1. Introduction
In Drosophila melanogaster, approximately 112 accessory gland proteins (Acps) are transferred from the male to the female during copulation. These Acps subsequently dramatically alter both female behaviour and physiology (Ram & Wolfner, Reference Ram and Wolfner2007; Findlay et al., Reference Findlay, Yi, MacCoss and Swanson2008), with functions that include: reducing female remating rate, increasing ovulation and egg laying, and facilitating sperm storage in females thereby increasing male reproductive success (Chen et al., Reference Chen, Stummzollinger, Aigaki, Balmer, Bienz and Bohlen1988; Neubaum & Wolfner, Reference Neubaum and Wolfner1999). Null mutants for D. melanogaster Acps such as Acp70A (sex peptide (SP)), Acp26Aa (ovulin) and Acp36DE suffer reduced fertility and/or perform poorly in sperm competition (Herndon & Wolfner, Reference Herndon and Wolfner1995; Neubaum & Wolfner, Reference Neubaum and Wolfner1999; Chapman et al., Reference Chapman, Herndon, Heifetz, Partridge and Wolfner2001). The genes encoding Acps evolve rapidly both within and between species (Swanson & Vacquier, Reference Swanson and Vacquier2002; Begun & Lindfors, Reference Begun and Lindfors2005; Mueller et al., Reference Mueller, Ram, McGraw, Qazi, Siggia, Clark, Aquadro and Wolfner2005; Schully & Hellberg, Reference Schully and Hellberg2006; Haerty et al., Reference Haerty, Jagadeeshan, Kulathinal, Wong, Ram, Sirot, Levesque, Artieri, Wolfner, Civetta and Singh2007) and variation in their amino acid sequence suggests strong positive selection on these genes (Swanson et al., Reference Swanson, Clark, Waldrip-Dail, Wolfner and Aquadro2001).
A major function of the most comprehensively studied Acp, SP, is to induce a female ‘refractory’ period (reduced acceptance of further matings). This occurs after SP microinjection (Chen et al., Reference Chen, Stummzollinger, Aigaki, Balmer, Bienz and Bohlen1988) or ectopic expression in females (Aigaki et al., Reference Aigaki, Fleischmann, Chen and Kubli1991). Additionally, SP reduces female fitness, possibly through the increased production of juvenile hormone (JH) it causes in vitro in females' corpora allata (Moshitzky et al., Reference Moshitzky, Fleischmann, Chaimov, Saudan, Klauser, Kubli and Applebaum1996; Wigby & Chapman, Reference Wigby and Chapman2005; Harshman & Zera, Reference Harshman and Zera2007), and because of this, SP has been implicated in sexual conflict, as have seminal proteins generally (Chapman et al., Reference Chapman, Liddle, Kalb, Wolfner and Partridge1995; Eberhard, Reference Eberhard1996; Wolfner, Reference Wolfner2002; Wigby & Chapman, Reference Wigby and Chapman2005). The gene encoding SP (Acp70A) also shows a strong signal of positive selection and evidence suggests that it is one of the most rapidly evolving Acp genes (Cirera & Aguade, Reference Cirera and Aguade1997). Sexual selection and/or sexual conflict are likely to be involved in promoting this rapid evolution.
