Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2019 Dec 17.
Published in final edited form as: Curr Biol. 2018 Nov 29;28(24):3969–3975.e3. doi: 10.1016/j.cub.2018.10.036

Tissue-specific cis-regulatory divergence implicates eloF in inhibiting interspecies mating in Drosophila

Peter A Combs 1, Joshua J Krupp 2, Neil M Khosla 1, Dennis Bua 1, Dmitri A Petrov 1, Joel D Levine 2, Hunter B Fraser 1
PMCID: PMC6322677  NIHMSID: NIHMS1511895  PMID: 30503619

Summary

Reproductive isolation is a key component of speciation. In many insects, a major driver of this isolation is cuticular hydrocarbon pheromones, which help to identify potential intraspecific mates [1–3]. When the distributions of related species overlap, there may be strong selection on mate choice for intraspecific partners [4–9], since interspecific hybridization carries significant fitness costs [10]. Drosophila has been a key model for the investigation of reproductive isolation; while both male and female mate choice have been extensively investigated [6,11–16], the genes underlying species recognition remain largely unknown. To explore the molecular mechanisms underlying Drosophila speciation, we measured tissue-specific cis-regulatory divergence using RNA-seq in D. simulans × D. sechellia hybrids. By focusing on cis-regulatory changes specific to female oenocytes, the tissue that produces cuticular hydrocarbons, we rapidly identified a small number of candidate genes. We found that one of these, the fatty acid elongase eloF, broadly affects the hydrocarbons present on D. sechellia and D. melanogaster females as well as the propensity of D. simulans males to mate with them. Therefore, cis-regulatory changes in eloF may be a major driver in the sexual isolation of D. simulans from multiple other species. Our RNA-seq approach proved to be far more efficient than QTL mapping in identifying candidate genes; the same framework can be used to pinpoint candidate drivers of cis-regulatory divergence in traits differing between any interfertile species.

Results and Discussion

D. simulans and D. sechellia are closely related sister species, which have accumulated ~0.018 substitutions per site in the several hundred thousand years since they last shared a common ancestor [17,18]. They are believed to have diverged in allopatry [19], though currently their ranges overlap and hybrids can be found in the wild [20]. In laboratory conditions, D. sechellia males will readily mate with D. simulans females, producing sterile male and fertile female hybrid offspring, while in the reciprocal cross D. simulans males will not readily mate with D. sechellia females [21]. Male mate choice has been estimated to account for over 70% of the reproductive isolation between these species [16], but the gene(s) accounting for this isolation are unknown. This divergence in male mate choice likely involves species-specific differences in female CHCs, as well as in the male response to these CHCs; in this work we focused on female CHCs.

Allele-specific expression identifies fatty acid elongases as a major differentiator between D. simulans and D. sechellia female oenocytes

While QTLs affecting CHCs have been mapped, these contain many CHC-related genes [22–24], and fine-mapping has not been reported. As a complementary approach, we reasoned that genes responsible for major changes in female CHCs may share three key characteristics: 1) Cis-regulatory divergence in female oenocytes; 2) Female-specific expression; and 3) Oenocyte-specific expression. Although these are not required, any genes meeting all three criteria would be promising candidates.

Cis-regulatory divergence can be measured genome-wide via high-throughput sequencing of cDNA (RNA-seq) in interspecific hybrids. Hybrids are required because comparisons between species reflect both cis- and trans-acting changes; measuring allele-specific expression (ASE) in F1 hybrids controls for trans-acting changes, since each allele experiences the same trans-regulatory environment within the hybrid nuclei [25]. Thus, differential expression of the two alleles in a hybrid can only be explained by cis-regulatory divergence.

In order to identify genes with cis-regulatory divergence specific to female oenocytes, we constructed RNA-seq libraries from hybrid D. simulans × D. sechellia tissues (Figure 1A, see Methods). To measure female-specificity, we included samples from both male and female oenocytes, and to measure oenocyte-specificity, we included samples from male and female fat bodies (an adjacent non-CHC producing tissue) [26]. We estimated the significance of each gene’s allele specific expression (ASE) using a negative-binomial test [27] for deviation from the average fraction of D. sechellia reads in a given sample.

Figure 1: RNA-seq of oenocytes and fat bodies from hybrid D. simulans × D. sechellia flies reveals a strong cis-regulatory component of CHC production.

Figure 1:

Open in a new tab

A) We dissected oenocytes (brown) and fat bodies (pink) from hybrid D. simulans × D. sechellia males and females and performed RNA-sequencing.

B) Genes are plotted by specificity of expression to female oenocytes (x-axis; mean of female oenocyte expression divided by maximum expression in female fat bodies, male oenocytes, and female oenocytes) and allele-specific expression p-value (y-axis). Green dots indicate genes with significant ASE compared to the distribution of reads in the female oenocytes, blue dots indicate those that have significantly higher expression in female oenocytes compared to female fat bodies and male oenocytes, and red dots indicate genes with both tissue-specific and species-specific expression.

C) Overlap of genes with ASE in female oenocytes (green circle), and differential expression in female oenocytes compared to other tissues (blue and cyan circles).

