1. A dose-dependent decrease in male fertility occurs in the Indian Meal Moth, Plodia interpunctella, when sub-lethally infected with granulovirus during the larval stage.
2. Here, the causes for this decline are investigated by examining eupyrene and apyrene sperm numbers and sizes produced by males across four levels of viral challenge.
3. The results could not explain how reduced male fertility is caused in this host–pathogen interaction. While a reduction in both eupyrene and apyrene sperm numbers from all virus-treated males was found, this was not significant and neither was there a difference in sperm lengths across the four treatments. There were also no differences in the variances of sperm numbers or lengths between the doses, and no associations between sperm numbers or lengths and body size were found.
4. A significant correlation between eupyrene and apyrene numbers was found, but this was independent of dose. Significant between-male variance in apyrene sperm lengths was found, indicating that individual males differ in the range of apyrene sperm sizes they produce.
5. It is suggested that further intracellular and behavioural study is needed to identify the causes of the granulovirus-induced reduction in fertility of P. interpunctella.
Diseases play an important role in the population dynamics of vertebrates (e.g. Anderson & May 1986, 1991) and invertebrates (e.g. Anderson & May 1981). In recent years, insect pathogens have received particular attention because of their potential as biocontrol agents (Payne 1988; Hochberg 1989). Insects are infected by a number of different pathogens and, of these, the baculoviruses (essentially the nucleopolyhedroviruses and granuloviruses) have been studied extensively (see Tanada & Kaya 1993). Typically, they infect Lepidoptera and Hymenoptera and often exhibit a high degree of species-specificity. In Lepidoptera only the larval stages become infected, after ingesting virus particles. Common features of the host–baculovirus interaction are (i) disease-induced mortality is dose-dependent and (ii) susceptibility to infection decreases as host age increases (see Tanada & Kaya 1993). Consequently, some hosts survive the pathogen challenge but they may nonetheless incur sub-lethal costs which are thought to have an important influence upon host population dynamics (e.g. Anderson & May 1981; Sait, Begon & Thompson 1994a).
Sait et al. (1994b) showed that after mating with sub-lethally affected male Plodia interpunctella (Hübner) (Lepidoptera: Pyralidae), the fertility of healthy female mates was significantly reduced. Furthermore, this infertility effect was greater as the dose of virus, and hence challenge to the host, was increased. Female fertility could not have been affected by viral cross-infection from males because the pathogen can only affect the host once it has been ingested in the larval stage (see Tanada & Kaya 1993). One possible explanation is that male reproductive capacity was affected by diverting resources away from spermatogenesis in order to combat the infection. Alternatively (or in combination), reduced male fertility may be the result of deleterious viral activity in the cells of tissues like the fat body, one of the primary targets of the virus (Tanada & Kaya 1993). Hormonal changes associated with virus infections (e.g. Subrahmanyam & Ramakrishnan 1980, 1981; O’Reilly & Miller 1989) may affect production or development of viable eggs or sperm (Riddiford & Williams 1967; Gelbic & Metwally 1981), but this has not been measured directly in sub-lethally affected individuals, nor related directly to observed sub-lethal effects.
A common reason for diminished male fertility in the animal kingdom is oligospermia, where insufficient numbers of spermatozoa are produced to effect full fertility. In humans, for example, fertility problems occur in males producing less than 20 million spermatozoa per ml of semen (Herlihy, Lee & Lipshulz 1987). Less is known about this phenomenon in insects, although there is evidence that low sperm numbers can cause a decline in female fertility (e.g. Linley & Hinds 1974). Gage & Cook (1994) also showed that dietary protein restrictions in the larval stage limited the production of sperm numbers in adult P. interpunctella since, in Lepidoptera, the acquisition of resources for spermatogenesis often occurs during the larval phase (Leviatan & Freidländer 1979). Resource restrictions may also cause the production of unusually large proportions of ‘abnormal’ spermatozoa, another potential cause of infertility. In rats, for example, protein deficiency leads to the production of a larger proportion of abnormal sperm (Vawda & Mandlwana 1990). If a sub-lethal pathogen infection limits resources available for spermatogenesis, an increase in the number of abnormal sperm may result. In Lepidoptera, resource restrictions could limit elongation of the flagellum and mitochondrial derivatives at later stages of spermatogenesis (Leviatan & Freidländer 1979) causing the production of shorter sperm.
