Stocks:

Two inbred lines, I4 and I10, were chosen because they display a high level of early onset adult mortality. Discovery of these lethal effects and the origin of the lines is described elsewhere (Vermeulen and Bijlsma 2004). Summarizing, I4 showed high mortality in the first week of adult life that was expressed only at high temperature (29°), whereas I10 showed the reverse effect, expressing high mortality only at low temperature (21°). The outbred control line, O6, is a randomly chosen line from the control lines used in the same experiment. We used line O6 as a control line for both inbred lines in all experiments described here. All these lines were established in 1997. The mortality effects of I4 and I10 were first demonstrated in 1999, 2 years after establishment, indicating that the genetic constitution underlying these effects is stable. Throughout this study, the line effects are described as though they are caused by single genes and alleles. This assumption will be tested in further research.

Maintenance and experimental rearing conditions:

Lines have been maintained in large numbers in quarter-pint bottles (30 ml standard medium: 26 g dead yeast, 54 g sugar, 17 g agar, and 13 ml nipagine solution per liter) at 25° and 40–60% relative humidity. To obtain experimental animals, eggs were collected from an overnight egg-laying session (lethal conditions and dominance experiment) or from a 4-hr egg-laying period (adult and preadult transfer experiments). If adult flies were required, these eggs were transferred to quarter-pint bottles [standard medium + ampicillin (100 mg/liter)] under noncrowding conditions (300 eggs/bottle). The bottles were then transferred to an incubator. The resulting flies were collected as virgins by use of carbon dioxide or low temperature (on ice) to immobilize the flies.

Adult survival:

To assess the degree of dominance, lifetime survival was determined. Dead flies were scored three times per week for this experiment. For all other experiments, survival during the first 2 weeks, hereafter called early adult survival, was assessed. Dead flies were scored every day for 14 days, by which the lethal phase (LP) could be established. Both for lifetime survival and for early adult survival, vials were refreshed twice a week for 29° and higher temperatures and once a week for 25° and lower temperatures. Mortality was determined at multiple intervals allowing for determination of mean life span. Lifetime survival data were analyzed with a nonparametric test (Kruskal-Wallis one-way analysis of variance by ranks). Life table analysis of mortality and survival patterns was performed as in Lee (1992). All data were analyzed in SPSS (SPSS for windows, Release 11.0) and in Statistix for Windows (Version 1.0 analytical software).

Lethal conditions:

For inferences about I4, flies of this line and the O6 control were reared at 25° and adult males were collected as virgins at eclosion using carbon dioxide. Flies were transferred to the treatment temperature and early adult survival was monitored. The temperature treatments were 21°, 29°, and 31° (±0.5°) and 50 males were used at each temperature. For inferences about I10, flies of this line and the O6 control line were reared at 25° and adults were collected as virgins at eclosion using carbon dioxide and then separated by sex and transferred to glass reaction tubes in groups of 10 flies. To assess sensitivity to cold shock, tubes with flies were suspended in water baths containing either ice water (0°) or water at 25°. Treatments consisted of suspension of 2 hr in the 25° bath, 2 hr in the 0° bath, or a succession of 90 min at 25° followed by 30 min at 0°, which required a total of 2 hr. Thereafter, flies were transferred to vials containing standard medium and transferred to one of three treatment temperatures. In total this resulted in nine treatments consisting of all combinations of cold shock (0, 30, and 120 min) and treatment temperature (15°, 20°, and 29° ± 0.5°). For each treatment 30 flies were used per sex. Early adult mortality was monitored for both experiments.

Adult shift experiments:

Successive transfers from the permissive temperature (PT) to the restrictive temperature (RT) were made to establish the end of the temperature-sensitive period (TSP) as described by Suzuki (1970). If flies were transferred before the TSP had ended, they died during the LP (depicted in Figure 1A). Only flies transferred after the TSP survived the LP. Experimental flies were obtained by collecting eggs of I4 and I10 that were transferred to an incubator where flies were allowed to develop at either 21° (the PT for I4 treatments and the RT for the I10 lethal control) or 29° (the PT for I10 treatments and the RT for the I4 lethal control). The resulting flies were collected as virgins (age between 0 and 4 hr) using ice to immobilize the flies. A total of 30 flies for each treatment and sex were distributed to vials with 9 ml of food. Due to extremely low viability of the I10 line, analysis of each treatment except the RT control had to be restricted to 15 individuals. Vials were transferred to 21° and 29° (±0.5°) to construct the following treatments: a positive control for the lethal effect that remained constantly at RT (RT control), an immediate transfer at eclosion from PT to RT (0 hr), a transfer to RT after 4 hr at PT and after 8, 12, 16, and 20 hr, and a negative control that remained at PT (PT control; see Figure 1B). Early adult survival was recorded. Vials were refreshed on day 1 and thereafter twice a week for 29° and once a week for 21°.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.—

