Wolbachia in European Populations of the Invasive Pest Drosophila suzukii: Regional Variation in Infection Frequencies

The invasive pest Drosophila suzukii is characterized by a specific fresh-fruit targeting behavior and has quickly become a menace for the fruit economy of newly infested North American and European regions. D. suzukii carries a strain of the endosymbiotic bacterium Wolbachia, named wSuz, which has a low infection frequency and no reproductive manipulation capabilities in American populations of D. suzukii. To further understand the nature of wSuz biology and assess its utility as a tool for controlling this pest’s populations, we investigated the prevalence of Wolbachia in 23 European D. suzukii populations, and compared our results with those available in American populations. Our data showed a highly variable infection frequency with a mean prevalence of 46%, which is significantly higher than the 17% found in American populations. Based on Multilocus Sequence Typing analysis, a single wSuz strain was diagnosed in all European populations of D. suzukii. In agreement with American data, we found no evidence of cytoplasmic incompatibility induced by wSuz. These findings raise two questions: a) why Wolbachia is maintained in field populations of D. suzukii and b) what are the selective forces responsible for the variation in prevalence within populations, particularly between European and American continents? Our results provide new insights into the D. suzukii-Wolbachia association and highlight regional variations that await further investigation and that should be taken into account for using Wolbachia-based pest management programs.


