Video 7

Video 7 Introgression 3 wing display behavior

Download video file (64.7KB, mp4)
Video 8

Video 8 D. gunungcola wing display behavior at Coban Rondo Waterfall in East Java, Indonesia (Version 1)

Download video file (1MB, mp4)
Video 9

Video 9 D. gunungcola wing display behavior at Coban Rondo Waterfall in East Java, Indonesia (Version 2)

Download video file (493.1KB, mp4)
Video 11

Video 11 D. gunungcola (Batu City, Indonesia) wing display behavior in the laboratory

Download video file (244.5KB, mp4)
Video 10

Video 10 D. elegans wing display behavior in Tumpang, Indonesia

Download video file (2.1MB, mp4)
Supplementary Figure 1

Supplementary Figure S1 ImageJ procedure for measuring maximum wing display angles Screenshots of each wing display were captured for every recombinant courtship video. The maximum wing display bout was identified for each fly by quickly comparing screenshots that varied in wing display angles (from wing tip to wing tip) and picking by eye the display with the largest angle. Next, for each fly, the maximum wing display angle was quantified in ImageJ by using 1) Find Edges function, 2) polygon tool to Fit Ellipse around the fly body, 3) Ellipse Macros (Supplementary File S1) to fit the major and minor axes of the ellipse, and 4) draw Angle tool, fitting the angle vertex at the major and minor axes intersection to calculate the wing display angle from wing tip to wing tip.

NIHMS1624996-supplement-Supplementary_Figure_1.pdf (257.8KB, pdf)
Supplementary Figure 2

Supplementary Figure S2 LOD scores estimated from a two-dimensional, two QTL scan of maximum wing display angles (A) For the D. elegans backcross, the Interaction LODi, which estimates the likelihood that the effect of genotypes at one marker depend on genotypes at another, is displayed in the upper left triangle; the Full LODf, which estimates the effect of both additive and non-additive interactions between genotypes (see Supplementary Table S1 for LOD thresholds), is displayed in the lower right triangle (Broman and Sen, 2009). The color scale on the right indicates LOD values for LODi (left) and LODf (right). (B) For the D. gunungcola backcross, the Interaction LODi is displayed in the upper left triangle; the Full LODf (see Supplementary Table S2 for LOD thresholds) is displayed in the lower right triangle. The color scale on the right indicates LOD values for LODi (left) and LODf (right).

NIHMS1624996-supplement-Supplementary_Figure_2.pdf (1.5MB, pdf)
Supplementary Figure 5

Supplementary Figure S5 Fine-mapping the wing spot locus D. elegans and D. gunungcola backcross recombinants containing X chromosome breakpoints immediately flanking the wing spot QTL peak were aligned to compare the effects of each on wing pigmentation. Regions in red represent D. elegans linked loci, and regions in blue represent D. gunungcola linked loci. All recombinants possessing D. gunungcola loci to the right of ~10.95 Mbp are spotless.

NIHMS1624996-supplement-Supplementary_Figure_5.pdf (2.5MB, pdf)
Supplementary Figure 6

Supplementary Figure S6 Effects of the yellow gene on wing spot size and wing display behavior in D. elegans (A) Loss-of-function D. elegans HK yellow mutants develop smaller wing spots than D. elegans HK wild-type males (Student’s t-test; t = 4.7759; df = 15.28; P = 0.0002; two-tailed) and (B) show lower maximum wing display angles (Student’s t-test; t = 3.0294; df = 50.82; P = 0.004; two-tailed).

NIHMS1624996-supplement-Supplementary_Figure_6.pdf (3.1MB, pdf)
Supplementary Figure 3

Supplementary Figure S3 In situ hybridization of D. elegans and D. gunungcola L3 wing discs Male D. elegans (left) and D. gunungcola (right) L3 wing discs were dissected and stained with probes targeting omb mRNA.

NIHMS1624996-supplement-Supplementary_Figure_3.pdf (8.9MB, pdf)
Supplementary Figure 4

Supplementary Figure S4 Effects of Muller Element E on wing spot divergence (A) Wing spot QTL map for D. elegans (red) and D. gunungcola (blue) backcross recombinants. Note, all recombinant individuals that lacked wing spots were removed from this QTL analysis to identify loci contributing to wing spot size variation independent of wing spot presence or absence. LOD (logarithm of the odds) is indicated on the y-axis. The x-axis represents the physical map of Muller Elements X, B, C, D, E, and F based on the D. elegans assembled genome (see Methods). Individual SNP markers are indicated with black tick marks along the x-axis. Horizontal red and blue lines mark P = 0.01 for the D. elegans and D. gunungcola backcross, respectively. (B) Images illustrating D. elegans and D. gunungcola body color differences. (C) Multiplexed Shotgun Genotyping (MSG) (Andolfatto et al., 2011) was used to estimate genome-wide ancestry assignments for a single introgression line generated by repeatedly backcrossing D. gunungcola into a D. elegans genetic background (see Methods). The posterior probability that a region is homozygous for D. elegans (red) or D. gunungcola (blue) ancestry is plotted along the y-axis. (D) Representative wing spot images of D. elegans and D. gunungcola species parents, the introgression line genotyped in (B), and an F1 heterozygote generated by crossing D. elegans females to introgression males. (E) Quantification of wing spot size differences between each genotype. Results of Tukey HSD post hoc tests following one-way ANOVA are shown (One-way ANOVA F2,88 = 78.6; P < 2.0 × 10−16; post-hoc Tukey HSD was significant between D. elegans and Introgression: P < 1.0 × 10−7, D. elegans and F1 Inrogression/D. elegans heterozygote: P = 0.02, and Introgression and Inrogression/D. elegans heterozygote: P < 1.0 × 10−7. Gray triangles represent individual replicates.