See also Figure SI, Table S1

Even at a stringent cutoff, we identified 239 genes with significant (negative binomial q-value < 0.001) ASE in female oenocytes. This is not surprising, since various Drosophila interspecific hybrids have also yielded large numbers of genes with strong ASE [28,29]. Of the 239 significant genes, 27 were annotated with the Gene Ontology term “Fatty acid biosynthetic process” (GO:0006633). Therefore we concluded that, even when combined with GO annotations, ASE in female oenocytes was insufficient to identify a manageable number of CHC-related candidate genes.

We reasoned that in addition to ASE, genes important to female CHC differences between D. simulans and D. sechellia would likely be expressed specifically in female oenocytes (Figure 1B and C). To identify candidate genes, we looked for genes that had significantly higher expression in the female oenocytes compared to male oenocytes and to female fat bodies (Sleuth q-value<0.001 for both comparisons) [30]. Only six genes passed these cutoffs (Table 1). Reassuringly, one of these was desatF (also known as Fad2), a fatty-acid desaturase which is known to be expressed in D. sechellia female oenocytes, but not in males or in D. simulans [31]; another was eloF, which has been previously observed to lack expression in D. simulans, although its expression in D. sechellia has not been studied [32].

Table 1: Genes with female oenocyte- and species-specific expression.

Genes with significant tissue-specific (sleuth q-value <0.001 in comparisons both between the two female tissues, and between the two oenocyte samples) and species-specific expression (negative binomial q-value < .001). Specificity is the ratio of the mean expression in female oenocytes to the highest expression among male oenocytes, female fat bodies, and male fat bodies. Gene ontology (GO) terms are annotated molecular function terms (see Table S2 for citations). GO terms without experimental evidence are in italics. Protein domains are InterPro annotated protein domains/motifs as listed on FlyBase v2017_06 [45,46].

See also Table S2

Gene Female Oenocyte Specificity (oenocyte/female Sleuth q-value) % D. sechellia reads in female oenocytes (negative binomial p-value) GO term(s) Protein Domain(s)
eloF 76.6 (4.5e-9/2.3e-5) 98.75% (1.2e-21) fatty acid elongase activity ELO family
Fad2 18.2 (4.5e-9/2.5e-5) 95.5% (2.5e-72) Catalysis of an oxidation-reduction (redox) reaction in which hydrogen or electrons are transferred from each of two donors… Fatty acid desaturase type 1, conserved site; Fatty acid desaturase domain; Acyl-CoA desaturase
CG8534 4.3 (8.2e-4/3.6e-5) 93.75% (1.6e-24) fatty acid elongase activity ELO family
FASN3 2.31 (2.7e-4/3.8e-5) 60.5% (.00026) fatty acid synthase activity; hydrolase activity, acting on ester bonds Ketoacyl synthase (N-terminal, C-terminal, and C-terminal extension), Acyl transferase, Polyketide synthase, Alcohol dehydrogenase C-terminal
lectin-22C 173.9 (3.0e-4/2.0e-4) 85.75% (6.7e-09) galactose binding C-type lectin-like/link domain superfamily
bond 47.0 (7.8e-4/6.3e-5) 25.5% (2.8e-14) fatty acid elongase activity ELO family
Open in a new tab

Among the six candidate genes, the only enriched Gene Ontology molecular function terms were related to “fatty acid elongase activity” (GO:0009922 and its parent GO terms), which describe the three genes eloF, CG8534, and bond (in all cases, we use the names of the D. melanogaster orthologs) [33]. All three of these have ELO family fatty acid elongase domains [34]. Both eloF and CG8534 were D. sechellia-biased, while bond was D. simulans-biased. We further detected a weak signal for FASN3, a putative acyl transferase (Table 1). No other gene that is both oenocyte- and species-specific in its expression has an annotated Gene Ontology term or protein domain that is clearly related to CHC production (Table 1).

Compared to the female oenocytes, the other three tissues we profiled (male oenocyte, male fat body, and female fat body) all had a much weaker signal of ASE among genes with sex- and tissue-specific expression (Figure S1A-D). Therefore, we chose to focus on changes in female CHC production that might drive speciation.

eloF has widespread effects on the hydrocarbon profile of D. sechellia and D. melanogaster

To explore the role of our candidate genes on CHC profiles of these species, we performed gas chromatography coupled to mass spectrometry (GCMS). Consistent with previous measurements of hydrocarbon profiles of Drosophila, we found that wildtype D. simulans has more short-chain hydrocarbons than D. sechellia (Figure 2A) [35]. In particular, D. sechellia has almost no 23-carbon CHCs, while the predominant D. simulans hydrocarbon is 7-tricosene, a 23-carbon monoene. Indeed, there was only one hydrocarbon shorter than 26 carbons with a greater representation in D. sechellia than D. simulans, the 25-carbon pentacosadiene (~2 fold higher in D. sechellia). In contrast, all nine CHC peaks longer than 26 carbons were more abundant in D. sechellia than in D. simulans.

Figure 2: eloF-flies have an overall shorter CHC complement.

Figure 2:

Open in a new tab

A) Total ion chromatographs of the hydrocarbon profile of wild-type D. sechellia (top) and D. simulans (bottom). Retention time and abundance is relative to the n-hexacosane (26C) normalization peak. Grey regions indicate number of carbons in CHC backbone. CHCs with more than a 3-fold change are marked with asterisks at the location of the peak in the genotype with lower production.

B-C) Total ion chromatographs of the hydrocarbon profile of wild-type (top) and eloF- (bottom) D. melanogaster (B) and D. sechellia (C).

