Infection effects of the new microsporidian species Tubulinosema suzukii on its host Drosophila suzukii

Microsporidian infections of insects are important natural constraints of population growth, often reducing lifespan, fecundity and fertility of the infected host. The recently discovered Tubulinosema suzukii infects Drosophila suzukii (spotted wing drosophila, SWD), an invasive pest of many fruit crops in North America and Europe. In laboratory tests, fitness effects on larval and adult stages were explored. High level infection after larval treatment caused up to 70% pupal mortality, a decreased lifespan and a 70% reduced oviposition of emerging adults in biparental infection clusters. A shift to higher proportion of female offspring compared to controls suggested a potential parthenogenetic effect after microsporidian infection. A clear sex-linkage of effects was noted; females were specifically impaired, as concluded from fecundity tests with only infected female parents. Additive effects were noted when both parental sexes were infected, whereas least effects were found with only infected male parents, though survival of males was most negatively affected if they were fed with T. suzukii spores in the adult stage. Although most negative effects on fitness parameters were revealed after larval treatment, infection of offspring was never higher than 4%, suggesting limited vertical transmission. For that reason, a self-reliant spread in natural SWD populations would probably only occur by spore release from cadavers or frass.

Survival rates of inoculated SWD larvae and adults. Kaplan-Meier survival analyses were performed with individuals from the same experiment, which survived to the adult stage (Fig. 4a). Eclosion occurred from day 13 to day 19 as a result to some age differences among inoculated L2 larvae. Survival analysis was performed using log rank test with Bonferroni adjustment and revealed significant differences in the life span of SWD adults for every treatment compared to the control (Survival formula: χ 2 = 534, DF = 3, P < 0.001) (Fig. 4a). It was striking that the survival curves of the control group and the group treated with 5 × 10 2 spores were similar until day 35, then mortality increased in the treated group, indicating a delayed effect on the survival of adults when inoculated with very low spore concentration as larvae.
To study the effect of late infection initiation, newly eclosed SWD adults were treated with 3 × 10 5 spores in 100 µl. In Kaplan-Meier analyses (Fig. 4b), survival rates of control males was significantly higher than infected males (P = 0.027), infected females (P = 0.001) and control females (P = 0.004), whereas control females had similar survival curves as infection treatments for both males and females (χ 2 = 13.1, DF = 3, P = 0.004).
Fecundity and fertility of SWD treated with T. suzukii. When SWD larvae were infected, the mean number (± SEM) of eggs oviposited per female was 135.97 ± 8.81 for the untreated control pairs (Table 1). Pairs with infected parental females deposited 51% less eggs compared to control. For pairs with infected parental males the reduction was 48% and approximately 71% reduction in oviposition when both parents were infected (Table 1). For pairs with at least one infected parent, oviposition differed significantly from the untreated control, indicating a strong effect of microsporidium infection on the fecundity of SWD (χ 2 = 48.84, DF = 3, P < 0.05, Dunn's test: F-C: P = 0.006, M-C: P = 0.019, MF-C: P < 0.001). The mean lifetime (Table 1) of pairs did not differ from each other (χ 2 = 2.683, DF = 3, P = 0.44), indicating parental longevity was not a factor in the differences in fecundity and fertility among infected and control SWD.
Comparing the number of eggs and resulting offspring within each treatment, no reduced hatching rates were observed (Table 1). Indeed, the number of viable offspring was very similar to number of eggs, thus the primary For control females, 45% of total eggs oviposited were laid in the first 10 days after eclosion (Table 1). Treatments of pairs including infected females differed significantly from the untreated control, but not pairs with only males infected (F = 13.96, DF = 3, P(F) < 0.001, Tukey HSD: F-C: P < 0.05, M-C: P > 0.05, MF-C: P < 0.001).
When SWD adults were inoculated with T. suzukii, mean oviposition was approximately 30% less in microsporidian-treated groups (MF) compared to the control groups (C) ( Table 2), but no significant difference between both treatments was noted. Also, the number of eclosed offspring recorded 18 days after oviposition did not differ among treatments (Table 2). In both treatments, fertility was about 15-20% lower than fecundity but this difference was not significant ( Table 2).
