Many insects, including mosquitoes, planthoppers, aphids and leafhoppers, are the hosts of bacterial symbionts and the vectors for transmitting viral pathogens1,
Most ancient obligate bacterial symbionts are transovarially transmitted by female insects through eggs3. Koga et al. show that the facultative bacterial symbiont Serratia can hijack the transovarial transmission route that is established for the more ancient obligate symbiont Buchnera in the aphid body9. Transovarial transmission also plays an important epidemiological role in the persistence of viral pathogens in insect vectors10,11. To be transovarially transmitted, both symbiotic bacteria and viral pathogens must pass through the follicular cells into the oocytes of the female insect ovary, but whether viral pathogens can hijack the oocyte entry route for ancient obligate bacterial symbionts is unknown3,10,11. The rice dwarf virus (RDV) is the first plant virus recorded to be transmitted transovarially by insect vectors12. The green rice leafhopper Nephotettix cincticeps (Hemiptera: Cicadellidae) (Supplementary Fig. 1a), the main vector of RDV (ref. 8), is associated with two types of obligate bacteria symbiont—Sulcia and Nasuia—which follow an ancient path to enter the eggs laid by female leafhoppers7,13. Sulcia and Nasuia correspond to the H-symbiont and L-symbiont, respectively, as noted by Nasu in 1965 (refs 7, 13). Using electron microscopy, Nasu observed that RDV particles are distributed on Nasuia envelopes but not on Sulcia envelopes in the N. cincticeps oocyte13, and thus we hypothesize that RDV may exploit the oocyte entry path of bacteria symbionts for transovarial transmission.
To confirm this hypothesis, we used electron microscopy to trace the oocyte entry paths of RDV, Sulcia and Nasuia in the ovaries of N. cincticeps females. In the female adult, Sulcia and Nasuia started to enter the follicular epithelium surrounding the posterior pole of the oocyte, namely the epithelial plug (Supplementary Fig. 1b). Next, Sulcia and Nasuia entered the oocytes from the epithelial plug, where they accumulated in the deep depression of the oocytes and formed a characteristic ‘symbiont ball’ (Supplementary Fig. 1c). RDV particles were directly associated with Sulcia envelopes, wherever they were located close to the epithelial plug (Fig. 1a), within the epithelial plug (Fig. 1b) or in the oocyte (Fig. 1c). We observed that RDV particles first attached to Sulcia envelopes (Fig. 1d). Viral particles were then seen within the invaginations of the Sulcia envelopes, which finally formed envelope-enclosed vesicles (Fig. 1d). Occasionally, RDV particles were also found close to or on Nasuia envelopes (Supplementary Fig. 2). Based on these observations, we deduced that RDV may have evolved to form a specific relationship with Sulcia rather than with Nasuia. Thus, Sulcia appears to mediate the invasion of RDV into the oocytes of N. cincticeps.
To observe the oocyte invasion process of RDV and Sulcia, we labelled the ovaries in viruliferous or non-viruliferous N. cincticeps females with virus-specific IgG conjugated to fluorescein isothiocyanate (FITC) using immunofluorescence and specific oligonucleotide probes targeting the 16S rRNA of Sulcia using fluorescence in situ hybridization (FISH)7,14,15. In young viruliferous females, before the invasion of Sulcia into the ovary, RDV was found only in the terminal filament and pedicel (Supplementary Fig. 1d). In adult viruliferous females, RDV accompanied Sulcia into the epithelial plug (Fig. 1e), and then into the oocyte from the epithelial plug (Fig. 1f). Finally, RDV gradually spread throughout the oocytes (Fig. 1g). Thus, RDV particles accompanied Sulcia from the haemolymph, to the epithelial plug and finally into the oocytes of N. cincticeps (Fig. 1h).
Our microscopic observations suggest that RDV particles and Sulcia envelopes may have formed a specific interaction relationship. The minor outer capsid protein P2 of RDV particles, which protrudes from the surface of the outer shell of virions, is responsible for initial contact with vector receptors16 (Fig. 2a). We thus deduced that PDV P2 may directly interact with the outer membrane protein (OMP) of Sulcia. To obtain the sequence of the Sulcia OMP gene, we first determined the complete genome sequence of Sulcia from N. cincticeps (GenBank accession no. CP016223). The amino-acid sequence of the Sulcia OMP gene (GenBank accession no. KY012242) was 1,008 amino acids (aa) long and contained a 280 aa bacterial surface antigen (BSA) domain (Fig. 2a), which is responsible for contact with ligands17. Structural analysis has shown that RDV P2 is an L-shaped, flexible structure with one 10-nm-long domain anchored to the viral surface and a 15-nm-long domain for binding cellular receptors16. A yeast two-hybrid assay revealed that the 15 nm domain of RDV P2 interacted directly only with the fragments containing the BSA domain of Sulcia OMP (Fig. 2b). A glutathione S-transferase (GST) pull-down assay confirmed that the GST-fused 15 nm domain of RDV P2 specifically bound to the His-fused BSA domain of the Sulcia OMP (Fig. 2c), indicating direct recognition and contact of the Sulcia OMP by RDV particles.
