1 Introduction

Microbiota is increasingly regarded as an important factor that contributes to the hosts’ physiology (McFall-Ngai et al. 2013). In arthropods, infection with endosymbiotic bacteria, such as Wolbachia, is among the most important. Wolbachia are maternally inherited endosymbiotic bacteria that infect a variety of terrestrial arthropods (Hilgenboecker et al. 2008; Duron et al. 2008; Werren et al. 2008; Serga et al. 2019). Evolutionary success of bacteria in populations of the host species relies on the mode of influencing the reproduction of the host, such as cytoplasmic incompatibility (CI), male killing (MK), feminization of genetic males and induction of parthenogenesis (O’Neill 1998). Both the nature and intensity of reproductive manipulations depend on host genetics and Wolbachia strain (Braig et al. 1994; Veneti et al. 2012). In Drosophila, Wolbachia are known to cause the CI and MK reproductive phenotypes. It has been shown that the wRi strain, which causes CI, spread rapidly in the populations of Drosophila simulans in California (Turelli and Hoffmann 1991, 1995) and in eastern Australia (Kriesner et al. 2013) with the current infection rate close to 100%. In D. innubila, the MK strain confers a selective advantage of about 5% and is maintained at 35% infection frequency in populations (Dyer and Jaenike 2004). However, the Wolbachia strains wMel and wSuz have low to no effects on reproductive phenotypes with infection levels often lower than 100% (Hoffmann 1988; Solignac et al. 1994; Hoffmann et al. 1996; Hamm et al. 2014). CI variability, at least in the case of D. melanogaster, can be explained by “young brothers” and “grandmother” effects (Yamada et al. 2007; Layton et al. 2019). In cases with low levels of CI, maintenance of the infection in a population is explained by particular fitness benefits (Serga and Kozeretskaia 2013), although the reasons for the success of bacteria in natural populations infected with non-CI or non-MK strains are not always clearly identifiable (Charlat et al. 2004; Harcombe and Hoffmann 2004; Fry et al. 2004).

wMel is the only strain presented in natural populations of D. melanogaster (Solignac et al. 1994; Verspoor and Haddrill 2011). This strain is classified into two groups of genotypes – wMel and wMelCS (Riegler et al. 2005). The frequency of the genotypes varies across natural populations of D. melanogaster (Riegler et al. 2005; Serga et al. 2014; Bykov et al. 2019). Presently, flies infected with the wMel genotype are predominant in most populations. wMelCS is a rather rare genotype, although there are populations in which it prevails, for example, in some populations of Portugal and Asia (Ilinsky and Zakharov 2007; Nunes et al. 2008). In addition, it has been shown that the wMelCS genotype is prevalent among laboratory lines collected before the 1950s, while the wMel genotype was predominantly identified in the lines collected in the second half of the twentieth century (Riegler et al. 2005). Based on this fact, it was hypothesized that the wMelCS genotype was replaced worldwide with wMel. The causes of the replacement of wMelCS with wMel are not completely clear. According to one hypothesis, wMelCS is more costly for the host because of its higher titers in the host organism compared to wMel (Chrostek et al. 2013). According to an alternative hypothesis, flies infected with wMelCS prefer colder environments than those infected with wMel. This leads to a higher rate of development of flies with wMel, which increases their fitness and increases the number of generations per year, eventually leading to a replacement (Truitt et al. 2019). However, there are populations where both rare (wMelCS-like) and common genotypes (wMel-like) are present at relatively high frequencies, for example, in a population from Uman in Ukraine (Serga et al. 2014).

