Wolbachia status of the five laboratory strains
All flies of the three Wolbachia infected medfly lines (60 out of 60 individuals) produced the expected 16S rRNA gene amplicon, while all screened flies of the two uninfected medfly lines (40 out of 40) were found negative (data not shown). The wsp based PCR screening that discriminates among the different wCer Wolbachia strains was performed on the same flies of the three Wolbachia-infected medfly lines and 56S2 plus 88.6 produced only the expected wCer2-specific wsp amplicon (40 out of the 40 flies), while all flies from the S10.3 line produced only the wCer4-specific wsp amplicon (data not shown). Three individuals (out of the 20) were selected per Wolbachia-infected medfly line and the MLST profile was analyzed through sequencing of the MLST genes. Again, all flies presented the expected MLST profile, as described in the Wolbachia MLST database.
Effects of Wolbachia infection on hatch rate
Logistic regression analysis revealed that both the genetic background of medfly and the Wolbachia infection, as well as their interaction were significant predictors of egg hatch (Wald’s t-test =55.68, df = 1, P < 0.0001; Wald’s t-test = 782.96, df = 1, P < 0.0001, and Wald’s t-test = 7.39, df = 1, P = 0.007, respectively). As shown in Fig. 1a, in both medfly genetic backgrounds, Wolbachia infection detrimentally reduced female fertility. Different Wolbachia strains exerted differential reduction in egg hatch rates on the same medfly genetic background (x2 = 1757.49, df = 2, P < 0.001). Chi-square test revealed significant differences between the infected 88.6 and S10.3 lines, and the uninfected Benakeio line (x2 = 1757.49, df = 2, P < 0.001). Both Wolbachia strains (wCer2 or wCer4) reduced hatch rates compared to the uninfected flies (x2 = 833.37 and 1666.67, df = 1, P < 0.001). Hatch rates were lower in wCer4 than in wCer2 infected lines (x2 = 174.72, df = 1, P < 0.0001) (Fig. 1b).
Effects of Wolbachia infection on larval and pupal survival
Wolbachia infection was not a significant predictor of larval survival (Wald’s t-test t = 0.521, df = 1, P = 0.470) (Fig. 2a). However, it increased survival rates of pupae (Wald’s t-test t = 7.805, df = 1, P = 0.005) in both medfly genetic backgrounds (Fig. 2b). The effect of the different medfly genetic backgrounds was also significant predictor of both larval and pupal stage survival (Wald’s t-test t = 11.842, df = 1, P = 0.001 and Wald’s t-test t = 48.016, df = 1, P < 0.001, respectively). The interaction between Wolbachia infection and medfly genetic background was a significant predictor of pupal survival indicating a differential response of the two medfly genotypes (Wald’s t-test t = 17.386, df = 1, P < 0.001) (Fig. 2a). As far as effects of different Wolbachia strains on survival during larval and pupal stages are regarded, chi-square test revealed significant differences between the 88.6 and S10.3 lines, and the uninfected Benakeio line (x2 = 94.159 and 25.642, df = 2, P < 0.0001) (Figs. 2c, d). The Wolbachia strain wCer2 increased both larval and pupal survival compared to the uninfected flies (x2 = 5.525, df = 1, P = 0.019 and x2 = 7.948, df = 1, P = 0.005). Conversely, the Wolbachia strain wCer4 reduced the survival in the larval stage whereas it increased the survival in the pupal stage when compared to the uninfected lines (x2 = 66.693, df = 1, P < 0.001 and x2 = 25.304, df = 1, P = 0.001). Both survival during the larval and pupal stage were lower in the wCer4 infected flies compared to wCer2 infected ones (x2 = 81.615, df = 1, P < 0.001 and x2 = 5.274, df = 1, P = 0.021, respectively) (Figs. 2c, d).
