- Top of page
- Materials and Methods
- Conflict of Interest
- Supporting Information
Mating between close relatives often leads to a decrease in fitness known as inbreeding depression (Lynch and Walsh 1998), which can be strong enough to drive small populations to extinction (Saccheri et al. 1998; O'Grady et al. 2006). Understanding the effects of inbreeding on reproductive success is becoming increasingly important, as many animal populations become smaller and more fragmented, thus increasing the likelihood of mating between close relatives (Frankham et al. 2002). Inbreeding depression can be caused by either the loss of high-fitness heterozygotes (the overdominance hypothesis) or by increased expression of deleterious recessives (the partial dominance hypothesis) (Lynch and Walsh 1998). Inbreeding increases homozygosity, and according to the overdominance hypothesis, the decrease in the frequency of high-fitness heterozygotes leads to a decline in fitness. The partial dominance hypothesis (that currently has the greatest support; Charlesworth and Willis 2009) proposes that increasing homozygosity unmasks deleterious recessive alleles leading to a fitness decline.
Directional dominance is required for inbreeding depression (Falconer and Mackay 1996; Roff 1997; Lynch and Walsh 1998). Directionality arises due to natural or sexual selection driving alleles that increase fitness to fixation and eliminating dominant deleterious alleles, whereas deleterious recessive alleles are hidden from selection and hence maintained at low frequencies (Falconer and Mackay 1996; Roff 1997; Lynch and Walsh 1998). These deleterious recessives are then expressed in inbred individuals that are more homozygous than outbred individuals (Falconer and Mackay 1996; Roff 1997; Lynch and Walsh 1998). Thus, traits closely linked to fitness are predicted to show strong inbreeding depression (Falconer and Mackay 1996; Roff 1997; Lynch and Walsh 1998) and several studies support this prediction (Falconer and Mackay 1996; Roff 1998; DeRose and Roff 1999; Wright et al. 2008; but see Ellmer and Andersson 2004). In contrast, directional dominance should be low for traits weakly associated with fitness or those under stabilizing selection because mutations moving trait values up or down will be selectively equivalent and hence such traits are predicted to exhibit little-to-no inbreeding depression (Lynch and Walsh 1998). Therefore, inbreeding depression in a trait is a signature of directional selection in the past, whereas a lack of inbreeding depression suggests either stabilizing or weak selection in the past (Falconer and Mackay 1996; Roff 1997; Lynch and Walsh 1998; Ketola et al. in press).Distinguishing stabilizing selection from weak selection based on the lack of inbreeding depression is not possible without further knowledge on selection acting on the trait in question. Also, it is important to bear in mind that recessive alleles in very important fitness traits may have been purged under small population size by strong directional selection, which can substantially lower inbreeding depression (Falconer and Mackay 1996; Ketola et al. in press). In addition, the degree of inbreeding depression is nonlinearly dependent on the frequency of the recessive alleles in the population: the strongest inbreeding depression occurs when the frequency of recessive alleles is intermediate (Falconer and Mackay 1996). Thus, the differences in the magnitude of the inbreeding depression can also depend on the numbers of loci coding for traits. As the number of loci involved in trait expression increases, selection per locus weakens, and this weakening maintains recessive alleles in higher frequencies and results in higher inbreeding depression in the trait (Falconer and Mackay 1996). Inbreeding depression is further affected by the genomic mutation rate U, with higher U causing stronger inbreeding depression (Roff 1997).