In the absence of high levels of CI and MK, the evolutionary success of Wolbachia is determined by the ratio of negative and positive effects on D. melanogaster fitness, as well as by imperfect maternal transmission (Kriesner et al. 2016). Fry et al., 2004 have shown that infection with Wolbachia leads to different fitness effects depending on the D. melanogaster line (Fry et al. 2004). In some lines, infection with Wolbachia leads to higher survival or fecundity, while in others to lower (Alexandrov et al. 2007; Maistrenko et al. 2015, 2016; Roshina et al. 2018; Capobianco et al. 2018). In particular, decreased lifespan has been reported from wild Drosophila strains collected in Russia (Roshina et al. 2018) and North America (Capobianco et al., 2018) and in inbred fly strains from Drosophila Genetic Reference Panel (Albertson et al., 2013), while an extend in lifespan has been observed in Wolbachia-positive laboratory lines (Alexandrov et al. 2007). The strain of bacteria can also be an important factor, but most of these studies did not perform genotyping of Wolbachia. wMelCS Wolbachia transferred from D. melanogaster via microinjection into D. simulans caused a reduction of lifespan and fecundity (Martinez et al. 2015). Wolbachia is also able to affect host sensitivity to physiological stress conditions (Brownlie et al. 2009; Wang et al. 2012; Albertson et al. 2013; Gruntenko et al. 2017), particularly oxidative stress (Wong et al. 2015; Capobianco et al. 2018) and viral infection (Hedges et al. 2008; Teixeira et al. 2008). The effect of the bacteria on the stress response has, however, not been detected in all lines and appears to depended on the flies’ genetic background (Capobianco et al. 2018). For example, in lines from Australian natural populations of D. melanogaster Wolbachia did not influence adult starvation resistance and also had no effect on adult heat resistance (Harcombe and Hoffmann 2004). However, in inbred lines from North Carolina, removing Wolbachia with tetracycline induced differential starvation survival (Albertson et al. 2013).

To investigate the effects of Wolbachia on fitness traits, it is important to obtain genetically identical lines that are infected and uninfected with the bacterium. A number of approaches are used for this, such as antibiotic or temperature treatments (Li et al. 2014), as well as injection of bacterial strains to non-natural hosts (Martinez et al. 2015). The most effective antibiotics are tetracycline and rifampicin (Li et al. 2014). Antibiotic treatment allows to quickly and efficiently obtain genetically identical lines infected and uninfected with Wolbachia. One problem is that the antibiotic itself can affect fitness traits (O’Shea and Singh 2015). In addition, broad spectrum antibiotics can affect the composition of the microbiota in general. The alternative approach is based on obtaining lines with the same genotypes infected with different Wolbachia strains by injection of bacteria (Martinez et al. 2015). In this case, the microbiome of the line is not disturbed, however, Wolbachia is introduced into a new genotype, which can also significantly affect fitness traits.

In this study, we analyzed fitness and stress related traits in isofemale lines from a unique natural population from Uman, in which persistence of the wMelCS genotype is observed from year to year with varying infection rates. We analyzed the effects of Wolbachia on important fitness parameters such as sensitivity to cold and oxidative stress, female fecundity and lifespan, depending on the bacterial genotype and on the host Drosophila genetic background. We found that, in the studied population, Wolbachia impacted fitness traits only in certain Drosophila genotypes. Positive effects were manifested in alterations of reproductive traits, but at the cost of reduced lifespan and lower resistance to stress. Based on these findings, we conclude that the effect of the bacteria on fitness and stress related traits depends on the host genetic background, and not on the Wolbachia genotype, at least in the studied Uman population.

2 Materials and methods

Drosophila lines

We used 6 isofemale lines of D. melanogaster (Um59, Um8, Um16, Um15, Um25, Um37) established from flies collected in 2012 in an apple garden near Uman, N 48°45′45.26“, E 30°14’38.97” (Serga et al. 2014). We also used laboratory lines Canton-S (provided by Lyudmila Zakharenko, Novosibirsk, Russia) and Oregon-R (Bloomington Drosophila Stock Center, USA). Lines Um59, Um8, and Um16 were infected with the wMel genotype of Wolbachia, lines Um25 and Canton-S were infected with wMelCS, and Oregon-R, Um15 and Um37 were not initially infected with Wolbachia. All isofemale lines had been cultivated in the laboratory for 3 years before experiments started.