The sex ratio of offspring from T. suzukii inoculated parents was shifted to female offspring (χ 2 = 9.09, DF = 1, P = 0.003), control treatments showed a sex ratio close to 1:1 (χ 2 = 0.46, DF = 1, P = 0.50) ( Table 2). Inspection of 223 F1 adults derived from T. suzukii treatment did not show any infected offspring, indicating that the microsporidium is not vertically transmitted if infection occurs in adult stage. Microscopic inspection of 31 Microsporidia-treated pairs resulted in one male fly with an established T. suzukii-infection.

Discussion
A novel microsporidium T. suzukii was discovered in 2015 in laboratory-reared SWD 39 . In this study, the potential impact of T. suzukii infection on SWD larvae and adults has been investigated and significant concentrationdependent effects on the mortality as well as the eclosion, survival, lifetime fecundity and effects on offspring were analysed when L2 larvae were inoculated with the microsporidium.
We succeeded in measuring the infection progress in individual SWD inoculated in larval stage by qPCR, which allowed us following the replication cycle of T. suzukii upon emergence of flies. Within the first 13 days post inoculation, DNA copies of microsporidian SSU rDNA increased up to 1000-fold, indicating strong microsporidian proliferation and explaining low pupal survival and adult eclosion after exposure to a high number  www.nature.com/scientificreports/ of spores. Although we counted a maximum of 10 6 spores when dissecting flies, quantification of DNA copies suggested even up to 10 8 spores per fly in late infection stages. We found high larval mortality, whereas eclosion declined dramatically (up to 70%) with increasing inoculated spore concentration. It has to be emphasized that the used amount of spores was experimentally motivated, but no data are available if such doses would be achieved in natural transmission scenarios when spores are released from dead individuals. SWD larvae surviving the T. suzukii infection would die during pupal or in early adult stages compared to controls. At high spore concentration, only 30% of flies survived the adult stage and these died within 40 days compared to 65 days in  www.nature.com/scientificreports/ control. At low spore concentration, differences between the inoculated and control groups were not seen until day 35, when mortality increased. There was no evidence that T. suzukii has such strong effect on eclosion of other drosophilids when L2 larvae were inoculated. Compared to controls, eclosion of D. willistoni was reduced about 33% when larvae were inoculated with 5 × 10 4 spores but D. melanogaster was not affected at all. Virulence of T. suzukii to these two drosophilids was low, but transmission efficiency as another important infection parameter 40 was not determined for both in this study. When adults of SWD were inoculated, male control flies had a significantly longer lifespan compared to other treatments. This effect could be due to the experimental design where one female and one male were kept together. As shown for other drosophilids, mating significantly reduces survival of females and the absence of rivals results in increased male lifespan [41][42][43] . The high larval and lower adult susceptibility of SWD are in line with findings from other microsporidia of drosophilids, such as Tubulinosema ratisbonensis and Tubulinosema kingi 23,[44][45][46][47][48][49] .
Fecundity and fertility were significantly reduced after larval inoculation resulting in about 70% reduction of egg deposits in biparental infections, but not after adult treatment. Females appeared to be the mainly weakened, as SWD pairs with only females infected had the second highest reduction of egg deposits. One important fitness parameter is the oviposition rate within the first 10 days after adult emergence, when drosophilids are laying the majority of eggs compared to the rest of the lifetime 50 . SWD pairs inoculated with T. suzukii laid about 46-57% of the lifetime-eggs within the first 10 days, whereas untreated controls deposited 45%. Selection studies with continuous exposure of D. melanogaster to microsporidia showed increased longevity and higher early-life fecundity rates in selection lines when infected with the pathogen. In contrast, longevity of controls challenged with microsporidia was significantly shorter compared to the selection line, whereas the lifetime fecundity was higher in absence of the pathogen 51 . In our study, the sublethal concentrations for this experiment (single exposure of 10 μl with 1.5 × 10 4 spores/μl) did not reduce adult fly survival. Hence, we observed no significantly higher oviposition than in controls within 10 dpi, comparable with results of earlier studies for T. kingi-infected D. melanogaster 44 . We found no change in fertility, all treatments showed only 10% reduced hatching of offspring assuming T. suzukii has no effect on fertility. No tissue-pathological evidence for T. suzukii infections in male gonads or sperms was found in a recent study, but sample size was low 39 . However, the combination of parental pairs with male-only infection (M: u♀ × i♂) resulted in reduced oviposition as well. We found a statistically significant shift to higher proportions of female offspring (0.64) compared to controls (0.49) in this experimental setup, suggesting a parthenogenetic effect due to microsporidian infection. Since this effect was found from only three infected pairs more replicates would be useful to further support this finding. To our knowledge, pathogen-induced parthenogenesis was never described for SWD, although Wolbachia bacteria are known to induce female-biased progeny for approximately 40 drosophilid species [52][53][54][55][56] . The number of progeny in male-only treatments was comparably low to female-only treatment. Previous studies reported unmated females producing unfertilized eggs have a reduced total number of progeny for some drosophilids species 57,58 .
