Venerose, a phospho-galactoside abundant in seminal fluid
We employed mass spectrometry to identify sugars in male ejaculate, but found no evidence of mass units corresponding to glucose, trehalose, or fructose. Instead, we observed a robust signal at a molecular weight of 377, [M-H]−= 376.10142 (Fig. 1A). In subsequent examinations of male and female reproductive tissue extracts, we found this peak was present exclusively in males and primarily localized to the MAG, the major source of seminal fluid substances (Fig. 1B, Fig. S1). Hereafter, we will refer to this substance as the Drosophila venereal sugar or venerose. Venerose was the major peak in the hydrophilic eluent of MAG extracts (Fig. 1B), with trace amounts also present in male ejaculatory ducts and testes (Fig. 1C).
After deducing the molecular formula of venerose—C11H24O11NP—via electrospray ionization high-resolution mass spectrometry ([M-H]− C11H23O11NP−, calculated m/z 376.10142, found 376.10139, Δ 0.08 ppm mass error), we proceeded to a structural identification via 1H, 13C, and 31P NMR and 2D NMR spectral analyses (Table S1; Fig. S2A-K). The combined 1H-1H COSY, 13C-1H HMQC, and HMBC spectra revealed the structure of venerose as 1-O-[4-O-(2-aminoethylphosphate)-β-D-galactopyranosyl]-x-glycerol (Fig. 1D). Although this sugar was previously identified in D. melanogaster male accessory glands, it had not yet been associated with a biological function37. Our 31P-NMR analysis of the MAG extracts indicated that a single male fly produces approximately 0.45 µg or 1.2 nmol of venerose. In a single mating, the male transfers 39 ± 3.6% of its venerose pool to the female, after which it restores its venerose pool within approximately 24 hours (Fig. 1E).
Glucosyltransferase is required for venerose production
In a search for genes involved in venerose biosynthesis, we identified several mutants with transposons inserted in genes encoding enzymes involved in diacylglycerol biosynthesis (Fig. S3A-C). The MAG is rich in lipids and lipases38, and we found that RNAi-mediated knockdown of specific lipases in the MAG significantly reduced venerose production (Fig. S3D). Venerose carries a β-galactose attached to glycerol, a moiety commonly found in glycoglycerolipids. This suggests venerose production requires a glycosyltransferase capable of transferring galactose or other hexoses to a lipid acceptor with a glycerol moiety (e.g., diacylglycerol). Thus, we searched the Carbohydrate-Active enZymes (CAZy) database39,40 and found that the GT28 gene family has 1,2-diacylglycerol 3-β-galactosyl-transferase activity41 (Table S2). The D. melanogaster genome contains 18 genes with a GT28 domain, 6 of which are expressed in the MAG (Fig. 2A; Table S3 and S4). By performing MAG-specific RNAi knockdown, we found that knockdown of UGT305A1, UGT302C1, UGT304A1, or UGT37D1 each significantly reduced venerose levels (Fig. 2B; Fig.S3E). This result suggests these enzymes contribute with some redundancy to venerose production. The MAG matures over 4–5 days after emergence42, and although the MAG contains minimal venerose at emergence, its venerose pool increases rapidly post-emergence (Fig. S3F). We found UGT305A1 expression in the MAG increases during this period (Fig. 2A; Table S4). We also found overexpression of UGT305A1 in the MAG increased venerose production by 33–40% compared to controls (Fig. 2C). Therefore, in subsequent experiments, we used UGT305A1 knockdown males as venerose-deficient males.
Females absorb venerose from the male ejaculate
Drosophila females incorporate phosphorus from the male ejaculate into both somatic tissues and developing oocytes22. Since venerose carrying its phosphoethanolamine moiety is highly abundant in the male ejaculate, we hypothesized that females absorb phosphorus from the ejaculate via venerose. We radio-labeled venerose by feeding newly emerged males 32P-containing medium for 5 days. When we did the same with UGT305A1 knockdown, we saw a roughly 40% reduction in venerose production and a corresponding 40% reduction in 32P radioactivity in the hydrophilic eluents of MAG extracts, suggesting most 32P radioactivity in the MAG extracts originates from labeled venerose (Fig. 2D). In contrast, UGT305A1 knockdown did not affect the 32P activity of venerose-deficient head extracts (Fig. 2E). When examined 24 hours after mating, the ovaries of females mated with UGT305A1 knockdown males also had approximately 40% less 32P activity than those of females mated with control males (Fig. 2F). These findings strongly suggest mated females absorb elemental phosphorus by metabolizing infused venerose. Furthermore, we detected venerose in hemolymph collected from mated females, but not from virgin females (Fig. 2G, Fig. S1C). Copulation in D. melanogaster induces physical damage to the female reproductive system, allowing components of the ejaculate to diffuse passively into the hemocoel43 and providing a plausible route by which venerose enters the female hemolymph.