To create genetically similar infected and uninfected fly lines from a single stock, we used an antibiotic treatment. We created tetracycline-treated (T) lines by adding 0.25 mg/ml of tetracycline to the cultivation medium (6 g agar-agar, 50 g semolina, 80 g yeast, 50 g sugar and 2 mL propionic acid per 1 L of water). All flies were reared for two generations on media with antibiotics and four generations on media without antibiotics to mitigate their effects (Fry et al., 2004). After treatment with antibiotics, all lines were tested for presence/absence of Wolbachia by PCR. Since antibiotics can potentially affect fly fitness, we also included lines initially uninfected with Wolbachia to account for the potential effect of tetracycline.

Wolbachia detection and genotyping

DNA was extracted from 20 whole adult flies of each strain by the high-salt method (Aljanabi and Martinez 1997). Wolbachia infection was tested by PCR using a published set of primers to bacterial 16S rDNA (O’Neill et al. 1992) and the wsp gene (Zhou et al. 1998). Each PCR was repeated twice. Wolbachia genotype was identified by the number of the minisatellite repeats VNTR-141, VNTR-105 and the presence of the insertion sequence IS5 in the loci WD0516/7 and WD1310 of the Wolbachia genome as described in Riegler et al. (2005).

Survival

Fly survival was estimated for lines Um59, Um8, Um16, Um25, Um37, Canton-S and Oregon-R treated and not treated with tetracycline. 1–3 day old male flies were placed into vials (14 cm in length and 2 cm in diameter) with the standard medium (10 flies per vial) and were reared at 24–25°С. Live flies were counted every 3 days and transferred into vials with a fresh medium. 100 flies were used for measuring lifespan in each line. Maximum lifespan was determined as the day when no flies remained alive.

Fecundity

We estimated fecundity for all tetracycline-treated and intact lines except Um16 and Oregon-R. For measuring fecundity, newly eclosed flies (mixed males and females) were kept up to 7 days of age in 200 mL bottles (50–60 flies per line). For each line, we selected 15–20 flies that were 5 days old and placed them in separate vials (10 cm length and 5 cm diameter) at 25°С. The flies were then removed from the vials after 22 h and the number of eggs was counted in each vial.

Cold tolerance

Cold tolerance was estimated via the chill-coma recovery time approach in all tetracycline-treated and intact lines from Uman and Canton-S. For this test, 10 males of each line were subjected to temperature stress at −9 °C for 45 s. After that, the flies were placed in 28 °C and the time in seconds until the beginning of the first movement was determined. Cold stress test was performed for male flies aged 3 and 21 days.

Oxidative stress

Oxidative stress tolerance was estimated for all the lines (except Oregon-R) treated and not treated with tetracycline according to Lander’s method (Jünger et al. 2003). Hydrogen peroxide was used as a prooxidant (5% solution). For positive control, we used a 10% sucrose solution. One hundred 3 days old males were placed in test tubes with the agarose medium (10 individuals in each) and treated with 200 μl of 5% H2O2 in 10% sucrose added to the filter. Every two days, flies were transferred into vials containing fresh medium. Sensitivity to oxidative stress was determined by the survival rate at 96 h after the exposure.

Statistical analysis

Kolmogorov-Smirnov test was used to verify the nature of the distribution of the lifespan data. Mann-Whitney test was used to test for differences in lifespan and fecundity. For cold stress tolerance, we used ANOVA followed by the post-hoc Tukey HSD test for pairwise comparisons (Tukey 1949). Differences in tolerance to oxidative stress were analyzed with Fisher’s exact test (Fisher 1922). Differences among the host genotypes were estimated by comparing tetracycline treated lines. All statistical analyses were performed using R v.3.4.4 (R Core Team 2018). The raw data for statistical analysis is available at https://github.com/omaistrenko/WolbachiaPhenotypesSymbiosis.

3 Results

Survival

Lifespan analysis of the host genotypes from the Uman population revealed variability (Supplementary Table 1), whereby the Um25 line was longer lived compared to Um59 and Um37 (Mann-Whitney test, p=0.009 and p=0.013, Benjamin-Hochberg correction). Lifespan was shorter in the Canton-S and Oregon-R laboratory lines compared to most of the other lines established from the Uman population (Mann-Whitney test, p<0.05, Benjamin-Hochberg correction).