Here, we investigate the known dose-dependent fertility-reducing effects of sub-lethal virus infections in male P. interpunctella by measuring ejaculate sperm numbers and sperm lengths. Plodia interpunctella, like almost all Lepidoptera, produces two distinct sperm types; a nucleate eupyrene form and an anucleate apyrene form (Meves 1902; Gage & Cook 1994). The function of apyrene sperm has not been established (Silberglied, Shepherd & Dickinson 1984) but its anucleate status precludes any direct function in fertilization. By investigating the effects of viral infection upon production of both sperm types, the importance of apyrene sperm in the reproductive dynamics of this lepidopteran is further examined.
Materials and methods
The stock culture of P. interpunctella had been maintained as an out-breeding colony for a number of years. The rearing diet consisted of 800 g wheat bran, 160 g dried brewer's yeast, 200 ml glycerol, 200 ml clear honey, 1 g sorbic acid, 1 g methyl paraben. Moths were maintained at 28 ± 1 °C, 60 ± 5% r.h. (ranges) and a 13/11-h light/dark cycle.
It is important to note that the materials and methodology used in the Sait et al. (1994b) study were replicated closely here; the same moth and virus stocks were used, as well as the same range of male P. interpunctella sample sizes. Consequently, given the significant reduction in fertility recorded by Sait et al. (1994b), we would expect to detect a significant difference in sperm numbers and sizes if there was a sub-lethal effect on these sperm characteristics. The granulovirus (GV) was extracted from infected hosts and purified through sucrose gradients and differential centrifugation (Smith & Crook 1988). Three virus doses (serial dilutions of a stock virus suspension of 2·403 × 105, 2·403 × 106 and 2·403 × 107 virus particles/male larva) and a control solution were used. The virus challenge here was greater than that in the study by Sait et al. (1994b), though a similar range of doses were administered (6·36 × 104, 6·36 × 105 and 6·36 × 106 virus particles/larva in that study). Since fifth instar larvae do not die from the viral disease, infection levels between this study and that of Sait et al. (1994b) cannot be compared directly. However, comparison of the dose–mortality responses for instars 1–4 between bioassays demonstrates that infection levels are closely maintained (Sait, Begon & Thompson 1994c). For example, in two different bioassays, log LD50 values for first instar larvae (with upper and lower log 95% CL) were 2·81 (+ 2·90, – 2·72) and 2·72 (+ 2·82, – 2·61) virus particles. Similarly, for third instar larvae the log LD50 values were 4·96 (+ 5·05, – 4·87) and 4·82 (+ 4·96, – 4·78) virus particles (see Sait et al. 1994c for full details of comparisons). Thus, since infection levels are similar across studies, we maintain that the virus challenge for fifth instar males in this study is comparable with Sait et al. (1994b).
Male larvae were selected randomly from mass culture when 21-days-old (approximately mid-fifth instar) and dosed using a droplet-feeding technique (Sait et al. 1994b,c). To reiterate, fifth instar hosts in this interaction do not die from the lethal disease (but do suffer sub-lethal effects –Sait et al. 1994b) so all treated insects were potentially affected sub-lethally here. Larvae were placed individually in 1-cm3 compartments of 25-compartment Petri dishes with excess food and allowed to develop to adult emergence.
