Resolving Mechanisms of Competitive Fertilization Success in Drosophila melanogaster

See allHide authors and affiliations

Science  16 Apr 2010:
Vol. 328, Issue 5976, pp. 354-357
DOI: 10.1126/science.1187096


Our understanding of postcopulatory sexual selection has been constrained by an inability to discriminate competing sperm of different males, coupled with challenges of directly observing live sperm inside the female reproductive tract. Real-time and spatiotemporal analyses of sperm movement, storage, and use within female Drosophila melanogaster inseminated by two transgenic males with, respectively, green and red sperm heads allowed us to unambiguously discriminate among hypothesized mechanisms underlying sperm precedence, including physical displacement and incapacitation of “resident” sperm by second males, female ejection of sperm, and biased use of competing sperm for fertilization. We find that competitive male fertilization success derives from a multivariate process involving ejaculate-female and ejaculate-ejaculate interactions, as well as complex sperm behavior in vivo.

Remating with different males by females generates sexual conflict over paternity (1) and sets the stage for postcopulatory sexual selection (24), which can drive diversification of both male and female biochemistry, physiology, morphology, and behavior (4, 5). Most investigations of postcopulatory sexual selection have focused on the pattern of sperm precedence, such as the proportion of progeny sired by the second of two males subsequent to female remating (P2). However, without knowledge of underlying mechanisms, these patterns reveal little about the intensity of selection or sex-specific adaptation (46). Consequently, even with Drosophila melanogaster, there is contention over the mechanisms giving rise to the roughly 80% last-male sperm precedence observed (712). Our understanding of these and other phenomena has been constrained by the technical challenge of directly observing sperm dynamics within the female reproductive tract and our limited ability to discriminate between sperm of different males (7, 13, 14).

We have overcome these challenges by transforming D. melanogaster to express a protamine labeled with green fluorescent protein (GFP) or red fluorescent protein (RFP) in sperm heads, which can be easily observed and unambiguously differentiated within the female reproductive tract (figs. S1 and S2 and movies S1 to S3). These lines enable direct visualization of sperm competition in vivo, in real time, and over extensive periods of time, allowing us to discriminate among alternative hypothesized mechanisms of sperm precedence. Multiple indices of male fitness relevant to sperm and/or ejaculate function were assayed, with transgenic males compared with each other and with a wild-type LHm strain (into which the GFP and RFP constructs were backcrossed for six generations). Although some significant differences were found, with transformed lines performing less well, equivalent to, or better than the wild type in different fitness assays, all three strains fell within the typical range of values reported in the literature for all assays (figs. S3 to S7), which suggested there was no dysfunction of transformant sperm. Moreover, results of the sperm precedence experiments reported below are unbiased, because GFP and RFP males (i) were competed using a reciprocal mating order design and (ii) did not differ in female remating interval, P1, or P2 in those experiments (15).

We quantified (i) spatiotemporal patterns of sperm storage and use by the female after remating, (ii) the extent and timing of sperm ejection by females, and (iii) the influence of remating on resident sperm motility. Unless otherwise specified, in all experiments, 3-day-old, virgin LHm females were randomly assigned to all treatment groups, initially mated to a GFP or an RFP male, and, beginning 3 days later (the typical remating latency for D. melanogaster), provided a daily, 6-hour opportunity to remate to a male of the alternative genotype (reciprocal male mating order balanced). All copulations were observed, and copulation durations recorded (15).

To quantify temporal dynamics of sperm fate, females in two experimental replicates were flash-frozen at randomly assigned times after the start of the second mating (ASM): 0 min (immediately after second male mounting), 5 min, 10 min, 15 min, 30 min, 45 min, 60 min, 2 hours, 8 hours, and 24 hours, with three additional time points in the second replicate: 90 min, 48 hours, and 72 hours. For the last three time treatments of the second replicate (24 hours, 48 hours, 72 hours), all progeny were reared, and the proportion sired by the second male (P2) was determined by assaying all sons for green or red sperm. Frozen females were dissected, and red and green sperm were counted in the bursa (the site of insemination), paired spermathecae (including spermathecal ducts), and the seminal receptacle (SR), distinguishing between the proximal and distal halves. For illustrative purposes, results from both replicates and male mating orders are combined in Fig. 1 [individual replicates and male mating orders shown in figs. S8 to S10; see (15) for tests of differences], with sample sizes per time treatment ranging from 30 to 50 females (total N = 543) (table S1).

Fig. 1

Numbers of first-male (red) and second-male (yellow) sperm in the (A) bursa, (B) seminal receptacle, and (C) spermathecae for 13 time points ASM, averaged over two experimental replicates and both reciprocal mating orders. Error bars represent SEM.