Two lines (Um8/wMel and Canton-S/wMelCS) had significantly longer lifespans under tetracycline treatment (Table 1, Fig. 1). These lines were infected with different Wolbachia genotypes. We compared the lifespan of the lines infected with the wMel Wolbachia genotype (Um8, Um16 and Um59) with that of the lines with the wMelCS genotype (Um25 and Canton-S). Lines Um8 and Um59 infected with wMel demonstrated a shorter lifespan compared to Um25 infected with wMelCS (Mann-Whitney test, p <0.00001 and p=0.00024). The lifespan of Canton-S/wMelCS laboratory line was shorter compared to Um59/wMel and Um16/wMel (Mann-Whitney test, p=0.00288 and p <0.00001) and did not differ from Um8/wMel (Table 1; Mann-Whitney test, p=0.0601). In other words, host genotype is a more important factor influencing lifespan compared to the presence of Wolbachia and its genotype.

Table 1 Lifespan statistics for tetracycline treated and untreated fly lines
Full size table
Fig. 1
figure 1

Survival curves for all fly lines. Wolbachia significantly lowered survival in the Um8 and Canton-S lines (P-values obtained from the “survfit” function that fits the Kaplan-Meier regression model)

Full size image

Fecundity

We did not observed differences in fecundity in lines established from the Uman population (Mann-Whitney test, p>0.05, Benjamin-Hochberg correction). However, the laboratory line Canton-S had lower fecundity when compared to other lines (except Um37) (Mann-Whitney test, p<0.05, Benjamin-Hochberg correction) (Supplementary Table 2).

We analyzed flies’ fecundity before and after treatment with tetracycline. In most lines, we did not find differences between tetracycline-treated and untreated flies (Mann-Whitney test, p> 0.05) (Fig. 2). In Um8/wMel and Canton-S/wMelCS, which were infected with different Wolbachia genotypes, the number of eggs laid after tetracycline treatment significantly decreased from (Mann-Whitney test, p=0.00228 and p=0.00096 respectively). Interestingly, Wolbachia increases fecundity in the same lines in which it reduces life expectancy.

Fig. 2
figure 2

Fecundity of all analyzed fly lines (treated with tetracycline (T) and untreated)). Um25 and Canton-S infected with the wMelCS Wolbachia genotype, Um8 and Um 59 infected with the wMel genotype, Um15 and Um37 not initially infected

Full size image

Cold tolerance

We observed an effect of Wolbachia on chill coma recovery time only in line Um8/wMel (Fig. 3.). This effect was only detectable in 21 days old flies. Flies infected with Wolbachia recovered slower from a chill coma (TukeyHSD, p=0.0001). The obtained results suggest that Wolbachia could only affect cold tolerance in certain host genotypes. However, comparison between host genotypes from different lines (after tetracycline treatment) did not show any differences in cold recovery (Supplementary Table 3).

Fig. 3
figure 3

Chill coma recovery time after exposure to extreme cold temperature of males in all analyzed lines (treated with tetracycline, T, and untreated). Canton-S, Um25 infected with the wMelCS Wolbachia genotype, Um8, Um16 and Um59 infected with the wMel genotype, Um15 and Um37 not initially infected

Full size image

Oxidative stress

We did not observe any effect of the host genotype on oxidative stress response. Only the Um15 line had significantly lower oxidative stress resistance in presence of tetracycline among the studied lines (Fisher’s exact test, p<0.05, Benjamin-Hochberg correction) (Supplementary Table 4).

We observed putatively Wolbachia-induced differences in oxidative stress tolerance only in line Um59/wMel, whereby flies treated with tetracycline had higher tolerance to oxidative stress (Fisher’s exact test, p= 0.001, Benjamin-Hochberg correction, Table 2). The effect of Wolbachia on oxidative stress is likely to be insignificant and/or dependent on host genetic background rather than Wolbachia genotype and requires larger datasets to obtain conclusive results.