For traits that have been under directional selection, the partial dominance hypothesis of inbreeding depression predicts that inbreeding moves trait values away from high fitness because deleterious recessive alleles will always change trait values in the direction opposing the long-term past selection (Falconer and Mackay 1996; Roff 1997; Lynch and Walsh 1998; Ketola et al. in press). Hence the direction of inbreeding depression can be used to identify trait values associated with high fitness, although this approach has not been widely applied (but see Mackay 1985; Mallet and Chippindale 2011; Ketola and Kotiaho 2012). Here, we aim to better understand the evolution of sperm length using the inbreeding method. Even though comparative investigations in a wide range of taxa have found relationships between sperm length and the level of sperm competition (reviewed by Snook 2005; Pitnick et al. 2009a; Pizzari and Parker 2009), selection acting on sperm length is not well understood in most taxa, despite sperm competition being widespread and credited with driving the rapid diversification of ejaculate traits (Parker 1970; Ravi Ram and Wolfner 2007; Pitnick et al. 2009a,b; Pizzari and Parker 2009). Thorough investigation of sperm length variation in Drosophila, however, using comparative, experimental, genetic, and functional approaches has demonstrated rapid diversification of sperm length (Pitnick et al. 1999, 2003; Manier et al. 2013). Longer sperm have a selective advantage because they are better at displacing, and resisting displacement, by competitor sperm from female sperm storage organs (Miller and Pitnick 2002; Pattarini et al. 2006; Lüpold et al. 2012; M. K. Manier, J. M. Belote, S. Lüpold, O. Ala-Honkola, K. S. Berben, W. T. Starmer and S. Pitnick, unpubl. data), and sperm length has a close association with the intensity of sexual selection (Bjork and Pitnick 2006). If there is strong directional selection for longer sperm, as experimental studies suggest, we would expect inbreeding to decrease sperm length. On the other hand, substantive energetic and life-history costs of manufacturing relatively long sperm have been demonstrated (Pitnick et al. 1995; Pitnick 1996; Immler et al. 2011), and thus net selection on sperm length may be stabilizing. Indeed, a hemiclonal analysis of D. melanogaster found high heritability but low evolvability (the coefficient of additive genetic variation; Houle 1992) of sperm length and concluded this pattern was consistent with stabilizing selection (Morrow et al. 2008). In this case, theory predicts that directional dominance for sperm length will be low and that there will be no inbreeding depression in sperm length (Roff 1997; Lynch and Walsh 1998).
Several recent studies have shown that inbreeding typically decreases sperm competitiveness (Hughes 1997; Konior et al. 2005; Zajitschek et al. 2009; Michalczyk et al. 2010; Simmons 2011). However, the mechanisms leading to the lower fertilization success of inbred males remain unresolved. Correlational studies documented that sperm number and quality (e.g., sperm motility or proportion of morphologically normal sperm in an ejaculate) are often lower in inbred populations (Wildt et al. 1982; Roldan et al. 1998; Gomendio et al. 2000; Margulis and Walsh 2002; van Eldik et al. 2006; Gage et al. 2006; Fitzpatrick and Evans 2009; Weeks et al. 2009). In addition, an experimental investigation of the guppy, Poecilia reticulata, showed that inbreeding decreases sperm numbers (Zajitschek and Brooks 2010), but such an effect was not found in three-spined sticklebacks, Gasterosteus aculeatus (Mehlis et al. 2012). Despite the obvious connection between sperm traits and sperm competition success, experimental studies that simultaneously measure the effects of inbreeding on competitive fertilization success and characteristics of the sperm themselves are currently lacking.
In addition to sperm, other male reproductive characters are closely linked to fitness and therefore should also be sensitive to inbreeding depression. Male mating success has been shown to be a major fitness component in D. melanogaster (Prout 1971a,b; Bundgaard and Christiansen 1972) and, as predicted, it decreases with inbreeding (Brittnacher 1981; Sharp 1984; Partridge et al. 1985; Miller et al. 1993; Hughes 1995; Enders and Nunney 2010). Reduced male mating success due to inbreeding has also been reported in the housefly Musca domestica (Meffert and Bryant 1991), the butterfly Bicyclus anynana (Joron and Brakefield 2003), two species of poeciliid fish (van Oosterhout et al. 2003; Mariette et al. 2006; Ala-Honkola et al. 2009) and in the decorated cricket Gryllodes sigillatus (Ketola and Kotiaho 2010). In the fly D. simulans, inbreeding reduced male attractiveness (Okada et al. 2011), as measured by copulation latency, a standard measure of attractiveness (attractive males mate faster: Ritchie et al. 1999; Barth et al. 1997; Taylor et al. 2008). Inbred males have also been shown to be less attractive in the house mouse Mus musculus (Ilmonen et al. 2009), the zebra finch Taeniopygia guttata (Bolund et al. 2010), the guppy P. reticulata (Zajitschek and Brooks 2010), and the meal worm beetle Tenebrio molitor (Pölkki et al. 2012).