Transfer protocols for the adult shift experiment. (A) Schematic of the procedure for the detection of the end of the TSP. Shifts from PT to RT before and in TSP will lead to death in the LP, whereas shifts beyond the TSP will not. (B) Successive adult transfers to the RT as performed for I4 and I10 adults to detect the end of the TSP (see text for details).

Preadult lethal effects:

Transfer experiments during developmental stages were performed to determine whether sensitivity to the restrictive conditions was limited to the adult phase. For I4 and the O6 control lines, five transfer treatments were generated: always kept at RT, always at PT, a transfer to RT in the larval stage at 3 days after egg laying, in the pupal stage at 7 days, and as a freshly eclosed adult at age 0–12 hr (see Figure 2A). For I10 and the O6 control lines, five other treatments were generated: a negative control with no cold shock or transfer and four treatments involving a cold shock followed by transfer to RT. These conditions were administered in the larval stage at 3 days, in the pupal stage at 7 days, at eclosion at age 0–12 hr, and at 3 days after eclosion (see Figure 2B). Cold shocks were administered by placing vials 2 hr in ice. Every treatment consisted of 10 vials with 100 eggs from which five adult males/vial were collected as virgins at eclosion using carbon dioxide, resulting in 50 adults/treatment. For I4, the RT was at 29° and the PT at 20°. For I10, the RT was 20° and cold-shock duration was 120 min at 0°. The PT was at 25°. The number of pupae, eclosing males and females, and early adult survival were determined. Egg-to-adult viability was taken as the percentage of eclosing flies and data were angular transformed prior to analysis with a t-test.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.—

Transfer protocols for detection of preadult lethal effects. (A) Shift experiments for the I4 line to the RT (29°) were performed on larvae, pupae, and freshly eclosed adults. (B) Shift experiments for the I10 line to the RT (20°) were accompanied by a cold shock (CS) at 0° and were performed on larvae, pupae, freshly eclosed adults, and 3-day-old adults. RT and PT control are control treatments. The treatment names at the right of A and B correspond to those in Figure 5 (see text for details).

Outcrosses to assess dominance:

For information on the amount of dominance of the lethals, crosses were performed between the inbred lines and the control line. Virgin females and males were obtained from the culture. For every cross, at least 100 virgin females and 100 males were used for inbred lines and 50 of each for O6. F₁ progeny of reciprocal mass crosses between the inbred lines and the control line were collected as virgins (0–10 hr old) using ice to immobilize the flies. A total of 40 flies for each cross and sex were distributed to vials with 9 ml medium (eight vials of 5 flies for both sexes) except for I10, because insufficient adults eclosed (20 females and 30 males for each temperature). Vials were transferred to 21° and 29° (±0.5°) and lifetime survival was scored.

Lethal conditions:

To demonstrate the conditional expression of the mortality pattern, both inbred lines were tested at 20° and 29°. Also, in an attempt to increase the expression, both lines were challenged with an additional temperature (31° for I4 and 15° for I10).

For I4, 31° was considered a likely temperature to maximize the contrast in survival between I4 and the control line O6. At this temperature males become sterile and at higher temperatures early adult survival of control flies is affected (David 1988). Only males were assessed, since previous investigations showed that females do not express the lethal effect at 29°. Figure 3A shows survival through the LP of I4 males at 20° and 29° (negative and positive control for expression, respectively) and 31°. Mortality at temperature regimes of 20° and 29° confirmed previous findings (Vermeulen and Bijlsma 2004). The O6 control line showed no premature mortality in this period (survival was 96% and higher; data not shown). The expression of the I4 lethal effect was diminished at 31°, although survival still was significantly lower than that at 20° (chi-square test of contingency P = 0.006). Adults appear to be capable of phenotypic adaptation, similar to that which occurs during development (see Adult shift experiments below). Possibly, exposure to 31° triggered a heat-shock response that increased survival (Dahlgaard et al. 1998).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.—

Male early adult survival under different sets of conditions. (A) I4 survival through the lethal phase for three adult temperatures. (B) I10 survival through the lethal phase for three adult temperatures following a cold shock at 0° for 30 min. The shaded area in A and B shows the location of the LP in these experiments (see also Figure 4, A and C).