As an alternative approach, we incubated purified viral particles of RDV or rice gall dwarf virus (RGDV) with live ovaries dissected from non-viruliferous N. cincticeps. RGDV, a plant reovirus, is also transovarially transmitted by N. cincticeps, but the mechanism is unknown8. This in vitro experiment showed that purified particles of RDV, but not of RGDV, can bind to Sulcia in live ovaries of N. cincticeps (Fig. 2d), suggesting that RGDV did not accompany Sulcia into the insect oocyte. Pretreatment with P2-specific antibodies prevented this binding in vitro (Fig. 2d). We also incubated the His-fused 15 nm or 10 nm domains of RDV P2 purified from Escherichia coli with live ovaries dissected from non-viruliferous N. cincticeps. We found that the 15 nm domain, but not the 10 nm domain of RDV P2, can bind to Sulcia (Fig. 2d). Thus, the direct attachment of RDV particles with the envelopes of Sulcia was mediated by specific recognition between the 15 nm domain of RDV P2 and the BSA domain of Sulcia OMP (Fig. 2e).
To determine the role of the direct interaction of RDV P2 with Sulcia OMP in transovarial transmission of RDV, purified RDV particles were microinjected together with OMP-specific antibodies (Supplementary Fig. 3) or pre-immune serum into adult N. cincticeps females. We found that treatment with OMP-specific antibodies efficiently prevented ovary invasion by RDV, but did not affect oocyte entry by Sulcia in the ovaries of female N. cincticeps (Fig. 2f and Table 1). In nature, 100% of the eggs laid by viruliferous female N. cincticeps contained Sulcia, but only about 50% contained RDV. In the eggs laid by viruliferous females that had been microinjected with OMP-specific antibodies, only about 17.6% contained RDV, even though 100% contained Sulcia (Table 1), suggesting that blocking the adsorption of viral particles to Sulcia envelopes by OMP-specific antibodies strongly inhibited the transovarial transmission of RDV. Thus, a direct interaction of RDV P2 and Sulcia OMP appears to mediate viral transmission from females to offspring.
On using an oral administration of rifampicin at an appropriate dosage to nymphal N. cincticeps via rice seedlings7, we found that the titres of Sulcia and Nasuia in the eggs laid by female N. cincticeps were decreased by ∼60% and 70%, respectively (Fig. 3a,b). Rifampicin treatment of N. cincticeps resulted in a retarded nymph stage and nymphal hatchability (Fig. 3c,d), but nymphal survival, female fecundity and lifespan did not change significantly (Fig. 3e–h). Thus, reduction in the titres of bacteria symbionts by an appropriate antibiotic treatment has a limited deleterious effect on the development of N. cincticeps. Notably, treatment with antibiotics also strongly inhibited transovarial transmission of RDV (Fig. 3i,j), further confirming a role of bacteria symbionts in carrying RDV into the oocytes of N. cincticeps.
In addition to N. cincticeps, RDV is also transmitted by the rice leafhoppers Nephotettix nigropictus and Recilia dorsalis8,18. As Sulcia is a common obligate symbiotic bacterium that persists in most leafhopper species19, we deduced that RDV may also exploit the ancient oocyte entry routes established by Sulcia in N. nigropictus and R. dorsalis. Immunofluorescence and FISH revealed that RDV accompanied Sulcia to invade the epithelial plug or oocyte of ovaries in female N. nigropictus and R. dorsalis (Supplementary Fig. 4a). Electron microscopy confirmed the enclosure of RDV particles in the vesicles formed by the envelopes of Sulcia of N. nigropictus and R. dorsalis (Supplementary Fig. 4b). Consistently, the amino-acid sequences of the BSA domains of Sulcia OMPs among the rice leafhoppers were the same, but they differed in 30 amino acids from that of the non-vector maize leafhopper Dalbulus maidis20 (Supplementary Fig. 4c). A yeast two-hybrid assay confirmed that the 15-nm-long domain of RDV P2 did not interact with the BSA domain of Sulcia OMP from the non-vector leafhopper (Supplementary Fig. 4d). Thus, exploiting Sulcia for the transovarial transmission of RDV is a conserved mechanism in rice leafhopper vectors.
