Introduction

Symbiosis with microbes, ranging from parasitism to mutualism, is prevalent in both plants and animals (1, 2). The ubiquity of microbial symbionts is likely attributed to their profound impact on host biology, including host survival, development, immunity, reproduction, and even behavior (2, 3). The transmission mode of symbionts is a key factor in shaping the ecology and evolution of both symbionts and their hosts (4–6). In addition to vertical transmission, incongruence between symbionts and host phylogenies indicates a large number of horizontal symbiont transfers across species (1). These events are important, as they allow symbionts to expand their host range and enable hosts to acquire new symbionts and alter their fitness. However, the transmission of symbionts has been relatively little studied compared to their function (1).

Wolbachia (Rickettsiales, Alphaproteobacteria) is intracellular gram-negative bacteria and one of the most famous endosymbionts that infests ∼50% arthropod and several filarial nematode species (7–9). On the one hand, Wolbachia can induce a range of fascinating phenotypes, including a variety of reproductive manipulations, provision of nutrients, and alteration of host behavior, thus facilitating its spread among populations (7). On the other hand, although Wolbachia is primarily maternally transmitted, there are widespread and frequent horizontal transfers across hosts (10, 11). Together, these characteristics make Wolbachia the most infectious microbe on Earth in terms of the number of species it infects (7, 8).

Wolbachia has also received much attention for its applications in controlling pests and vector-borne diseases (8). Specific strains of Wolbachia are artificially transfected into target pests and subsequently released into the field to either suppress pest populations or replace populations to depress the spread of vector-borne diseases (12, 13). Understanding how Wolbachia spreads horizontally is critical in assessing its successful application and potential risks (14). This is because the released Wolbachia may leak into natural pest populations, frustrating population suppression strategies based on the cytoplasmic incompatibility (CI) of Wolbachia. It may also spread to nontarget organisms, potentially disrupting their population dynamics, reducing genetic diversity, and even leading to extinction. Therefore, the lack of a thorough understanding of Wolbachia transmission and its consequences could hinder its broader application (14).

Despite the extensive interest in the horizontal transmission of Wolbachia, our understanding of this subject remains incomplete (15). Similar to other symbionts, Wolbachia host shifts may occur through three main routes: parasitism, predation, and shared plant or other food sources (15). The relative contributions of these three routes to the transmission process remain unclear. Multiple surveys report a significant similarity in Wolbachia sequences between parasitoid wasps and their respective hosts, suggesting that parasitism may serve as a primary route for Wolbachia’s host shift (16–21). However, without quantitative tests, this observation could simply reflect a bias in research focus. For the intertrophic transmission of Wolbachia between parasitoid wasps and their hosts, experimental evidence has shown that parasitoid wasps can acquire hosts’ Wolbachia and vertically transmit them for several generations (22). However, it remains uncertain whether Wolbachia can be transferred from parasitoid wasps to their hosts. Some have argued that the transfer of Wolbachia between parasitoid wasps and their hosts is unidirectional, from host to wasp, as all parasitoid wasps emerge from their hosts, but parasitized hosts are eventually killed or castrated if not killed immediately (17, 23, 24).

In this study, our objective was to elucidate the role of parasitism and other potential routes in Wolbachia horizontal transmission and to investigate whether Wolbachia can be transferred from parasitoids to their hosts. We conducted a systematic survey of Wolbachia surface protein (wsp) sequences from the NCBI database and executed experiments using the whitefly Bemisia tabaci and its parasitoid wasp Encarsia formosa. Our study illuminates the crucial role of parasitism in Wolbachia intertrophic transmission, demonstrates the bidirectional nature of Wolbachia transfer between parasitoids and their hosts, and thus expands the current understanding of Wolbachia horizontal transmission.

