The Effects of Host Age, Host Nuclear Background and Temperature on Phenotypic Effects of the Virulent Wolbachia Strain popcorn in Drosophila melanogaster

Corresponding author: K. Tracy Reynolds, Animal and Population Biology, University of Edinburgh, King's Bldg., W. Mains Rd., Edinburgh EH9 3JT, Scotland., tracy.reynolds{at}ed.ac.uk (E-mail)

Communicating editor: M. A. ASMUSSEN

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Because of their obligate endosymbiotic nature, Wolbachia strains by necessity are defined by their phenotypic effects upon their host. Nevertheless, studies on the influence of host background and environmental conditions upon the manifestation of Wolbachia effects are relatively uncommon. Here we examine the behavior of the overreplicating Wolbachia strain popcorn in four different Drosophila melanogaster backgrounds at two temperatures. Unlike other strains of Wolbachia in Drosophila, popcorn has a major fitness impact upon its hosts. The rapid proliferation of popcorn causes cells to rupture, resulting in the premature death of adult hosts. Apart from this effect, we found that popcorn delayed development time, and host background influenced both this trait and the rate of mortality associated with infection. Temperature influenced the impact of popcorn upon host mortality, with no reduction in life span occurring in flies reared at 19°. No effect upon fecundity was found. Contrary to earlier reports, popcorn induced high levels of incompatibility when young males were used in tests, and CI levels declined rapidly with male age. The population dynamics of popcorn-type infections will therefore depend on environmental temperature, host background, and the age structure of the population.

WOLBACHIA is a widespread, maternally inherited bacterium of a wide variety of insects (WERREN et al. 1995 ; JEYAPRAKASH and HOY 2000 ). It can influence the reproductive systems of its hosts so that the production of infected females is favored, thus increasing the number of individuals responsible for its transmission. The most widespread of these changes is cytoplasmic incompatibility (CI). In diploid insects CI is manifested as a reduction in the viability of eggs from matings between infected males and uninfected females. Infected females remain unaffected and, as a result, contribute relatively more offspring to succeeding generations than uninfected females do. The result is an increase in the frequency of the Wolbachia infection (TURELLI and HOFFMANN 1995 ).

Apart from the direct manipulation of reproduction, Wolbachia appears to have few other fitness effects upon insect hosts. This may in part be due to the limited number of species and traits tested; however, it does appear that large deleterious effects in particular are rare. This is probably a consequence of the intimate linkage between host fitness and Wolbachia transmission. However, a strain of Wolbachia that causes the premature death of its host was recently discovered in Drosophila melanogaster during an unrelated screen for gene mutations (MIN and BENZER 1997 ). This strain, named popcorn, appears to be quiescent in the developing fly, but once the adult emerges it enters into a stage of massive overproliferation throughout the body, including the brain. This results in the rupture of cells, which may join other cells to form large masses, a process that results in the early death of the fly. popcorn is maintained in laboratory populations because its detrimental effects occur well after the fly has begun to reproduce. The origin of popcorn is unclear; however, it is possible that it has been generated in the laboratory and does not occur in nature.

On the basis of sequence data, popcorn appears to be very similar to the only other reported strain of Wolbachia, designated wDm, that infects D. melanogaster (MIN and BENZER 1997 ; MCGRAW et al. 2001 ). This suggests that the overreplication phenotype of popcorn has resulted from a mutation in the native D. melanogaster strain rather than from a recent infection event. In addition to overreplication, the popcorn strain differs from the native strain in its inability to induce incompatibility in D. melanogaster (MIN and BENZER 1997 ; MCGRAW et al. 2001 ). However, this seems to be due to a host effect rather than the popcorn strain. If popcorn is transferred to D. simulans, CI is expressed (MCGRAW et al. 2001 ). In D. melanogaster the infection is seen only in the testis sheath cells, but in D. simulans popcorn infects both the testis sheath and the sperm bundles. This suggests that in D. melanogaster a lack of contact between popcorn and the developing sperm prevents it from being modified.