SP is found bound to sperm in females (Peng et al., Reference Peng, Chen, Busser, Liu, Honegger and Kubli2005), and it is then cleaved from the sperm and is thought to interact with both the female genital tract and the nervous system (Ottiger et al., Reference Ottiger, Soller, Stocker and Kubli2000; Yapici et al., Reference Yapici, Kim, Ribeiro and Dickson2008). SP's effect on females are prolonged, with SP affecting female behaviour for over 5 days, whereas ovulin, for example, induces shorter term effects on egg laying that last for only one day (Herndon & Wolfner, Reference Herndon and Wolfner1995; Heifetz et al., Reference Heifetz, Lung, Frongillo and Wolfner2000, Reference Heifetz, Vandenberg, Cohn and Wolfner2005; Chapman et al., Reference Chapman, Bangham, Vinti, Seifried, Lung, Wolfner, Smith and Partridge2003). Other Acps are removed quickly from the female reproductive tract (Monsma et al., Reference Monsma, Harada and Wolfner1990; Coleman et al., Reference Coleman, Drahn, Petersen, Stolorov and Kraus1995; Bertram et al., Reference Bertram, Neubaum and Wolfner1996; Ram et al., Reference Ram, Ji and Wolfner2005), but the prolonged occupation of the female reproductive tract by SP is probably because it is bound to sperm (Peng et al., Reference Peng, Chen, Busser, Liu, Honegger and Kubli2005). SP is likely to be a strong determinant of male fitness. This is because female multiple mating is common in D. melanogaster and sperm dumping and sperm displacement of the first male's sperm when females remates (Gromko et al., Reference Gromko, Gilbert, Richmond and Smith1984a, Reference Gromko, Newport and Kortierb; Snook & Hosken, Reference Snook and Hosken2004) mean the last male to mate often sires ~80% of subsequent offspring. Therefore, a male's ability to prevent females from remating will be an important male fitness component and SP affects the duration of this delay (Fiumera et al., Reference Fiumera, Dumont and Clark2007).
Variation in transcript levels of protein-coding genes is thought to be responsible for many of the phenotypic differences within and between populations of D. melanogaster, including sexual dimorphism (Baker et al., Reference Baker, Meadows, Wang, Dow and Russell2007), exemplifying how important transcriptional regulation can be. Additionally, studies of variation in Acp70A gene sequence show several polymorphisms either within or just upstream of the Acp70A coding region (Fiumera et al., Reference Fiumera, Dumont and Clark2007). Such polymorphisms in Acp70A gene sequence occur in natural populations, yet the natural variation in expression levels of Acp70A has not been previously examined, nor is it known if variation in expression levels leads to variation in female responses, such as delaying remating.
Endosymbionts are also known to have drastic effects on host sexual behaviour by manipulating their reproductive physiology and/or behaviour (Folstad & Karter, Reference Folstad and Karter1992; Min & Benzer, Reference Min and Benzer1997; Champion de Crespigny et al., Reference Champion de Crespigny, Pitt and Wedell2006; Negri et al., Reference Negri, Franchini, Mandrioli, Mazzoglio and Almai2008). One such obligate intracellular organism is the bacterium Wolbachia (Jeyaprakash & Hoy, Reference Jeyaprakash and Hoy2000), which is widespread in insects (Werren et al., Reference Werren, Zhang and Guo1995) and occurs at frequencies as high as 30–75% in both wild and laboratory populations of D. melanogaster (Corby-Harris et al., Reference Corby-Harris, Pontaroli, Shimkets, Bennetzen, Habel and Promislow2007). Wolbachia is present in almost all Drosophila tissues with highest infection levels in the ovaries of females (Dobson et al., Reference Dobson, Bourtzis, Braig, Jones, Zhou, Rousset and O'Neill1999), where they infect the eggs and are transmitted to any offspring subsequently produced (reviewed in Tram et al., Reference Tram, Ferree and Sullivan2003). Crosses between infected males and uninfected females cause reduced egg-hatching success due to cytoplasmic incompatibility (CI) (Werren, Reference Werren1997). In Nasonia and Drosophila, CI appears to occur because the two sister sets of chromosomes do not align synchronously at meiosis (Tram & Sullivan, Reference Tram and Sullivan2002; Tram et al., Reference Tram, Ferree and Sullivan2003), but the molecular mechanism by which Wolbachia induces the CI phenotype is still unknown. Although evidence for fecundity costs associated with Wolbachia infection is inconsistent (Hoffmann et al., Reference Hoffmann, Clancy and Merton1994; Min & Benzer, Reference Min and Benzer1997; McGraw et al., Reference McGraw, Merritt, Droller and O'Neill2002; Weeks et al., Reference Weeks, Turelli, Harcombe, Reynolds and Hoffmann2007), in Drosophila simulans Wolbachia infection causes reduced sperm production in males (Snook et al., Reference Snook, Cleland, Wolfner and Karr2000) and poor competitive ability of sperm when competing with other males' sperm for fertilization of ova within females (Champion de Crespigny et al., Reference Champion de Crespigny, Pitt and Wedell2006). Fewer sperm and low sperm competitive ability may generate selection on other ejaculate components to compensate for these detrimental effects. One target of compensating selection could be SP as this influences females' remating propensity and hence reduces sperm competition risk.