Results

Wolbachia is frequently transmitted between parasitoid wasps and their hosts

To investigate potential horizontal transmission of Wolbachia, we retrieved 4685 wsp sequences from the NCBI database. Out of these 4685 sequences, 4253 could be assigned to 1377 species. We constructed a phylogenetic tree of wsp sequences (Fig. 1a) and extracted the minimum genetic distances of wsp between every species pair. Based on the relationships between species, we defined the species pairs into categories “Parasitism”, “Plant-sharing”, “Predation” and “Others” (for details, see Methods and Materials). Among these species pairs, 16.5% (53 out of 321) in “Parasitism” pairs had the minimum interspecific wsp distances less than 0.01 (i.e., > 99% identity). This proportion is significantly greater than the 6.3% (145 out of 2294) in “Plant-sharing” pairs, the 1.1% (13 out of 1146) in “Predation” pairs, and the 1.5% (14120 out of 943315) in “Others” pairs (χ² test; all comparisons: p < 1e^-5). Consistently, the minimum interspecific wsp distances in “Parasitism” relationships were significantly shorter than those in “Plant-sharing”, “Predation”, and “Other” relationships (Fig. 1b; Mann‒Whitney U test (MWUT); all comparisons: p < 1e-12). The minimum wsp distances in “Plant-sharing” were also significantly smaller than those in “Others” (Fig. 1b; MWUT; W = 730215564, p < 2.2e^-16). However, the minimum wsp distances showed no significant difference between “Predation” and “Others” (Fig. 1b; MWUT; W = 699285646, p = 0.096).

Fig. 1.

To test whether these effects were merely due to sampling bias, we further obtained divergent times from TimeTree for 95.2% (901,615 out of 947,376) of the species pairs. As expected, the minimum interspecific wsp distance increased as the divergent time increased (Fig. 1c; Spearman’s correlation; ρ = 0.14, p < 2.2e^-16). Considering the impact of divergent time, linear regression analyses were conducted with the wsp distance as the dependent variable and divergent time as the independent variable. We found that both “Parasitism” and “Plant-sharing” had significant effects (both tests: p < 2e^-16). The estimated effect of “Parasitism” (-0.28) was more profound than that of “Plant-sharing” (-0.13) (p = 3.9e^-16). Given that we cannot obtain divergence times for all species pairs, especially for those closely related, we also classified the species pairs according to their last common ancestor. We divided all species pairs into six categories based on whether the two species belonged to the same genus, family, order, class, phylum, or kingdom. Similar results were observed (Supplementary Note 1). To further rule out potential influences of sampling bias, subsampling analyses were conducted using three methods: 1) shuffling species relationships, 2) subsampling by controlling divergent time, and 3) subsampling by controlling the last common ancestor (for details, see Methods and Materials). For all three methods, both “Parasitism” and “Plant-sharing”, but not “Predation”, exhibited significantly shorter minimum interspecific distances of wsp than randomly generated controls (Fig. S1; both “Parasitism” and “Plant-sharing”: p < 0.001; “Predation”: p > 0.05). These results confirmed that parasitism and plant-sharing promote interspecific transfers of Wolbachia, rather than being due to sampling bias.

An example of frequent Wolbachia transfers between parasitoids and their hosts is observed in Trichogramma wasps and their lepidopteran hosts (Fig. 1d). Trichogramma is a genus of generalist egg parasitoids, targeting mainly lepidopterans such as moths and butterflies (25). Fig. 1d displays representative wsp sequences from Trichogramma and lepidopterans, illustrating potential interspecific transitions of Wolbachia. In the analyzed dataset, 16 out of 23 (69.6%) Trichogramma species exhibited the minimum interspecific wsp distance of less than 0.01 with at least one lepidopteran species. Similarly, 79 out of 254 (31.1%) surveyed lepidopteran species displayed a minimum interspecific wsp distance of less than 0.01 with at least one Trichogramma species. These results suggest frequent Wolbachia transitions between Trichogramma wasps and lepidopterans.

Collectively, these findings support that Wolbachia are frequently transmitted between parasitoid wasps and their hosts.

Wsp phylogeny suggests transfer directions of Wolbachia between the whitefly B. tabaci and its parasitoid wasps

However, the interspecific identities of wsp between parasitoids and their hosts typically offer no clues about the direction of Wolbachia transmission. We noticed frequent Wolbachia transfers between the whitefly B. tabaci and its parasitoid wasps (Fig. 2a). Notably, one of the juvenile parasitoids of B. tabaci is En. formosa. Wolbachia induces parthenogenesis and has evolved into an obligate symbiont in En. formosa, exhibiting a 100% infection rate (26, 27). This system presents a unique opportunity to infer the directions of Wolbachia transmission between parasitoids and their hosts.

Fig. 2.