MCGRAW et al. 2002 have demonstrated that the host can also influence some of the fitness effects caused by popcorn. In its native host, D. melanogaster, no reductions in fecundity and hatch rate occur, and this appears to be associated with a lack of overreplication in the ovaries. However, if popcorn is transferred to D. simulans, overreplication in the ovaries does occur and is accompanied by reductions in both of the aforementioned traits. Over a number of generations the density of Wolbachia in the ovaries falls and fecundity and hatch rates increase, suggesting that selection acts on the host to ameliorate these fitness deficits via the suppression of replication. The popcorn infection is therefore considered to reduce longevity in its original host without other deleterious effects and without causing CI and on this basis may provide information about the cytological and physiological basis of these phenotypic effects.

However, recent findings suggest that both CI and fitness effects have not yet been sufficiently rigorously tested (WEEKS et al. 2002 ). One factor that can confound the testing of incompatibility is declining CI expression with male age. This effect has been demonstrated in a number of species, including the D. melanogaster sibling species D. simulans. In this host, infection by the wRi Wolbachia strain results in almost complete embryonic death when young males are used in incompatible crosses, followed by a period of 2–3 weeks when egg hatch increases to normal levels (HOFFMANN et al. 1990 ). Until recently, the effect of male age on CI had not been carefully examined in D. melanogaster. In this host species, incompatibility has been reported as varying from partial to nonexistent. However, when 1-day-old males are used in tests, CI is much stronger, and in some cases >95% of eggs fail to hatch (REYNOLDS and HOFFMANN 2002 ). The rapid rate of decline of CI with age, with CI levels approaching the hatch rates of uninfected control crosses by 5 days of age, appears to have contributed to the variability of CI levels reported for the wDm strain, including cases where CI has been reported as being absent (e.g., HOLDEN et al. 1993 ).

Recent findings have also highlighted the importance of host background on fitness effects associated with Wolbachia. Wolbachia strains can have varied fitness effects, both in degree and nature, in different host species (BOYLE et al. 1993 ; BORDENSTEIN and WERREN 1998 ; GRENIER et al. 1998 ; POINSOT et al. 1998 ; FUJII et al. 2001 ). However, it is becoming increasingly evident that intraspecific host variation can also influence the effects of Wolbachia. In a comparison of two D. melanogaster populations OLSEN et al. 2001 demonstrated that population background can alter the effects of Wolbachia on at least two fitness traits when tested under field conditions. Differences were found for fecundity and for viability in the egg and larval stages. These effects were in turn influenced by environmental conditions. For example, under tropical conditions infected flies from both populations exhibited a fecundity deficit, but under cool temperate conditions one of the populations exhibited enhanced fecundity. Interestingly, no effect of Wolbachia on fecundity has been detected under laboratory conditions in D. melanogaster, further emphasizing the impact of the environment. Two recent studies have also found intraspecific variation in CI expression (CLARK and KARR 2002 ; REYNOLDS and HOFFMANN 2002 ). These findings highlight the importance of carefully investigating phenotypes associated with Wolbachia under a range of conditions (WEEKS et al. 2002 ).

In this study we investigate background and age effects of the popcorn infection in D. melanogaster. The first aim is to test 1- to 5-day-old D. melanogaster males to determine if CI is really absent in the popcorn strain. The second aim is to assess if the popcorn effect on longevity is expressed in different host backgrounds. In addition, the potential effects of popcorn on two other fitness traits (fecundity and development time) in different backgrounds are considered. For both fecundity and longevity, we consider the influence of temperature on any phenotypic effects.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Fly stocks:
The popcorn-infected D. melanogaster stock (designated w¹¹¹⁸) used in these experiments was kindly provided by Scott O'Neill and was originally described in MIN and BENZER 1997 . The Oregon-R line is a well-known laboratory stock. The Mossman and Huonville lines originated from flies collected in the field in tropical Mossman, Queensland, Australia, and temperate Huonville, Tasmania, Australia, in 2000. Both strains were mass-bred from 10 isofemale lines in 2000. Each isofemale line had been treated with tetracycline immediately following collection and hence these stocks were not infected with Wolbachia. These lines represent different genetic backgrounds due to laboratory culture and different selection histories. All stocks were maintained on a yeast-sugar-agar medium that contained the preservatives Nipigin and propionic acid and the antibacterial agents streptomycin and penicillin.