Introduction
Drosophila suzukii Matsumura (Diptera: Drosophilidae), first described in Japan in the early 1900s [1], is an invasive pest of Southeast Asian origin.Since its early detection in California (USA), Spain and Italy (Europe) in 2008, D. suzukii has rapidly spread through these two continents aided by global trading and absence of niche competitors [2,3,4,5,6].Although the vast majority of Drosophila species are not fruit pests (larvae developing only in damaged or rotten fruits), D. suzukii is able to lay eggs on healthy ripening fruits thanks to female's serrated ovipositor [7].A wide range of soft and stone fruits including raspberry, strawberry, blueberry, plums and grapes come under D. suzukii's damage range [4,8].Damage is caused by developing larvae inside the fruit, leading to millions of dollars of annual economic losses worldwide [9,10].Control of D. suzukii populations in the field mainly relies on the use of chemical pesticides, a practice with serious drawbacks because of its use close to harvest and the consequent risk of high residues left on fruits.Management agendas are therefore in permanent search for alternative strategies including those based on bio-control [4,11,12,13,14,15].
Previous studies revealed the presence of the bacterium Wolbachia in D. suzukii (strain named wSuz) [16,17,18,19,20] but of no other heritable bacterial symbionts [21].This is not surprising since Wolbachia is widespread among terrestrial arthropods with an estimate of 52% of species infected at variable prevalence [22].Wolbachia is a maternally inherited bacterium that has established a wide range of relationships with their hosts, from mutualism to parasitism.It is particularly known for its ability to manipulate host reproduction through different strategies to maximize its spread and maintenance in host populations [23].The most common strategy is cytoplasmic incompatibility (CI), a sperm-egg incompatibility expressed in crosses between infected males and uninfected females leading to the death of embryos in diploid species, thereby facilitating Wolbachia to spread throughout host populations [24,25].However, in the absence of CI or other means of reproductive manipulation, Wolbachia can also provide direct host fitness benefits (e.g.increasing survival time, enhancing reproduction, provisioning nutrition or protection against viruses), which may sometimes allow infected individuals to outcompete uninfected counterparts [26].For example, in D. mauritiana, Wolbachia infection leads to a decrease of apoptosis and an increase of mitotic divisions in germ line stem cells of infected female ovaries, resulting in more eggs produced than in uninfected ovaries [27].Lack of any of the suggested effects on the host can potentially lead to loss of infection from a population [28,29,30].Overall, the infection dynamics of Wolbachia depend on the bacterial strain, host species, genetic background, environmental conditions and interactions among these factors [31].Recently, a field survey in North America indicated a low prevalence of Wolbachia in field populations of D. suzukii from four localities sampled over two years (prevalence ranged between 7 to 58%), with a mean infection rate of 17% [19].This study showed no signs of CI or any other reproductive phenotype induced by wSuz in D. suzukii.Moreover, the imperfect transmission of Wolbachia from wild-caught females to their offspring (20 to 95%) suggested possible direct fitness benefits provided by wSuz to its American D. suzukii host but the phenotypic nature of these benefits has not yet been determined.
In order to improve our understanding of the global infection dynamics of wSuz in natural D. suzukii populations and thereby assess the utility of Wolbachia as a tool for controlling this insect pest population in the field, we investigated the prevalence of Wolbachia in 23 localities among eight European countries, typed Wolbachia strains through a Multilocus Sequence Typing approach and performed mating experiments on two different European populations.1).Adults were caught directly in orchards or from attractive traps, placed alive in ethanol (96%) and stored at -20°C until DNA extraction.For Wolbachia screening, we gave priority to females because they are known to be responsible for Wolbachia transmission.DNA from whole individuals was extracted using the Nucleo Spin Tissue kit (Macherey-Nagel).The quality of extraction was checked by PCR targeting the arthropod-specific locus Internal Transcribed Spacer ITS2 gene (Its2U/Its2L) [32].Wolbachia detection was performed using Wolbachia specific primers for two different genes: ftsZ, a bacterial cell division gene (F2/ R2) [33], and wsp (Wolbachia surface protein) (81F/691R) [34].PCR reactions were performed in 25 μL volumes containing 100 μM dNTP, 200 nM primers, 20mM Dream Taq Green Buffer, 0.5 IU Taq DNA polymerase (Eurobio) and 1μL of DNA template.Cycling conditions were 94°C (2 min), 94°C (30 sec), 55°C (ITS2 and ftsZ primers) or 52°C (wsp primers) (30 sec), 72°C (45 sec), 72°C (10 min) for 34 cycles.All the primer sequences were indicated in S1 File.PCR products were visualized in 1% agarose gels.An individual was considered infected when the two Wolbachia-specific primers produced fragments of the appropriate size.To reduce the possibility of false negatives, all samples for which we failed to amplify Wolbachia signal with normal PCR were verified by a more sensitive real-time quantitative PCR method using the wsp primers 81F/691R.The 10μL reaction mixture contained 200 nM of each primer, 5 μL of Light Cycler 1 480 SYBR Green I Master (Roche) and 1 μL of DNA sample.The amplification consisted of 10 min at 95°C followed by 40 cycles of 10 sec at 95°C, 20 sec at 53°C and 30 sec at

MLST approach
PCRs targeting six genes of Wolbachia (wsp, gatB, coxA, hcpA, ftsZ, fbpA) [34,35] were carried out on two randomly selected wSuz-infected samples per locality (Fig 1 and Table 1).All PCR reactions were performed as described before.Annealing temperature was 55°C except for wsp (52°C) and all the PCR products were sequenced.Sequences were aligned using MUSCLE [36] algorithm implemented in CLC DNA Workbench 6.9.1.(CLC Bio) and inspected by eye.On an average, we sequenced and compared approximately 3200 nucleotides per individual (between 470 and 610 nt per gene).All primers used in this study are presented in S1 File.Before starting the CI experiment, both lines were acclimatized under the same laboratory conditions for three generations in Cambridge Lab at a constant temperature of 22°C and 65% relative humidity with 12:12 light/dark cycle, thereby avoiding any bias due to lab environments.