One clade of the wsp phylogeny contained 99 sequences from B. tabaci, 14 sequences from Encarsia and Eretmocerus parasitoid wasps of B. tabaci, and 8 other sequences (Fig. 2a). The prevalence of B. tabaci sequences interspersed with those of the parasitoids suggests that the transmission direction of Wolbachia in this clade may be primarily from B. tabaci to its parasitoid wasps. Notably, in another clade of wsp, nine En. formosa and five B. tabaci wsp sequences are clustered together, along with one wsp sequence detected in cotton and 12 sequences from other parasitoid wasps of B. tabaci, such as En. inaron, En. lutea, En. bimaculata, Er. mundus, and Aimtus hesperidum (Fig. 2a). Given that Wolbachia is obligate with a 100% infection rate in En. formosa, it is reasonable to infer that the transmission direction of Wolbachia in this clade was from En. formosa to B. tabaci.

Wolbachia can be transmitted from the parasitoid wasp En. formosa to its whitefly host in cage experiments

Bemisia tabaci is a species complex of at least 40 cryptic species (28). Infection rates of Wolbachia reported across multiple populations within these cryptic species exhibit dramatic fluctuations, ranging from 0% to 100% (28–31). To investigate whether Wolbachia can be transmitted from parasitoid wasps to their whitefly hosts, we first established a Wolbachia-free iso-female line of B. tabaci Q biotype. Subsequently, we conducted outdoor cage experiments using the Wolbachia-free B. tabaci and its parasitoid En. formosa, which was 100% infected with Wolbachia (Fig. 2b). After introducing En. formosa, the Wolbachia infection rate in whiteflies increased from zero to 4.17% after 20 days and reached 10.83% after 100 days (Fig. 2c; ANOVA; F_5,17=11.44, p < 0.001). Correspondingly, the female ratio of whiteflies decreased from 69.17% to 60.08% (Fig. 2d; ANOVA; F_5,17=3.14, p = 0.048). In contrast, the whitefly population with no exposure to wasps maintained a zero infection rate of Wolbachia (Fig. 2c), and the female ratio showed no significant changes (Fig. 2d; ANOVA; F_5,17 = 0.076, p = 0.99). Additionally, the Sanger sequencing results showed that wsp sequences from both whitefly and En. formosa were identical (Fig. S2). These results indicate that Wolbachia from En. formosa can be rapidly transmitted to its whitefly host.

Parasitism failure transmits Wolbachia from the parasitoid wasp En. formosa to its whitefly host

We hypothesized that Wolbachia was transmitted from parasitoid wasps to their hosts through unsuccessful parasitism. To test this hypothesis, we performed parasitism experiments on wasp individuals and applied irradiation treatment to wasps to reduce their parasitism success rate. After 60 Gy irradiation, the fecundity of En. formosa in 12 h significantly decreased (Fig. 3a; Student’s t test; t = 12.91, p < 0.001), and the parasitism success rate on whiteflies drastically declined from 78.8% to 9.9% (Fig. 3b; Student’s t test; t = 13.09, p < 0.001). Correspondingly, the Wolbachia infection rate increased from 7.5% to 78.3% in whiteflies survived from parasitism (Fig. 3c; Student’s t test; t = 6.61, p < 0.001), despite a decrease in the Wolbachia titer in En. formosa post irradiation (Fig. S3). More details can be found for irradiation treatments at different dosages (Supplementary Note 2). Fluorescence in situ hybridization (FISH) assays further showed that Wolbachia was injected into the host nymphs along with En. formosa eggs (Fig. 3 d-f and S4). These results indicate that parasitism failure can transfer Wolbachia from the parasitoid wasps En. formosa to their whitefly hosts.

Fig. 3.

Vertical transmission and fitness cost of Wolbachia in whiteflies after horizontal transfer from En. formosa

Next, we investigated the vertical transmission and fitness cost of Wolbachia in its new host, B. tabaci, following its horizontal transfer from En. formosa. The vertical transmission rate varied from 22.2% to 33.3% across generations in whiteflies, with a slight but statistically insignificant increase from G₁ to G₅ (Fig. 4b; ANOVA; F_4,40 = 2.09, p = 0.10). Moreover, Wolbachia was also detected in the G₃ whiteflies, specifically in the nymph’s bacteriocytes, the abdomen of male adults, and the ovaries of female adults (Fig. 4 and S5). MLST typing confirmed that the Wolbachia strain introduced into the whiteflies was identical to the strain from En. formosa (Fig. S6).