The ability to be efficiently transmitted from females to their offspring is the key feature shaping associations between insects and their inherited symbionts or transmitted viral pathogens, but, so far, little has been known about the mechanisms involved. We have shown that rice stripe virus, a tenuivirus, uses the vitellogenin as a transfer vehicle to enter the developing oocytes of its planthopper vector10. A similar study has shown that an endogenous retrovirus of Drosophila melanogaster is transmitted from follicular cells to the oocyte by a mechanism that is linked to endocytic yolk uptake21. Here, we have discovered a new mode of transovarial transmission of viruses where RDV exploits the ancient oocyte entry path established for Sulcia in rice leafhoppers through a direct virus–symbiont bacterium interaction, allowing the virus to easily overcome vertical transmission barriers. Thus, viral pathogens have evolved to use the existing oocyte entry paths for their own purpose. Given the widespread coexistence of viral pathogens and symbiotic bacteria in nature, and especially in invertebrates22,23, our study demonstrates the need to investigate the interactions between viral pathogens and heritable symbiotic bacteria in more organism groups, as well as their role in the transmission of pathogens.
Rice leafhoppers are associated with two types of obligate bacteria symbiont, Sulcia and Nasuia7. Fifty years ago, Nasu13 shows that RDV particles are localized on Nasuia envelopes but not on Sulcia envelopes in the oocyte of N. cincticeps. Here, we show that the direct interaction between RDV P2 and the Sulcia OMP can induce invagination and the formation of membrane-enclosed vesicles. Similarly, the interaction between RDV P2 and receptors on the plasma membrane of N. cincticeps cultured cells also induces the formation of such invagination or vesicles24. By contrast, the attachment of RDV particles to Nasuia envelopes seems not to induce significant membrane changes. The evidence that Sulcia OMP-specific antibodies can strongly prevent viral invasion into the oocyte of N. cincticeps further confirms that the oocyte entry route established for Nasuia may not be used by RDV. Thus, RDV P2 may have evolved to interact with OMPs from Sulcia but not from Nasuia. In this case, the Sulcia OMP can be regarded as the receptor for RDV P2. It seems that RDV is a more recent and independent acquisition by Sulcia to form a parasitism, and the evolutionary fate of all partners is tightly linked.
Transmission electron microscopy
For subcellular localization of RDV, Sulcia and Nasuia within the ovaries of leafhoppers over time, ovaries from viruliferous leafhoppers were excised at different days post emergence, fixed, dehydrated and embedded. Thin sections were examined with an H-7650 Hitachi transmission electron microscope, as described previously14.
Immunofluorescence microscopy and FISH
To trace the oocyte entry paths of RDV and Sulcia in the ovaries of rice leafhoppers, 200 second-instar nymphs were allowed a two-day acquisition access period on rice plants infected with RDV, then allowed to feed on rice seedlings. At different days post emergence, 30 insect ovaries were fixed and immunolabelled with virus-specific antibodies conjugated to FITC (virus–FITC), as described previously14. The ovaries were then fixed again for 20 min, pretreated in hybridization buffer (20 mM Tris-HCl, 0.9 M NaCl, 0.1% sodium dodecyl sulfate, 5 mM EDTA, 10× Denhardt's sloution) for 15 min, and incubated in hybridization buffer containing 10 pmol ml−1 fluorescent dye HEX-labelled oligonucleotide DNA probe (Sangon Biotech), targeting the 16S rRNA sequence of N. cincticeps Sulcia (NcSul_16S/r1-HEX)7, as described previously with slight modifications15. After 4 h incubation at 50 °C, the samples were thoroughly washed in a washing buffer (0.15 M NaCl, 0.015 M sodium citrate) and then observed with a Leica TCS SP5 confocal microscope. As controls, excised ovaries or salivary glands from rice leafhoppers that fed on healthy plants were prepared in the same way.
