Insect Bacterial Symbiont-Mediated Vitellogenin Uptake into Oocytes To Support Egg Development

Many insects harbor obligate bacterial symbionts that can be vertically transmitted to offspring by female insects through eggs. Here, we report that leafhopper vitellogenin (Vg) recognizes and binds a surface channel molecule (porin) on the envelope of obligate bacterial symbiont Nasuia, which potentially induces the opening of porin channels for Vg to access the cytoplasm of Nasuia. Thus, Vg can exploit bacterial symbionts as the independent carriers into the oocytes. Such Nasuia-carried Vg contents support efficient insect egg development. Thus, our findings indicate that insects have evolved strategies to exploit the symbionts for carrying additional Vgs to guarantee optimal insect reproduction.

spp. of ants (2)(3)(4). They tend to supplement the host's diet with amino acids or vitamins that are rare or absent in the food source. Virtually all plant sap-feeding insects have obligate bacterial symbionts providing essential amino acids (4). For example, Buchnera aphidicola synthesizes several amino acids that are required for the aphid's metabolism (5). When experimentally deprived of the symbiont by antibiotics, host insects suffer retarded growth and high mortality (5). Thus, obligate bacterial symbionts are crucial to improve overall insect host fitness such as providing essential nutrients for survival and reproduction.
Almost all obligate bacterial symbionts in insects are vertically transmitted from a female host to her offspring following their transmission into the developing oocytes at the vitellogenic stage (1,6,7). Similarly, the yolk protein precursor vitellogenin (Vg), the essential nutrient for oocyte development in female insects, is also mainly transported into the developing oocytes at the vitellogenic stage (8,9). Thus, a potential correlation of the oocyte entry for symbionts and Vg might exist. Insect ovaries consist of several ovarioles, each of which contains a germarium, vitellarium, and pedicel. Oocytes produced by the germarium are linearly arranged within the vitellarium and surrounded by a layer of follicular cells (10). Vg is typically biosynthesized extraovarially by the insect fat body, secreted into the hemolymph, taken up specifically by the germarium, and deposited into the oocyte cytoplasm in its storage form, vitellin (9). The Vg protein precursors of most insect species are regulated at the posttranslational level and cleaved into large (140-to 190-kDa) and small (ϳ50-kDa) subunits before being processed in the ovary (11). It is believed that the large Vg subunits are transported into the rapidly growing insect oocytes through Vg receptor (VgR)-mediated endocytosis (12). Some facultative bacterial symbionts have evolved to hitchhike the existing pathways for Vg to enter insect oocytes. For example, it is thought that Spiroplasma in fruit flies (Drosophila) or Wolbachia in small brown planthoppers utilizes the Vg or yolk entry systems to enter the germarium of the ovary and transport into the oocyte (29). However, whether bacterial symbiont-mediated Vg uptake into the oocytes occurs remains undetermined. We recently reported a new oocyte entry route for viral pathogens by hitchhiking the obligate bacterial symbionts "Candidatus Sulcia muelleri" (hereafter Sulcia) and "Candidatus Nasuia deltocephalinicola" (hereafter Nasuia) from the hemolymph into the posterior pole of the terminal oocyte in leafhopper vectors (13,14). Given that the female insects need to supply the developing eggs with nutrients, it may be possible that the symbionts could also provide a means of transport for host nutrients such as Vg into the eggs.
Leafhoppers (Cicadellidae), a large insect family containing approximately 22,000 described species, are cosmopolitan (15). They feed on various crop plants and fruit trees and cause direct damage to agriculture production. Around 200 species of leafhoppers reported can transmit numerous pathogenic microbes (15,16). Leafhoppers are generally associated with the obligate bacterium Sulcia and a betaproteobacterial partner, such as Nasuia, which are transmitted via ancient oocyte entry paths from the mother to the offspring (17,18). Sulcia and Nasuia provide at least 10 essential amino acids for the survival of host insects (4). In general, Sulcia and Nasuia leave the bacteriocytes and then simultaneously enter the follicular cells surrounding the posterior poles of the terminal oocytes (epithelial plug) at the vitellogenic stage (17,18). From there, they directly squeeze and break through the distinct microvilli into the oocytes to form a characteristic "symbiont ball" (17,18). The rice green leafhopper Nephotettix cincticeps is widely distributed in rice growing countries in Asia. Here, we report that a new oocyte entry route for Vg of N. cincticeps (NcVg) by hitchhiking the obligate bacterial symbiont Nasuia from the hemolymph into the posterior pole of the terminal oocyte. Such Nasuia-carried NcVgs provide at least 20% of the total Vgs in the developing eggs. Thus, our findings indicate that insects have evolved strategies to exploit the symbionts for carrying additional Vgs to support egg development, possibly improving host fitness.