Venerose is essential for mating-induced GSC proliferation
To assess the role of venerose in female reproduction, we examined egg-laying activity in females mated with venerose-deficient UGT305A1 knockdown males. In these females, cumulative egg-laying activity over a period of 7 days was reduced by 37–66% (Fig. 3A). Venerose deficiency in the male ejaculate did not, however, appear to affect sperm fertility or offspring viability (Fig. S4A-C).
Females inseminated by venerose-deficient males began showing reduced egg-laying activity roughly three days post-mating (Fig. 3A). It is therefore possible that venerose affects mating-induced germline stem cell (GSC) proliferation, then influencing egg production 2–3 days post-mating10. Females mated with control males showed an increase in GSCs, whereas females mated with UGT305A1 knockdown venerose-deficient males showed reduced GSC proliferation (Fig. 3B-C). This impairment was restored by injecting 0.8 nmol of synthetic venerose (Fig. 3D). Venerose injection into virgin females also stimulated GSC proliferation to the levels observed in mated females (Fig. 3E). This effect appeared to be dependent on the amount of injected venerose, with amounts as low as 0.16 nmol inducing GSC proliferation (Fig. S4D). This amount corresponds to roughly 35% of the total venerose transmitted during mating.
Venerose stimulates GSC proliferation through a brain-derived factor
Next, we asked whether venerose acts directly on the ovary to stimulate GSC proliferation. First, we isolated ovaries from virgin females and incubated them in culture medium supplemented with or without venerose. As a positive control, we used 1 mM octopamine (OA) to stimulate GSC proliferation to the level observed in mated females14. Unlike with OA, however, we found 10 mM venerose did not affect GSC numbers in cultured ovaries (Fig. 3F). We therefore formulated another hypothesis in which venerose acts indirectly via another tissue, such as the brain or gut, by triggering the release of an external factor(s) that stimulates GSC proliferation10,15,44. Indeed, we found significantly more GSCs in ovaries co-cultured with brains in medium supplemented with venerose, but not in medium without it (Fig. 3G; Fig. S4E). A similar co-culture with gut in venerose-containing medium, however, did not stimulate GSC proliferation (Fig. 3G). These observations suggest venerose acts on the brain to induce the secretion of an endocrine factor that then stimulates GSC proliferation in the ovaries.
Brain-derived Dh44 acts on Dh44R2 in the ovaries to trigger venerose-induced GSC proliferation
Dh44R2, one of two GPCR receptors sensitive to Dh44, is expressed in terminal filament (TF) cells in larval ovaries45. We verified its expression in TFs and discovered its expression in transition cells (TCs), which are TF cells that contact cap cells in the adult ovary (Fig. 4A-E, Fig. S5A). TF cells, cap cells, and escort cells comprise the GSC niche, which provides a physical anchor for GSCs and controls their maintenance and proliferation46,47. Dh44R1, a paralogue of Dh44R2, does not appear to be expressed in GSC niche cells or other ovarian cells (Fig. S5B). The expression of Dh44R2 in GSC niche cells suggests that its ligand Dh44, which is secreted by the brain in response to circulating venerose, may be involved in mating-induced GSC proliferation. We cultured virgin ovaries in medium containing varying concentrations of Dh44 and found that levels of synthetic Dh44 as low as 10 nM stimulate GSC proliferation as much as 1 mM OA (Fig. 4F). In ex vivo culture experiments using bab1-Gal4 to knockdown Dh44R2 in TF but not escort cells, we found a complete inhibition of Dh44-induced GSC proliferation (Fig. 4G; Fig. S5C). This suggests a direct mode of action for Dh44 on the ovary. Moreover, we found Dh44R2 knockdown prevented the increase in GSC proliferation we observed in mated females (Fig. 4H). In contrast, Dh44R1 deficiency had no effect on Dh44-induced GSC proliferation (Fig. S5E, F).
Mating-induced GSC proliferation is mediated by Decapentaplegic (Dpp), the fly counterpart of bone morphogenetic protein (BMP)10,13–15,48,49. Venerose injection into virgin females increased phosphorylated Mad (pMad) in GSCs, indicating Dpp activity (Fig. 4I). Conversely, Dh44R2 knockdown suppressed pMad activity in mated females (Fig. 4J). From these results, we conclude that after entering the female’s hemolymph from the male ejaculate, venerose stimulates Dh44+ neurons in the brain to secrete Dh44. This Dh44 then acts on GSC niche cells to promote GSC proliferation (Fig. 4A).