Lowering the adult temperature to 21° was not sufficient to trigger the lethal effect, but in addition required the cold shock that was applied during collection of recently eclosed adults by immobilizing them on ice. The results of experiments that demonstrate conditional mortality (both sexes pooled) are shown in Table 1. Survival of males through the lethal phase after a 30-min cold shock is shown in Figure 3B. None of the treatment affected the O6 control line (survival 95% and higher; data not shown). A cold shock and a RT of 20° or lower had a large effect on I10 (Table 1). A cold shock of 30 min was sufficient to trigger the lethal effect in full, but penetrance also depended significantly on subsequent adult temperature (Table 1). Adult exposure to 20° or lower caused significantly high mortality, whereas at 29° flies did not express the lethal effect (Figure 3B, Table 1). Both males and females of I10 were affected, but females experienced a lower mortality rate (data not shown, but see Figure 4D).

• Download figure
• Open in new tab
• Download powerpoint

Figure 4.—

The LP and the end of the TSP as determined by shift experiments. (A) The LP and (B) survival after the LP of I4. Treatments include six successive transfers after eclosion to the restrictive temperature and the controls that remain at the RT and PT. (C) The LP and (D) survival after the LP of I10. Flies of all treatments were chilled at 0° once at eclosion (<4 hr after eclosion). Error bars represent the standard error and are based on binomial variances.

Adult shift experiments:

Conditional lethals traditionally are characterized by two biologically relevant periods, the LP and the TSP (Suzuki 1970). These periods can be determined by using shift experiments, in which cultures are shifted between the permissive and restrictive temperature. The end of the TSP was determined by shifting freshly eclosed flies to the RT at successive intervals. Figure 1A illustrates how transfer to the RT before or during the TSP will lead to the death of the flies in the LP. There is, however, a point after the TSP has ended after which the transfer will no longer harm the flies (Figure 4).

For I4, daily mortality pooled over all transfer times (0–20 hr, Figure 4B) during the first 2 weeks reveals a large mortality peak from 6 to 13 days after eclosion (Figure 4A). This is the LP during which mortality caused by the ts lethal effect occurs. Figure 4B depicts the survival after the LP and thus, of all respective transfer times, the number of flies still alive after 13 days. Flies in the PT treatment experienced high survival throughout the LP (100% survival), but the RT treatment failed to fully show the lethal effect (77% survival). However, in all other treatments the lethal effect was prominent (survival between 17 and 43%). This pattern can be explained by two alternatives: either the flies phenotypically adapt to the RT during development or they experience a LP prior to eclosion, killing all sensitive flies. The following section on preadult lethal effects describes experiments indicating that the latter explanation is incorrect. The absence of a point between treatments after which transfer to the RT no longer leads to mortality during the LP indicates that the TSP is continuous throughout adult life.

In the shifts of the I10 line, pooled survival over all treatments during the first two weeks shows a large mortality peak from day 2 to day 4 after eclosion (Figure 4C). Survival after the LP is depicted in Figure 4D. Here, too, the RT fails to show the lethal effect. Since the lethal conditions include a cold shock (see Lethal conditions in results), this is unlikely to be caused by differential viability before eclosion. Therefore, this insensitivity to the lethal conditions is caused by phenotypic adaptation, as in the I4 line. Both males and females show waning of the lethal effect at later transfers to the RT, until at 20 hr after eclosion they no longer die in the LP (Figure 4D). This shows that I10 flies can overcome the deleterious effects of the cold shock by spending a day at the PT (29°).

Preadult lethal effects:

The beginning of the TSP can be determined by a set of experiments in which fly cultures are successively shifted from RT to PT (Suzuki 1970), but because flies can phenotypically adapt to the RT during development, this was not feasible (Figure 4, B and D). To investigate preadult mortality in response to the lethal conditions, we subjected larvae and pupae to the lethal conditions (Figure 2, A and B). Flies of the I4 inbred line that were transferred to the lethal conditions (29°) in the larval or pupal stage experienced no large increases in mortality prior to eclosion (Figure 5A). For the larval transfer, I4 had significantly lower viability as compared to O6 control flies (P = 0.033), but was only marginally affected (−2.7%, Figure 5A). For the pupal transfer, viability was not significantly different (P = 0.819). Timing of adult mortality over all treatments was not different (Kruskal-Wallis for 14-day mortality, P > 0.1) and occurred during the usual adult LP, indicating that induction of the lethal effect had in fact occurred in adults. Therefore, during preadult development I4 is not amenable to lethality. The egg-to-adult viability data suggest that insensitivity to the lethal conditions of I4 flies reared at the RT (see previous paragraph) is not caused by mortality of sensitive flies during development. Flies of larval and pupal transfers display progressively decreasing resistance as compared to the RT control (Figure 5B), but adult survival is significantly higher for larval and pupal transfer treatments than for the adult transfer (chi-square test of contingency, P < 0.001 and P = 0.001, respectively), indicating that phenotypic adaptation is dependent on the time spent at the RT.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5.—