To obtain popcorn-infected flies with different nuclear backgrounds, a backcrossing scheme was used. Virgin popcorn-infected females from the w¹¹¹⁸ stock were mated to virgin males from the Oregon-R, Mossman, and Huonville lines. Virgin F₁ females were collected from each of these crosses and again mated to virgin males from the above lines. This procedure was carried out for a total of five generations. Larvae from each stock, including the original w¹¹¹⁸ stock, were then treated with tetracycline to create uninfected controls. This was done by adding the tetracycline to normal food medium at a concentration of 0.03%. Each stock was then reared for a further generation on medium without tetracycline before experiments were performed to avoid any carryover effects of the tetracycline treatment. All stocks were then tested for their infection status via PCR using the Wolbachia primers "76-99 forward" and "1012-994 reverse" described in O'NEILL et al. 1992 . The D. melanogaster primers "su (s) forward 724-753" and "su (s) reverse 1113-1092" were also included in each reaction as a control for the DNA extraction (VOELKER et al. 1991 ).

Cytoplasmic incompatibility:
Incompatibility was tested using flies from the w¹¹¹⁸ stock. A single virgin male was placed with a single virgin female in a glass vial along with a spoon holding 1.5 ml of a yeast-treacle-agar medium. The medium also contained food dye to aid the visualization of eggs, and yeast dissolved in water was brushed over each spoon to encourage egg laying. The pairs of males and females were monitored for mating. Any pairs that did not mate were removed from the experiment. Following mating the males were removed and the females allowed to lay eggs for a period of 24 hr at 25°. The females were then removed and the eggs counted. Spoons were placed at 25° for a further 24 hr and the number of unhatched eggs counted. Spoons with <10 eggs were not included in the analyses. This resulted in 19–56 replicates per treatment. Confidence limits for incompatibility levels were determined from angular transformed data and these were then transformed back to proportions.

To assess the effects of male age upon incompatibility, males were held together on standard laboratory medium and aged at 25°. These males were then mated to virgin females and the same procedure as above was followed to assess CI levels. Males aged from 1 to 7 days of age were tested, with 18–48 replicates per treatment and age class.

Fecundity:
Fecundity was scored on F₁ offspring from the following three crosses: uninfected female x uninfected male (UF x UM), uninfected female x infected male (UF x IM), and infected female x uninfected male (IF x UM). The progeny of the last cross carried the popcorn infection, while the progeny of the other two crosses did not. Fecundity was scored over three periods of the females' life span using a method similar to that detailed above for scoring incompatibility, except in this case males and females were held together for the entire period of the experiment. During a scoring period, each pair of flies was placed with spoons containing yeast-treacle-agar medium. The spoons were replaced every 24 hr. Between each scoring period pairs of flies were held in vials containing laboratory medium. All host lines were tested at 25°, and the w¹¹¹⁸ line was also tested at 19°. In the comparisons made at 25° we tested 17–24 replicates for each treatment at each scoring period, and for the comparisons at 19° we tested 51–70 replicates for early and middle fecundity and 26–36 replicates for late fecundity.

Egg-to-adult development time:
To measure the effect of popcorn upon development time, the period between egg lay and pupal emergence was monitored using eggs obtained from the three crosses described above (UF x UM, UF x IM, and IF x UM). Groups of 10–15 eggs that were laid in a 24-hr period were transferred to vials containing 25 ml of laboratory medium. Each vial contained a small card for larval pupation. The vials were placed at 25° until pupal development was complete. The adults that emerged were then counted and removed from the vials on a daily basis. Between 30 and 40 replicate vials were used for each cross-type.

Longevity:
The effect of popcorn upon the longevity of flies was assessed at 25° by holding groups of 10 female flies together in vials containing 25 ml of laboratory medium and monitoring the number of individuals that died every 4–5 days. The flies were moved to new vials with fresh medium at 7-day intervals. The females used were F₁ offspring obtained from the crosses mentioned above (UF x UM, UF x IM, and IF x UM), with 10 replicate vials per cross-type. This experiment was also conducted at 19° using the w¹¹¹⁸ line, but in this case the females assayed were those that were also used in the 19° fecundity tests, with the flies held in male-female pairs rather than in groups. In this test 70–100 F₁ females from each cross were assayed.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Cytoplasmic incompatibility:
In contrast to the findings of MIN and BENZER 1997 and MCGRAW et al. 2002 , we found that the popcorn Wolbachia strain is capable of causing strong CI in D. melanogaster. In crosses between 1-day-old infected males and uninfected females, mean egg hatch failure was 0.77 (95% confidence limits of 0.67, 0.86, N = 56) compared to a mean failure rate of 0.02 (95% confidence limits of 0, 0.04, N = 19) in crosses where both sexes were uninfected. This difference was significant by a nonparametric Mann-Whitney test (Z = -5.950, P < 0.001). Egg hatch in the incompatible cross also differed significantly from that in the reciprocal uninfected male x infected female cross (Z = -7.695, P < 0.001). No significant difference was found between the reciprocal cross (mean hatch failure rate 0.06, 95% confidence limits of 0.03, 0.11, N = 56) and the control uninfected male x uninfected female cross (Z = -1.445, P = 0.148), indicating that popcorn is capable of restoring compatibility.