Cytoplasmic Incompatibility (CI) assays
To obtain Wolbachia-infected and uninfected fly lines with the same genetic background, antibiotic treatments were performed on infected lines using 0.25mg mL -1 tetracycline for three consecutive generations in mass populations.After treatments, flies were fed on normal standard diet for two further generations to recover them from any side effect caused by antibiotic treatment.Due to different lab practices, flies from Italian populations were further allowed to feed on natural fecal material from infected individuals to re-acquire their loss of gut-associated microbiota but flies from France were not gut-flora restored.Ten isofemale lines were established and the presence of Wolbachia was checked by PCR in mothers after laying eggs.The absence of Wolbachia was re-confirmed by real-time PCR as described above.This was repeated for three generations and then, for each population, one isofemale line was retained for crossing experiments.Their infection status was also checked and confirmed just before the CI assays (n = 20 for each line), which were performed on the 10 th generation after antibiotic treatments were stopped.
All types of crosses (infected female x infected male, infected female x uninfected male, uninfected female x infected male and uninfected female x uninfected male) were performed for each genetic background at Cambridge.Freshly hatched individuals were sexed and placed separately into cornmeal diet tubes to ensure the virginity of flies.A 3-days old virgin male and a 5-days old virgin female were allowed to mate in food vial for 24h.Females were then individually allowed to oviposit for 48h on grape-juice agar petri dish.Every 48h, females were transferred to a new petri dish, following 48h period for oviposition, and the total number of eggs along with the number of hatched larvae per female were recorded.Experiment was performed in replicates for each type of cross as mentioned in Table 2.The incompatibility relationship was determined by the rate of eggs that hatched.Females that laid less than 10 eggs were not included in the analysis (percentages are indicated in S2 File).

Statistical analysis
We tested the difference in Wolbachia prevalence between American and European continents by fitting a generalized linear model with a quasi-binomial error (data for North America was taken from [19]).Samples originated from the same population but collected in several years were pooled.We tested for the existence of differences between samplings by using Fisher's exact test.Estimates of the mean prevalence on each continent were transformed from logits into percentages.Statistical analysis of the CI data was done using a generalized linear model (quasi-binomial family).An ANOVA was done on this model and Tukey's (HSD) tests were used for two by two comparisons.R 3.2.2version [37] was used to perform all the analysis.

Ethics Statement
Sampling of D. suzukii was carried out on private lands with owners' permission.All individuals were sampled in experimental stations and did not require specific permission.The field studies did not involve endangered or protected species.

Variable wSuz infection frequencies in D. suzukii populations
We detected Wolbachia in 22 populations out of 23 surveyed from eight European countries (Fig 1).Out of the 561 D. suzukii flies tested, 258 were found infected with Wolbachia (Table 1).We found a highly variable infection prevalence (from 0 to 100%) between localities (χ 2 = 192, d.f = 42, P<0.001).The two extreme values (0 and 100) have been observed only in two populations (Bellegarde in France and Vigolo Vattaro in Italy respectively); in all the other localities, prevalence ranged between 15% and 85% with a mean of 46%.
In two European localities, Mougins (in France) and San Michele (in Italy), we sampled for two consecutive years at the same period (October, 2012-13 and September, 2013-14 respectively) and found an increasing trend of the Wolbachia prevalence over time for both localities (Mougins: from 0.46 to 0.57; San Michele: from 0.22 to 0.65; Table 1) though the trend was not significant (Fisher's exact test, P = 0.74 and P = 0.05 respectively).
Hamm [19] screened 929 individuals sampled from four American localities of D. suzukii over two years and found infection frequencies ranging from 7 to 58% with a mean of 17%.Therefore, in both continents, the prevalence of wSuz is highly variable.Overall, the European infection frequencies is found almost three times significantly higher than the ones in North America (GLM: t = 4.54, d.f = 26, P<0.001).
European D. suzukii flies carry one single sequence type of Wolbachia "wSuz" MLST analysis, assayed from two individuals per population (Fig 1 and Table 1), indicated that all European D. suzukii individuals carry the same Wolbachia sequence type (100% nucleotides identity for all genes).Sequences were identical to those of wSuz previously found and characterized in D. suzukii individuals originated from Italy [18] and North America [19].The sequences obtained in the present study are recorded in Genbank as KS308222-7.