Effects of Wolbachia infection on immature development
The effect of Wolbachia infection on the embryonic and larval developmental duration of the five medfly laboratory lines is depicted in Fig. 3. Cox regression analysis revealed that both the genetic background of medfly and Wolbachia infection were significant predictors of egg to pupae developmental duration (Wald’s t-test = 290.51 and 30.12, df = 1, P < 0.0001, respectively), as well as their interaction (Wald’s t-test = 9.36, df = 1, P < 0.0001). The infection reduced egg to pupae duration on VIENNA 8 GSS genetic background whereas it prolonged the respective duration on the BENAKEIO flies (Fig. 3a).
Survival analysis revealed significant differences in egg to pupae duration as well among uninfected Benakeio, and the infected 88.6 and S10.3 lines (log rank test: x2 = 82.19, P < 0.0001). Specifically, Wolbachia infection, either wCer2 or wCer4, prolonged the egg and larval developmental duration when compared to the uninfected flies (x2 = 74.115, 38.014: P < 0.0001for 88.6 and S10.3, respectively). No differences were found between the Wolbachia infected lines (x2 = 0.102, P = 0.750), (Fig. 3b).
Cox regression analysis revealed that both Wolbachia infection and sex were significant predictors of the pre-pupa duration when the GSS lines (Wolbachia infected 56S2 GSS and the uninfected Vienna 8 GSS) were compared (Wald’s t-test = 55.58 and 99.11, df = 1, P < 0.0001, respectively). The interaction between Wolbachia infection and sex was not significant (Wald’s t-test = 0.88, df = 1, P = 0.348) indicating that the bacterium affected the developmental duration of both sexes similarly in the VIENNA 8 GSS genetic background, (Fig. 4).
Effects of Wolbachia infection on adult sex ratio
Two-way ANOVA revealed that neither the genetic background of medfly (F = 0.046, df = 1,9, P = 0.835) nor Wolbachia infection (F = 0.793, df = 1,9, P = 0.396) and their interaction (F = 0.010, df = 1,9, P = 0.924) were significant predictors of the sex ratio (male/female) of the emerged adults. Likewise, the Wolbachia strain (wCer2 and wCer4) was not a significant predictor of the sex ratio of emerged adults (F = 0.073, df = 2,6, P = 0.931; see Additional file 2).
Effects of Wolbachia infection on adults’ longevity
Neither Wolbachia infection nor medfly genetic background were significant predictors of adult lifespan (Wald’s t-test = 1.07 and 3.75, df = 1, P = 0.300 and 0.053, respectively) in contrast, sex was significant as males outlived females (Wald’s t-test = 6.491, df = 1, P = 0.011), (Figs. 5a, b). Neither Wolbachia strain nor sex were significant predictors of adult longevity on the BENAKEIO flies (Wald’s t-test = 1.65 and 3.47, df = 1, P = 0.199 and 0.062, respectively), (Fig. 5c, d). The wCer2 infected females suffered reduced survival rates compared to uninfected and the wCer4 infected ones, but this observation was not significant (Fig. 5c).
Effects of Wolbachia infection on fecundity
Two-way ANOVA [on ln(x) transformed lifetime fecundity rates to achieve normality and equal variance] revealed that neither the medfly genetic background (F = 2.388, df = 1,38, P = 0.131) nor Wolbachia infection (F = 0.310, df = 1,38, P = 0.581) affected lifetime fecundity rates. Similarly, the interaction between medfly genetic background and Wolbachia infection was not significant as well (F = 0.367, df = 1,38, P = 0.548). In contrast, fecundity rates were significantly different among Wolbachia infected medfly lines S10.3, and both the 88.6 and uninfected BENAKEIO (F = 9.451, df = 2,28, P = 0.001) (Fig. 6).