Male age and incompatibility:
Following the finding of a high level of incompatibility in D. melanogaster, we tested the effects of male age upon egg hatch rates. The high CI induced by 1-day-old males declined rapidly with increasing age, with egg hatch rates approaching those of the control uninfected x uninfected cross by day 7 (Fig 1), although the differences between the crosses remain significant for all ages tested (Mann-Whitney tests, P < 0.01 for all days).

View larger version (19K): In this window In a new window Download PPT slide   Figure 1. The effect of male age on the level of CI induced by the popcorn infection in the w1118 D. melanogaster line. Error bars indicate 95% confidence intervals.

Longevity:
The popcorn-infected flies appeared to die faster than their uninfected counterparts in all of the backgrounds (Fig 2). To compare mortality rates, survival curves using the Kaplan-Meier method were created and compared with log-rank tests. At 25° there were significant differences between the survival curves of infected flies and uninfected controls for all nuclear backgrounds (Oregon-R, ² = 7.62, d.f. = 1, P < 0.01; Huonville, ² = 67.17, d.f. = 1, P < 0.001; w¹¹¹⁸, ² = 41.52, d.f. = 1, P < 0.001; Mossman, ² = 9.27, d.f. = 1, P < 0.01). Thus a reduction in longevity as a result of infection by popcorn was evident in the four backgrounds. However, comparisons between the uninfected control survival curves revealed significant differences in some cases, suggesting that nuclear background also influenced survival time. As a result, direct comparisons between the curves using this method were not possible. Instead, the Cox-Hazard regression model was used to calculate the relative risk ratio of infection for each population. Essentially, this model is used to determine relative differences in survival between groups. The difference is expressed as a relative risk ratio. In this instance the relative risk ratio represents the difference in the risk of death associated with infection status. In the Oregon-R and Mossman backgrounds, the increased risk of death associated with infection was small with risks 1.45 times (95% confidence intervals of 1.04, 2.03, P < 0.05) and 1.51 times (95% confidence intervals of 1.09, 2.10, P = 0.01) that of the uninfected controls respectively. The risk was slightly greater in the w¹¹¹⁸ background with a risk 2.33 (95% confidence intervals of 1.70, 3.19, P < 0.001) times that of the uninfected controls. In the Huonville background, the risk was greater than twice that found in the Oregon-R and Mossman backgrounds with an increased risk of death of 3.92 times (95% confidence intervals of 2.67, 5.75, P < 0.001) that of the control flies. For unknown reasons, UF x IM F₁ flies with Huonville and Mossman backgrounds had lower rates of mortality than their UF x UM F₁ counterparts (² = 11.84 and 14.18, respectively, d.f. = 1, P < 0.001 in both cases). No differences in mortality rates were found between w¹¹¹⁸-infected and uninfected flies at 19° (Fig 3).

View larger version (27K): In this window In a new window Download PPT slide   Figure 2. The effect of popcorn on mortality rate in four D. melanogaster nuclear backgrounds at 25°. In the case of the Oregon-R, Mossman, and Huonville backgrounds, the F1 progeny of three crosses (indicated in parentheses) were tested. The progeny from two of the crosses were uninfected while the offspring from the third cross were infected. In the case of the w1118 line, mortality was tested on the offspring of two of the crosses. Un M, uninfected male; Un F, uninfected female; In F, infected female; In M, infected male. Error bars indicate standard errors.

View larger version (22K): In this window In a new window Download PPT slide   Figure 3. The effect of popcorn on mortality rate in the w1118 line at 19°. The data points represent absolute values rather than means.