Fig. 4.

We then examined the impact of the introduced Wolbachia on the fitness of whiteflies. Compared to uninfected females, Wolbachia-infected females showed decreased fecundity (Fig. 5a; Student’s t test; t = 8.51, p < 0.001). Moreover, offspring from infected mothers exhibited a diminished egg hatching rate (Fig. 5b; Student’s t test; t = 8.33, p < 0.001), a lower survival rate among nymphs (Fig. 5c; Student’s t test; t = 13.54, p < 0.001), and a decreased ratio of females in adults (Fig. 5d; Student’s t test; t = 12.29, p < 0.001), although there was no significant difference in the development time from egg to adult (Fig. 5e; Student’s t test; t = 1.51, p < 0.001). We also investigated the effects of paternal infection with Wolbachia. Regardless of whether the female was infected with Wolbachia, the infection status in males showed no significant influence on their fecundity, the hatching rate of their offspring’s eggs, the survival rate of their nymph offspring, or the female ratio of their adult offspring (Table 1; Student’s t test; all comparisons: p > 0.05).

Fig. 5.

Table 1.

Collectively, these results indicate that Wolbachia from En. formosa, when shifted into the new host B. tabaci, exhibits a low rate of vertical transmission and a substantial fitness cost, without apparent reproductive manipulation phenotypes.

Discussion

The effect of parasitism on Wolbachia horizontal transmission

As with previous studies, we utilized the sequence similarities of wsp to infer potential horizontal transfers of Wolbachia. Here, we systematically investigated all wsp sequences in the NCBI database, enabling us to examine the effect of potential factors such as parasitism, plant-sharing, and predation. Our findings clearly indicate that parasitism exhibits shorter interspecific wsp distances than plant-sharing and predation.

Moreover, the fact that herbivores sharing plants have identical wsp sequences does not necessarily imply plant-mediated horizontal transfer of Wolbachia. This is because species that share the same plants often have recent divergent times (Fig. 1b), which increases the potential for hybridization and also sharing common parasitoid wasps. Therefore, the plant-sharing category overestimates the extent of plant-mediated Wolbachia horizontal transmission, further supporting the notion that parasitism is the primary route of Wolbachia horizontal transmission.

Directions of Wolbachia transmission between parasitoids and their hosts

However, investigations based on Wolbachia sequence similarity have significant limitations. First, determining the direction of transfer is challenging based solely on the identity of Wolbachia strains. In certain exceptional cases, inferences of transfer direction might be drawn from the prevalence of Wolbachia infection in the two species (15). For example, Wolbachia is obligate in En. formosa and exhibits a 100% infection rate across all populations (27). This unique instance of horizontal Wolbachia transfer between En. formosa and its whitefly hosts provides compelling evidence, suggesting that Wolbachia is likely transmitted from the parasitoid wasp to its host, rather than the reverse. Second, the PCR detection of wsp does not necessarily indicate horizontal transfer of Wolbachia (24). Instead, it could merely represent contamination that arises during predation or parasitism. The contamination can be Wolbachia-infected tissues or even just fragmented Wolbachia DNA, which could be found on the surface, within the gut, or inside the body cavity (as in the case of parasitized hosts).

To verify Wolbachia transmission from parasitoid wasps to their hosts, we conducted outdoor cage experiments and indoor tests using Wolbachia-free whitefly B. tabaci and its parasitoid En. formosa. Wolbachia was detected by nested PCR in whitefly adults (G₀) that survived parasitism. Encarsia formosa parasitizes B. tabaci nymphs, which undergo a pseudo-pupal stage to reach adulthood (26, 32). In our experiments, PCR detection of Wolbachia typically occurred 3∼7 days post-parasitism, minimizing risks of contamination from parasitism. Wolbachia was also detected by PCR in subsequent generations (G1–G5) of whiteflies and induced notable fitness costs. PCR sequencing confirmed that the Wolbachia strain in the B. tabaci matched that of En. formosa. Furthermore, FISH assays revealed a tissue-specific distribution of Wolbachia in both nymphs and adults of the whiteflies, matching previously reported patterns (33). Collectively, these findings provide compelling evidence that Wolbachia from En. formosa can be horizontally transmitted to B. tabaci, beyond mere DNA contamination.