Egg-to-adult viability and early adult survival of males for different timing of exposure to the lethal conditions (see text for an explanation of the treatments). (A) Viability of I4 and O6 and (B) adult survival of I4. (C) Viability of I10 and O6 and (D) adult survival of I10. Error bars represent the standard error. The SE of viability is based on replicates and that of adult survival on binomial variances.

I10 flies were exposed to lethal conditions involving cold shocks at 0° and transfer to the RT (20°; Figure 2B). A 2-hr cold shock was thought to be relatively harmless to control flies. Unexpectedly, however, the O6 control line showed a large viability reduction after the larval treatment while I10 larvae were less sensitive to this treatment (Figure 5C). To address this point, we repeated this part of the experiment, including an additional control line (O2 outbred line, described in Vermeulen and Bijlsma 2004). Egg-to-adult viability was significantly reduced in all lines (survival of O2, 11 ± 1.2%; O6, 21 ± 4.5%; I10, 18 ± 3.8%), indicating that larvae of all lines were sensitive to a cold shock during this stage of their development. However, no significant differences between lines were found in this experiment (ANOVA, P = 0.178). O6 flies further were resistant to all other treatments. I10 pupae show a large reduction in survival during the pupal stage (most died as pharate adults; data not shown) when compared to the O6 line within the same treatment (P < 0.001) as well as when compared to the survival of I10 during the pupal stage in the larval treatment (P < 0.001). Cold shock on pupae of Sarcophaga crassipalpis leads to a similar lethal effect in pharate adults (Denlinger et al. 1991). Three-day-old flies were completely resistant to the treatment, showing that adult lethality could be induced only in young flies after eclosion and before day 3 (Figure 5D). Although this ts lethal was discovered by its effect on life span, its adult lethality seems to be a side effect of a malfunction in a developmental process.

Dominance:

To obtain insight into the degree of dominance in these inbred lines, we reciprocally crossed them to the O6 control line and the life span of the resulting F₁ progeny was measured. In such a design, maternal effects also can be assessed. Furthermore, if the lethal resides on the X chromosome, the male mortality will show only in one of the reciprocal crosses. The survival curves at the RT are depicted in Figure 6.

• Download figure
• Open in new tab
• Download powerpoint

Figure 6.—

Cumulative survival curves of all crossing products at the RT. (A) I4 male and (B) I4 female survival curves were established at 29°. (C) I10 male and (D) I10 female survival curves were established at 21°.

At the RT (29°), >80% of the males of the I4 inbred line died within 2 weeks, whereas control males displayed almost 100% survival (Figure 6A). This proved to be a sex-specific effect, since I4 females did not show premature death during this period (Figure 6B). The adult lethal effect shows intermediate expression in heterozygotes, as shown by the high early adult mortality of the males of both reciprocal crosses. A Kruskal-Wallis test indicated significant differences between males of these lines at 29° (P < 0.001). Multiple comparison at the 0.05 significance level indicated that the reciprocal crosses differed from both parental lines but not from each other. The reciprocal crosses did not differ in life span, indicating that the line effect is not associated with the X chromosome.

In the I10 line, >70% of the males died within the first week at the RT (21°; Figure 6C). Lethality was also expressed in females, but to a lesser extent (60% mortality in week 1, Figure 6D). The survival curves of flies from the crosses to the control line did not show any early adult mortality at the RT, indicating that the mortality effect in this line has a recessive nature. Males from both reciprocal crosses do not express the lethal phenotype, indicating that the effect is not associated with the X chromosome.

Line I4:

I4 exhibits a temperature-sensitive lethal effect that responds to restrictive conditions only during the adult phase and is incessant throughout adult life. Survival at 29° of I4 flies through the lethal phase was never <10% so the allele should be designated as a semilethal at this temperature (Simmons and Crow 1977). Since the TSP comprises the entire adult phase and since 29° is well within the thermal range of Drosophila, it is possible that this gene plays a role in aging in outbred populations. The gene could be involved in a wide range of endogenous functions that affect adult survival. It will be interesting to determine if a single gene is responsible and if that gene has previously been identified as playing a role in life-span determination on the basis of model systems for genetic studies.