D) Average log2 fold changes of the measured compounds between D. simulans and D. sechellia vs log2 fold changes between wild-type and knockout of eloF in D. sechellia. Points are colored by the number of carbons in the backbone.

E) Principal components analysis of wild-type and eloF-D. melanogaster, simulans, and sechellia. Principal components were calculated for the wild-type data, then eloF-data were projected onto the same coordinates.

See also Figure S2

To explore the effects of our candidate genes on CHC profiles, we studied the phenotypic effects of their RNAi knockdowns in D. melanogaster. We did not pursue desatF, which already has a well-established role in Drosophila speciation [31,36,37], FASN3, which is essential for viability [38], or lectin-22C, which has relatively weak ASE and no obvious connection to CHC production. For the remaining three CHC-related candidates, we created RNAi knockdowns in D. melanogaster females for each of these genes specifically in oenocytes by crossing PromE(800)-gal4 males with UAS-shRNA females from the TRiP project [39,40], then screened the CHC profiles of the progeny by GCMS.

Of our three candidate genes (CG8534, bond, and eloF), we found that one (CG8534) was essential for viability, while knockdown of our second candidate (bond) in females led to ~60% increases in levels of pentacosadiene (a 25-carbon hydrocarbon) and ~60% decrease in levels of heptacosadiene (27 carbon) (Figure S2A). However other hydrocarbons were not significantly affected.

We observed the most pronounced effects for RNAi knockdown of our third candidate, eloF. We found that female flies with eloF knocked down had significantly fewer long-chain CHCs and more short-chain CHCs than wildtype flies (>3-fold change between CHCs with longer vs. shorter than 26 carbons; Figure 2B), consistent with previous work [32]. Interestingly, eloF also had the strongest ASE among the six candidate genes (79-fold higher expression from D. sechellia alleles).

To examine the effect of eloF on CHCs in D. sechellia, we used CRISPR/Cas9 genome editing to create two independent lines of D. sechellia with eloF knocked out (Figure S3). Nearly all of the CHCs whose levels changed after eloF knockdown in D. melanogaster showed a similar difference in D. sechellia (Figure 2C). Thus, we conclude that the molecular substrates and products of eloF are similar between D. melanogaster and D. sechellia.

We noticed that there was a strong correlation between the CHC changes observed between the sister species D. simulans and D. sechellia and the changes between wild-type and eloF knockout D. sechellia females (Figure 2D). We found a similar result for eloF-depleted D. melanogaster females (Figure S2B). These correlations suggest that loss of eloF phenocopies the natural interspecific divergence in CHC profiles.

To visualize entire CHC profiles, we performed principal components analysis, which showed that 94% of the total variation was captured by the first two components. The first principal component of variation separated D. simulans from both D. melanogaster and D. sechellia (Figure 2E). While knockdown or knockout of eloF did not completely transform the profiles of either species to match D. simulans, it did shift the profiles significantly closer. Since the CHC profile after knockdown is much more similar to D. simulans, whose CHCs inhibit mating with D. simulans males, we reasoned that one or more of the CHCs produced by eloF may contribute to this inhibition of inter-species mating.

eloF is necessary for inhibition of inter-species mating

To determine whether the change in eloF expression (and concomitant CHC changes) could be responsible for sexual isolation between the species, we performed mate choice assays by video recording and noting the time of various mating behaviors (Figure 3A-C). We first tested whether eloF might drive the behavioral isolation of D. simulans and D. sechellia. As expected, D. simulans males courted wild-type D. sechellia females at a significantly lower rate than D. simulans females. Remarkably, D. simulans males courted eloF-D. sechellia females at the same rate as conspecific females (Figure 3D). We observed no significant difference in the courtship rate between the two independently generated D. sechellia knockout lines.

Figure 3: D. simulans males court interspecific eloF-females at significantly higher rates.

Figure 3:

Open in a new tab

A-C) We recorded between 42 and 80 pairs of single D. simulans males courting single females of each indicated genotype. We recorded the time between male’s first tapping the female (A, and ostensibly sampling the female CHCs) and either singing behavior (B) or licking of the female’s posterior prior to copulation (C). N indicates the number of assays that proceeded to courtship (numerator) and the total number of assays recorded.

D) Female flies bearing a functional copy of eloF (D. melanogaster WT and D. sechellia WT) were courted by D. simulans males at significantly lower rates than D. simulans conspecific females and interspecific females without eloF. We performed the indicated Fisher’s exact tests for differences in courtship rate (as measured by rate of proceeding to precopulatory licking), with Bonferroni-corrected p-values above each bar when significant.

E) Violin plots of the delay between first contact between males and females and initiation of licking courtship behavior. Black lines indicate mean time to courtship. Gray ticks indicate the underlying data. Although the D. simulans males were slower to court D. melanogaster WT females, this represents only 5 cases of courtship (out of 60 trials), and no comparisons were significant by t-test at even a nominal p=0.05 cutoff.

See also Figure S4

We then asked whether eloF might also mediate mate discrimination between D. melanogaster and D. simulans. As expected, when D. simulans males were presented with wildtype D. melanogaster females they rarely proceeded to courtship (Figure 3D and Figure S4A). However, when we knocked down eloF expression in D. melanogaster females using oenocyte-specific RNAi, males courted them at rates only slightly lower than conspecifics.