Although previous research has demonstrated that parasitoids can acquire Wolbachia from their hosts, the reverse direction of transmission, from host to parasitoid, has been largely overlooked and lacks supportive experimental evidence. One possible reason for this oversight is that all parasitoids emerge from their hosts, but hosts are eventually killed by the parasitism of parasitoid wasps (17, 23, 24). However, parasitoid wasps’ success in parasitizing their hosts does not always reach 100% (34). Some hosts can manage to survive after parasitism. Various factors can influence the outcome of parasitoid-host interactions, including environmental conditions, the species and genotype of both wasps and hosts, the host’s age, and the presence of symbiotic bacteria within the host (35–37). Moreover, we used radiation to reduce the parasitism success rate, which notably enhanced the transfer of Wolbachia from En. formosa to its whitefly host.

Potential intertrophic transmission network of Wolbachia

A previous study reported that parasitoid wasps can act as vectors to transmit Wolbachia, without the necessity of being infected themselves (38). Through the probing actions of the Eretmocerus parasitoid, Wolbachia can be transmitted among whitefly hosts (38). This is often referred to as the ’dirty needle’ model. Conversely, hosts can also serve as vectors for Wolbachia transmission among parasitoid wasps. Wolbachia can be transferred from infected to noninfected Trichogramma wasps through superparasitism (39, 40). However, the Wolbachia transmission of these two modes is restricted within the same trophic level. In contrast, the transmission of Wolbachia between parasitoid wasps and hosts can cross trophic levels. Our findings, when combined with existing knowledge, suggest that the intertrophic transmission of Wolbachia is bidirectional between parasitoid wasps and their hosts. This greatly enhances our understanding of the horizontal transmission of Wolbachia.

Interestingly, we found on NCBI that a strain of Wolbachia detected in the cotton plant was identical to the Wolbachia from En. formosa, based on the wsp sequence and MSLT typing (Fig. 2 and S6). This Wolbachia strain was probably transmitted to the cotton plant from the feeding of whiteflies (41). Given that En. formosa is 100% infected with its obligate Wolbachia strain, a possible transmission route could be from En. formosa to whiteflies via parasitism and then to the cotton plant through the feeding of whiteflies. This finding indirectly supports our conclusion that Wolbachia can be transmitted from parasitoid wasps to their hosts. This suggests that once the Wolbachia of parasitoids is transmitted to herbivorous hosts, it may further spread to host plants. The reverse transmission route can also be possible. It is likely that Wolbachia’s widespread and complex horizontal transmission network is established through such bidirectional transmissions across multiple trophic levels, e.g., "plant-herbivore-parasitoid-hyperparasitoid". Further investigations are needed to test these hypotheses.

Wolbachia establishment after host transfers

Moreover, the physical transfer of Wolbachia often represents merely the first step of its establishment in a new host (15). Several subsequent steps are required for Wolbachia to establish itself within new species, e.g., entry into germ cells, vertical transmission, and mechanisms that promote its spread within the population (15). First, we found that Wolbachia from En. formosa was enriched in the ovaries of whiteflies and vertically transmitted after entering B. tabaci. However, the vertical transmission efficiency is low, ranging from 22.2% to 33.3%. We also noted an absence of reproductive manipulations by the newly introduced Wolbachia in whiteflies. The reduced female ratio after infection does not support the induction of parthenogenesis, feminization or male-killing by Wolbachia. We neither observed cytoplasmic incompatibility, where the mating of infected males and uninfected females resulted in reduced offspring hatching.

In contrast, the introduced Wolbachia from En. formosa reduced whiteflies’ egg laying and hatching, larval survival, and female proportion, demonstrating significant fitness costs in the new host. This is likely due to Wolbachia’s coevolution with its host in En. formosa, which may have led to the loss of its ability to colonize new host species. Given the low vertical transmission rate, high fitness cost, and lack of clear reproductive manipulations, it is reasonable to predict that the spread of Wolbachia in its new host population will be limited. Finally, these factors, together with the frequency of Wolbachia introductions by parasitoids and its spread via parasitoid or plant vectors, shape the dynamics and equilibrium of Wolbachia. These dynamics could shift with the emergence of reproductive manipulation or other beneficial phenotypes that promote Wolbachia spread, probably through gene mutation, recombination or horizontal gene transfer within Wolbachia (2). There are still many questions waiting to be further studied in these steps of Wolbachia host shifts.