Cytoplasmic incompatibility assays
To test whether wSuz is able to induce CI, which could facilitate its spread throughout European D. suzukii populations, we performed mating experiments on flies originating from France and Italy (Table 2).In diploid species, CI leads to the death of embryos when infected males mate with uninfected females; therefore in case of wSuz induced CI expression, we expected a lower egg hatch rate in these crosses.We found a significant effect of the type of cross on the hatching rates only for the French population (F = 3.86, d.f = 3, P = 0.01; for Italian populations: F = 2.14, d.f = 3, P = 0.10).Globally, the hatching rate was higher in crosses that involved individuals of the same infection status (control crosses) than that of different infection status (infected and uninfected).The "incompatible" crosses (infected male x uninfected female) tended to have the lowest hatching rate.For the French population, the difference was significant only when the incompatible cross was compared to the two control crosses: infected male and female (Tukey HSD test, P = 0.02) and uninfected male and uninfected female (Tukey HSD test, P = 0.01).For the Italian population there is no significant difference between the incompatible cross and the three others.Therefore, like in the American populations, these results do not allow us to conclude that wSuz can induce CI in European D. suzukii populations.Mean eggs hatch rate (only females that lay at least 10 eggs) ± standard deviation (sd).N: number of crosses.IN: infected with Wolbachia, UN: uninfected.Statistical analysis was performed on lines from the two origins separately.One-way ANOVA indicated that the hatching rate differed significantly between the mating types.One-way ANOVA was followed by Tukey's test to compare between crossing types (means marked with the same letter are not significantly different; P = 0.05).