Two days after emergence, virgin treatment males were mated once with randomly selected virgin females that had recently emerged from healthy stock and were sexually receptive. Immediately after mating, females were dissected and ejaculate size determined by established protocols (Gage & Cook 1994). Briefly, the intact spermatophore was removed from each mated female, ruptured in modified Barth saline (Gurdon 1991) to release the ejaculate, and dispersed under magnification by gentle stirring. The eupyrene bundles, each consisting of 256 spermatozoa (Gage & Cook 1994), were counted under × 40 magnification. The dispersed solution was washed into a 30-ml specimen tube with Barth saline and diluted with distilled water. Six 10-μl subsamples of each diluent were air-dried on glass slides and the numbers of apyrene sperm in each dried 10-μl smear were counted under × 100 magnification using dark-field phase contrast microscopy. Numbers of apyrene sperm per ejaculate were calculated by multiplying the average 10-μl subsample sperm count by its dilution factor.
The eupyrene and apyrene sperm on the slides were measured using video image analysis under dark-field phase contrast microscopy (see Gage & Cook 1994 for further details). Essentially, images of sperm were processed using Visilog (Noesis Ltd, France) 4.1.3 image analysis software, reduced to a single line of pixels to represent the backbone of the original sperm image, and the area of this line calculated. Ten apyrene sperm and three to ten eupyrene sperm per male were counted. Sperm lengths were converted to means per male for analysis.
Male forewing length was measured as a function of body size in order to examine the potential effect of male size on sperm numbers or size, and whether a sub-lethal infection affected this relationship. There is a significant correlation between body mass and forewing length regardless of treatment with virus (Sait 1992), and testis size shows positive allometry with body mass in P. interpunctella (Gage 1995).
There were no differences in male size between any of the treatments. Mean winglength (± SD, n = 20 for each treatment): 6·0 ± 0·3, 6·1 ± 0·2, 6·2 ± 0·2 and 6·0 ± 0·3 mm for control, low, medium and high dose treatments, respectively; F3,76 = 1·7, P > 0·05.
There were no significant differences between mean eupyrene sperm numbers across the four doses (F3,150 = 1·72, P > 0·05; Table 1), nor any correlation between sperm number and dose. Comparisons of untreated males with virus-treated males (pooling the three doses) showed that eupyrene sperm numbers were reduced in infected males, but this difference was not quite significant (8835·7 ± 509·2 and 7876·5 ± 237·9 for control and treated males, respectively (mean (SE); F1,152 = 3·44, P = 0·066). If the 33% (maximum) reduction in fertility observed by Sait et al. (1994b) was caused by a decline in sperm numbers alone, we would expect the reduction measured in this study to be of a similar magnitude. Therefore, the power of our ANOVA was determined to detect a minimum difference of 30% between the two most different means of the four treatments. A power test (Sokal & Rohlf 1995) was performed at the 5% level of significance with sample sizes of 35 in each category. To detect a decline of 2651 eupyrene sperm (a 30% reduction compared with the control) the ANOVA has a power of 93%.
Table 1. . Mean numbers (± SE) of eupyrene and apyrene sperm from males treated with a control solution and three levels of virus challenge (low = 2·403 × 105; medium = 2·403 × 106; high = 2·403 × 107 virus particles per male)
As above, there were no significant differences in mean apyrene sperm numbers between treatments (F3,150 = 1·74, P > 0·05; Table 1), nor any correlation with dose. Infected males (pooled) again produced fewer apyrene sperm but the difference was not quite significant (77 791·5 ± 4799·1 and 68 608·9 ± 2350·6 for control and treated males, respectively (mean ± SE); F1,152 = 3·3, P = 0·071). A power test (5% significance level, sample sizes of 35) showed a 90% probability of detecting a 30% difference in apyrene numbers between the two most different means; a reduction of at least 23 337 apyrene sperm compared with the control.
The proportions of eupyrene:apyrene sperm did not differ between treatments (F3,150 = 0·14, P > 0·05) nor between infected and uninfected males (F1,152 = 0·16, P > 0·05).