Fecundity:
Our initial fecundity test was conducted at 19° and involved only flies from the w¹¹¹⁸ line. Fecundity was scored when females were 1–5 (early fecundity), 10–14 (middle fecundity), and 30–34 (late fecundity) days of age (Table 1). ANOVAs indicated significant differences among the F₁ offspring of the three crosses for early (F = 6.80, d.f. = 2, 186, P = 0.001) and middle (F = 5.506, d.f. = 2, 164, P = 0.005) fecundity. Comparisons via Tukey B tests revealed that there was no consistent pattern for these two periods. For early fecundity, the uninfected F₁ offspring from the IM x UF cross had a significantly higher mean fecundity than the offspring of the other two crosses. However, for middle fecundity, the popcorn-infected offspring had a higher mean fecundity than the uninfected offspring from the IM x UF cross, but this was not significantly different from the fecundity of the uninfected offspring from the UM x UF cross. The reversal in the relationship between the UM x IF and IM x UF F₁ crosses may be indicative of a trade-off between early and middle fecundity. No significant differences were found for late fecundity.

The popcorn infection had little effect upon fecundity in any of the four nuclear backgrounds when tested at 25°. Fecundity was scored when females were 1–3 (early fecundity), 11–13 (middle fecundity), and 18–20 (late fecundity) days of age (Table 1). Significant differences were found in only two instances: for early fecundity in the Mossman (F = 5.17, d.f. = 2, 63, P < 0.01) and Huonville (F = 4.823, d.f. = 2, 62, P < 0.05) backgrounds, but these differences were not attributable to infection status. For the Huonville background, the uninfected offspring from the UM x UF cross had a lower early fecundity than the offspring from the other two crosses, which did not differ from each other. For the Mossman background, the uninfected offspring of the IM x UF cross had a significantly higher mean fecundity than the infected F₁'s from the UM x IF cross, but neither differed from the uninfected offspring of the UM x UF cross. No differences were found in the w¹¹¹⁸ line at this temperature. These results and those from the first test suggest no consistent effect of infection status upon fecundity.

Egg-to-adult development time:
A significant effect of the popcorn infection upon egg-to-adult development time was found in two of the nuclear backgrounds tested. In the Oregon-R and Huonville backgrounds (Fig 4), popcorn had a delaying effect upon emergence. This effect was most pronounced in the Huonville background, where the proportion of infected adults that emerged was significantly lower than the proportion of uninfected adults on days 9–12 (Kruskal-Wallis, P < 0.001 on all days). By day 10, >85% of the uninfected adults had emerged, but no infected adults had yet emerged. Most of the infected adults emerged between days 11 and 12. A similar effect was also found in the Oregon-R background, where the proportion of popcorn-infected adults that emerged was significantly lower on days 11–13 (Kruskal-Wallis, P < 0.001 on all days). On day 12, >85% of uninfected adults had emerged, in contrast to only 41% of popcorn-infected adults.

View larger version (26K): In this window In a new window Download PPT slide   Figure 4. The effect of popcorn on development time in four D. melanogaster nuclear backgrounds at 25°. As was done for mortality rate, the progeny of three crosses were tested, with the offspring of two of the crosses being uninfected and the other infected. Abbreviations are as in Fig 2. Error bars indicate standard errors.

For the other two backgrounds, significant differences were found only on day 9 (Kruskal-Wallis, P < 0.001 for both backgrounds). However, these differences were not consistent with the above results. For the Mossman flies, the uninfected F₁'s from IM x UF cross had the slowest development rate, while for the w¹¹¹⁸ flies all offspring types differed significantly from each other on day 9, with the infected flies having the fastest rate of development.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Contrary to earlier reports we have found that popcorn does not differ from the wDm Wolbachia strain in its ability to cause CI in D. melanogaster. The high level of CI induced by young males and its rapid decline with age is consistent with the behavior of the native nonvirulent wDm strain (REYNOLDS and HOFFMANN 2002 ). The similar rates of CI decline for both strains in D. melanogaster also suggest a common host response to both infection types. The ability of popcorn to induce high levels of CI when transferred to D. simulans (MCGRAW et al. 2001 ) is also in keeping with the behavior of wDm, and together these findings suggest that popcorn does not differ from the native strain in its CI characteristics. The failure of previous studies to detect CI in popcorn-infected D. melanogaster may be due to its rapid decline with age. This decline is further accelerated if the males are allowed to repeatedly mate (KARR et al. 1998 ; REYNOLDS and HOFFMANN 2002 ). MIN and BENZER 1997 do not mention the age of the males used or the methodology of their original tests, so it is possible that both factors could have affected their findings. Age and repeated mating may also have contributed to the lack of CI found by MCGRAW et al. 2002 . Although young males were used, mating pairs were allowed to remain together for up to 3 days.