Venerose stimulates GSC proliferation via the brain Dh44-PI neurons
Many brain neurons express Dh44, but the 6 endocrine Dh44-expressing Dh44-PI neurons in the pars intercerebralis express the highest levels of Dh4431. Dh44-PI neurons secrete Dh44 in response to hemolymph-born sugars35,36. We therefore hypothesized that Dh44-PI neurons may also secrete Dh44 in response to venerose, subsequently stimulating GSC proliferation. When we injected venerose into Dh44-RNAi virgin females whose Dh44-PI neurons no longer produce Dh44, they did not show the venerose-induced increase in GSCs observed in controls (Fig. 4K). Next, we activated Dh44-PI neurons with the light-sensitive cation channel CsChrimson, which requires all-trans-retinal (ATR) for activation. After 24 hours of light exposure, females fed an ATR-containing diet but not those fed an ATR-free diet showed enhanced GSC proliferation (Fig. 4L). We also performed ex vivo culture experiments to validate the necessity of Dh44 secreted by Dh44-PI neurons. Virgin ovaries co-cultured with control brains but not Dh44-knockdown brains showed a significant increase in GSCs in response to venerose (Fig. 4M).
Venerose stimulates Dh44-PI neurons to secrete Dh44
Thus far, our results indicate that venerose enters the female hemolymph after mating. Because males transfer ~ 40% of the contents of their MAG in a single mating session (Fig. 1E) and because MAG extracts from individual males contain ~ 1.2 nmol of venerose, we estimate that venerose levels could reach 6 mM in the female hemolymph. Consistent with this estimate, we found that treatment of isolated brains with as low as 2 mM synthetic venerose induced a significant decrease in Dh44 protein levels in Dh44-PI neurons. This indicates that venerose induces Dh44 secretion at physiologically relevant concentrations (Fig. 5A). To further examine the role venerose plays in Dh44 secretion, we asked whether venerose can trigger Ca2+ oscillations in Dh44-PI neurons expressing the Ca2+ reporter GCaMP6m. In brain preparations containing tetrodotoxin (TTX) to eliminate any synaptic input on the Dh44-PI neurons, we found the addition of 2 mM synthetic venerose induced a Ca2+ response in ~ 40% of Dh44-PI (Fig. 5B-D). Venerose produced longer-lasting Ca2+ dynamics than glucose (Fig. 5B, Video S1), with an oscillation duration of 128 ± 45 seconds (for 2 mM). Notably, 2 mM venerose stimulated robust Ca2+ oscillations in Dh44-PI neurons with a latency of 8.3 ± 1.4 minutes. In a copulation episode lasting ~ 15 minutes, ejaculate transfer is completed within 7–8 minutes of copulation initiation50. This suggests the Dh44-PI neurons begin Dh44 secretion roughly by the end of copulation.
Venerose stimulates Dh44-PI neurons via the sugar transporter Tret1-1
Dh44-PI neurons regulate post-ingestive nutrient selection, with Ca2+ oscillations in response to nutritive D-glucose, but not to nonnutritive L-glucose35. Since the sugar transporter blocker phlorizin can suppress nutritive sugar-induced Ca2+ oscillations35, we reasoned that sugar transporters in Dh44-PI critical for post-ingestive nutrient selection may also be important for responsiveness to venerose. After knocking down 23 Drosophila sugar transporters in Dh44-PI neurons, we found that knockdown of trehalose transporter 1–1 (Tret 1–1) impaired D-glucose selection and reduced D-glucose-induced Ca2+ oscillations (Fig. S6; Fig. 5E). Furthermore, Tret1-1 knockdown significantly reduced venerose-induced Ca2+ oscillations, Dh44 secretion, and GSC proliferation (Fig. 5E-G).
Energy-deprived females absorb more venerose
Males transfer considerable amounts of venerose to females during mating. The elemental phosphorus females absorb from the male ejaculate is predominantly derived from venerose (Fig. 2G). This suggests venerose transfer may function as a form of seminal feeding, akin to courtship feeding. To further explore this, we investigated the relationship between female energy state and absorption of venerose. When mated with males producing 32P-labelled venerose, the ovaries of starved females exhibited higher 32P activity than their well-nourished counterparts (Fig. 6A), indicating that energy-deprived conditions lead to increased female venerose absorption. As females can only absorb venerose while retaining the male ejaculate in their uterus, energy states that enhance venerose absorption likely extend the EHP.