The here discussed high larval and low adult infection effects were similar to previous results of Vijendravarma et al. on stage-specific susceptibility of D. melanogaster treated with T. kingi 47 . For invertebrate hosts, several resistance mechanisms after microsporidian exposure were reported, showing induced cellular immune response like encapsulation, melanisation, hemocyte production, and phagocytosis 65,66 . In Diptera (i.e. drosophilids and mosquitoes), microsporidia infection induced up-regulation of lysozym genes and AMP production 65 .
In selection experiments with D. melanogaster challenged with the microsporidium T. kingi or the fungus Beauveria bassiana, similar effects were observed: Flies selected on the pathogens had increased fitness and higher intra-specific competitiveness under pathogen pressure compared to non-selected control lines in presence of pathogens 51,67 . Due to the relatedness of T. kingi and T. suzukii 39 , similar outcomes may hold for SWD. Hence, combined effects of potential resistance to T. suzukii as well as over-represented fitness costs with low adult susceptibility implicate that this microsporidium might be a difficult tool for SWD pest management. Above that, commercial field application of microsporidia for insect biocontrol is still a matter of debate. In the 1970s, Antonospora (Paranosema) locustae (formerly: Nosema locustae) was used for locust and grasshopper management 25 but low and slow mortality, inefficiency for some target species, and insignificant efficacy resulted in an economic failure of this agent. Table 2. Lifetime, fecundity (oviposition) and fertility (offspring) of SWD pairs inoculated with T. suzukii spores at adult stage. Sex ratio of resulting offspring was determined after eclosion of progeny. ID = treatment abbreviation (C = untreated control, MF = both infected), treatment: ♀ = female, ♂ = male, u = uninfected, i = infected, R/N = number of replicates and total pairs, ( §) = significance letter (< 0.05), SEM = standard error of mean, N (sex ratio) = total number of offspring analysed for sex. Lifetime, eggs/pair/lifetime, eggs/pair/10 days, offspring are shown as mean. www.nature.com/scientificreports/ Furthermore, we conclude T. suzukii is not transmitted transovarially as offspring infection was rarely observed, but is horizontally transmitted via cadavers or frass. Transovum transmission is also possible. In former studies of the authors, T. suzukii spores were found inside maturing SWD ovaries of early adult stage females, possibly influencing egg development 39 . Spores were not found inside mature ovaries but infecting adipose tissue surrounding the ovaries, which can be transferred via egg deposition 39 . Other studies revealed transmission rates of 1-11% in Drosophila showing a similar transmission pattern for Tubulinosema microsporidia infecting closely related drosophilids 23,45 . In experiments where adults were inoculated, no oviposition reduction or vertical transmission was observed, suggesting a delayed infection process 59,60 .
In our experiments, T. suzukii had a very strong impact on longevity and fecundity of SWD in the lab when early larval stages were infected. Based on the negative effects, SWD infection could have a population reducing effect but adult flies are no suitable targets of T. suzukii. As this microsporidium was only found in a laboratory population so far, further field populations of SWD should be screened for natural occurrence of T. suzukii and other potential pathogens for biological control of this fruit damaging fly.

Material and methods
Insect host rearing. Microsporidia-free D. suzukii (SWD) flies were maintained in 30 × 30 × 30 cm cages (BUGDORM, MEGAVIEW SCIENCE CO., Taiwan) with water, a diet for adult flies (1 g brewer's yeast and 1 g sucrose) and an artificial oviposition medium as described elsewhere ( 39 , modified from Chabert, et al. 30 ). The oviposition medium was replaced weekly. If larvae with synchronous development were needed, the medium was replaced every 4 h. Insect rearing and subsequent biotests were performed under the following conditions: 22 ± 1 °C, 50% relative humidity (r.H.), 16:8 h light:dark photoperiod. Microsporidia-free D. melanogaster (DM) and D. willistoni (DW) were maintained under same conditions as described for SWD.