Here, we report on an investigation of the influence of inbreeding on male attractiveness (mating latency) and competitive fertilization success in D. melanogaster. We also assess the impact of inbreeding on several characters that affect sperm competition success, such as sperm length, ejaculate size, in vivo sperm swimming speed, as well as sperm viability and sperm storage in the female reproductive tract, with the goal of inferring the mode of past selection on all these traits. These characters are all key determinants of male reproductive success and hence should be closely linked to fitness. Indeed, previous work has shown that male reproductive success is the most important meta-trait determining male fitness (Prout 1971a,b; Bundgaard and Christiansen 1972), and sperm competitiveness and sperm length have been shown to directly determine male fitness in D. melanogaster (Miller and Pitnick 2002; Pattarini et al. 2006; Fricke et al. 2010; Lüpold et al. 2012). Ejaculate size has been shown to positively correlate with the amount of previous male's sperm displaced in the same population as in this study (Manier et al. 2010) suggesting selection for larger ejaculate size. Also, relatively slow and/or long sperm have been shown to be better at displacing resident sperm from storage (Lüpold et al. 2012) suggesting selection for slower sperm. By assessing inbreeding depression or lack thereof in these traits, we can infer the history of past selection acting on them: inbreeding depression implies a history of directional selection (to generate the directional dominance needed to cause inbreeding depression), whereas a lack of inbreeding depression implies stabilizing or weak selection (as mutations up or down are selectively equivalent).
The use of transgenic flies with either red or green fluorescently tagged sperm heads (Manier et al. 2010) allowed us to distinguish between the ejaculates of two males in competition within the female reproductive tract and to quantify aspects of ejaculate quality and fate (Fig. 1, Manier et al. 2010; Lüpold et al. 2011, 2012). As the severity of inbreeding depression in a population depends (among other things) on historical population size and ancestral inbreeding, its level is difficult to predict a priori (Falconer and Mackay 1996; Frankham et al. 2002). Both Zajitschek et al. (2009) and Robinson et al. (2009) did not find inbreeding depression in sperm competition success at low levels on inbreeding (f = 0.25 or less), so we used two levels of inbreeding in this study: a theoretical f = 0.25 (one generation of full-sibling breeding) represents a realistic level of inbreeding in nature (Keller and Waller 2002) and f = 0.5 (three generations of full-sibling breeding) represents severe inbreeding.
Figure 1. Drosophila melanogaster lines with red or green fluorescently tagged sperm heads allowed us to distinguish between the ejaculates of two males in competition within the female reproductive tract (here inside female seminal receptacle).
Download figure to PowerPoint
- Top of page
- Materials and Methods
- Conflict of Interest
- Supporting Information
Our demonstration of strong inbreeding depression (87% for highly inbred lines and 65% for moderately inbred lines) in male attractiveness (i.e., mating latency) and mild inbreeding depression in male remating latency (6% for highly inbred lines) is consistent with earlier studies on D. melanogaster that have documented strong inbreeding depression in male mating ability (Brittnacher 1981; Sharp 1984; Partridge et al. 1985; Miller et al. 1993; Hughes 1995; Enders and Nunney 2010). In these studies, male mating ability was measured in competitive mating trials (N inbred males competing against N outbred males for N females and the proportion of offspring sired used as a measure of male mating ability). Our finding that mating latency is increased due to inbreeding may provide the explanatory mechanism for the low male mating ability observed in these earlier studies. This relationship appears to be widespread, as mating latency or male sexual motivation has similarly been shown to suffer inbreeding depression in several other species (Joron and Brakefield 2003; van Oosterhout et al. 2003; Mariette et al. 2006; Ala-Honkola et al. 2009; Ketola and Kotiaho 2010; Okada et al. 2011). Our findings are also consistent with previous studies that have inferred directional selection on male attractiveness (Hosken et al. 2008; Ilmonen et al. 2009; Bolund et al. 2010; Zajitschek and Brooks 2010; Okada et al. 2011; Pölkki et al. 2012).
