Desiccation resistance: effect of cuticular hydrocarbons and water content in Drosophila melanogaster adults

Abstract

Background: The insect cuticle covers the whole body and all appendages and has bi-directionnal selective permeability: it protects against environmental stress and pathogen infection and also helps to reduce water loss. The adult cuticle is often associated with a superficial layer of fatty acid-derived molecules such as waxes and long chain hydrocarbons that prevent rapid dehydration. The waterproofing properties of cuticular hydrocarbons (CHs) depend on their chain length and desaturation number. Drosophila CH biosynthesis involves an enzymatic pathway including several elongase and desaturase enzymes.

Methods: The link between desiccation resistance and CH profile remains unclear, so we tested (1) experimentally selected desiccation-resistant lines, (2) transgenic flies with altered desaturase expression and (3) natural and laboratory-induced CH variants. We also explored the possible relationship between desiccation resistance, relative water content and fecundity in females.

Results: We found that increased desiccation resistance is linked with the increased proportion of desaturated CHs, but not with their total amount. Experimentally-induced desiccation resistance and CH variation both remained stable after many generations without selection. Conversely, flies with a higher water content and a lower proportion of desaturated CHs showed reduced desiccation resistance. This was also the case in flies with defective desaturase expression in the fat body.

Discussion: We conclude that rapidly acquired desiccation resistance, depending on both CH profile and water content, can remain stable without selection in a humid environment. These three phenotypes, which might be expected to show a simple relationship, turn out to have complex physiological and genetic links.

Keywords: Cuticle; Dehydration; Desaturase; Drosophila; Fat body; Insect; Selection.

Figure 1. Experimental selection of desiccation resistance line.

(A) Flies were kept in single sex groups of approximately 50 individuals in empty glass vials. Six or seven of these glass vials were packed inside an airtight transparent plastic box that was seeded with a layer of silicagel crystals to maintain a low relative humidity (20 ± 1%). The box was placed on a hot plate at 25 ± 0.2°. Four boxes were simultaneously tested. Every two hours, the number of dead flies was counted, providing a measure of survival over time. This allowed us to estimate when 50% flies died (lethality time 50% = LT50), using logistic regression. The slope of the lethality curve was also determined to evaluate the relative lethality per hour. At the end of each experiment, the few surviving flies (maximum 1–2% of all flies) were transferred into fresh food vials and mated with siblings to produce the next generation. (B) Arrows indicate the generations of the experiment. Plain arrows indicate the generations during which experimental selection for desiccation (light blue) and/or phenotypic measurements (CH, cuticular hydrocarbons; W, weight; Fec, fecundity) were carried out. Dashed arrows indicate the generations during which no measurement or selection was carried out.

Figure 2. Survival in females selected for desiccation resistance over the first six generations.

Female flies were selected using the experimental procedure described in Fig. 1. (A) For each generation (F1 to F6), the curves represent the survival measured in various genotypes (dashed line, Di2 control line; cyan line, 77S selected lines pooled; magenta line, backcross between 77S females and unselected sibling males; at F1, three selected lines are shown). (B) At each generation, the two box-plots represent the LT50 and the lethality slope using colors similar to those of the corresponding genotypes. Data are shown as box plots representing the 50% median data (the small horizontal bar indicates the median value while the plain dot represents the mean). The whiskers shown below and above each box represent the first and third quartiles, respectively. Stars or different letters indicate significant differences. After excluding extreme outliers using Tukey’s method, LT50 and slopes were tested using a Kruskall–Wallis test completed by a Conover-Iman multiple pairwise comparisons at level p = 0.05 (with a Bonferroni correction) or with a Mann–Whitney test. ***, p < 0.001; **, p < 0.01; *, p < 0.05. The absence of a letter or stars indicates that no significant difference was detected. N = 5–17 (except 5S line at F1 and 77s × Di2 line at F3 where N = 3). A similar selection procedure was carried out on males between the F1 and F6 generations (Fig. S2).

Figure 3. Survival in females of selected lines between F18 and F57.

Female flies were selected at the indicated generations using the procedure described in Figs. 1 and 2. More precisely, we show the survival curve (A, C) and the LT50 with the slope (B, D) for the generations F18–F28 (A, B) and F29–F57 (C, D). N = 5–22 (except Di2 line at F18: N = 4, and at F25, 28 & 29: N = 3). For parameters and statistics, see Fig. 2 legend.

Figure 4. Principal cuticular hydrocarbons in flies of selected lines following relaxation of selection.