The suggestion of MCGRAW et al. 2001 that the ability of popcorn and, by implication, other strains to induce CI is dependent on its ability to infect sperm bundles may still prove to be correct. These authors observed popcorn to be present in the sperm bundles of D. simulans males, but absent in the sperm bundles of D. melanogaster. On the basis of the presumption that popcorn did not cause CI in D. melanogaster, MCGRAW et al. 2001 concluded that popcorn could not infect the sperm bundles of this host and hence was unable to modify sperm. However, if the link between sperm bundle infection and CI induction is correct, then popcorn must infect the sperm bundles of young D. melanogaster males at some stage. Perhaps infection of sperm bundles is prevented in older D. melanogaster males. This could account for the decline in CI with male age in both D. melanogaster and D. simulans, with the effect taking longer to occur in the latter. Alternatively, it is possible that infection of sperm bundles is not associated with CI induction.

The absence of a reduction in life span in infected flies at 19° was somewhat surprising. This suggests that either overreplication of popcorn did not commence or bacterial multiplication was too slow at this temperature to reach lethal levels prior to the flies dying of other causes. Bacterial multiplication within the host might occur more rapidly with increasing temperature, reducing the time taken for lethal levels of cell rupture to occur. At 25° the rate of mortality began to diverge from that of the uninfected flies from 20 days of age onward, compared to 8 days of age at 29° as found by MIN and BENZER 1997 . It has been suggested that the effect of popcorn upon life span could prove useful in reducing populations of pest insects. However, the temperature dependency of this trait may restrict its use to certain climatic conditions.

The lack of a fecundity effect in any of the nuclear backgrounds may reflect the distribution of the infection in reproductive tissues. MCGRAW et al. 2002 found that rather than increasing over time as in other tissues, the density of popcorn within the ovaries of D. melanogaster females remained low and similar to that of the wDm infection. As was the case with the wDm infection, no effect on fecundity was detected. In contrast, popcorn density did increase in the ovaries of D. simulans females and this was associated with a dramatic decline in fecundity. MCGRAW et al. 2002 suggested that the lack of overreplication may be the result of the evolution of host control overreplication in the ovaries to reduce any fecundity effect. However, these findings need to be interpreted cautiously, as environmental influences alter the effect of Wolbachia upon fecundity; OLSEN et al. 2001 found that under some field conditions wDm can have a negative impact on fecundity.

It is possible that the effect of popcorn on development time is strain specific and is linked in some way to the overreplication phenotype. OLSEN et al. 2001 found no differences in development time between infected and uninfected D. melanogaster flies carrying the wDm strain under field conditions. Similarly, the Riverside Wolbachia strain does not appear to affect development time in D. simulans under laboratory conditions (HOFFMANN et al. 1990 ) and a Wolbachia infection in Leptopilina heterotoma also has no impact (FLEURY et al. 2000 ). However, our results suggest that host background can influence the effect of Wolbachia on development. Perhaps other Wolbachia strains could also influence development if tested in other backgrounds and under a variety of conditions. It would also be of interest to test for differences in developmental delays between males and females, as it is possible that popcorn may behave differently in each sex.

The host differences in popcorn-induced mortality could be due to delays in the onset of popcorn overreplication or the rate of replication of popcorn may be slowed down in some hosts. The evolution of the host toward an amelioration of the effect on life span would be expected; however, the two host lines that had the lowest risk of mortality associated with infection have had no previous contact with popcorn. Perhaps some D. melanogaster lines have evolved greater control of the replication of the wDm strain, and because of the similarities between wDm and popcorn, this ability can be transferred to popcorn. We recently demonstrated, for example, that host-specific differences exist in D. melanogaster for the expression of incompatibility (REYNOLDS and HOFFMANN 2002 ). Alternatively, perhaps some host environments are simply less suitable than others for Wolbachia replication. The host strain in which the delay of egg-to-adult development was most pronounced, Huonville, also suffered the greatest increase in mortality rate. However, in the case of the Oregon-R background, both development time and the onset of increased mortality were delayed. As we did not test directly the onset of overreplication we cannot ascertain if differences in the timing of this event are responsible for these effects. Perhaps Wolbachia can exert other unknown effects upon the host prior to emergence that can delay pupal development.