Line I10:

The TSP of this I10 cold-sensitive lethal effect is restricted to the pupal phase and early adult phase, most probably only during the first 24 hr. Therefore it might be involved in developmental processes and is excluded from playing a role in aging.

Flies of the I10 inbred line that were exposed to lethal conditions displayed a typical set of symptoms that eventually led to their death. Flies that received a cold shock at age 0–12 hr after eclosion and were kept at the RT became sluggish at 2 days after eclosion and started staggering. At 3 days most flies entered a state of total inertia, from which only a few flies recovered after several days. Although chill coma appears to be a prerequisite for expression of the lethal effect, it differs significantly from this second comatose state, which is entered several days after flies have recovered from the coma caused by low temperature. The occurrence of a comatose state during the onset of the LP makes it feasible that a malfunction in the nervous system causes the lethal effect. There has been, to our knowledge, no report of conditional lethals that require a cold shock for expression of the lethal effect. This allele may provide insight into natural variation for cold resistance and the physiological basis of cold shock, such as exemplified by chill coma variation in natural populations (David et al. 1998).

Conditional lethality and inbreeding depression:

Inbreeding depression changes the temperature sensitivity of life span, presumably through a decreased impact of the aging process (Vermeulen and Bijlsma 2004), and can also have dramatic line-specific effects on adult survival in Drosophila (Hughes 1995). The ts lethal effects reported here affected male adult survival after an inbreeding event.

Conditional lethality in Drosophila has been extensively studied, mostly using induced mutants (Suzuki 1970; Tarasoff and Suzuki 1970; Tasaka and Suzuki 1973; Arking 1975). These studies also included lethals affecting adult survival, although these are only rarely reported, since developmental processes in adults are less obvious (see Baird and Liszczynskyj 1985; Homyk et al. 1986). However, several aspects of behavior and reproduction still must develop in a freshly eclosed adult. In addition, fundamental cellular and tissue activities need to be maintained in the adult, so disturbances in these processes may also yield adult lethals (Homyk et al. 1986). Homyk et al. (1986) found that most adult lethals also exhibit semilethality in preadult stages, such as is the case with the I10 lethal. Naturally occurring ts lethals affecting adult survival, such as the lethals in this study, also are rarely reported but have been found before (e.g., Morrison and Milkman 1978) and their frequent expression upon inbreeding shows that they are not rare (Bijlsma et al. 1999). The phenotypic adaptation in our study may help explain their frequent occurrence. The buffering of genetic pathways by modifier genes potentially can diminish fitness effects of lethals and thus allow relatively high allele frequencies (Rutherford and Lindquist 1998). These ts lethals presumably are part of the genetic variation in natural populations, but it is unresolved whether these lethals map to loci having an effect on life-span variation in outbred populations.

There are preliminary indications that the I4 lethal shows allelism to l(2)hs (our unpublished results), a heat-sensitive lethal that previously has been shown to segregate in our base stock (Oudman 1991). Thus, there is a suggestion that the effect is largely due to a single gene. However, there is additional evidence that lethality in I4 might also be the result of epistatic interactions of l(2)hs with the I4 inbred background. If such epistatic interactions prove to be common during inbreeding depression, this warrants a reconsideration of the relative contribution of a few major genes vs. that of many genes with small effect in inbred populations. Epistatic interactions may become crucial in relationship of the effect of inbreeding depression on life span prospectively due to a disruption of homeostasis (Templeton and Read 1994). In any case, the identity of genes that cause inbreeding depression is a worthwhile goal in the context of agriculture and conservation biology.

Inbreeding depression and the genetics of aging:

The conditional lethals described here presumably represent cases of inbreeding effects. Thus, they are prospectively important in setting the stage for a deeper understanding of the identity of genes that cause inbreeding depression. Moreover, quantitative trait loci causing life-span variation among recombinant inbred lines are typically environment and sex dependent (Nuzhdin et al. 1997; Leips and Mackay 2000), which might be related to the survival effects documented in the present study. Deleterious alleles at relatively high frequency in populations could have a significant effect on aging. Thus, it is a worthwhile goal to identify the genes and alleles that confer accelerated mortality upon inbreeding to investigate corresponding allele frequencies in outbred populations. In such populations, one can address questions about the evolutionary forces that shape life span and identify how specific proximate mechanisms of aging operate when defined by natural genetic variation.