Results
Sub-lethal temperature effects on behaviour and male reproductive output
On the basis that males of temperate Drosophila melanogasterstrains become sterile at 30°C (Chakir et al., 2002; Petavy et al., 2001), we investigated the recovery dynamics for males of a temperate Portuguese D. melanogaster strain (Martins et al., 2014) developed at mid (29°C) or sub-lethal (31°C) temperatures. First, we documented viability of larvae under the different developmental temperatures and found those to be clearly challenging as only 45.5% respective 23.7% of flies eclosed when exposed to 29°C or to 31°C, in contrast to 95% of control flies eclosed successfully. Remaining at 29 or 31°C was stressful to adults as well, as males clearly did not recover fertility in contrast to males allowed to recover (Fig. 1A, C and table 1).
We next assessed the fecundity of heat-challenged males and found females mated once to a heat-challenged male laid significantly fewer eggs compared to females mated to control males (Fig. 1B, table 1). Males kept at the challenging temperature after eclosion induced lower numbers of eggs laid throughout, while a small improvement was seen for males allowed to recover. We independently repeated this experiment with the addition of sampling also day 1 after eclosion and found the pattern to be robust (Supplementary Fig. S1A, B and Supplementary table S2).
The offspring counts revealed that recovering males were initially sterile and recovered at different rates (see Fig. 1C, interaction term table 1), with offspring numbers approaching those of control males on day 6. This is matched by a steep increase in egg-to-adult survival between day 2 and 4 (see Fig. 1D), while males remaining at the challenging conditions produced very few offspring and had a low egg to adult survival (Fig. 1D, table 1). No adult flies were produced by males grown and kept at 31°C. Males grown and kept at 29°C produced on average 82% fewer offspring on day 6, with respect to control males, while those allowed to recover after eclosion had 36% fewer offspring. The slight increase for males recovering from development at 31°C still resulted in a mean reduction of 63% in offspring produced. Hence, at this temperature males suffer a severe fitness loss even if allowed to recover, while persistent heat stress results in near complete sterility even at the moderate challenge.
When considering the mating behavior for individual pairs, we overall found little effect of a moderate heat-challenge on mating success and only males experiencing a severe heat-challenge were negatively impaired. Males raised at 31°C gained few copulations compared to males from the other treatment groups (χ2 = 66.135,df = 4, P < 0.0001; Fig. 2A). Males raised at 31°C had longer mating latencies (GLM with gamma distribution:Deviance = 7.582, F = 3.002, df = 4, P = 0.021; Fig. 2B) while there was no effect on copulation duration (GLM with poisson distribution: Deviance = 8.867, df = 4,P = 0.065; Supplementary Fig. S2A).
In addition to single mating productivity, we also tested male sperm competitive ability after developmental heat-exposure. We document a severe negative impact of heat on male sperm defense ability (GLM with a quasibinomial distribution: Deviance = 5173.8, F = 22.177,df = 4, P < 0.0001, Fig 2D), even after we allowed males to recover for 5 days.
Additionally, heat-challenged males were unable to prevent female remating, regardless of the possibility for recovery, as all females remated when the second male was present, while only 35.1% of females first mated to a control male remated (χ2 = 82.624; df = 4, P < 0.0001). This pattern was also reflected in second mating latencies (GLM with quasibinomial error distribution; Deviance = 73.482, F = 181.55, df = 1, P < 0.0001; see Fig. 2C).