The choice by males seems to be nearly binary. In the cases when D. simulans males did court eloF-bearing females, they did so approximately as quickly as they did for D. simulans females (Figure 3E and Figure S4B). There was no significant difference in time between first contact between the flies and any of the steps in courtship.

Outlook and future work

In this study, we have found that RNA-seq in F1 hybrids is a rapid, efficient means of identifying genes potentially involved in phenotypic divergence. Neither comparisons of expression across tissues nor of ASE within a single tissue was able to sufficiently narrow the list of candidate genes (Figure 1C); however, the combination of these orthogonal filters, together with gene annotations, allowed us to focus on only three candidate genes. This can be compared with the most widely used alternative for studying the genetic basis of phenotypic divergence, QTL mapping. In QTL mapping, hundreds of progeny from genetic crosses must be genotyped and phenotyped, requiring significant effort even for rapidly reproducing species. Moreover, this effort leads to QTLs that typically span over a hundred genes, since resolution is limited by infrequent recombination events. Therefore, studies to test specific genes within those regions are often prohibitive. We envision that our approach of intersecting filters based only on RNA-seq in F1s may be widely applicable to other tissue-specific, sex-specific, stage-specific, or condition-specific traits that differ between interfertile populations or species. For example, we previously found that in yeast, intersecting ASE with genes induced by a specific toxin pinpointed several genes whose cis-regulatory divergence contributed to toxin resistance [41]. This approach can also be combined with QTL mapping, though in many cases this may not be necessary (as exemplified by the present study, where QTL data played no role in our selection of candidate genes).

Consistent with other recent observations [16], we found that CHC differences seem to be the major source of sexual isolation between D. simulans from both D. sechellia and D. melanogaster, and we also showed that ablating eloF alleviates nearly all of the isolation from both D. sechellia and D. melanogaster. The magnitude of this effect is comparable to the reduction in barriers between D. simulans males and D. melanogaster females by ablating oenocytes entirely, a much more radical intervention (eloF appears to represent ~85% of the barrier in this study, compared to ~100% in [39]).

We can hypothesize a parsimonious evolutionary scenario to explain our observations. D. sechellia and D. melanogaster both express eloF in female oenocytes; therefore this is likely to be the ancestral state for these species, with the lower eloF expression in D. simulans being a derived change specific to this species. Our results, together with QTLs for CHC differences and mate discrimination between D. simulans and D. sechellia that contain eloF [23,24], suggest that this dramatic cis-regulatory divergence may have led to the sexual isolation of D. simulans; however further evidence, such as a reciprocal hemizygosity test [42], would be required to prove this.

This work raises several important questions. For example, because eloF affects so many CHCs, it is not clear which CHC(s) act as the discriminative signal. One candidate is the 27-carbon CHC 7,11-heptacosadiene, which is involved in male D. melanogaster and D. simulans preference [39], although other CHCs could also contribute. More generally, the ecological relevance of sexual isolation measured in the lab remains unknown, considering that rates of hybridization in the wild can be strikingly higher than laboratory predictions [43].

Another open question regards the sequence change(s) that have led to the expression divergence of eloF. The transcription factor Doublesex has been implicated in the evolution of other Drosophila species’ CHC profiles [31], but we have not found any divergent Doublesex binding sites near eloF. However, the noncoding region around eloF and CG8534 contains 136 SNPs and 10 small indels where D. simulans has a derived allele differing from both D. sechellia and D. melanogaster (thus matching the parsimonious evolutionary scenario described above) and several nonsynonymous changes in eloF.

Our finding that D. simulans males prefer mates lacking eloF suggests that male preferences have co-evolved with CHC profiles in D. simulans. An intriguing question for future work will be whether the gene(s) responsible for this co-evolved male preference could be identified with a similar cell type-specific ASE approach as demonstrated here. The small population of neurons recently shown to be responsible for key differences in the male neural circuit that evaluates a species-specific CHC [44] would be a logical focus for such a study.

STAR Methods

CONTACT FOR REAGENT AND RESOURCE SHARING

Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Hunter Fraser (hbfraser@stanford.edu).

EXPERIMENTAL MODEL AND SUBJECT DETAILS

Fly rearing and generation

For RNAi flies, virgin females of the shRNA driver were isolated within 18 hours of eclosion, then kept isolated from males for 3 days on standard cornmeal media to ensure virgin status. We used Bloomington Stock IDs 34676 (bond), 53947 (eloF), and 53299 (CG8534). We combined approximately 25 UAS-shRNA females with approximately 10 PromE(800) Gal4 driver males [39]. As negative controls, we crossed PromE(800)-gal4 males with females of Bloomington stock #32186, which carries 10 copies of UAS-driven mCD8-tagged GFP. Adults were moved to fresh vials every 3 days to ensure separation of the parents and the Gal4+UAS offspring.

Knockout D. sechellia flies were created using CRISPR/Cas9 mediated editing. We designed guides to cut at the 55th nucleotide downstream of the ATG and the 114th nucleotide upstream of the stop codon of GM23846 (the D. sechellia ortholog of eloF). We used sense oligos CTTCGCAGCGATCCATGGGTCCCCA (gene 5’-ward cut site) and CTTCGATCCGCATCCGTAGGTCAA (gene 3’-ward cut site). Embryos were injected (WellGenetics, Taipei, Taiwan) with both guides and a dsDNA donor containing ~1000bp homology arms and RFP driven by a 3xP3 promoter and flanked by LoxP sites.