Here, we test for variation in Acp70A expression levels in 15 isofemale lines of D. melanogaster and then assess the effects this variation has on the duration of the female refractory period, a key target of SP, and an important male fitness component. We also test for the effects of Wolbachia on Acp70A gene expression patterns as the negative impact of this parasite on other male fitness components (e.g. sperm number) has the potential to select for compensatory increases in SP production.
2. Materials and methods
(i) Rearing conditions
D. melanogaster isofemale lines were collected by Trudy Mackay in North Carolina in 2004, donated to us by Frank Jiggins and continually maintained by full sib mating. They arrived in our lab in February 2007 and were reared in 7·5×2·5 cm glass vials with approximately 50 individuals per vial. Vials containing flies were kept at 25°C on a 12:12 light:dark light cycle with 15 ml standard food mix (10 g agarose, 85 g granulated sugar, 60 g maize, 10 g yeast, 1 litre deionized H2O and 1 g nipagin). Before the experiment six individuals from each line were diagnosed for Wolbachia infection by PCR following Snook et al. (Reference Snook, Cleland, Wolfner and Karr2000) after DNA extraction using EDNA kits (Fisher Biotech). Due to the isoline status of the flies and high transmission fidelity of Wolbachia (Hoffmann et al., Reference Hoffmann, Turelli and Harshman1990) this was taken as an indication that all flies are either infected or uninfected within lines.
To generate experimental flies reared under standardized conditions, populations laid eggs on 2×1 cm laying caps of food mix in 9×2·5 cm universal tubes for 24 h. Low-density vials were set-up with 40 eggs placed on approximately 7 ml standard food mix vials to reduce larval competition. Upon emergence virgin adults were collected under CO2 anaesthesia and sexed on ice every 8–12 h. Males were placed individually in standard food vials and females were discarded. Five-day-old males were subsequently frozen in liquid nitrogen and stored at −80°C.
Acp70A expression
RNA was extracted from 2 to 5 males per isofemale line using Tri reagent (Sigma) and treated with DNase (Sigma). PCR was used to confirm complete DNA removal using Acp70A specific primers FP: 5′-CGTTTGCGTACTCGGCTTGGTC, RP 5′-CCCCAAATTAAGACGGCACCACT. (PCR cycle: 95°C for 3 min, followed by 39 cycles of 95°C for 1 min, 58°C for 1 min and 72°C for 1 min, followed by 72°C for 5 min). 10 μl reactions were used containing 1 μl RNA sample, 1 μl 10 pM primers and 7 μl 1·1×ReddyMix™ (ABgene). 5 μl of sample was run on 1·4% agarose gel at 120 V for 30 min and viewed under a UV lamp.
