Proteomics reveals novel Drosophila seminal fluid proteins transferred at mating

Abstract

Across diverse taxa, seminal fluid proteins (Sfps) transferred at mating affect the reproductive success of both sexes. Such reproductive proteins often evolve under positive selection between species; because of this rapid divergence, Sfps are hypothesized to play a role in speciation by contributing to reproductive isolation between populations. In Drosophila, individual Sfps have been characterized and are known to alter male sperm competitive ability and female post-mating behavior, but a proteomic-scale view of the transferred Sfps has been missing. Here we describe a novel proteomic method that uses whole-organism isotopic labeling to detect transferred Sfps in mated female D. melanogaster. We identified 63 proteins, which were previously unknown to function in reproduction, and confirmed the transfer of dozens of predicted Sfps. Relative quantification of protein abundance revealed that several of these novel Sfps are abundant in seminal fluid. Positive selection and tandem gene duplication are the prevailing forces of Sfp evolution, and comparative proteomics with additional species revealed lineage-specific changes in seminal fluid content. We also report a proteomic-based gene discovery method that uncovered 19 previously unannotated genes in D. melanogaster. Our results demonstrate an experimental method to identify transferred proteins in any system that is amenable to isotopic labeling, and they underscore the power of combining proteomic and evolutionary analyses to shed light on the complex process of Drosophila reproduction.

Figure 1. Whole-Gene, Pairwise Estimates of d _N and d _S for Transferred Sfps of D. melanogaster

Each D. melanogaster Sfp was aligned to its ortholog in either D. simulans or D. sechellia, and a pairwise, whole-gene estimate of each rate was made. Those genes with ω > 0.5 (dashed line) and/or d _N > 0.05 were analyzed for sites under selection using multiple species alignments from the genome sequencing project. Open dots represent genes with significant evidence for positive selection acting on a subset of sites in the M8 versus M8a model comparison in codeml (see Table S3). The solid line represents ω = 1.

Figure 2. Transferred Sfps Cluster in Small Groups throughout the D. melanogaster Autosomes

(A) Euchromatic genomic locations of the autosomal Sfps are shown, based on the version 5.2 genome assembly. Open points indicate Sfps that occur in a cluster, filled points indicate those that do not. We observed significantly more clusters of genes than were found when randomly selecting 138 genes with the observed chromosomal distribution (1,000 simulations; median number of clusters: 3; range: 0–9). (B) Genomic map of Chromosome 2L containing the tandem duplicates CG17472 and CG31680. CG17472, which is present in all five species of the melanogaster subgroup, contains several residues under positive selection. Orange bars indicate naturally occurring transposon remnants. A pseudogenized copy of the locus is also indicated (ψ). (C) Phylogenetic analysis of the tandem gene duplication. Numbers above branches are branch estimates of ω in the free-ratio model. Italicized numbers under nodes indicate percent bootstrap support based on 1,000 replicates. Numbers at the tips of the non-melanogaster branches of the tree are the GLEANR-predicted gene numbers for the indicated species.

Figure 3. Relative Quantification of Transferred Sfps Found in Mated Female D. melanogaster in a Total of Six Canton S and tudor Experiments

The NSAF was used to estimate the relative abundance for each identified Sfp within each mating experiments, and values for each protein were averaged across all experiments in which the protein was found. Proteins were sorted by abundance and plotted to show the dynamic range of proteins detected (about 10³). The color of each point indicates the number of experiments in which the protein was identified; note that sperm-specific proteins could only be found in a maximum of three experiments (Canton S). Error bars represent one standard error of the mean NSAF.

Figure 4. Discovery of Unannotated Genes in D. melanogaster Using a Six–Reading Frame Database Search and RACE

(A) Mass spectrum identifying one of three peptides (AKGCTVGGNYFR) detected for Sfp24Ba. Numbers above the peaks indicate mass/charge (m/z) ratios used to identify individual residues. Red color indicates b ions, blue color indicates y ions. (B) Genomic map of Chromosome 2L showing a cluster of two newly identified Sfps (orange) and other annotated genes in the region (blue), including Acp24A4, a transferred Sfp. (C) Predicted three-dimensional structure of Sfp24Ba (with the SignalP-predicted signal sequence removed), which shows significant homology to bikunin and Kunitz-type protease inhibitors (PHYRE e-value = 3.1 × 10⁻¹⁴). The N and C termini are indicated by blue and red shading, respectively.