The effect of parental rearing conditions on offspring life history in Anopheles stephensi

Mosquitoes originated from a longstanding laboratory stock of An. stephensi and were reared under standard laboratory conditions at 27 ± 2°C, 70% humidity and in a 12:12 light: dark cycle. Eggs were hatched in three plastic trays (25 cm × 25 cm) filled with 1.5 L of distilled water. Two days after hatching, larvae from the three trays were mixed and 200 transferred to 30 ml vials containing 5 mls of distilled water, where they were reared individually (Figure 1a). Half of these larvae were then randomly allocated to the low food treatment group (1 mg of Tetrafin^® food per day) and half to the high food treatment group (10 mg of Tetrafin^® food per day). When individuals pupated, their vial was covered with fine nylon gauze until emergence. Food levels were determined from an earlier pilot study and include the range where 100% emergence occurred. The position of the vials within the insectary was fully randomized at this and all subsequent stages.

On emergence, the mosquitoes were pooled into three mesh cages (30 × 30 × 30 cm) per treatments groups and provided with 10% glucose solution ad libitum (Figure 1b). Approximately one week after emergence, one anaesthetized mouse was placed on each cage from which the mosquitoes were allowed to feed for 20 minutes. One day later bowls of water were introduced into the cage for oviposition.

These eggs were then pooled and hatched in six plastic trays for each treatment and 400 individual larvae (200 from each parental treatment group) transferred to a standard vials as above (Figure 1c). The larvae were allocated to either the high or low food treatment groups (as above) and monitored daily for pupation. When individuals pupated, they were transferred to a vial that was covered with fine nylon gauze and maintained without food for survival analysis (Figure 1d). On death, mosquitoes were transferred to a 1 ml tube and refrigerated until the end of the experiment. At the end of the experiment the wings of each mosquito were dissected and measured from the distal to dorsal points using microscopy.

This experiment was repeated eight weeks later. In the repeat experiment, offspring were separated into two groups for either survival analysis as in the previous experiment, or for fecundity analysis after a blood meal (Figure 1d and 1e). Fecundity analysis involved transferring both male and female offspring into mating cages (30 × 30 × 30 cm), with four mating cages per offspring group. Five days after emergence, anaesthetized mice were placed on each of the 16 mating cages and the mosquitoes allowed feed for 20 minutes. These mice had all been infected 14 days before the feed with Plasmodium chabaudi, as the original intention was to determine whether vectorial capacity was affected by parental environmental conditions. However, most unexpectedly on the day of the blood-meal, neither asexual nor sexual parasites were present in any of the mice and on dissection of the female mosquitoes after egg-laying, no transmission had occurred. As all infected mice were parasite negative on the day of the feed and were randomly assigned treatment groups which were replicated, any treatment effect should be unrelated to any effect of infection. Immediately after the feed, each female mosquito was transferred to a clean vial covered with fine nylon gauze for three days to allow all haematin (a by-product of decomposition of haemoglobin) to be excreted (Figure 1e), from which blood meal size was estimated [30]. Excreted haematin collected in the bottom of the vial was dissolved in 1 ml of 1% LiC0₃ solution. The absorbance of the resulting mixture was read at 387 nm in a spectrophotometer using LiC0₃ solution as a blank and compared with a standard curve made with porcine serum haematin (Sigma Aldrich). Solutions that were within the error range of the LiC0₃ blanks (absorbance < 0.01) were eliminated from the analysis and classified as non-feeders. After the 3-day haematin collection period, mosquitoes were moved to new 30 ml tubes containing 3 mls of water to allow oviposition (Figure 1f).

The fitness components of emergence time, survival, adult size, blood meal size and fecundity were measured. Both emergence time and survival were measured as the number of days required for either a) the larvae to emerge as an adult, post hatching, or b) taken to die post emergence. As an indicator of size, the length of one wing per mosquito from the distal to dorsal points using microscopy was measured. Haematin mass was used as an indicator of blood meal size, while fecundity was determined by counting the number of eggs laid over the three days following the blood meal.

The life history traits of emergence time and survival were analysed using Proportional Hazards (JMP in 5.1). Adult size, blood meal size and fecundity were analysed using General Linear Models. The explanatory variables were parental food (two levels), offspring food (two levels), gender (two levels) and where relevant, experimental block (two levels). For all models, a maximal model including all two and three-way interactions was fitted first. Models were then minimized by removing non-significant terms beginning with the highest-level interaction. In no cases were any of the three-way block*parent*offspring condition interactions significant and these are therefore not reported. Significant block*parent or block*offspring interactions did occur, but as they only reflected differences in magnitude between blocks, they are not reported. For the blood meal analysis as well as the fecundity analysis, the average blood meal size per replicate cage and the average number of eggs per replicate cage were used as response variables (n = 16), with average adult size as a covariate. This is a conservative approach to deal with the issue of pseudoreplication of treatments arising from mosquitoes fed on the same mouse. The statistical significance of the main effect of mouse and of the interactions between mouse and the covariates (adult size and blood meal size) was also tested. For both of the response variables of blood meal size and fecundity, the main effect of mouse and the interactions were non-significant.