We revealed two different mechanisms by which resident sperm are displaced from the sperm-storage organs after remating and which contribute to last-male sperm precedence. First, there was an early release of some resident sperm from storage into the bursa, a process that does not involve second-male sperm. Second, resident sperm appear to have been physically displaced over time from both the SR and the spermathecae by incoming second-male sperm. Wild-type females store ~400 sperm at maximum in their SR and 130 sperm in the paired spermathecae (Fig. 1). Upon remating (i.e., time 0 ASM), resident sperm were depleted from this maximum capacity, because sperm were used for fertilization over the preceding days to 149 (± 12) sperm in the SR and 108 (± 11) sperm in the spermathecae. Resident sperm were rarely observed in the bursa (Fig. 1) or in the spermathecal ducts at time 0. Insemination of second-male sperm began as early as 5 min ASM (25 out of 46 cases or 54% of the time), and in 19% of cases where second-male sperm had not yet been transferred (5 out of 26), some resident sperm (17 ± 5) were observed in the bursa 5 to 10 min ASM (Figs. 1 and 2). We also saw first-male sperm in the bursa in 14% of observations where second-male sperm had been transferred but had not yet entered storage (5 out of 37). This release of resident sperm was presumptively female mediated and triggered by male accessory gland proteins (Acps) and/or mechanical stimulation of copulation.

Fig. 2

Representative female reproductive tracts at (A) 5 min, (B) 15 min, (C) 90 min, and (D) 8 hours ASM, showing relative abundances and locations of first-male (GFP) and second-male (RFP) sperm.

First-male sperm continued to be displaced as second-male sperm entered storage, soon after insemination (the SR earlier than the spermathecae) (Fig. 1) (16). Sperm transfer peaked at 15 min ASM (1412 ± 66 sperm, or ~2.7 times the female’s storage capacity), after which second-male sperm steeply declined in the bursa, while resident sperm steadily increased. Although sperm storage leveled off around 45 min ASM, the first- and second-male sperm proportions continued to change, with displacement peaking 60 to 90 min ASM (Fig. 1), when 26% of resident stored sperm had been relocated into the bursa. Although this process is likely to displace some of the second male’s own sperm (17), it will also increase the proportion of second-male sperm in storage (as observed) because of their overrepresentation in the bursa. This outcome also depends on active motility within the storage organs (movies S1 to S3) and adequate mixing of first- and second-male sperm, consistent with our observations (fig. S11 and movies S4 and S5). Competing sperm were highly stratified during and immediately after copulation, with resident sperm concentrated in the distal half of the SR (Fig. 2, B and C), but dynamic mobility of sperm in storage increased mixing over time (Fig. 2D and fig. S11).

The question of why males ejaculate many more sperm than can be stored has long puzzled researchers, given that the (1.8 mm) sperm of D. melanogaster are many times longer than any distance they travel within the female and that storage appears to be a rapid and efficient process (18). We postulate that excess sperm and, perhaps, their mobility (movie S6) are adaptations to sperm competition in this species, specifically to physical displacement of resident sperm. Indeed, the size of the second male’s ejaculate was significantly correlated with the amount of resident sperm displaced from storage, and the strength of this relation increased substantially between 45 and 60 min ASM (fig. S12) [45 min ASM: n = 45, F1, 43 = 6.85, coefficient of determination (R2) = 0.137, P = 0.012; 60 min ASM: n = 34, F1, 32 = 30.95, R2 = 0.49, P < 0.0001]. These results also suggest that ejaculate size is evolving via sexual selection mediated by sperm competition.

Sperm storage and displacement dynamics cease when the female ejects excess second-male and displaced first-male sperm from her reproductive tract. To establish that females are ejecting sperm, as previously inferred (13), we doubly mated wild-type females to GFP and RFP males and, within 30 min after copulation, transferred females in groups of up to five (mean = 4.7) into 20 cubes constructed of glass coverslips. Cubes were deconstructed 3 hours later, and the coverslips were examined under fluorescence. We found several ejected sperm masses per cube, highly stereotypical in form (Fig. 3), consisting of mixed (but predominantly second-male) sperm in association with the gelatinous mating plug. To better quantify ejection, wild-type females were isolated in glass three-well spot plates beneath coverslips immediately after mating to GFP males, and we checked for ejection every 10 min for up to 5 hours using a stereomicroscope. Ejected masses were immediately transferred to saline on slides and examined under epifluorescence to confirm the presence of GFP sperm, and females were dissected to examine the bursa for sperm. We found that 84% of females (n = 61) ejected sperm within five hours of mating (181 ± 11 min, mean ± SEM), and none of these females had any sperm in their bursa, nor had they laid any eggs. Additionally, on examination, first eggs laid in media vials in the time-series experiment (above) were never associated with large amounts of sperm, which suggests that active sperm ejection by females is typical, if not universal.