Yeast two-hybrid assay
To test the interaction between the OMP of Sulcia and outer capsid proteins P2 or P8 of RDV, a yeast two-hybrid assay was performed using the Matchmaker Gal4 Two-Hybrid System 3 (Clontech). We first sequenced the complete genome of Sulcia from N. cincticeps using the Miseq PE300 sequence technique and obtained the complete OMP gene, which contained an open reading frame (ORF) of 3,027 bp and a typical bacterial surface antigen (BSA) domain (Fig. 2a). The respective N-terminal (bp 1–2,115, OMPN) and C-terminal (bp 2,116–2,955, BSA domain, OMPC) segments of the OMP gene were constructed in the prey plasmid pGADT7. The respective N-terminal (bp 1–2,064, P2N) and C-terminal (bp 2,065–3,444, P2C) segments of the RDV P2 gene were constructed in the bait plasmid pGBKT7. Meanwhile, the RDV major outer capsid protein P8 gene was constructed in the bait plasmid pGBKT7. The conserved BSA domains of Sulcia OMP genes from other rice leafhoppers (R. dorsalis and N. nigropictus) were amplified and constructed in the prey plasmid pGADT7. In addition, the conserved BSA domain of the Sulcia OMP gene (OMPC) from maize leafhopper (D. maidis) was obtained from NCBI (GenBank accession no. CP010105.1), synthesized by Invitrogen and constructed in the prey plasmid pGADT7. The bait and prey plasmids were used to co-transform yeast strain AH109, and β-galactosidase activity was detected on SD/Leu-Trp-His-Ade-/X-a-gal culture medium. The positive control pGBKT7-53/pGADT7-T and negative control pGBKT7-Lam/pGADT7-T were transformed in the same way.
GST pull-down assay
To confirm the interaction between the 15 nm domain of RDV P2 and the BSA domain of Sulcia OMP of N. cincticeps, a GST pull-down assay was performed as described previously2. Briefly, the GST-fused 15 nm domain of RDV P2 (P2N) or GST were first bound to GST–Sepharose 4B beads (GE) for 3 h at 4 °C, then the mixture was centrifuged for 5 min at 100g and the supernatant discarded. The His-MBP-fused BSA domain of Sulcia OMP or His-MBP was added to the beads and incubated for 2 h at 4 °C. The mixture was centrifuged for 5 min at 100g and washed with wash buffer (300 mM NaCl, 10 mM Na2HPO3, 2.7 mM KCl, 1.7 M KH2PO4). Immunoprecipitated proteins were detected using western blotting with the respective His-tagged and GST-tagged antibodies (Sigma).
Preparing antibodies against Sulcia OMP and neutralizing RDV–Sulcia binding in vivo
Mouse polyclonal antisera against Sulcia OMP were prepared as described previously25. Briefly, the conserved BSA domain (bp 2,116–2,955, OMPC) of the Sulcia OMP gene of N. cincticeps was amplified and inserted into vector pH4. The recombinant plasmid was used to transform E. coli strain Rosetta to express the targeted protein, which was then injected into mice to produce antibodies. To test the specificity of the antibodies against Sulcia OMP, we extracted total proteins from the midgut, fat body or bacteriome of viruliferous N. cincticeps and OMP protein expressed by E. coli. Samples were separated by SDS–PAGE and probed with antibodies against Sulcia OMP or pre-immune serum. To determine whether antibodies against Sulcia OMP specifically recognized Sulcia OMP in vivo, at 12 days post emergence the ovaries of viruliferous N. cincticeps were fixed, immunolabelled with virus–FITC and OMP-specific antibodies conjugated to rhodamine (OMP-rhodamine; Invitrogen), and then examined by immunofluorescence microscopy as described previously14.
To determine whether the antibodies against Sulcia OMP neutralized the direct interaction between RDV P2 and Sulcia OMP, we microinjected OMP antibodies into female adult leafhoppers following the method reported for the brown planthopper26. Briefly, at 1 day after emergence, the abdomens of 30 adult female leafhoppers were microinjected with 1 µl of a mixture of purified viruses (0.01 µg µl−1) and OMP antibodies (0.5 µg µl−1), and the insects were then allowed to feed on rice seedlings27,28. At 8 days post emergence, the ovaries were immunolabelled with RDV–FITC and hybridized with the probe NcSul_16S/r1-HEX, and then processed for confocal microscopy. As controls, adult female leafhoppers were microinjected with purified viruses (0.01 µg µl−1) and pre-immune serum (0.5 µg µl−1) and treated in the same way. To further detect any effects of OMP antibodies on transovarial transmission, we microinjected female leafhoppers with a mixture of RDV particles and OMP antibodies and maintained them on rice seedling. At 15 days post emergence, 50 eggs were collected from rice seedlings and tested for mRNA transcripts of the RDV P8 gene by reverse transcription polymerase chain reaction(RT-PCR) assay. As controls, the female leafhoppers were microinjected with RDV particles and pre-immune serum and treated in the same way.