In experiment 1, two hundred larvae were split into either a high food group or low food group, to form the parental generation. Once adults, these parents were mated within groups (three per treatment) and 400 eggs from the parental high food generation and 400 eggs from the parental low food generation were used for the offspring experiments. Of these, complete records of the emergence time, survival, gender and size were noted for 460 mosquitoes which were included in the analysis. In the repeat experiment, 120 larvae formed the parental generation, from which, 400 eggs from the parental high food generation and 400 eggs from the parental low food generation were used for the offspring analysis. Of these, emergence time was noted for 478 mosquitoes, with 140 of used in the survival trial and 287 used for blood meal and fecundity analysis. Altogether, 149 females were included the fecundity trials in which 19,844 eggs were counted.

The time taken for larvae to emerge as adults post-hatching was determined by offspring food conditions and to a lesser degree by offspring gender (Figure 2a, Table 1-1). Larvae reared in high food conditions emerged from pupation up to four days earlier than larvae reared in low food conditions and males emerged on average, one day before females. The food levels experienced by the parental generation did not influence emergence time and this lack of parental influence was constant across offspring food levels (Figure 2a, Table 1-1).

Offspring larval food level was also the main factor influencing the size of offspring once they reached adulthood (Fig 2b, Table 1-2). Larvae emerging from high food conditions were on average 16% larger than those emerging from low food conditions. Gender was also an important determinant of offspring size, with females being 7% larger than males (Figure 2b, table 1-2). The larval food levels of the parental generation did not influence offspring adult size (Figure 2b, Table 1-2).

The food level experienced by the offspring was again the only factor found to be influencing adult survival (Figure 2c, Table 1-3a). As expected, offspring reared in high food conditions survived for longer than offspring reared in low food conditions. Survival did not differ between the males and the females and was unaffected by parental rearing conditions (Figure 2c, Table 1-3a). Larger offspring survived for longer than smaller offspring (χ² = 148, df = 1, p = <0.0001), but offspring food level was still a major determinant of survival even when controlling for offspring size (Figure 2c, Table 1-3c).

Nineteen out of the 149 females that were given access to a blood meal were classified as non feeders. No effect of parental food level, offspring food level or an interaction between the two influenced female propensity to fed (parent: χ² = 2.8, df = 1, p = 0.09; offspring: χ² = 2.7, df = 1, p = 0.1; parent*offspring: χ² = 2.5, df = 1, p = 0.1, respectively).

Of the 130 females that did feed, only the level of food they themselves experienced as larvae influenced the size of the blood meal they took as adults (Figure 3a, Table 1-4a). Daughters reared in high food conditions took 25% larger blood meals than the daughters reared at low food conditions. Controlling for adult size, it was again found that blood-meal size was influenced only by the food level experience of the offspring (Figure 3a, table 1-4b). However, genetic and phenotypic variance is often greater in stressful environments [5, 28, 29]. Indeed, parental effects were apparent when the offspring reared on low food only were considered: blood meal size was influenced by the larval food level of parents, even when controlled for adult size, with the offspring of parents reared in low food conditions taking larger blood meals (parent: F_1,6 = 27.4, p = 0.002, parent controlled for size: F_1,5 = 138.3, p < 0.0001, respectively).

Parental effects influenced offspring fecundity. Offspring of parents reared in low food conditions produced more eggs than the offspring from parents reared in high food conditions (Figure 3b, Table 1-5a). Parental food level influenced offspring egg number, even when controlling for offspring size (Figure 3b, Table 1-5b). Considering the offspring reared on low food only, the fecundity of low food daughters was influenced by the larval food level of their parents, even when controlled for adult size (F_1,6 = 13.6, p = 0.01, F_1,5 = 12.3, p = 0.017, respectively).

Blood meal size and egg number were positively correlated, with larger blood meal sizes resulting in increased egg production (F_1,14 = 34.1, p < 0.0001). The number of eggs produced for a given blood meal size was influenced by offspring as well as parental larval food level (Figure 3b, Table 1-5c). Parental effects influenced the fecundity of daughters reared at low food conditions but not of daughters reared at high food (F_1,5 = 16.1, p = 0.01, F_1,5 = 4.5, p = 0.09, respectively). For the same size blood meal, the daughters reared at low food conditions were 20% more fecund if their parents had also been reared at low food levels.