Analysis of covariance revealed no correlations between body size and ejaculate sperm numbers or proportions, either across 154 matings or among the four treatments. Eupyrene sperm numbers did not correlate with body size (common slope of four treatments: VR (1 df) = 0·0, P > 0·05; difference between slopes: VR (3 df) = 1·13, P < 0·05). Similarly, apyrene numbers did not correlate with body size (common slope: VR (1 df) = 0·79, P > 0·05; difference between slopes: VR (3 df) = 1·72, P < 0·05). There were no correlations between male body size and the proportions of eupyrene (to apyrene) sperm in the ejaculate (common slope: VR (1 df) = 0·1, P > 0·05; difference between slopes: VR (3 df) = 2·1, P < 0·05).
Across 154 matings, however, there was a strong correlation between eupyrene and apyrene ejaculate numbers (Fig. 1; common slope: VR (1 df) = 35·8, P < 0·001) and no differences between treatments (VR (3 df) = 1·28, P > 0·05).
There were no differences across the four doses in eupyrene length (mean sperm lengths per male are independent data for analysis: F3,32 = 0·6, P > 0·05) or apyrene length (F3,37 = 0·4, P > 0·05; Table 2). Furthermore, there were no differences in sperm lengths from infected (combined) vs uninfected males (eupyrene, F1,34 = 0·0, P > 0·05; apyrene, F1,39 = 0·45, P > 0·05).
Table 2. . Mean lengths (± SE) of eupyrene and apyrene sperm from males treated with a control solution and three levels of virus challenge
There was no difference in the variances of apyrene sperm length across the four treatments (F3,36 = 0·55, P > 0·05). However, there was significant between-male variance in apyrene sperm length (F40,369 = 1·5, P = 0·04), suggesting that individual males differ in the specific distribution of apyrene sperm lengths they produce. We could not detect sufficient eupyrene sperm for repeated measurement of between-treatment or between-male variance because of dilution for apyrene counting.
No correlation was found between male body size and mean eupyrene (r2 = 0·015, P > 0·05, n = 20) or apyrene lengths (r2 = 0·0, P > 0·05, n = 25).
Our principal finding was that the significant decline in fertility of healthy females mated to sub-lethally infected P. interpunctella males (Sait et al. 1994b) could not be explained by a reduction in sperm number alone. Sait et al. (1994b) observed mean dose-dependent reductions of 20%, 29% and 33% in female fertility across the three increasing sub-lethal male doses, but here, using higher doses, we found no associated decline in sperm numbers or sizes, nor evidence of a dose-dependent effect. Despite the tendency for infected males to produce fewer sperm, the power test emphasized that this non-significant decline in sperm number does not explain the combined 27% drop in the fertility of females when mated to infected males (Sait et al. 1994b). Sub-lethal effects are notoriously difficult to observe because individuals will vary in the extent to which they are affected (Sait et al. 1994b; Goulson & Cory 1995). However, we also detected no differences in the variances of sperm numbers or apyrene sperm lengths between the four male categories. On the other hand, these results do suggest that one possible mechanism of infertility, resource restriction as a result of the host combating the sub-lethal challenge, may not be a factor since it is known that sperm numbers are reduced in dietary protein-restricted P. interpunctella (Gage & Cook 1994).
It is possible that sub-lethal viral doses disrupt male fertility via intracellular infection of sperm organelles. ‘Virus-like particles’ (VLPs), thought to be retrotransposons, have been observed within the testes and spermatocyst cell cytoplasm in Heliothis species (Lepidoptera; Noctuidae) (Degrugillier 1989; Degrugillier & Newman 1993). The VLPs were detected in both eupyrene and apyrene spermatocyst cells and there was an association between their presence and the degree of eupyrene abnormality (Degrugillier & Newman 1993). While the origin of VLPs is unclear, GV-derived particles in sub-lethally infected P. interpunctella males may also be incorporated into spermatozoa, causing structural or genetic disruption. At the gross level no evidence of sperm abnormality, such as split or paired flagella, was seen, but detailed ultrastructural examination may reveal a potential source of male infertility at the subcellular level.