Would popcorn have an impact on field populations of Drosophila? This would depend on the average life span of flies and the expression of fitness effects under field conditions. It is not known how long flies live in the field, although evidence from studies of CI levels suggests that D. simulans males can live for at least 2 weeks (HOFFMANN et al. 1990 ). However, on the basis of the results here, if they do not live beyond this age, then the impact of popcorn may not be particularly strong. In addition, substantial periods of reproduction would have taken place prior to the death of most individuals, further weakening its impact. It is possible, however, that increases in mortality may occur at a much earlier age in the more stressful field environment.

Could popcorn prove useful in pest control strategies? Although our tests were carried out under laboratory conditions, they indicate that under some conditions and in some host backgrounds popcorn can have negative effects upon fitness that may restrict its usefulness. For example, delayed emergence could be detrimental to the spread of popcorn in environments where resources are a limiting factor. The rapid decline in CI may also inhibit its usefulness, as could any delay in the onset of increased mortality in adults. In particular, the mortality effect may not be expressed under certain conditions. However, the effects of Wolbachia strains need to be tested in the host background and under the environmental conditions in which they are to be used. Negative fitness effects that occur under laboratory conditions may not necessarily occur in the field and vice versa. Similarly, simple relationships between fitness effects and their actual impact cannot be assumed. For example, the relationship between incompatibility levels and infection frequencies in field populations is complex. Studies of the wDm strain in Australian populations of D. melanogaster have shown that infection frequencies can be close to fixation in tropical climates (HOFFMANN et al. 1998 ) despite an apparent rapid decline in CI.

In summary, our findings emphasize the need to rigorously test the effects of Wolbachia before ascribing strains with particular effects (or lack thereof). Traits that are commonly used to describe Wolbachia strains, such as incompatibility levels and fecundity effects, are often not an invariable property of a particular strain, but appear to be the result of an interaction of both Wolbachia and host genotypes and the environment. Here we have shown that host background and the environment can determine not only the degree to which an effect is manifested, but also whether it is exhibited at all. If Wolbachia strains are to be used in control strategies for pest insects, these factors need to be taken into account.

	ACKNOWLEDGMENTS

This work was supported by the Australian Research Council via their Large Grant and Special Research Centre programs.

Manuscript received July 17, 2002; Accepted for publication March 24, 2003.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

BORDENSTEIN, S. R. and J. H. WERREN, 1998  Effects of A and B Wolbachia and host genotype on interspecies cytoplasmic incompatibility in Nasonia. Genetics 148:1833-1844.[Abstract/Free Full Text]

BOYLE, L., S. L. O'NEIL, H. M. ROBERTSON, and T. L. KARR, 1993  Interspecific and intraspecific horizontal transfer of Wolbachia in Drosophila.. Science 260:1796-1799.[Abstract/Free Full Text]

CLARK, M. E. and T. L. KARR, 2002  Distribution of Wolbachia within Drosophila reproductive tissue: implications for the expression of cytoplasmic incompatibility. Integ. Comp. Biol. 42:332-339.

FLEURY, F., F. VAVRE, N. RIS, P. FOUILLET, and M. BOULETREAU, 2000  Physiological cost induced by the maternally-transmitted endosymbionts Wolbachia in the Drosophila parasitoid Leptopilina heterotoma.. Parasitology 121:493-500.

FUJII, Y., D. KAGEYAMA, S. HOSHIZAKI, H. ISHIKAWA, and T. SASAKI, 2001  Transfection of Wolbachia in Lepidoptera: the feminizer of the adzuki bean borer Ostrinia scapulalis causes male killing in the Mediterranean flour moth Ephestia kuehniella.. Proc. R. Soc. Lond. Ser. B 268:855-859.[Medline]

GRENIER, S., B. PINTUREAU, A. HEDDI, F. LASSABLIERE, and C. JAGER et al., 1998  Successful horizontal transfer of Wolbachia symbionts between Trichogramma wasps. Proc. R. Soc. Lond. Ser. B 265:1441-1445.