Embryos were from the D. sechellia genome strain #14021–0248.25. Out of 200 injections over 2 rounds, 2 offspring showed positive RFP expression in the eyes, as expected for the 3xP3 promoter. As shown in Figure S3, we screened for presence of the inserted sequence using primers CTCCCAGCGATCATTCATTT and GCTGCTACACTTGCCACAAA (170bp product), and for absence of eloF using TCTGCAGGTTCTGATGGCAG and ACTGTGGAAAGGCAACACCA (306bp product).

All flies, either wild type, RNAi, or CRISPR edited were separated by sex within 18 hours of eclosion, then kept isolated for 5-7 days to ensure virgin status. Any vials with larvae after 5 days were discarded. Since the PromE(800)-gal4 construct is balanced with Tm3.5b, we selected straight-winged flies as RNAi positive.

METHOD DETAILS

RNA extraction and sequencing

We mated D. sechellia males to D. simulans females and dissected both oenocytes and fat bodies from the progeny, pooling material from 20 individuals from each sex. Oenocyte and fat body dissections were performed as described in [47]: briefly, flies were pinned to a Sylgard dissection plate (Dow-Corning) and covered with chilled Shields and Sang M3 medium (Sigma). The oenocytes and fat body of 10-day-old D. simulans/D. sechellia hybrid flies were isolated separately from the dorsal abdominal segments of both adult male and female abdomens using a fine tungsten dissecting needle. Each tissue sample represented the pooled material collected from 20 flies. Hybrid flies were reared in a 12hr light: 12 hr dark cycle and tissues dissected at equal time intervals across a 24hr period.

Immediately following dissection tissues were placed into cell lysis buffer to aid in preserving the integrity of the RNA. Total RNA was isolated using the RNeasy Micro kit (Qiagen). We collected two independent samples of each tissue from each sex, which in our experience is adequate for identifying strong patterns of allele specific expression.

We prepared libraries from the RNA using the NextFLEX RNA-seq library preparation kit (BioO Scientific, Austin, TX), and sequenced the libraries using 101bp paired end reads on an Illumina HiSeq 2000.

Gas chromatography–mass spectrometry

We performed GCMS by anesthetizing 5 females at 4°C for 3–5 minutes, then washing them for 5 minutes with 50μL of hexane spiked with 10mg/mL of n-hexane as a standard. Spectra were obtained using an Agilent (HP) 7890/5975 single quadrupole GC-MS instrument with a split ratio of 1:20, injector temperature of 280°C, and an oven temperature program of 35°C hold for 3.75min, 20°C/min ramp from 35°C to 320°C, and a 320°C hold for 7 min. We collected spectra for at least 3 sets of 5 flies for each genotype (but 5 spectra for D. melanogaster WT and 6 for D. simulans). Identities of different hydrocarbon peaks were inferred by inspecting the singly-ionized mass spectrum bin.

Mating assays

We performed mating assays by anesthetizing separate vials of males and females at 4°C for 3-5 minutes, then used a paintbrush to transfer one male and one female to each well of the mating chamber. The mating chamber was 3D printed from acrylic plastic and has 18 separate 2cm diameter × 5mm circular wells, with a removable clear plastic lid. We allowed flies to acclimate at room temperature and ambient light for 10-15 minutes, then recorded 30m of video with bright lights, which we found were required for D. simulans males to initiate courtship. The mating light was a 75W, 14” circular fluorescent bulb placed approximately 30cm above the mating chamber. Video of mating assays was recorded using a Dino-Lite digital microscope, then analyzed by two separate graders (PAC and NMK), who recorded the time of first contact by the male, the time of the male first following the female, the time of the first wing song by the male, and the time of first licking by the male of the female’s abdomen [11]. Prior to analyzing videos, we estimated that sample sizes of 50 assays would be sufficient to identify a change from 90% successful mating to 55% successful mating with 90% power [48]. Graders were blinded to the fly identities in each video, which had a uniform, random 4-digit number as the file identifier. We noted the time of the first instance of various copulatory behaviors: tapping, male wing song, and licking. With the exception of licking, these behaviors are not subject to rejection by females (the mating chambers are small enough that females are effectively unable to escape, while tapping is very rapid and wing song does not involve contact), and thus primarily represent choice by the males.

QUANTIFICATION AND STATISTICAL ANALYSIS

RNA-seq analysis

We used 2 independent samples of each tissue, each consisting of material from 20 flies.

In order to minimize the number of sequence mismatches between the strains we used and the reference genome, we created a corrected D. simulans genome sequence by using bowtie2 version 2.2.5 with arguments --very-sensitive to map genomic DNA reads from D. simulans and D. sechellia to the FlyBase 2.01 D. simulans reference genome [29,49]. Polymorphisms were called using GATK (HaplotypeCaller -- genotyping_mode DISCOVERY -fixMisencodedQuals -stand_emit_conf 10 - stand_call_conf 30) [50], then the ~34,000 SNPs that were fixed in both D. simulans and D. sechellia were replaced with the consensus sequence (this step was more important for creating a simulans/sechellia version of the D. melanogaster genome for Figure S1E-F).

