Despite holding a central role for fertilisation success, reproductive traits often show elevated rates of evolution and diversification. The rapid evolution of seminal fluid proteins (Sfps) within populations is predicted to cause mis-signalling between the male ejaculate and female reproductive tract between populations resulting in postmating prezygotic (PMPZ) isolation. Crosses between populations of Drosophila montana show PMPZ isolation in the form of reduced fertilisation success in both noncompetitive and competitive contexts. Here we test whether male ejaculate proteins deriving from either the accessory glands or the ejaculatory bulb differ between populations using liquid chromatography tandem mass spectrometry. We find more than 150 differentially abundant proteins between populations which may contribute to PMPZ isolation. These proteins include a number of proteases and peptidases, and several orthologs of D. melanogaster Sfps, all known to mediate fertilisation success and which mimic PMPZ isolation phenotypes. Males of one population typically produced greater quantities of Sfps and the strongest PMPZ isolation occurs in this direction. The accessory glands and ejaculatory bulb have different functions and the ejaculatory bulb contributes more to population differences than the accessory glands. Proteins with a secretory signal, but not Sfps, evolve faster than non-secretory proteins although the conservative criteria used to define Sfps may have impaired the ability to identify rapidly evolving proteins. We take advantage of quantitative proteomics data from three Drosophila species to determine shared and unique functional enrichments of Sfps that could be subject to selection between taxa and subsequently mediate PMPZ isolation. Our study provides the first high throughput quantitative proteomic evidence showing divergence of reproductive proteins implicated in the emergence of PMPZ isolation between populations.

LC-MS/MS was performed by nano-flow liquid chromatography (U3000 RSLCnano, Thermo FisherTM) coupled to a hybrid quadrupole-orbitrap mass spectrometer (QExactive HF, Thermo ScientificTM). We performed quantitative proteomic analysis for label free quantification using the MaxLFQ algorithm to generate relative peptide and protein intensity information (Cox et al. 2014) in MaxQuant (Tyanova et al. 2016). For protein identification we matched mass spectra to the D. montana predicted proteome (Parker et al. 2018), generated using gene predictions from the Maker2 pipeline (Holt and Yandell 2011) reciprocally blasted against D. virilis proteins (Parker et al. 2018). Additional identifiers were added from Parker et al. 2018 and via FlyBase.org

Supplementary materials and methods describing complete data acquisiton and processing can be found here: https://github.com/MartinGarlovsky/Dmon_ejaculate_proteomics