Preparation of T. suzukii spores for SWD inoculation. T. suzukii was first isolated from a D. suzukii
rearing originating from flies collected in Oregon, USA 38 . Adult D. suzukii were homogenized with a micro pestle, dissolved in distilled sterile water and filtered through four layers of gauze, then additionally filtered through a cotton filter disc with 12-15 μm particle retention (Grade 1288, Ø 90 mm, SARTORIUS AG, Göttingen, Germany). Spores were precipitated by centrifugation at 10,000×g (Centrifuge 5424 R, EPPENDORF, Hamburg, Germany) and resuspended in 500 μl sterile tap water. Spore concentration and purity was determined with a Thoma hemocytometer under phase contrast microscopy (DMRB, LEICA, Wetzlar, Germany), followed by dilution in sterile tap water to final concentrations as required for subsequent biotests.

Preparation of standards in real-time quantitative PCR (qPCR). T. suzukii spores were extracted
from adult SWD carrying an infection with spores (about 3 weeks after initial inoculation of L2 larvae with 10 μl spore suspension containing 1.5 × 10 4 T. suzukii spores). To produce standard curves, ten flies were homogenized with a micro pestle. About 1 × 10 7 spores were purified with one filtration step through four layers of gauze mesh and a final purification with Percoll. For this purpose, 400 μl spore suspension was overlaid on 1.6 ml 75% Percoll (MERCK, Darmstadt, Germany) dissolved in 1 × PBS in a 2-ml reaction tube and spun down for 20 min at 12,900×g and 15 °C in an EPPENDORF centrifuge (5424R, EPPENDORF, Hamburg, Germany). The spores formed a band close to the bottom of the reaction tube. The spore band was washed twice in 1 × PBS at 15,000×g for 5 min, and the resulting pellet was resuspended in distilled water. Afterwards, spores were inspected for purity under phase contrast microscopy (DMRB, LEICA, Wetzlar, Germany). For preparation of a qPCR standard, serial dilutions of purified spores were prepared with 10 1 -10 6 spores in 100 μl distilled water. SWD inoculation and DNA extraction for qPCR. L2 larvae of SWD were exposed to 10 μl spore suspension with 4 × 10 4 spores in a microtiter plate overlaid on 440 μl pureed apple. Every 2-3 days three larvae and later pupae or adults were removed and euthanized with ethyl-acetate and surface sterilized with 0.05% sodium hypochlorite. One individual was examined visually for infection by light microscopy with 400 × magnification (DMRB, LEICA, Wetzlar, Germany) and a smear was stained with modified Giemsa-stain according to Eberle et al. 61 . Two larvae per replicate were used for genomic DNA extraction as described above and then qPCR. This was repeated for adult SWD which were starved for 3 h followed by bulk feeding for 18 h with a spore suspension containing in total 5 × 10 5 T. suzukii spores mixed with blue food colour (modified droplet feeding method from Hughes and Wood 62 ). To each group of ten flies, 10 μl spore suspension were provided. The time frame for euthanizing four flies per replicate (day 3,5,10,18,28,38) was longer than for larval inoculation, as infection was sometimes delayed in adults. Two flies were prepared for microscopy and Giemsa-staining and two other were used for genomic DNA extraction after surface sterilization with sodium hypochlorite.