Effects of Wolbachia infection on male mating competitiveness
We used the RI index (Relative Index, analogous to the Relative Sterility Index (RSI), see FAO/IAEA/USDA 2014) to compare mating competitiveness of males of the five medfly lines tested against wild males in competition for wild females (Fig. 7). Overall, Wolbachia infection did not affect mating competitiveness (F = 0.553, df = 1,21, P = 0.465) in contrast to medfly genetic background (F = 45.849, df = 1,21, P < 0.0001). The interaction between medfly genetic background and Wolbachia infection was marginally significant (F = 4.636, df = 1,21, P = 0.043) indicating a rather differential impact of the Wolbachia infection on the two different medfly genetic backgrounds (Fig. 7a). One way ANOVA revealed significant differences in male mating competitiveness among BENAKEIO uninfected, S10.3 and 88.6 lines (F = 9.450, df = 1,12, P = 0.003). wCer2 and wCer4 infections reduced and increased male mating competitiveness, respectively (Fig. 7b).
Effects of Wolbachia infection on flight ability
Wolbachia infection was a significant predictor of adult flight ability (F = 70.42, df = 1,16, P < 0.0001), in contrast to the medfly genetic background (F = 0.10, df = 1,16, P = 0.754). The significant interaction between Wolbachia infection and medfly genetic background highlights the differential effect of Wolbachia infection on the two medfly genetic backgrounds resulting in positive and negative effects on flight performance on VIENNA 8 GSS and BENAKEIO flies respectively (F = 173.49, df = 1,16, P < 0.0001) (Fig. 8a).
Wolbachia infection, regardless of the bacteria strain, significantly reduced the flight ability of the BENAKEIO flies (F = 216.34, df = 2,12, P < 0.0001). Significant differences between the two infected lines were also observed (P < 0.05) (Fig. 8b).
Considering only the VIENNA 8 GSS medfly genetic background, two way ANOVA revealed that neither Wolbachia infection nor sex were significant predictors of the flight ability (F = 25.00, df = 1,16, P = 0.126 and F = 85.05, df = 1,16, P = 0.069, respectively). Similarly, the interaction between Wolbachia infection and adult sex was not significant (F = 0.172, df = 1,16, P = 0.684), (Fig. 9).
Effects of Wolbachia infection on longevity under food and water deprivation
Medfly genetic background, Wolbachia infection and sex were significant predictors of adult longevity under food and water deprivation (Wald’s t-test = 224.17, 37.28 and 30.25, df = 1, P < 0.0001). The significant interaction between medfly genetic background and Wolbachia infection (Wald’s t-test = 39.72, df = 1, P < 0.0001) highlights the differential effect of the Wolbachia infection on the two medfly lines. Specifically, Wolbachia infection reduced VIENNA 8 GSS longevity under water and food deprivation, whereas it increased the BENAKEIO longevity under the same stress conditions (Fig. 10a, b).
For the BENAKEIO flies, Cox regression analysis revealed that the Wolbachia strain significantly affected adult longevity under water and food deprivation (Wald’s t-test = 62.01, df = 2, P < 0.0001). Nevertheless, neither sex (Wald’s t-test = 2.26, df = 1, P = 0.133) nor the interaction between the different medfly lines (BENAKEIO genetic background) and sex were significant predictors of adult longevity under food and water deprivation (P > 0.05). Adults of the uninfected Benakeio showed similar stress tolerance compared to S10.3 (Wald’s t-test = 1.58, df = 1, P = 0.208) but much lower compared to the wCer2 infected ones (Wald’s t-test = 35.15, df = 1, P < 0.0001). Adults of the 88.6 line expressed longer survival rates under food and water deprivation compared to S10.3 ones (Wald’s t-test = 54.13, df = 1, P < 0.0001), (Fig. 10c, d).
Focusing on the VIENNA 8 GSS medfly genetic background, statistical analysis showed that Wolbachia infection was not a significant predictor of longevity under stress (Wald’s t-test = 2.734, df = 1, P = 0.098), in contrast to sex (Wald’s t-test = 22.52, df = 1, P < 0.0001) (Fig. 10a, b).