RESULTS
Correlation of the oocyte entry paths for obligate bacterial symbionts and NcVg in N. cincticeps. To observe the distribution of the two types of obligate bacterial symbionts Sulcia and Nasuia, the female adults of N. cincticeps were labeled with specific oligonucleotide probes targeting the 16S rRNA of Sulcia or Nasuia using fluorescence in situ hybridization (FISH). We observed that Sulcia and Nasuia were harbored within the two bacteriocytes in insect abdomens ( Fig. 1A to C). FISH and electron microscopy showed that the two symbionts simultaneously entered the epithelial plug ( Fig. 1D and F; see also Fig. S1A to G in the supplemental material) and then directly squeezed and broke through the microvilli into the terminal oocyte ( Fig. 1E and F and Fig. S1F to I). We found the almost simultaneous appearance of the two symbionts and yolk granules within the oocytes at the vitellogenic stage ( Fig. S1J to M), suggesting a potential correlation of the oocyte entry paths for symbionts and NcVg. Furthermore, reverse transcription-quantitative PCR (RT-qPCR) analysis revealed that NcVg transcript levels were significantly increased during the vitellogenic stage and decreased during the post-vitellogenic stage (Fig. 1G). The 16S rRNA levels of Nasuia and Sulcia in the ovaries were positively correlated with this NcVg accumulation ( Fig. 1H and I). These observations indicate that the entry of NcVg into the oocytes may be related to the transovarial transmission of the obligate bacterial symbionts.
To further investigate the correlation of the oocyte entry paths for these two symbionts and NcVg in N. cincticeps, we obtained the full-length NcVg cDNA sequence by RT-PCR process (GenBank accession no. KX022097.1). The 6,121-bp NcVg cDNA contained a 6,036-bp open reading frame encoding a deduced 2,011-amino-acid protein (ϳ220 kDa) and four conserved domains, including two vitellogenin_N domains (NcVg1 and NcVg2), a domain of unknown function (NcVg3), and a von Willebrand factor type D domain (NcVg4) (Fig. 1J and Fig. S2). Furthermore, an aminoterminal sequence analysis demonstrated that the pro-NcVg precursor contained consensus RXXR cleavage site sequences (9) (Fig. 1J and Fig. S2). Vg precursors are generally cleaved into large (140-to 190-kDa) and small (ϳ50-kDa) subunits before being transported into the ovary (11). The predicted cleavage of NcVg at the consensus RXXR sites in the N-terminal region may generate these large and small subunits (Fig. 1J). Since the NcVg2 domain stretches across the two putative cleaved subunits, we thus prepared NcVg2-specific antibody to investigate the cleavage pattern of pro-NcVg in N. cincticeps ( Fig. 1J and Fig. S2). We identified intact NcVg (ϳ220 kDa) in the fat body and two cleaved subunits, 35 and 178 kDa, in the hemolymph, whereas only the 178-kDa subunit was detected in the ovary (Fig. 1K). We then purified the 35and 178-kDa subunits of NcVg from the hemolymph of the female insects for a mass spectrometry analysis and found that they corresponded to the N and C regions of NcVg, respectively (Fig. S2). These data indicated that pro-NcVg was cleaved at the N-terminal consensus RXXR cleavage site to generate the 35-and 178-kDa subunits in the hemolymph, and only the large 178-kDa NcVg subunit was sequestered by the developing N. cincticeps oocytes.