Sample and standard spore preparations (see "Preparation of standards in real-time quantitative PCR (qPCR)") were centrifuged at 15,000×g for 10 min. The spore pellet was dissolved in 200 μl CTAB lysis buffer (APPLICHEM, Darmstadt, Germany). www.nature.com/scientificreports/ AGC CAT GCA TGCT-3′) and 1 μl 10 mM reverse primer Tn562R (5′´-CCG CTT CGA ATA TAA GCA TTGA-3′) 39  (a) LC 50 determination: application of five different spore concentrations in logarithmic scale (10 1 to 10 5 spores per μl) to L2 larvae in a single microtiter plate. For each concentration one microtiter plate was prepared to avoid contamination and spill-over of larvae to another treatment. Two to nine independent replicates were performed. (b) Eclosion and survival experiments: application of either 5 × 10 2 , 5 × 10 3 or 5 × 10 4 /10 μl T. suzukii spores to each larva in one single microtiter plate. For each concentration one microtiter plate was prepared to avoid contamination. Four to eleven replicates were performed. For each independent replicate, a new spore suspension was prepared. (c) Fecundity tests: application of 1.5 × 10 4 /10 μl T. suzukii spores to each larva in a single microtiter plate. (d) Host range testing with D. melanogaster and D. willistoni: application of 5 × 10 4 /10 μl T. suzukii spores to each one L2 larva of one species in a single microtiter plate. Three to nine replicates were prepared.
Untreated controls were prepared in separate microtiter plates and contained the equivalent amount (10 μl) of sterile tap water overlaid on 440 μl pureed apple. All implemented microtiter plates of one replicate for infection treatment and controls were prepared identically using the same batch of L2 larvae and kept under same rearing conditions. Experimental design for LC 50 , eclosion and survival tests. Microtiter plates containing inoculated larvae were transferred into cylindrical cages (30 cm height, 25 cm diameter, closed with a nylon membrane) containing a water source, adult diet and oviposition medium (see 2.1) were changed twice a week. Mortality in LC 50 tests until 18 days post inoculation (dpi), eclosion until 19 dpi and survival of adults until 63 dpi were recorded daily until all SWD died. The LC 50 at 18 dpi was calculated using probit analysis. Mortality data were corrected for control mortality by Abbott formula 63 .
Experimental design for fecundity tests. For fecundity experiments, inoculated larvae pupated and eclosed and the adults were separated into male and female groups directly after eclosion for 3 days to avoid premature mating. One 3-day-old naive male and female adult were then placed together for mating and oviposition They were held in boxes (6 cm height, 10 cm diameter) containing a Petri dish (3 cm diameter) with oviposition medium prepared as described in 2.1. After 48 h, the Petri dish with oviposition medium was replaced and eggs were counted from the oviposition medium. The oviposition medium was then placed in a separate box and eggs were maintained until hatch. This procedure was repeated every 2 days until the last SWD pair inoculated with T. suzukii died. Number of hatched offspring was determined 18 days after oviposition and sex ratios were recorded. Treated flies were post-hoc inspected for established infections and assigned to different groups in the analyses: MF: i♀ × i♂ = female and male infected, F: i♀ × u♂ = female infected/male uninfected, M: u♀ × i♂ = female uninfected/male infected. Transmission was examined in two separate trials (18 pairs) with both infected parents. Oviposition medium was changed every 2 days as described above and the oviposition rate was recorded. Eggs were transferred to fresh medium for development. Eggs were not surface sterilized to avoid manipulation of the respiratory filaments resulting in higher mortality. Experiments were carried out in an incubator (RUMED 3501, RUBARTH, Laatzen, Germany) under rearing conditions: 22 ± 1 °C, 60% rH, 16:8 h light-dark photoperiod.

Inoculation of adults and design of survival and fecundity experiments. SWD adults at 3 days
post-eclosion (survival experiment) or one day post-eclosion (fecundity experiment) were placed together in groups of four (survival experiment) or separated into groups of male-only and female-only groups (fecundity experiment) into a plastic box where they were starved for 3 h followed by bulk feeding with 100 μl tap water droplets containing in total 3 × 10 5 T. suzukii spores mixed with blue food colour over night (about 18 h) (modified droplet feeding method from Hughes and Wood 62 ). Flies with a blue abdomen were selected for the experiments, whereby one male and one female were placed together in a cage (containing diet and water). For each independent replicate, a new spore suspension was prepared. Four replicates were prepared in total. In survival experiments, daily survival was recorded until all adults died. For fecundity experiments, egg deposition was recorded every 2 days and number of hatched offspring was determined 18 days post oviposition. Treated flies were post-hoc microscopically inspected to determine infection status.
Statistical analyses. Estimation of the median lethal concentration (LC 50 ) and slope of the concentrationmortality curve were calculated by probit analysis using TOXRAT software 64  www.nature.com/scientificreports/