Cuticular hydrocarbon levels (CHs) were measured in F7, F8 and F9 females and in F8 males separately in the six 77S lines (77S₀–77S₅) experimentally selected for desiccation resistance (F1–F6; see Fig. 2). Here, we show the total absolute amount of CHs (∑CH in µg, A) and the ratio of Desaturated: linear saturated CHs (D:L ratio; B) This ratio was calculated using the formula ([D—L]/[D + L]). N = 5–20 for females and N = 9–14 for males. For statistics, see Fig. 2 legend. We also determined the absolute (Q) and relative (%) amounts of desaturated CHs (alkenes) and of linear saturated CHs (alkanes) (Fig. S3).

Figure 5. Principal cuticular hydrocarbon levels in females of selected lines between F18 and F57.

∑CH (A) and D:L ratio (B) were measured in F18, F19, F55 and F57 females. As well as control unselected Di2 females, six 77S lines (77S₀–77S₅) were tested in F18 and F19; only four of these lines (77S₁–77S₄) survived to F55 and F57. The 77S-Sel lines (77S-Sel1–77S-Sel4) that were tested at F57 were the offspring of F55 reselected females from their respective 77S lines (e.g., 77S₁ females yielded the 77S-Sel1 line). N = 7–38. For more information on parameters, lines and statistics, see legends of Figs. 2 and 4. The absolute (Q) and relative (%) amounts of desaturated CHs (alkenes) and of linear saturated CHs (alkanes) determined in these flies are shown in Fig. S4.

Figure 6. Fresh:dry weight ratio in females of selected lines.

Groups of 10 freshly killed females were weighed (fresh weight) and after 24 h desiccation were weighed again (dry weight). The fresh:dry weight ratio of each group was calculated. Females were weighed at F19 (A) and F57 (B). N = 6–20. For more information on genotypes and statistics, see legends to Figs. 2, 4 and 5.

Figure 7. Desiccation resistance in various desat1 transgenic females.

To test the effect of desat1 knock-down expression in various desat1-expressing tissues, we used the female progeny of matings between transgenic females carrying the UAS-desat1-IR transgene (IR) and transgenic males either carrying each desat1 putative regulatory region fused with Gal4 (PRR-Gal4) corresponding to each desat1 transcript (RA, RC, RE, RB, RD, RDiO), or the complete desat1 regulatory region (6908 bp = 6908). Di2 control females and female progeny resulting of matings between IR females and Di2 males were also tested (Di2; Di2-IR; left box plots). The LT50 (A) and the lethality slope (B) of all these genotypes were determined. To control for the effect of each desat1 PRR-Gal4 transgene on the LT50 (C) and lethality slope (D), we used flies from matings between Di2 females and PRR-Gal4 males, alongside Di2-IR and Di2 control females. N = 5–13. For more information on parameters and statistics, see Fig. 2 legend.

Figure 8. Principal cuticular hydrocarbons in various desat1 transgenic females.

CH levels were measured in all transgenic and control female flies tested for desiccation resistance (see Fig. 7). Transgenic females combining a maternal IR transgene with a paternal PRR-Gal4 or 6908 transgene were tested at both F55 (A) and F57 (B). Di2, Di2/w and Di2-IR control females were tested (left box-plots) either at F55 (A) or at F57 (B). Control genotypes carrying a paternal copy of each PRR-Gal4 transgene or of the 6908 transgene combined with a maternal Di2 genome were also tested (C). Alongside these control genotypes, we also tested the effect of the IR transgene in the Di2 background (second box-plot from the left). N = 7–16. For more information on CHs, genotypes and statistics, see legends to Figs. 4 and 7.

Figure 9. Fresh:dry weight ratio in various desat1 transgenic females.

The fresh: dry weight ratio were analyzed in transgenic (A) and in control genotypes (B) N = 5–15. For more information on genotypes and statistics, see legends to Figs. 2 and 7.

Figure 10. Survival to desiccation in various cuticular hydrocarbon variant females.

The LT50 (A) and lethality slope (B) of various CH variants was measured. We also assayed the resistance of flies from these lines: control Di2, Di2/w (combining the Di2 genome with the w¹¹¹⁸ white-eye mutation), Zimbabwe 30 line (Z30), ricket^1∕4 double trans-heterozygote mutant (rk¹∕rk⁴), rk heterozygotes carrying the SM2 balancer (rk∕CyO = rk¹∕CyO and rk⁴/CyO), and Di2/w flies carrying the SM2 balancer (CyO/w). N = 10–30 (except for CyO/w line: N = 3). For more information on statistics, see legend to Fig. 2.