RNA-seq reads were mapped to the D. simulans reference genome using STAR with arguments --outFilterMultimapNmax 1 --outSAMattributes MD NH -- clip5pNbases 6 --sjdbGTFfile [49,51]. Following the WASP pipeline, duplicate reads were discarded randomly, then filtered based on whether reads with the alleles swapped in silico to create artificial transcripts from the other species mapped to the same position [52]. Reads were assigned to a species only if both paired ends mapped unambiguously to one species, and allele-specific expression negative binomial p-values were calculated from aligned read counts using DESeq2 with model ~Replicate + AlignsToSpecies [27]. Despite the use of a D. simulans reference genome, we found a majority of reads were assigned to D. sechellia (Table S1), possibly indicating a widespread species-specific bias in the strength of cis-regulatory elements, as has been observed in other interspecific Drosophila crosses [53]. Alternatively, the pattern could be explained if 3-4 of the 20 dissected flies in each sample were actually non-hybrid D. sechellia, though this seems unlikely given that >99.1% of reads from the mitochondrial genome were assigned to D. simulans.

Default DESeq settings were used to correct for multiple hypothesis testing. Transcript abundances were estimated using kallisto with default arguments [54]. We used sleuth to identify differentially expressed genes between samples with matched sex and tissue type [30]. Since kallisto has not been extensively tested for suitability allele-specific expression, we opted to use the conservative mapping-based approach outlined above.

Mating Assays

Video was analyzed by two separate graders (PAC and NMK), who recorded the time of first contact by the male, the time of the male first following the female, the time of the first wing song by the male, and the time of first licking by the male of the female’s abdomen [11]. Graders were blinded to the fly identities in each video. Rates of proceeding to pre-copulatory licking were compared using Fishers exact test. Time between tapping and courtship behaviors were compared using a t-test.

DATA AND SOFTWARE AVAILABILITY

An interactive tool to explore the RNA-seq dataset is available at http://combsfraser-oenocytes.appspot.com/. Sequencing data have been deposited at the Gene Expression Omnibus under access number GSE114478. Video data are available at https://drive.google.com/a/stanford.edu/file/d/1m3K0vsW2qSqFCHtnrBj7ZVHNflOLaogl/view. Analysis scripts are available at https://github.com/TheFraserLab/CombsOenocytes2018 (doi: 10.5281/zenodo.1436010).

Supplementary Material

1

  • RNA-seq on Drosophila hybrids identifies tissue-specific cis-regulatory changes

  • eloF, a fatty acid elongase, is a major determinant of hydrocarbon pheromones

  • Ablation of eloF increases inter-species mating nearly to intra-species levels

  • This approach can be applied broadly to rapidly pinpoint key evolutionary drivers

Combs et al use RNA-seq in hybrid Drosophila to identify genes related to cuticular hydrocarbon (CHC) differences, including eloF. Ablation of eloF alters CHC profiles and increases inter-species mating nearly to intra-species levels. This approach can be applied broadly to rapidly pinpoint key evolutionary drivers in a wide range of species.

Acknowledgements

We thank Nirao Shah, Osama Ahmed, Mark Siegal, and the Fraser lab for helpful discussions, and David Stern for helpful discussions and the D. simulans tsimbazaza strain. The Bloomington Drosophila Stock Center (NIH P40OD018537) provided the RNAi strains and the Drosophila Species Stock Center at UCSD (NSF CSBR grant 1351502) provided the D. sechellia strains. GCMS was performed at the Vincent Coates Foundation Mass Spectrometry Laboratory at Stanford University, which is supported in part by NIH P30 CA124435. This work was supported by NIH grant 2R01GM097171-05A1 to H.B.F and by grants from NSERC and CIHR to J.D.L and J.J.K.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Declaration of Interests

The authors declare no competing interests.