NIHMS1624996-supplement-Supplementary_Figure_4.pdf (13.5MB, pdf)
Supplementary Figure 9

Supplementary Figure S9 Comparison of genital morphology between D. elegans and D. gunungcola strains Male genitalia were dissected and imaged for comparison between D. elegans and D. gunungcola strains. The medial region of D. gunungcola clasper exhibits two groups of black teeth, which consist of three claw-like teeth and two short, pointy teeth. D. elegans has two groups of 2–3 short black rounded teeth at the medial margin of its clasper. The anal plate of both species has a concave medial margin. One difference between both species is visible on the anal plate’s ventromedial tooth: The corner of the D. gunungcola ventromedial tooth exhibits a sharp angle (>90°), while D. elegans exhibits a smoother curve. The parameres of both species are elongated, however D. gunungcola’s parameres are round at the posterior end and finely serrated, while the posterior end of D. elegans’ parameres are curved and pointy. The aedeagus of both species is expanded at the apical end, narrowed at the mid region, and broadened at the basal end.

NIHMS1624996-supplement-Supplementary_Figure_9.pdf (5.4MB, pdf)
Supplementary Figure 10

Supplementary Figure S10 omb protein alignment between D. elegans and D. gunungcola strains Genomic DNA isolated from field-collected D. gunungcola from Coban Rondo, Indonesia was sequenced to compare with the D. gunungcola SK and D. elegans HK laboratory strains (see Methods - Library preparation and sequencing). We were only able to recover part of the omb coding sequence from our raw reads. A nonsynonymous coding change (asterisk) that distinguished the laboratory D. gunungcola SK strain from D. elegans HK also distinguished the new D. gunungcola Coban Rondo strain from D. elegans HK, matching the D. gunungcola SK sequence.

NIHMS1624996-supplement-Supplementary_Figure_10.pdf (332.8KB, pdf)
Supplementary Figure 11

Supplementary Figure S11 Comparison of D. elegans and D. gunungcola wing display behavior in the field and laboratory On flowers in the field, D. elegans males from Taiwan and Indonesia perform wing displays qualitatively and quantitatively similar to the D. elegans HK laboratory strain (Figure 1B,E). We recorded D. gunungcola performing wing displays for the first time on flowers in Indonesia, and these displays appear to be innate given their performance in a laboratory environment without any previous social experience. Measurements of wing display angles, however, suggest that newly collected D. gunungcola from Coban Rondo cannot perform wing displays as wide as D. elegans display angles. Instead, D. gunungcola wing displays look qualitatively similar to F1G hybrid displays (Figure 1C), although these flies showed no other signs of being hybrids with D. gunungcola.

NIHMS1624996-supplement-Supplementary_Figure_11.pdf (899.1KB, pdf)
Supplementary Figure 8

Supplementary Figure S8 In situ hybridization of D. elegans and D. gunungcola pupal wings probed for omb mRNA omb mRNA (purple) was probed at 30 and 48 h after pupal formation (APF) in both males and females.

NIHMS1624996-supplement-Supplementary_Figure_8.pdf (10.1MB, pdf)
Supplementary Figure 7

Supplementary Figure S7 Male wings from new D. gunungcola isolates Male D. gunungcola from five newly collected isofemale lines from Coban Rondo, Indonesia do not develop wing spots.

NIHMS1624996-supplement-Supplementary_Figure_7.pdf (18.2MB, pdf)
Supplementary File 1
NIHMS1624996-supplement-Supplementary_File_1.txt (1.2KB, txt)
Supplementary File 3
NIHMS1624996-supplement-Supplementary_File_3.csv (2KB, csv)
Supplementary File 2
NIHMS1624996-supplement-Supplementary_File_2.csv (1.9KB, csv)
Supplementary File 4
NIHMS1624996-supplement-Supplementary_File_4.txt (37.3KB, txt)
Supplementary File 5
NIHMS1624996-supplement-Supplementary_File_5.docx (13.9KB, docx)
Supplementary Table 2
NIHMS1624996-supplement-Supplementary_Table_2.docx (16.6KB, docx)
Supplementary Table 1
NIHMS1624996-supplement-Supplementary_Table_1.docx (16.2KB, docx)
Supplementary Table 3
NIHMS1624996-supplement-Supplementary_Table_3.docx (13KB, docx)

RESOURCES