Pathogen infection could also influence sperm function, causing decreases in motility or longevity, but direct in vitro and in vivo behavioural examinations of sperm are necessary to test this. Although sperm longevity can be extensive in insects owing to female maintenance of sperm in the specialized spermatheca (Tschinkel 1987), there is also some evidence for sperm death or loss during storage (Tsubaki & Yamagishi 1991). Viral challenge during spermatogenesis could induce premature sperm mortality in Lepidoptera and we are currently investigating egg hatch-rate, and whether infertility occurs constantly throughout the period of oviposition, or whether it increases with female, and hence sperm, age.
In addition to sperm, male Lepidoptera provide their mates with accessory secretions in the ejaculate. Nutrients from these secretions are incorporated into the eggs (Boggs & Gilbert 1979; Greenfield 1982) and may be used to increase female fecundity (Wiklund et al. 1993). Most adult lepidopterans accrue all resources for reproduction during the larval stage and adult females may therefore be protein-limited for reproduction. Recently, it has been shown that female lepidopterans may forage actively for male-derived nutrients by mating multiply to realize their full reproductive potential, despite receiving sufficient sperm to fertilize all the eggs from a single mating (Kaitala & Wiklund 1994). Sub-lethal virus infections may affect production or quality of these protein constituents which, in combination with the decrease in sperm numbers demonstrated here, could account for the observed reduction in fertility.
Sub-lethal viral infection reduces male reproductive efficiency (Sait et al. 1994b), possibly through genetic disruption, or via resource restriction for reproduction and spermatogenesis. Despite these challenging effects, there was no effect upon apyrene production (they still constitute almost 90% of the ejaculate in P. interpunctella; see Table 1, and Gage & Cook 1994; Cook & Gage 1995), lending further support to the idea that apyrene sperm play a central role in reproduction. Moreover, it is found that apyrene numbers correlate closely with eupyrene numbers (Fig. 1), suggesting a functional link between the two. Cook & Gage (1995) suggested that apyrene sperm may function as a ‘cheap filler’, negating female sexual receptivity while the eupyrene sperm fertilize. Thus, apyrene sperm function as ‘ejaculate protectors’ and their numbers may correlate with eupyrene numbers accordingly. No difference in the ratios of eupyrene:apyrene sperm at different body sizes was found, suggesting that different-sized males are not playing different reproductive strategies. Similarly, no relationship between total sperm numbers and body size was found. Across male P. interpunctella, testis size shows positive allometry with body size (Gage 1995), so it seems likely that total sperm output will be constrained by body size. However, here it appears that individual ejaculate size (at least for a male's first mating) is maintained, possibly trading against the total number of ejaculates that can be produced in subsequent matings.
Interestingly, significant between-male variance in apyrene sperm length was found, a similar phenomenon being found in Dung Flies (Ward & Hauschteck-Jünger 1993) and Atlantic Salmon (M. J. G. Gage, unpublished data). However, since no relationship has been described between sperm length and body size, either here or in the other studies, the cause or significance of this phenomenon is not known. Such variance, though, may have implications for between-male differences in fertility or sperm competitiveness.
Despite detailed examinations of sperm sizes and numbers across four levels of pathogen challenge, we could not explain how sub-lethal viral infections cause male infertility in P. interpunctella. Since total sperm complement per ejaculate, and the sizes of individual spermatozoa, have not been decreased, we suggest that disruption may have occurred directly at the intracellular level, rather than via resource restrictions for spermatogenesis which we know induces decreases in sperm number (Gage & Cook 1994). Sperm function may have been affected, perhaps causing increased age-dependent mortality of sperm, or inefficiencies in motility. We suggest that investigations of sperm internal structure and sperm behaviour from virally challenged male moths may indicate how sub-lethal infections cause male infertility in this host–pathogen interaction.
We would like to thank the two referees, Carol Boggs and Dave Goulson, for their constructive criticism of the manuscript. This work was conducted while SMS was financed by NERC grant GR3/8039, MJGG was in receipt of a Royal Society Fellowship and PAC was funded by a BBSRC studentship.