References

  • 1.Steiger S, Franz R, Eggert A-K, and Müller JK (2008). The Coolidge effect, individual recognition and selection for distinctive cuticular signatures in a burying beetle. Proc. Biol. Sci 275, 1831–1838. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Büllesbach J, Gadau J, Beukeboom LW, Echinger F, Raychoudhury R, Werren JH, and Schmitt T (2013). Cuticular hydrocarbon divergence in the jewel wasp Nasonia: evolutionary shifts in chemical communication channels? J. Evol. Biol 26, 2467–2478. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Nunes TM, Oldroyd BP, Elias LG, Mateus S, Turatti IC, and Lopes NP (2017). Evolution of queen cuticular hydrocarbons and worker reproduction in stingless bees. Nat Ecol Evol 1, 0185–3. [Google Scholar]
  • 4.Coyne JA, and Orr HA (1989). PATTERNS OF SPECIATION IN DROSOPHILA. Evolution 43, 362–381. [DOI] [PubMed] [Google Scholar]
  • 5.Noor MA (1995). Speciation driven by natural selection in Drosophila. Nature 375, 674–675. [DOI] [PubMed] [Google Scholar]
  • 6.Coyne JA, and Orr HA (2004). Speciation (Sinauer Associates Incorporated) [Google Scholar]
  • 7.Servedio MR, and Noor MAF (2003). The Role of Reinforcement in Speciation: Theory and Data. Annu. Rev. Ecol. Evol. Syst 34, 339–364. [Google Scholar]
  • 8.Turelli M, Lipkowitz JR, and Brandvain Y (2014). On the Coyne and Orr-igin of species: effects of intrinsic postzygotic isolation, ecological differentiation, × chromosome size, and sympatry on Drosophila speciation. Evolution 68, 1176–1187. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Turissini DA, McGirr JA, Patel SS, David JR, and Matute DR (2018). The Rate of Evolution of Postmating-Prezygotic Reproductive Isolation in Drosophila. Molecular Biology and Evolution 35, 312–334. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Barbash DA (2010). Ninety years of Drosophila melanogaster hybrids. Genetics 186, 1–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Sokolowski MB (2001). Drosophila: genetics meets behaviour. Nat. Rev. Genet 2, 879–890. [DOI] [PubMed] [Google Scholar]
  • 12.Spieth HT (1952). Mating behavior within the genus Drosophila (Diptera). Bulletin of the American Museum of Natural History 99, 399–474. [Google Scholar]
  • 13.Partridge L, and Farquhar M (1981). Sexual activity reduces lifespan of male fruitflies. Nature 294, 580–582. [Google Scholar]
  • 14. Greenspan RJ, and Ferveur JF (2000). Courtship in Drosophila. Annu. Rev. Genet 34, 205–232. [DOI] [PubMed] [Google Scholar]
  • 15.Edward DA, and Chapman T (2011). The evolution and significance of male mate choice. Trends Ecol. Evol. (Amst.) 26, 647–654. [DOI] [PubMed] [Google Scholar]
  • 16.Shahandeh MP, Pischedda A, and Turner TL (2018). Male mate choice via cuticular hydrocarbon pheromones drives reproductive isolation between Drosophila species. Evolution 72, 123–135. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Garrigan D, Kingan SB, Geneva AJ, Andolfatto P, Clark AG, Thornton KR, and Presgraves DC (2012). Genome sequencing reveals complex speciation in the Drosophila simulans clade. Genome Res 22, 1499–1511. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Schrider DR, Ayroles J, Matute DR, and Kern AD (2018). Supervised machine learning reveals introgressed loci in the genomes of Drosophila simulans and D. sechellia. PLoS Genet 14, e1007341. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Kliman RM, Andolfatto P, Coyne JA, Depaulis F, Kreitman M, Berry AJ, McCarter J, Wakeley J, and Hey J (2000). The Population Genetics of the Origin and Divergence of the Drosophila simulans Complex Species. Genetics 156, 1913–1931. Available at: http://www.genetics.org/content/156/4/1913.long. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Matute DR, and Ayroles JF (2014). Hybridization occurs between Drosophila simulans and D. sechellia in the Seychelles archipelago. J. Evol. Biol 27, 1057–1068. [DOI] [PubMed] [Google Scholar]
  • 21.Lachaise D, David JR, Lemeunier F, and Tsacas L (1986). The reproductive relationships of Drosophila sechellia with D. mauritiana, D. simulans, and D. melanogaster from the Afrotropical region. Evolution [DOI] [PubMed] [Google Scholar]
  • 22.Coyne JA, Crittenden AP, and Mah K (1994). Genetics of a pheromonal difference contributing to reproductive isolation in Drosophila. Science 265, 1461–1464. [DOI] [PubMed] [Google Scholar]
  • 23.Gleason JM, Jallon J-M, Rouault J-D, and Ritchie MG (2005). Quantitative trait loci for cuticular hydrocarbons associated with sexual isolation between Drosophila simulans and D. sechellia. Genetics 171, 1789–1798. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Gleason JM, James RA, Wicker-Thomas C, and Ritchie MG (2009). Identification of quantitative trait loci function through analysis of multiple cuticular hydrocarbons differing between Drosophila simulans and Drosophila sechellia females. Heredity (Edinb) 103, 416–424. [DOI] [PubMed] [Google Scholar]
  • 25.Wittkopp PJ, and Kalay G (2012). Cis-regulatory elements: molecular mechanisms and evolutionary processes underlying divergence. Nat. Rev. Genet 13, 59–69. [DOI] [PubMed] [Google Scholar]
  • 26.Lawrence PA, and Johnston P (1986). Observations on cell lineage of internal organs of Drosophila. J Embryol Exp Morphol 91, 251–266. [PubMed] [Google Scholar]
  • 27.Love MI, Huber W, and Anders S (2014). Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol 15, 550. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Graze RM, McIntyre LM, Main BJ, Wayne ML, and Nuzhdin SV (2009). Regulatory divergence in Drosophila melanogaster and D. simulans, a genomewide analysis of allele-specific expression. Genetics 183, 547–61-1SI–21SI. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Coolon JD, McManus CJ, Stevenson KR, Graveley BR, and Wittkopp PJ (2014). Tempo and mode of regulatory evolution in Drosophila. Genome Res 24, 797–808. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Pimentel H, Bray NL, Puente S, Melsted P, and Pachter L (2017). Differential analysis of RNA-seq incorporating quantification uncertainty. Nat Meth 14, 687–690. [DOI] [PubMed] [Google Scholar]