Also in line with earlier studies (Hughes 1997; Konior et al. 2005; Zajitschek et al. 2009; Michalczyk et al. 2010; Simmons 2011), we found that inbreeding decreases the competitive fertilization success of males, which may indicate a history of directional selection for higher P2 in the LHM population of D. melanogaster. Our results are further consistent with those of Zajitschek et al. (2009) and Robinson et al. (2009) in that they similarly found no decrease in male competitive fertilization success with moderate levels of inbreeding (theoretical f = 0.25) and a significant decline only associated with a higher level of inbreeding (theoretical f = 0.5 in this study and 0.59 in Zajitschek et al. 2009). Our P2 results were further consistent with an earlier study on the same base population (LHM) that found no response to directional selection on P2 (Bjork et al. 2007) indicating relatively little additive genetic variance for this trait due to past selection having fixed alleles that increase it. Furthermore, Hughes (1997) concluded that most of the genetic variation in this trait is dominance variance, which renders P2 a mostly nonheritable trait (e.g., Bjork et al. 2007), the variation of which is maintained predominantly by antagonistic pleiotropy and ejaculate × ejaculate and ejaculate × female interactions (Clark et al. 1999; Clark 2002; Bjork et al. 2007; Fiumera et al. 2007). Hughes (1997) and Fiumera et al. (2007) suggested P2 was under stabilizing selection, but our findings may suggest otherwise, at least historically, as traits under stabilizing selection should show little-to-no inbreeding depression (Lynch and Walsh 1998).
Of course, 3% inbreeding depression in P2 is not severe, but note that P2 in this study was lower in highly inbred males despite females taking longer prior to remating with such males compared to females remating with outbred control males (Table 1). This difference means that our test of the influence of inbreeding on P2 was particularly conservative, because females mated to highly inbred males should have used more of the first male's sperm prior to remating, a prediction that was supported by slightly greater numbers of offspring produced prior to remating by females that were remating with highly inbred compared to outbred control males (pair-wise difference, P = 0.053; Table 2). Hence, all else being equal, highly inbred males would be expected to have had higher P2 values than the outbred control males. Indeed, an earlier study using the same experimental material did find a positive correlation between the number of eggs produced prior to remating and P2 (see Table 2 in Ala-Honkola et al. 2011). We further consider the demonstrated inbreeding effect on P2 to be conservative as P2 values were unusually high experiment-wise (Fig. 2; 0.95 to 0.98, compared to previous reports of about 0.8: e.g., Morrow et al. 2005; Bjork et al. 2007; but see Clark et al. 1999). In our previous study on the same LHM-RFP population, P2 values were about 0.8 when competitor males were from a brown-eyed line (LHMbwD; Ala-Honkola et al. 2011). Hence, our standard competitor males were unusually uncompetitive, and it is likely that differences among inbreeding treatments would have been magnified had we used competitor males from a line better in sperm competition.
In contrast to male attractiveness and P2, we did not find inbreeding depression for sperm length, suggesting length has a history of stabilizing selection in this population. This result is surprising as sperm competition studies with D. melanogaster lines selected for increased or decreased sperm length (Miller and Pitnick 2002; Pattarini et al. 2006) or using isogenic lines derived from the same LHM population as this study (Lüpold et al. 2012), have consistently demonstrated a long-sperm advantage, as manifested by a superior ability to displace, and to resist displacement by, competitor sperm. Although such a long-sperm advantage suggests that sperm length might be under directional selection, a hemiclonal analysis of the LHM population concluded that sperm length is under stabilizing selection (Morrow et al. 2008). The lack of inbreeding depression in this study is consistent with that finding. The sperm length selection experiment (with a different population of D. melanogaster) showed that sperm length responds to intense bidirectional artificial selection (after 18 generations, increasing by 5.3% in the high line and decreasing by 7.9% in the low line), so there is substantial additive genetic variation for this character (a mean realized heritability of 0.48; Miller and Pitnick 2002). The lack of strong asymmetry in response to selection also suggests that this trait has not been under strong directional selection (Falconer and Mackay 1996).