Conclusions

By investigating wsp sequences from the NCBI database, we found frequent intertrophic transmission of Wolbachia by parasitism but not predation. Combining bioinformatics and experimental approaches, we demonstrated that Wolbachia can be transmitted from the parasitoid wasp En. formosa to the host B. tabaci. To our knowledge, this is the first compelling evidence that Wolbachia can be transmitted from parasitoid wasps to their hosts, thus revealing the bidirectional nature of Wolbachia transfers between parasitoids and their hosts. These findings enrich our knowledge of the Wolbachia transmission network and have significant implications for understanding the ecology of Wolbachia, as well as for evaluating the release of Wolbachia in pest control.

Materials and Methods

Wsp sequence retrieval and phylogenetic analyses

tblastn was conducted against the NCBI “nr/nt” database using the wsp protein AAS14719.1 from Wolbachia of Drosophila melanogaster (July 2023). Default settings were used with a maximum target sequence of 5000 and the organism limitation of Wolbachia (taxid:953). The sequences were filtered to remove those shorter than 300 bp or containing premature stop codons. Wsp sequences were translated into proteins, aligned using MAFFT v7.475 (42), and reverse translated into codons using PAL2NAL v14 (43). The phylogenetic trees were constructed using IQ-Tree v2.2.0 (44), with the best model selected by the built-in ModelFinder (45). The phylogenetic trees were visualized using iTOL v6 (46).

Statistics for genetic distances of wsp

Genetic distances of wsp sequences were extracted from the wsp phylogeny using a custom Python script. Species interaction relationships were extracted from the GloBI database (August 2023) (47). Parasitic associations were extracted using interaction types of “parasiteOf” or “parasitoidOf”, excluding social parasitism to focus on direct biological parasitism. Predation associations were extracted using interaction types of “preysOn” or “eats”. Given the relatively broad dietary range of predators, a genus-to-genus expansion was adopted for the predation relationships. Herbivorous interactions were extracted using the interaction types “hasHost” or “eats”, specifically targeting taxa within the kingdom Plantae. Extracted relationships were manually curated to verify the accuracy. Divergent times between species were extracted from TimeTree v5 via its application programming interface (API) (September 2023) (48). To test the effects of special features of sampled species pairs, three subsampling methods were employed using a custom Python script. For each method, 1000 replicates were randomly generated. For species pair shuffling, in every replication, the pairs of species were randomly rearranged to create new combinations. To control divergent time, pairs of species were randomly sampled from the background to match the divergent times in the tested category (i.e., parasitism, plant-sharing or predation). The sampling background pools were all species pairs excluding the specific tested category. A similar process was applied to control the last common ancestor. Linear model analyses, statistical tests, and data visualization were executed using R v4.3.1.

Insect rearing

Both whiteflies B. tabaci and the parasitoid wasps En. formosa were maintained in nylon cages at 25±1°C, 70±5% RH, and a L:D photoperiod of 16:8 h. Bemisia tabaci was originally collected from the campus of Nanjing Agricultural University in 2012 and determined to be the Q biotype. A Wolbachia-free iso-female line of B. tabaci was then established. En. formosa was initially acquired from Beijing Ecoman Biotechnology Co., Ltd. in 2013 and was maintained on B. tabaci on tomato plants.

PCR and Sanger sequencing

Bemisia tabaci or En. formosa were initially washed using ethanol and ultrapure water three times each. DNA was extracted from individuals using the STE method (49). To avoid false-negative results, a nest-PCR targeting the wsp gene was employed for detecting Wolbachia in whiteflies (Fig. S7) as previously described (49). PCRs were performed using DreamTaq Green PCR Master Mix (Thermo Scientific) according to the manufacturer’s protocol. To confirm whether the Wolbachia present in whiteflies and their parasitoid wasps belonged to the same strain, PCR and Sanger sequencing were performed on Wolbachia genes, namely, wsp, gatB, coxA, hcpA, ftsZ, and fbpA, for MSLT typing (50). The primers used for wsp nest-PCR and MSLT typing are listed in Table S1.