HOFFMANN, A. A., M. TURELLI, and L. G. HARSHMAN, 1990  Factors affecting the distribution of cytoplasmic incompatibility in Drosophila simulans. Genetics 126:933-948.[Abstract]

HOFFMANN, A. A., M. HERCUS, and H. DAGHER, 1998  Population dynamics of the Wolbachia infection causing cytoplasmic incompatibility in Drosophila melanogaster.. Genetics 148:221-231.[Abstract/Free Full Text]

HOLDEN, P. R., P. JONES, and J. F. Y. BROOKFIELD, 1993  Evidence for a Wolbachia symbiont in Drosophila melanogaster.. Genet. Res. 62:23-29.[Medline]

JEYAPRAKASH, A. and M. A. HOY, 2000  Long PCR improves Wolbachia DNA amplification: wsp sequences found in 76% of sixty-three arthropod species. Insect Mol. Biol. 9:393-405.[Medline]

KARR, T. L., W. YANG, and M. E. FEDER, 1998  Overcoming cytoplasmic incompatibility in Drosophila.. Proc. R. Soc. Lond. Ser. B 265:391-395.[Medline]

MCGRAW, E. A., D. J. MERRITT, J. N. DROLLER, and S. L. O'NEILL, 2001  Wolbachia-mediated sperm modification is dependent on the host genotype in Drosophila. Proc. R. Soc. Lond. Ser. B 268:2565-2570.[Medline]

MCGRAW, E. A., D. J. MERRITT, J. N. DROLLER, and S. L. O'NEILL, 2002  Wolbachia density and virulence attenuation after transfer into a novel host. Proc. Natl. Acad. Sci. USA 99:2918-2923.[Abstract/Free Full Text]

MIN, K. T. and S. BENZER, 1997  Wolbachia, normally a symbiont of Drosophila, can be virulent, causing degeneration and early death. Proc. Natl. Acad. Sci. USA 94:10792-10796.[Abstract/Free Full Text]

OLSEN, K., K. T. REYNOLDS, and A. A. HOFFMANN, 2001  A field cage test of the effects of the endosymbiont Wolbachia on Drosophila melanogaster.. Heredity 86:731-737.[Medline]

O'NEILL, S. L., R. GIORDANO, A. M. E. COLBERT, T. L. KARR, and H. M. ROBERTSON, 1992  16S rRNA phylogenetic analysis of the bacterial endosymbionts associated with cytoplasmic incompatibility in insects. Proc. Natl. Acad. Sci. USA 89:2699-2702.[Abstract/Free Full Text]

POINSOT, D., K. BOURTZIS, G. MARKAKIS, C. SAVAKIS, and H. MERCOT, 1998  Wolbachia transfer from Drosophila melanogaster into D. simulans: host effect and cytoplasmic incompatibility relationships. Genetics 150:227-237.[Abstract/Free Full Text]

REYNOLDS, K. T. and A. A. HOFFMANN, 2002  Male age, host effects and the weak expression or non-expression of cytoplasmic incompatibility in Drosophila strains infected by maternally transmitted Wolbachia.. Genet. Res. 80:79-87.[Medline]

TURELLI, M. and A. A. HOFFMANN, 1995  Cytoplasmic incompatibility in Drosophila simulans: dynamics and parameter estimates from natural populations. Genetics 140:1319-1338.[Abstract]

VOELKER, R. A., W. GIBSON, J. P. GRAVES, J. F. STERLING, and M. T. EISENBERG, 1991  The Drosophila suppressor of sable gene encodes a polypeptide with regions similar to those of RNA-binding proteins. Mol. Cell. Biol. 11:894-905.[Abstract/Free Full Text]

WEEKS, A. R., K. T. REYNOLDS, and A. A. HOFFMANN, 2002  Wolbachia dynamics and host effects: What has (and has not) been demonstrated? Trends Ecol. Evol. 17:257-262.

WERREN, J. H., D. M. WINDSOR, and L. GUO, 1995  Distribution of Wolbachia among Neotropical arthropods. Proc. R. Soc. Lond. Ser. B 262:197-204.

This article has been cited by other articles:

				Y. Peng, J. E. Nielsen, J. P. Cunningham, and E. A. McGraw Wolbachia Infection Alters Olfactory-Cued Locomotion in Drosophila spp. Appl. Envir. Microbiol., July 1, 2008; 74(13): 3943 - 3948. [Abstract] [Full Text] [PDF]