RDV–Sulcia binding in vitro
To detect RDV binding with Sulcia in vitro, ovaries were excised from non-viruliferous N. cincticeps and incubated with purified RDV or RGDV particles (0.5 µg µl−1) for 4 h. Alternatively, purified RDV pre-incubated with P2 or GST antibodies (IgG) for 4 h was used to incubate ovaries from 30 non-viruliferous N. cincticeps. The ovaries were then immunolabelled with RDV–FITC or RGDV–FITC and hybridized with the probe NcSul_16S/r1-HEX and processed for confocal microscopy. To detect the binding of P2 of RDV with Sulcia in vitro, we expressed the His-tag-fused 15 nm or 10 nm domains of P2 in E. coli strain Rosetta and then purified the proteins using nickel-nitrilotriacetic acid (Ni-NTA) resin (Qiagen), as described previously25. Ovaries from 30 non-viruliferous N. cincticeps were incubated with these purified proteins (0.5 µg µl−1) for 4 h, then immunolabelled with P2–FITC and hybridized with the probe NcSul_16S/r1-HEX and processed for confocal microscopy.
Effects of antibiotics on Sulcia, Nasuia, RDV or N. cincticeps
To determine whether antibiotics can inhibit Sulcia and Nasuia in vivo, 100 first-instar nymphs of leafhoppers were treated with antibiotics, as described previously7,29. Briefly, rice seedlings were incubated with 25 µg ml−1 rifampicin, and then the first-instar nymphs of non-viruliferous leafhoppers were kept on the rice seedings. The adult leafhoppers were then selected for oviposition. Total RNAs of 30 three-day-old eggs were extracted using a TaqMan Gene Expression Cells-to-CT Kit (Life Technologies) according to the manufacturer's instructions. Quantitative RT-PCR (RT-qPCR) was performed using the SYBR Green PCR MasterMix kit (Promega) as described previously30. Cycle thresholds (CT) were obtained by the RT–qPCR assay. The Sulcia OMP and Nasuia 16S copies were calculated as the log of the copy number per egg by mapping the CT value to the standard curve of the Sulcia OMP gene (y = –3.17x + 58.68) and Nasuia l6S gene (y = –3.86x + 35.69).
We then evaluated the effects of antibiotics on the leafhoppers. One hundred non-viruliferous neonates were collected and reared on the rice seedlings, and were incubated with 25 µg ml−1 rifampicin. When the neonates developed into adults, the survival rate and nymphal stage duration were calculated. After emergence, pairs consisting of one female and one male newly emerged leafhopper were selected for oviposition, with single rice seedlings in glass tubes. The female fecundity rates were calculated according to the average number of eggs laid by females that had copulated. The female lifespan was calculated after they died. The egg hatch rates were calculated according to the total number of neonates versus the total number of neonates plus the number of non-hatched eggs. Egg duration was calculated at the same time. The control was 100 non-viruliferous neonates reared on rice seedlings in the presence of distilled water under the conditions described above. These experiments were also conducted with three replications.
To further monitor the correlation between Sulcia and RDV during their joint entry into the oocyte, 30 viruliferous leafhoppers were treated with the antibiotics or distilled water as described. Fifty eggs of F1 and F2 leafhoppers were collected for relative RT–qPCR assay to measure the RDV genome copy in individual leafhopper eggs under different treatments, as above. The level of leafhopper actin gene was used as the internal control for each RT–qPCR assay. Relative RDV P8 gene expression level was detected using the 2-ΔΔCT method31.
All data were analysed with SPSS, version 17.0. Percentage data were arcsine square-root transformed before analysis. Multiple comparisons of the means were conducted based on Tukey's honest significant difference (HSD) test using a one-way analysis of variance (ANOVA). The data were back-transformed after analysis for presentation in the text, figures and tables.
How to cite this article: Jia, D. et al. Insect symbiotic bacteria harbour viral pathogens for transovarial transmission. Nat. Microbiol. 2, 17025 (2017).
This work was supported by the National Science Foundation for Outstanding Youth (grant no. 31325023), the National Basic Research Program of China (973 Program, no. 2014CB138400), the National Natural Science Foundation of China (grant no. 31571979) and the FAFU Foundation for Outstanding Youth (grant no. XJQ201507).
Supplementary Figures 1–5