Quantitative reverse transcriptase PCR (Q-RTPCR) was carried out using a DNA engine Opticon 2 with FullVelocity® SYBR® Green one-step Q-RTPCR Reagents (Stratagene). The housekeeping gene RP49 (primers FP: 5′-ATCCGCCCAGCATACAG, RP: 5′-TTCGACCAGGTTACAAGAA) was used to normalize overall expression levels and Acp70A primers used to quantify expression levels (5′-GAATGGCCGTGGAATAGGAA, RP 5′-GGCACCACTTATCACGAGGATT (Chapman et al., Reference Chapman, Bangham, Vinti, Seifried, Lung, Wolfner, Smith and Partridge2003)). Standard curves for both primer pairs were established using serial dilutions of total RNA concentration across four orders of magnitude (Acp70A efficiency: 93·7%, RP49 efficiency: 104·5%). A sub-sample of individual flies were run twice on different Q-RTPCR runs and found to be highly repeatable across PCRs (regression of PCR1 on PCR2 n=52, mean r=0·70 and mean β>0·65, P<0·0001).
Relative Acp70A expression was calculated by taking the difference between the cycle threshold values (Cts) for the housekeeping gene and the Acp70A gene. All reactions were carried out in triplicate and a melt curve produced after each run to check priming specificity. Any Cts that were not within 0·5 cycles of the other triplicates were removed from final analysis. To normalize the data, the largest relative expression level was taken from an individual fly and given the value 1. All other values were converted to a value relative to this.
Log transformed Pfaffle (Pfaffl, Reference Pfaffl2001) and ΔΔCt (Livak & Schmittgen, Reference Livak and Schmittgen2001) were used for analysis and results were essentially identical. As a result only ΔΔCt data will be presented here. Data and residuals were normally distributed (Z=1·141, P=0·148, Z=1·341, P=0·055 respectively). Analyses were conducted using SPSS (SPSS Inc. Version 11 for Mac).
To test for effects on differences in Acp70A gene expression on females, we conducted a separate experiment assessing the duration of the female mating delay. Here males from nine of the experimental lines were collected as virgins and aged as before. Virgin females from a non-related isoline were collected and housed in vials containing up to 40 individuals. Between 1 and 11-five-day-old virgin males from each line were then mated to 3-day-old virgin females of the non-related line. All females were then housed individually and after 48 h they were exposed to virgin males of another non-related line every day for 4 h until all females had remated.
3. Results
We used isofemale lines of D. melanogaster to investigate natural genetic variation in Acp70A expression levels. Q-RTPCR was used to measure the transcript levels of Acp70A in individual males from each line. We also examined if Wolbachia infection affects expression of Acp70A. Using general linear mixed models (GLMMs) with isoline nested within Wolbachia infection status (infected v. uninfected), we found a significant effect of isoline, indicating genetic variation in Acp70A transcript levels across the 15 isolines (F 13, 43=2·64, P<0·01, Fig. 1), with a 5-fold difference in mean Acp70A expression levels across lines. In contrast, Wolbachia infection status was not associated with differences in Acp70A expression levels (F 1, 14=0·34, P=0·57).
Ordinary least-squares regression was used to test for an association between Acp70A transcript levels and female remating rates across isolines. Initial viewing of this association suggested a polynomial relationship and a polynomial regression revealed a significant polynomial association between mean Acp70A expression levels and median time for remating by females mated to males from each line (n=9 lines, r=0·81, F 2, 6=5·64, P=0·04, Fig. 2). The model lost significant explanatory power if either the linear or the quadratic term was removed so both were retained (linear term: t=−2·8, P=0·03, quadratic term: t=3·1, P=0·02).
Discussion
Our major findings were that there was significant genetic variation in Acp70A transcript levels and that this variation had a non-linear effect on female remating. Previously Fiumera et al. (Reference Fiumera, Dumont and Clark2005, Reference Fiumera, Dumont and Clark2007) have shown that DNA sequence variation in some Acp genes is associated with male fitness and that Acp70A polymorphisms are associated with varying refractory periods in females. Sequence polymorphisms may be one mechanism by which transcriptional variation occurs. Here, we show that genetic variation also exists in expression levels of a particular Acp gene, Acp70A with approximately 5-fold differences apparent across isolines. This variation is obviously a pre-requisite for the evolution of Acp70A expression differences, and variation could be maintained by condition dependence, as proposed for other sexually selected traits (Rowe & Houle, Reference Rowe and Houle1996). This possibility remains to be tested.