FISH detection of Wolbachia

The location of Wolbachia in whiteflies was determined using fluorescence in situ hybridization as previously reported (27). To enhance the detection signal, two 5’ rhodamine-labeled probes, W1: 5′-AATCCGGCCGARCCG ACCC-3′ and W2: 5′-CTTCTGTGAGTACCGTCATTATC-3′, were used to target Wolbachia 16S rRNA (51). Stained samples were examined under a Zeiss LSM 700 confocal microscope (Carl Zeiss, Germany). The specificity of the method was confirmed using Wolbachia-free whiteflies as negative controls.

Outdoor cage experiments for Wolbachia transmission

To assess the potential transmission of Wolbachia from En. formosa to B. tabaci, six cages (80 cm x 80 cm x 80 cm, enclosed with a 120 mesh nylon screen) were established on the Nanjing Agricultural University campus. Initially, six 20-cm tall tomato plants were placed in each cage, with two additional plants introduced at 30-day intervals. Each cage was populated with 100 female and 100 male Wolbachia-free whiteflies. After a 20-day period, ten En. formosa wasps were introduced into each of three randomly selected cages, following the random sampling of 40 adult whiteflies from each cage. Subsequently, an additional 40 adult whiteflies were sampled from each cage at 20-day intervals. These sampled whiteflies were sexed based on their morphologies and screened for Wolbachia infection using wsp nest PCR.

Wolbachia transmission from irradiated En. formosa to B. tabaci

To explore the possibility of Wolbachia transmission from En. formosa to B. tabaci due to unsuccessful parasitism, we subjected En. formosa wasps to radiation at the Nanjing Aerospace Irradiation Center. Wasps that had emerged within 24 h were exposed to Co⁶⁰ radiation at doses of 60, 80 and 100 Gy (dose rate of 2 Gy/min). A single irradiated wasp was subsequently introduced into a Petri dish, which contained a tomato leaf infested with Wolbachia-free 3^rd or 4^th instar whitefly nymphs. The parasitic behaviors of En. formosa were monitored under a Nikon SMZ800 stereomicroscope (Nikon Instrument Inc., Tokyo, Japan). Wasps were removed after 12 h of parasitism. Only whitefly nymphs that had been punctured by wasps were kept until the emergence of either parasitoid wasps or whiteflies. The newly emerged whiteflies were then collected for Wolbachia detection using wsp nested PCR. In each replication, five irradiated En. formosa were randomly selected for parasitization, with four replications performed in total. Nonirradiated wasps were used as controls.

Wolbachia transmission across generations in whiteflies

The parasitized and subsequently emerged whiteflies were denoted as the initial generation (G₀). To investigate the transmission rate of Wolbachia across whitefly generations, a pair of female and male newly emerged whitefly adults (G₀) were randomly chosen and introduced into a Petri dish containing a tomato leaf. After 5 d of oviposition, the whitefly adults (G₀) were removed and sampled for Wolbachia detection using wsp nested PCR. Only the eggs laid by Wolbachia-infected females were allowed to develop into adults. In a similar manner, upon the emergence of G₁ adults, whiteflies were paired and placed into individual Petri dishes to produce the G₂ generation and then sampled for Wolbachia detection. This procedure was repeated from the G₀ to G₅ generations. A parallel control experiment was conducted using Wolbachia-free whiteflies. Nine replications were conducted for each generation.

Effects of Wolbachia on whitefly fitness

Given the inability to obtain a whitefly strain with 100% Wolbachia infection, we selected individuals from the infected whitefly population, which initially acquired Wolbachia through parasitism of irradiated En. formosa and subsequently maintained for over five generations. Whitefly nymphs were individually isolated into PCR tubes at their 4^th instar stage. After emergence, female and male whitefly adults were paired and allowed to oviposit for five days within a Petri dish containing a tomato leaf. These adults were then removed for Wolbachia detection via PCR. For the Wolbachia-infected female adults, we recorded several fitness parameters. These included the total number of eggs laid, the number of first instar nymphs, the developmental time from egg to eclosion, and the number of eclosed male and female offspring. This procedure was replicated with 60 pairs, and whiteflies from the original Wolbachia-free population were used as controls. Additionally, we examined the effects of various paternal and maternal infection status combinations on whitefly fitness. A similar procedure was employed, with the exception that all whitefly adults were sourced from the infected population, and their infection status was determined through wsp nested PCR.