Energy-dependent effects of venerose on EHP and sperm storage
Given the essential role of Dh44 from Dh44-PI neurons in establishing the standard EHP (~ 90 minutes) and facilitating adequate sperm storage31, our finding that venerose stimulates Dh44 secretion suggests a potential regulatory role for venerose in EHP and sperm storage dynamics. Dh44-PI neurons function as nutrient sensors, particularly during energy-deprived states36, indicating heightened responsiveness to venerose under starved conditions. Starved females, when mated with control males, exhibited significantly longer EHP durations than their well-nourished counterparts (Fig. 6B). But pairings with UGT305A1 knockdown males, which produce ~ 40% less venerose than control males, resulted in markedly shorter EHPs and reduced sperm storage (Fig. 6B, C). Importantly, this effect was exclusive to starved females; we did not observe any notable effect in their well-nourished counterparts. These results suggest venerose transferred by males induces energy-deprived females, but not well-nourished ones, to extend their EHP, leading to enhanced venerose absorption and sperm storage.
Energy state determines Dh44 pool size
When we monitored Ca2+ oscillations in Dh44-PI through a small window on the head capsule of live flies, well-nourished females produced spontaneous oscillations with larger amplitudes than starved females (Fig. S7A). Thus, well-nourished females secrete more Dh44 peptide, leaving less in the Dh44-PI neurons. We corroborated this finding when we observed a significant increase in anti-Dh44 immunoreactivity in Dh44-PI neurons under starvation conditions (Fig. 6D). Next, we further investigated the relationship between Dh44-PI activity, Dh44 pool, and EHP. Optogenetic inhibition of Dh44-PI neurons before mating increased Dh44 peptide levels by 67 ± 21.3% and EHP by 24 minutes compared to controls (Fig. 6E), while thermal activation of Dh44-PI neurons reduced Dh44 peptide levels by 42 ± 9.6% and EHP by 56 minutes (Fig. 6F). These results suggest increased Dh44-PI activity and reduced Dh44 levels before mating reduce the EHP of female flies, whereas decreased Dh44-PI activity and elevated Dh44 levels prolong the EHP.
We observed a marked reduction in Dh44 peptide levels in mated females compared to virgins, regardless of their energy state prior to mating (Fig. S7B, C). This indicates that mating and venerose stimulate Dh44-PI neurons to secrete large amounts of Dh44. We infer from these results that the energy state of females before mating determines the size of the Dh44 pool available for secretion triggered by mating or venerose (Fig. 6G). Under well-nourished conditions, Dh44-PI neurons have a limited Dh44 pool because of their high level of spontaneous activity, probably induced by elevated hemolymph sugars. Conversely, under energy-deprived conditions, Dh44-PI neurons possess a much larger Dh44 pool. Consequently, when mated with control males that transfer a substantial amount of venerose, energy-deprived females with larger Dh44 pools secrete more Dh44, exhibit extended EHP, and absorb a greater amount of venerose compared to well-nourished females with limited Dh44 pools.
Females assess the energy status of male partners via venerose
Males under low-energy conditions produce and allocate less nutrients in their ejaculate51. We found starved males produced only half the venerose of well-nourished males (Fig. 7A), indicating that a male’s energy state significantly influences the venerose levels in his seminal fluid. Moreover, females mated with starved males exhibited no signs of mating-induced GSC proliferation (Fig. 7B).
Last, since male energy status influences the level of venerose in their seminal fluid, we asked whether females adjust EHP and sperm storage in response to the energy state of their male partners. As anticipated, we found starved females mated with males starved for 24 hours exhibited an EHP 23 minutes shorter than those mated with well-nourished males and stored 19.4% less sperm (Fig. 7C, D). In contrast, well-nourished females did not exhibit any discernible changes in EHP or sperm storage when exposed to the seminal fluid of starved males.
Together, these findings provide insight into a mechanism by which females, particularly those malnourished prior to mating, can discern males with distinct energy states by adjusting EHP and sperm storage (Fig. 7E). In energy-deprived conditions, females possess a larger Dh44 pool. In this situation, as the quantity of venerose in seminal fluid increases, Dh44 secretion increases, leading to increased EHP and consequently enhanced sperm storage. Well-nourished females, in contrast, have limited a Dh44 pool that is easily depleted by the smaller amounts of venerose transferred by starved males. For these females, Dh44 secretion cannot increase further in response to the seminal fluid of well-nourished males, which leads to a resistance to changes in the EHP and sperm storage regardless of variations in the quantity of venerose transferred by males.