Table 2 Comparisons of tetracycline treated and untreated flies in the oxidative stress assay, F-test (n=100)
Full size table

Combined effects of Wolbachia on host fitness-related traits

In the present study, we analyzed fitness-related traits of the same fly lines, which allows us to draw conclusions about trade-offs between phenotypes and presence of Wolbachia (Table 3). Both studied genotypes of Wolbachia (wMel and wMelCS) had an effect on the host’s phenotype. But for the wMel genotype of Wolbachia, we observed alterations in all the studied phenotypes of the host: lifespan, fecundity, cold and oxidative stress response. The strongest effect of the bacteria was observed in the Um8/wMel line, in which Wolbachia significantly increased fecundity, but at the cost of a lower life expectancy and sensitivity to cold stress. The effect of the bacteria under the influence of oxidative stress was observed only for a single line, Um59/wMel. The study also included lines (Um15, Um37 and Oregon-R) that were not initially infected with the bacterium but were also treated with tetracycline. We did not observe any effect of the antibiotic treatment on these lines, indicating that antibiotic is unlikely to be a confounding factor in this study. The obtained data indicate that the effect of Wolbachia depends more on the Drosophila genotype rather than the genotype of Wolbachia itself.

Table 3 Effects of Wolbachia on different fitness-related traits
Full size table

4 Discussion

In this study, we evaluated effects of Wolbachia on various life history traits (fecundity, lifespan, cold and oxidative stress) in D. melanogaster isofemale lines established from a unique Uman population that is stably infected by the cosmopolitan wMel genotype and the rare wMelCS genotype of Wolbachia (Serga et al. 2014). The infection frequency varied during the many years of monitoring (Ilinsky and Zakharov 2007; Serga et al. 2014). So, this population is likely a good model to investigate the mechanisms underlying the greater overall success of the wMel genotype compared to wMelCS worldwide, whereby the former genotype has replaced the latter in most fly populations, but still co-exists with it in some. In the absence of significant levels of CI or other reproductive manipulation phenotypes, the success of the bacteria of a certain strain in Drosophila populations is thought to be determined by the transmission rate and fitness benefits (Gundel et al. 2011). Transmission fidelity does not differ between the wMel and wMelCS genotypes and, depending on the fly genotype, reaches 90–100% (Serga et al. 2014). We analyzed the impact of Wolbachia on fitness related traits, such as fecundity, lifespan, survival under the influence of cold and oxidative stress response. The effects of Wolbachia on Drosophila fitness have been investigated repeatedly in multiple studies before, but many studies had conflicting results (Fry et al. 2004; Alexandrov et al. 2007; Roshina et al. 2018; Capobianco et al. 2018). In this paper, we analyzed several fitness related traits in the same lines. We found moderate effects of Wolbachia infection on fitness that depended on the fly’s genotype rather than the Wolbachia genotype.

It has been previously shown that Wolbachia infection can decrease the lifespan of D. melanogaster (Min and Benzer 1997). Significant shortening of lifespan was observed as a result of infection with the wMelPop strain and moderate for the wMelCS genotype (Chrostek et al. 2013). Pathogenic genotypes of Wolbachia are likely to stimulate the immune system. It has been shown previously that bacteria can over-activate the immune system which in turn is associated with decreased lifespan (Libert et al. 2006). Moreover, infected flies showed a decrease in lifespan compared to tetracycline-treated wild Drosophila strains collected in Russia (Roshina et al. 2018), North America (Capobianco et al., 2018), and inbred lines from Drosophila Genetic Reference Panel (Albertson et al. 2013), however, the bacteria were not genotyped in these studies. In our study, we show that in two lines infected with two different Wolbachia genotypes, there is also a decline in lifespan in infected individuals compared to those from which the bacterium was removed using antibiotics treatment. At the same time, Albertson et al. (2013) have shown that a similar effect of lifespan increase after treatment with tetracycline is observed in both Wolbachia-infected and uninfected lines. This finding indicates that presence of other bacteria might be affecting lifespan (Albertson et al. 2013).