  • 31.Shirangi TR, Dufour HD, Williams TM, and Carroll SB (2009). Rapid evolution of sex pheromone-producing enzyme expression in Drosophila. PLoS Biol 7, e1000168. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Chertemps T, Duportets L, Labeur C, Ueda R, Takahashi K, Saigo K, and Wicker-Thomas C (2007). A female-biased expressed elongase involved in long-chain hydrocarbon biosynthesis and courtship behavior in Drosophila melanogaster. Proc. Natl. Acad. Sci 104, 4273–4278. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Boyle EI, Weng S, Gollub J, Jin H, Botstein D, Cherry JM, and Sherlock G (2004). GO::TermFinder--open source software for accessing Gene Ontology information and finding significantly enriched Gene Ontology terms associated with a list of genes. Bioinformatics 20, 3710–3715. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Szafer-Glusman E, Giansanti MG, Nishihama R, Bolival B Jr, Pringle J, Gatti M, and Fuller MT (2008). A role for very-long-chain fatty acids in furrow ingression during cytokinesis in Drosophila spermatocytes. Curr Biol 18, 1426–1431. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Jallon J-M, and David JR (1987). Variation in Cuticular Hydrocarbons Among the Eight Species of the Drosophila melanogaster Subgroup. Evolution 41, 294–302. [DOI] [PubMed] [Google Scholar]
  • 36.Legendre A, Miao X-X, Da Lage J-L, and Wicker-Thomas C (2008). Evolution of a desaturase involved in female pheromonal cuticular hydrocarbon biosynthesis and courtship behavior in Drosophila. Insect Biochem. Mol. Biol 38, 244–255. [DOI] [PubMed] [Google Scholar]
  • 37.Fang S, Ting C-T, Lee C-R, Chu K-H, Wang C-C, and Tsaur S-C (2009). Molecular evolution and functional diversification of fatty acid desaturases after recurrent gene duplication in Drosophila. Molecular Biology and Evolution 26, 1447–1456. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Chung H, and Carroll SB (2015). Wax, sex and the origin of species: Dual roles of insect cuticular hydrocarbons in adaptation and mating. Bioessays 37, 822–830. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Billeter J-C, Atallah J, Krupp JJ, Millar JG, and Levine JD (2009). Specialized cells tag sexual and species identity in Drosophila melanogaster. Nature 461, 987–991. [DOI] [PubMed] [Google Scholar]
  • 40.Perkins LA, Holderbaum L, Tao R, Hu Y, Sopko R, McCall K, Yang-Zhou D, Flockhart I, Binari R, Shim H-S, et al. (2015). The Transgenic RNAi Project at Harvard Medical School: Resources and Validation. Genetics 201, 843–852. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Naranjo S, Smith JD, Artieri CG, Zhang M, Zhou Y, Palmer ME, and Fraser HB (2015). Dissecting the Genetic Basis of a Complex cis-Regulatory Adaptation. PLoS Genet 11, e1005751. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Stern DL (2014). Identification of loci that cause phenotypic variation in diverse species with the reciprocal hemizygosity test. Trends Genet 30, 547–554. [DOI] [PubMed] [Google Scholar]
  • 43.Coyne JA, Elwyn S, and Rolán-Alvarez E (2005). Impact of experimental design on Drosophila sexual isolation studies: direct effects and comparison to field hybridization data. Evolution 59, 2588–2601. [PubMed] [Google Scholar]
  • 44.Seeholzer LF, Seppo M, Stern DL, and Ruta V (2018). Evolution of a central neural circuit underlies Drosophila mate preferences. Nature. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45.Finn RD, Attwood TK, Babbitt PC, Bateman A, Bork P, Bridge AJ, Chang H-Y, Dosztányi Z, El-Gebali S, Fraser M, et al. (2017). InterPro in 2017-beyond protein family and domain annotations. Nucleic Acids Research 45, D190–D199. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46.Gramates LS, Marygold SJ, Santos GD, Urbano J-M, Antonazzo G, Matthews BB, Rey AJ, Tabone CJ, Crosby MA, Emmert DB, et al. (2017). FlyBase at 25: looking to the future. Nucleic Acids Research 45, D663–D671. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47.Krupp JJ, and Levine JD (2010). Dissection of oenocytes from adult Drosophila melanogaster. J Vis Exp, e2242–e2242. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 48.Chow S-C, Wang H, and Shao J (2007). Sample Size Calculations in Clinical Research, Second Edition (CRC Press; ). [Google Scholar]
  • 49.Hu TT, Eisen MB, Thornton KR, and Andolfatto P (2013). A second-generation assembly of the Drosophila simulans genome provides new insights into patterns of lineage-specific divergence. Genome Res 23, 89–98. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50.DePristo MA, Banks E, Poplin R, Garimella KV, Maguire JR, Hartl C, Philippakis AA, del Angel G, Rivas MA, Hanna M, et al. (2011). A framework for variation discovery and genotyping using next-generation DNA sequencing data. Nat Genet 43, 491–498. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 51.Dobin A, Davis CA, Schlesinger F, Drenkow J, Zaleski C, Jha S, Batut P, Chaisson M, and Gingeras TR (2013). STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29, 15–21. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 52.van de Geijn B, McVicker G, Gilad Y, and Pritchard JK (2015). WASP: allele-specific software for robust molecular quantitative trait locus discovery. Nat Meth 12, 1061–1063. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 53.Fontanillas P, Landry CR, Wittkopp PJ, Russ C, Gruber JD, Nusbaum C, and Hartl DL (2010). Key considerations for measuring allelic expression on a genomic scale using high-throughput sequencing. Mol. Ecol 19 Suppl 1, 212–227. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 54.Bray NL, Pimentel H, Melsted P, and Pachter L (2016). Near-optimal probabilistic RNA-seq quantification. Nat Biotech 34, 525–527. [DOI] [PubMed] [Google Scholar]

Associated Data

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

Supplementary Materials

1
NIHMS1511895-supplement-1.pdf (1MB, pdf)

RESOURCES