In D. melanogaster, male Acp stocks become depleted after mating (Monsma et al., Reference Monsma, Harada and Wolfner1990; Linklater et al., Reference Linklater, Wertheim, Wigby and Chapman2007) and are a limiting factor to mating success in another non-Drosophilid fly species (Rogers et al., Reference Rogers, Chapman, Fowler and Pomiankowski2005). This suggests that depletion of Acps is likely to directly influence male fitness (Hihara, Reference Hihara1981), although the full benefits of having larger stocks of SP may only become apparent after repeated mating. High Acp70A expression levels may enhance a male's ability to replenish accessory gland stores of SP more quickly or directly influence male SP-store volume. This remains to be established.
While genetic variation is needed for SP evolution, if there is no phenotypic variation in its effects, there will be no selection on that variation. To that end, SP induces a refractory period in females for up to 5 days post copulation and here we documented a significant association between Acp70A expression levels and the time taken for females to remate after a single copulation with a male from an experimental line. This association was non-linear, with a longer delay for low- and high-expression levels of Acp70A. Both the linear and quadratic effects were significant in our analysis, with the negative linear term evidently explaining some proportion of the variation in the left had section of Fig. 2 (when relative expression was less than c. 0·2). Although we have not investigated all potential effects, this first assessment suggests that there is disruptive selection acting on Acp70A expression as males with intermediate Acp70A expression levels suffered a relative cost in terms of female propensity to remate. As yet, it is not known whether variation in Acp70A expression between isolines directly relates to difference in the amount of SP transferred to females at mating, and/or whether there is variation in the ‘potency’ of the transferred SP as a suppressor of female receptivity. Precisely how this relates to other potential SP effects is unknown, as are associations between this and other Acps, but it appears that there is genetic variation in and selection on expression levels of Acp70A.
We find no effect of Wolbachia infection on expression levels of Acp70A. We acknowledge that with these sample sizes our power is relatively limited, but at this point in time, we must conclude there is no obvious interaction between Wolbachia and Acp70A expression in virgin males. Similarly, in D. simulans, Snook et al. (Reference Snook, Cleland, Wolfner and Karr2000) found no difference in the amount of other Acp proteins (ovulin and Acp36DE) transferred to females by infected and uninfected males. Sperm production, however, was lower in infected males (Snook et al., Reference Snook, Cleland, Wolfner and Karr2000). This sperm deficit is exacerbated as males mate repeatedly, resulting in reduced sperm competitive ability (Champion de Crespigny & Wedell, Reference Champion de Crespigny and Wedell2006). Similarly, Wolbachia infection may only affect Acp70A expression after several matings when either sperm and/or Acp stocks are depleted. Alternatively, there may be less need for Wolbachia-infected males to produce SP because there are fewer sperm to which it can bind. Studies examining the plasticity of SP binding to sperm are needed to test these ideas. Additionally it is as yet unknown whether Wolbachia infection affects D. melanogaster sperm production in the same manner as D. simulans.
In conclusion, we have shown natural variation in Acp70A expression in field-collected isolines of D. melanogaster corresponding to a 5-fold difference in RNA levels. With this variation in Acp70A expression levels we also expected to see phenotypic differences in its effect. However, the observed association was not a simple linear relationship. Instead we found evidence for disruptive selection on Acp70A expression levels through its effects on female remating delays. How this relates to other Acps and male fitness components remains to be investigated.
We would like to thank F. M. Jiggins for donating the fly lines, M. Hares, N. Chamberlain and A. J. Bretman for help with the Q-RTPCR assay, M. F. Wolfner, L. K. Sirot and two anonymous referees for very helpful comments on the manuscript, and the Biotechnology and Biological Science Research Council and Natural Environment Research Council for funding.