The increased fecundity of one wMel-infected line from the Uman population and a wMelCS-infected laboratory line observed in the present study is consistent with our previous observations (Serga et al. 2014) and contradicts findings from Australian populations, where the effect of Wolbachia on fecundity was not detected (Hoffmann et al. 1994). Similarly to our previous results, we did not observe any effect of the wMelCS genotype from the Uman population on fecundity (Serga et al. 2014). It is possible, that the wMel genotype is more capable of affecting the phenotype of the host and consequently promoting itself in the population. In our study, the wMel genotype of Wolbachia affected all of the studied phenotypic traits in the host: lifespan, fecundity, and cold and oxidative stress responses. Interestingly, higher fecundity and shorter lifespans in our study were observed in the same lines infected with the wMel genotype. Reproductive activity is known to be one of the key factors that affect life expectancy in D. melanogaster (Piper and Partridge 2018). Increased reproductive activity is usually associated with reduced lifespan (Partridge and Harvey 1988; Flatt 2011). According to the “cost of reproduction” concept, a trade-off between longevity and reproduction may be likely explained by a re-allocation of nutritional and other resources from somatic maintenance to reproduction (Fowler and Partridge 1989; Adler et al. 2013). So, it may be assumed that Wolbachia infection can promote early reproductive success at the cost of lifespan; these phenotype alterations are dependent on the host genotype. Thus, the interactions between the host and Wolbachia genotypes may potentially lead to context-dependent fitness effects that cause incomplete replacement of wMelCS by the wMel genotype in natural population.

Another explanation for the success of bacteria in a natural population is modulation of the response to stress factors in infected individuals. Temperature is one of the most important environmental abiotic factors that affect the physiology and life history traits. Response to temperature can be affected by host-symbiont interactions (Corbin et al. 2017). Firstly, temperature affects the Wolbachia titers in the host organism and its transmission rate (Ross et al. 2017). Secondly, Wolbachia can influence Drosophila thermal preferences. wMel/wMelsCS-infected flies prefer warmer conditions than uninfected flies (Truitt et al. 2019). Thirdly, the effect of Wolbachia on the timing of recovery from a chill coma has been shown for inbred lines earlier (Albertson et al. 2013). In our study, we found that the wMel-infected Um8 line demonstrated poorer cold tolerance. However, this effect likely depended on the age of the flies and on the genetic background. In an earlier study, the influence was inconsistent and the presence of Wolbachia either increased or decreased the recovery rate depending on the Drosophila line (Albertson et al. 2013). These findings together with our work suggest that cold recovery potentially depends on the interactions between host genetic background and Wolbachia.

Oxidative stress is another important factor that substantially influences aging and longevity. Wolbachia is known to induce excess ROS and, as a result, higher superoxide dismutase activity (Brennan et al. 2012). In our study, tolerance to external oxidative stress did not depend significantly on the Wolbachia genotype. Only the Um59 line, originally infected with the wMel genotype, had higher oxidative stress tolerance after tetracycline treatment. So, for this line, Wolbachia might have had a negative effect on the oxidative stress tolerance.

In this work, we used an approach in which fitness traits in the Wolbachia-host system can be investigated without transferring Wolbachia to the same fly genotype by using tetracycline antibiotic treatment. The advantage of this approach is the preservation of the natural interaction between Wolbachia and its host. However, tetracycline is a broad-spectrum antibiotic and can affect the composition of microbiota, as well as it can have negative effects on fitness traits (Li et al. 2014; O’Shea and Singh 2015). After antibiotic treatment, we conducted experiments on the 4th generation of files and we did not observe any effects of antibiotics on control lines that were initially not infected with Wolbachia.

Based on the results of this study and previous publications, we conclude that Wolbachia infection can cause both deleterious and beneficial effects on different fitness components in Drosophila, and these effects dependent on the host genetics rather than Wolbachia genotype. Moreover, the effects are not present in all fruit fly lines. Beneficial effects are often manifested in improved reproduction, which however likely comes at the cost of shortened lifespan and lower resistance to stress. Replacement of the supposedly more deleterious Wolbachia isolate wMelCS with the more neutral/beneficial wMel appears to be in line with the overall directionality in the evolution of host-symbiont relationships from parasitic toward more neutral/mutualistic interactions.

In conclusion, we show that Wolbachia may affect fitness related traits in Drosophila, such as fecundity, lifespan and stress tolerance. Further analysis is required of fruit fly lines originating from multiple populations to disentangle the effects of D. melanogaster and Wolbachia genotypes on the fitness of this host-symbiont system.