Discussion
We found that the Wolbachia infection in D. suzukii is highly variable in European populations with a mean prevalence of 46%.This infection rate differs significantly from the one found in the North America (17% in average) [19].Several parameters could explain this variability such as the efficiency of vertical transmission or the costs/benefits of infection, which might differ between continents.Moreover, variability can also be due to differences in the host and/ or Wolbachia genetic backgrounds as indicated in other insect species by trans-infection experiments using a given Wolbachia strain injected into different hosts (for example, see [38]).Apart from different host-Wolbachia association dynamics, the variability can also be dependent on environmental factors (i.e.temperature, diet, larvae density. ..) [39,40,28,41,42] that might vary locally and across time with their colonization dynamics, hence providing higher fitness advantage for Wolbachia persistence in European D. suzukii populations than in the American ones.If this is true, we should expect persistence at most/all localities to a certain infection rate representing the fitness advantage based on that particular locality's environmental and habitat specificity.Although we found a (non-significant) increasing trend for the Wolbachia prevalence in two localities, more sampling time points over longer durations are needed to draw a clearer picture.
Infection dynamics of vertically transmitted symbionts depend on the phenotypic effect induced on the host, the fitness cost and/or benefits of the infection, and how reliably symbionts are inherited.Our results clearly showed that wSuz is not fixed in D. suzukii populations as otherwise expected in cases of obligatory mutualism.We did not detect any significant evidence for CI induction by wSuz in Italian or French D. suzukii population.An imperfect vertical transmission of endosymbiotic bacteria (20 to 95% in American populations) [19] associated with an absence of strong reproductive manipulation would theoretically lead to elimination of the symbiont from host populations unless it provides the other host fitness advantages [29].The stable presence of wSuz in all but one of the populations studied in America and Europe would thus suggest that wSuz may provide direct benefits to its host allowing its maintenance in natural populations of D. suzukii.Another possibility would be frequent intraspecific horizontal transfers of Wolbachia [43].However, we favored the first hypothesis because the second one would imply that horizontal transfers occur in all populations from two continents, which seems unlikely.In addition, Wolbachia are also known as a mutualist in insects [44] providing beneficial effects to their native hosts, such as an increase of host survival [45] or fecundity under nutritional stress [46], or protection against viral, microbial, and fungal pathogens [47,48,49,50,51,52].While no fitness benefits have been detected in American populations [19], a recent study suggested a beneficial effect of Wolbachia infection on female fecundity in an Italian population of D. suzukii [53].Since both of the studies used different protocols to perform the assays, different lab practices might impair the significance of this difference.For example, Hamm [19] crossed native Wolbachia infected and uninfected flies from the wild: thereby we cannot exclude the confounding effect of different host genetic backgrounds on the fly fecundity.On the other hand, Mazzetto [53] compared infected individuals to uninfected ones that were immediately obtained after antibiotic treatments.Therefore the reduced fecundity of uninfected females can also be explained by side effects of the antibiotics on mitochondria and/or host metabolism [54].Although with caution, excluding any of these confounding factors suggest that the beneficial effect induced by wSuz may depend on the fly genetic background and/or that D. suzukii from Europe and North America are infected by different Wolbachia strains.Highly sophisticated approaches, for example, Next Generation Sequencing (NGS) based tools are needed to have more insights into the Wolbachia strain diversity in D. suzukii populations.Moreover, future experiments involving exchange or introgression of European and American Wolbachia strains to study the life history traits of host organisms in the presence and absence of Wolbachia would help to unveil whether differences in wSuz-mediated fitness benefits in different continental D. suzukii populations are Wolbachia strain-dependent or host dependent.
There is a growing interest in exploring symbiotic microorganisms for biocontrol management programs [55,56,57]; in the case of Wolbachia, the idea is to exploit its capability to induce CI to control natural populations of arthropod pests in an Incompatible Insect Technique (IIT) fashion [58,59], where males infected with a CI-inducing strain of Wolbachia are released in the field to mate with incompatible females leading to embryo mortality and ultimate population suppression.This procedure is somewhat analogous to the Sterile Insect Technique (SIT), a species-specific method of insect control that relies on the release of large numbers of sterile males instead of incompatible males [60], but the advantage is that insects do not have to be irradiated or genetically modified before release.The inconvenience is, however, that IIT and SIT require solid methods allowing efficient males and females separation, but D. suzukii males are easily recognizable because of their spotted wings [1].Considerable research efforts have already demonstrated that Wolbachia-inducing CI could be used as a tool for population control [61,62,63].However, it might be more complicated if Wolbachia naturally infects the insect pest like in the case of D. suzukii since it would require finding an incompatible CI-inducing Wolbachia strain, which is not rescued by wSuz in its original host.Hamm [19] and our results are concordant in showing that D. suzukii is infected by only one strain of Wolbachia, which is present in almost all populations screened and does not induce CI.Transinfection experiments using closely related Wolbachia strains such as the highly CI-inducing wRi strain from D. simulans will thus be needed to assess modification and rescue capabilities of different strain combinations towards a bidirectional CI-based control strategy.
561 individuals were sampled between 2010 and 2014 in 23 locations from eight different countries (Austria, France, Germany, Italy, Slovenia, Spain, Switzerland and United Kingdom) of the European continent (Fig 1 and Table

Fig 1 .
Fig 1. Wolbachia prevalence and collection sites for D. suzukii individuals.A number was assigned for each sampling site and details for each locality are given in the Table 1.Reprinted from http://d-maps.com under a CC BY license, with permission from Daniel Dalet, original copyright 2007-2016.doi:10.1371/journal.pone.0147766.g001

Table 1 .
Sampling information and percentage of individuals infected by Wolbachia.
The sample number corresponds to the code of the sampling indicated in Fig 1. CI: 95% confidence interval.doi:10.1371/journal.pone.0147766.t00172°C.Only for two individuals, Wolbachia infection was not detected by classic PCR but detected by real-time quantitative PCR.

Table 2 .
Cytoplasmic incompatibility assays for D. suzukii from France and Italy.