Fig. 3

Ejected sperm mass containing both first-male (GFP) and second-male (RFP) sperm under (A) differential interference contrast and (B) fluorescence.

After ejection, remaining stored sperm make up the “fertilization set” (19) or the population of sperm directly engaging in sperm-sperm competition. To delineate which storage organs house the fertilization set, we compared simple linear regressions of P2 on the proportion of second-male sperm in storage (S2) (Fig. 4) (15). S2 was found to be a highly significant predictor of P2 for the SR (F1, 73 = 25.55, P < 0.0001, R2 = 0.259) but not the spermathecae (F1, 70 = 0.26, P = 0.614, R2 = 0.004), reinforcing previous indirect evidence (20) that sperm in the SR constitutes the more immediate fertilization set. Proportional representation in the proximal versus distal regions of the SR may further contribute to differential fertilization success and has been shown to be influenced by sperm length (21, 22). However, sperm length here did not systematically vary between first and second males, and sperm proportions did not differ regionally in the SR in the hours immediately after mating (fig. S11).

Fig. 4

Relation between P2 and S2 (both variables arcsine square root transformed) in the (A) seminal receptacle and (B) spermathecae. Seminal receptacle: y = 1.203x – 0.191 (reduced major axis regression); spermathecae: y = 0.0383x + 1.144 (simple linear regression).

We finally tested if each male’s sperm in the SR post ejection was equally competitive, akin to a “fair raffle” (19), with sperm used for fertilization in direct proportion to their prevalence. Two results suggest that they were: (i) in vivo sperm motility did not change upon remating and (ii) sperm were used in direct proportion to their abundance. First, we examined if female remating induces a change in motility of GFP resident sperm in the SR, by quantifying their velocity and percent motility (i) 5 min ASM to an RFP male, (ii) 60 min ASM to an RFP male, (iii) 60 min after the first mating, and (iv) 3 days after the first mating (15). Percent motility and sperm velocity showed no significant changes, which suggested that in the SR, neither several days of storage nor female remating had any detrimental effect on sperm survival or motility (fig. S13 and movies S5 and S6). Second, if sperm competition was disproportionately and systematically biased toward one male over another (“loaded raffle”), the slope of the regression of P2 on S2 would be either significantly less than 1 (favoring the first male) or greater than 1 (favoring the second male). However, the slope of the reduced major axis regression was not significantly different from 1 [slope = 1.203, t(68) = 0.81, P = 0.422] (Fig. 4A), which supported the fair-raffle hypothesis.

Our observations of in vivo motility and fine-scale spatiotemporal patterns of sperm fate corroborate some previously conjectured mechanisms underlying competitive fertilization success, including sperm displacement (10, 12, 23) and sperm ejection by females (13), but fail to support other proposed mechanisms, such as sperm incapacitation (911). We cannot exclude the possibility that Acp-mediated incapacitation contributes to the early release of some resident sperm from storage. However, we would then expect to see dead or dying sperm in storage and in the bursa, yet neither sperm velocity nor sperm motility in storage changed upon remating. Our results complement other investigations of sperm mortality of competing ejaculates in vivo (13) and in vitro (24), which suggested that incapacitation makes little to no contribution to sperm precedence in D. melanogaster. The present investigation reveals that postcopulatory sexual selection in D. melanogaster includes the early release of resident sperm from storage, the heterogeneous distribution of competing sperm proportions in different storage organs, fair-raffle sperm use in the SR, and the surprising level of sperm mobility within female sperm-storage organs (movies S1 to S5). Most important, a complex and dynamic array of processes underlying patterns of sperm precedence (e.g., P2) can now be dissected in targeted, detailed genetic analyses to reveal specific behavioral, physiological, and biochemical mechanisms relevant to postcopulatory sexual selection.

Supporting Online Material

Materials and Methods

SOM Text

Figs. S1 to S13

Table S1


Movies S1 to S6

References and Notes

  1. Materials, methods, supplementary analyses, and discussion are available as supporting material on Science Online.
  2. Supported by the NSF (grants DEB-9806649 and DEB-0814732). We thank S. McClear for assistance with the protamine constructs and M. Wong for technical assistance.
View Abstract

Stay Connected to Science

Navigate This Article