Statistics of the experimental data

Statistical analysis of experimental data was conducted using SPSS version 18.0 (SPSS Inc., Chicago, USA). The proportions were arcsine-square root transformed. Two independent samples were compared using Student’s T test. One-way ANOVA was performed for analysis of multiple samples. Means were then compared using Honest Student’s Tukey test. Data were visualized using GraphPad Prism 6 software (GraphPad Software, San Diego, USA).

Data and code availability

The alignment files, phylogenetic trees, and custom scripts can be accessed on FigShare (https://doi.org/10.6084/m9.figshare.24718119).

Acknowledgements

This research was funded by “Shuangchuang Doctor” Foundation of Jiangsu Province (202030472), Nanjing Agricultural University startup fund (804018), the Hainan Major Science and Technology Project (ZDKJ2021007), and the Special Fund for Agro-scientific Research in the Public Interest of China (201303019). Bioinformatic analyses were supported by the high-performance computing platform of Bioinformatics Center, Nanjing Agricultural University.

Author contributions

Y.X.L., X.Y.H., and Z.C.Y. conceptualized and designed the research; Z.C.Y. carried out the bioinformatics analysis; L.D.Q., H.L.J., and X.X.W. conducted the experiments; Y.X.L. and Z.C.Y. interpreted the results; Z.C.Y. and L.D.Q. prepared the visualization; Z.C.Y. wrote the initial manuscript; Y.X.L. and X.Y.H revised the manuscript.

Competing interests

The authors declare no competing interests.

Supplementary Information

Supplementary Note 1.

We divided all species pairs into six categories based on whether the two species belonged to the same genus, family, order, class, phylum, or kingdom. These categories represent different divergent levels between species pairs. Considering the impact of the divergent level, linear regression analyses were conducted with the wsp distance as the dependent variable and the category of divergent levels as the independent variable. We found that both “Parasitism” and “Plant-sharing” had significant effects (both tests: p < 2e^-16 11 ), while “Predation” showed no significant effect (p = 0.58). The estimated effect of “Parasitism” (-0.28) was more profound than that of “Plant-sharing” (-0.14) (p = 1.32e^-15).

Supplementary Note 2.

Encarsia formosa wasps that had emerged within 24 h were exposed to Co⁶⁰ radiation at doses of 60, 80 and 100 Gy. On day 15 after irradiation treatments, the survival rate of adult wasps was 73.33 ± 3.33% in the nonirradiated group (Fig. S8). In contrast, the survival rates were 45.56 ± 1.92%, 7.78 ± 6.94%, and 3.33 ± 3.33% after exposure to 60 Gy, 80 Gy, and 100 Gy radiation, respectively. The survival rate of adult wasps decreased with increasing irradiation doses (Fig. S8; ANOVA; F3,8 = 34.18, p < 0.001). After irradiation, the ovaries were dissected, and changes in egg dimensions were recorded (Fig. S9). Following 80 and 100 Gy irradiation, almost no eggs were detected starting from day 4, while after 60 Gy irradiation, eggs were nearly absent starting from day 9. Fig. S10 shows the ovaries on days 1, 4, and 9 post-exposure to various irradiation doses. In the nonirradiated En. formosa, there were no obvious changes in the morphology of the ovarioles or oocytes from days 1 to 9 (Fig. S10). However, 4 days post-irradiation, the ovarioles of wasps exposed to 60 Gy appeared thinner, and some oocytes had dissolved (Fig. S10). In addition, the ovarioles of wasps exposed to 80 Gy and 100 Gy had become transparent, with the majority of the oocytes disappearing and becoming invisible (Fig. S10). Parasitism experiments demonstrated that the nonirradiated En. formosa laid 7.3 ± 1.1 eggs in 12 h. This number significantly dropped to 3.2 ± 0.9 eggs for wasps exposed to 60 Gy radiation (Fig. 3a; Student's t test; t = 12.91, p < 0.001). The wasps exposed to 80 Gy and 100 Gy radiation almost completely ceased egg laying, resulting in insufficient data collection.

Table S1.

Fig. S1.

Fig. S2.

Fig. S3.

Fig. S4.

Fig. S5.

Fig. S6.

Fig. S7.

Fig. S8.

Fig. S9.

Fig. S10.