True stabilizing selection tends to reduce genetic variation (Robertson 1956; Falconer and Mackay 1996), but given that empirical studies suggest that sperm length harbors ample genetic variance (e.g., Miller and Pitnick 2002; Morrow et al. 2008; Lüpold et al. 2012), it is likely that an intermediate optimum for sperm length arises from pleiotropic effects. For example, the production of longer sperm has been associated with substantive energetic and life-history costs (Pitnick and Markow 1994; Pitnick et al. 1995; Pitnick 1996), which could explain the net stabilizing selection (Falconer and Mackay 1996, pp. 344–347). The inbreeding method alone cannot distinguish weak selection from stabilizing selection as both scenarios predict little-to-no inbreeding depression for a trait (Lynch and Walsh 1998). However, we suggest that there is enough evidence for selection on sperm length from previous research (see above) to support our conclusion of the history of stabilizing selection on this trait in D. melanogaster.
In addition to sperm length, we simultaneously assayed numerous ejaculate characteristics with the dual goal of quantifying the extent of their inbreeding depression and of discerning the mechanisms underlying any treatment differences in competitive fertilization success. Although we were unable to confidently resolve the underlying causes of the reduced fertilization success in highly inbred males, we did demonstrate that it was not attributable to (1) fewer sperm per ejaculate, (2) a reduction in the number of sperm stored, or (3) lower sperm viability in the female reproductive tract but our statistical power to detect changes in sperm numbers was very low. Viability of inbred males’ sperm was measured indirectly from offspring production (measured over 10 days) by outbred females. Offspring production or offspring viability of females singly mated to inbred males did not differ from that of females mated to outbred males, suggesting that inbred males’ sperm survive equally well in the female reproductive tract and fertilize eggs as efficiently as sperm of outbred males. This contrasts with findings of inbreeding depression for male fertility in D. simulans (Okada et al. 2011), although Michalczyk et al. (2010) similarly found no difference in fertility or offspring viability of inbred (eight generations of full-sibling mating) flour beetle Tribolium castaneum males in a single-mating situation, despite inbred males having decreased competitive fertilization success.
When female Drosophila remate, some of the first male's sperm are released or displaced from the storage organs and eventually ejected by the female along with excess second-male sperm (Snook and Hosken 2004; Manier et al. 2010). Relatively slow and/or long sperm have been shown to be better at displacing resident sperm from storage (Lüpold et al. 2012), indicating selection for slower sperm. As no inbreeding depression was found in sperm swimming speed, either weak or stabilizing selection has acted on it (Lynch and Walsh 1998). We suggest that sperm swimming speed is likely to be under stabilizing selection. However, this requires confirmation.
Sperm competition success is a complex trait that may be affected by the number of sperm ejaculated (e.g., Boschetto et al. 2011), sperm mobility (e.g., Gage et al. 2004), sperm morphology (e.g., Oppliger et al. 2003; Lüpold et al. 2012), sperm viability (e.g., García-González and Simmons 2005), seminal fluid proteins (e.g., Chapman et al. 2000), female genotype (e.g., Wilson et al. 1997; Clark and Begun 1998; Birkhead et al. 2004), and ejaculate-female interactions (see review in Pitnick et al. 2009b). As we measured most of these parameters and found no differences between inbred and outbred males, we cannot presently explain the mechanism(s) underlying the reduced sperm competitiveness of highly inbred males. Unexamined mechanisms that could explain these findings include seminal fluid proteins and unknown aspects of female sperm choice (Eberhard 1996; Birkhead 1998; Pitnick and Brown 2000).
Directly assessing the history of selection on male ejaculate characteristics has proven difficult in most instances due to the large number of potential selective bouts that could define net selection, and because much of this occurs cryptically within the female reproductive tract. We have employed a novel method to assess selection and see this as a new way to provide insight into these key male fitness determinants. This study suggests that male attractiveness and possibly also competitive fertilization success (i.e., P2) have been under directional selection, whereas sperm length has a history of stabilizing selection.