We then observed the entry dynamics of NcVg in the vitellogenic-stage ovary of the female N. cincticeps using immunofluorescence and the NcVg2-specific antibody. NcVg was not visible in the pre-vitellogenic-stage ovary (Fig. 1L); it was initially accumulated in the nurse cells of the germarium and then moved into the oocyte via a nutritive cord in the vitellogenic-stage ovary ( Fig. 1M and N) (19). Surprisingly, dense NcVg signals were also observed in the epithelial plug surrounding the posterior pole of the oocyte, where the obligate bacterial symbionts Sulcia and Nasuia generally accumulated (Fig. 1O). As the oocytes developed, NcVg appeared to form a "ball" and accumulated within the posterior pole of the oocyte (Fig. 1P), before finally moving into the oocytes themselves ( Fig. 1P and Q). Collectively, these data indicated that NcVg can be taken up by the germarium and then transported into the oocytes via a nutritive cord and that it may also be transferred directly into the terminal oocytes through the epithelial plug in association with symbiont transmission (Fig. 1R). NcVg spreads from the hemolymph into the terminal oocyte in association with Nasuia. To trace the entry of NcVg, Nasuia, and Sulcia into the oocytes of N. cincticeps adult females, the hemocytes, bacteriocytes, and ovaries were excised and labeled with NcVg2-specific antibody conjugated with fluorescein isothiocyanate (NcVg2-FITC) and specific oligonucleotide probes targeting the 16S rRNA of Sulcia or Nasuia. Confocal microscopy revealed that NcVg was absent within the bacteriocytes ( Fig. 2A) and that it was found to colocalize with Nasuia rather than with Sulcia in the hemocytes ( Fig. 2B and Fig. S3A). Furthermore, NcVg, Sulcia, and Nasuia were not visible in the pre-vitellogenic-stage ovary (Fig. 2C); however, NcVg accompanied Nasuia and was transported into the epithelial plug (Fig. 2D, panels I, and Fig. S3B, panels I) and then into the posterior pole of the terminal oocyte (Fig. 2D, panels II, and Fig. S3B, panels II) in the vitellogenic-stage ovary. Careful observations revealed that NcVg was almost completely colocalized with Nasuia signals, suggesting that NcVg was localized inside the cytoplasm of Nasuia (Fig. 2D, panels I and II, and Fig. S3B, panels I). Finally, NcVg appeared to be released from the "symbiont ball" into the oocyte cytoplasm (Fig. 2D, panels III, and Fig. S3B, panels III). We examined at least 25 insects with Nasuia-associated hemocytes (see Table S1 in the supplemental material) or ovaries ( Table S2) and found that NcVg always colocalized with Nasuia rather than with Sulcia. A parallel immunoelectron microscopy confirmed that NcVg was absent within the cytoplasm of Nasuia in the bacteriocytes ( Fig. 2E and F) but was present within the cytoplasm of Nasuia in the epithelial plug ( Fig. 2G and H and Fig. S3C to F). In general, the outer envelopes could form the invaginations within the cytoplasm of Nasuia. Immunoelectron microscopy further revealed that NcVg could distribute along the envelope invaginations of Nasuia ( Fig. 2I and J and Fig. S3G to J). Occasionally, NcVg could be released from the broken envelopes of Nasuia ( Fig. S3K and L). Thus, we showed the possible pathways for NcVg to release from Nasuia by overcoming the bacterial envelope barriers in the terminal oocyte. Finally, the released NcVg can be absorbed by the developing yolk granules in the terminal oocyte ( Fig. 2K and Fig. S3M). Among at least 20 insects with Nasuia-or Sulcia-associated ovariole samples, we calculated that about 80% possessed NcVg-containing Nasuia, while none had NcVgcontaining Sulcia (Table S2). These findings suggest that Nasuia-NcVg association may facilitate their simultaneous joint entry into host oocytes, either by hitchhiking the NcVg uptake machinery or being exploited as a transfer vesicle for NcVg.
The joint movement of NcVg and Nasuia into the oocyte is independent of the NcVgR system. Given the natural role of VgR as a receptor for Vg, we next asked whether the simultaneous joint entry of Nasuia and NcVg into host oocytes is dependent on the VgR-dependent system. We prepared the NcVgR-specific antibody and observed the distribution of NcVgR using immunofluorescence. NcVgR was not visible in the pre-vitellogenic-stage ovary (Fig. 3A); however, it accumulated in the germarium, but not in the epithelial plug surrounding the posterior poles of the terminal oocytes in the vitellogenic-stage ovary ( Fig. 3B to F). We thus deduced that the joint movement of NcVg and Nasuia into the terminal oocyte is independent of the NcVgR system. Due to the joint entry of NcVg and Nasuia into the terminal oocyte at the vitellogenic stage, we then knocked down the in vivo expression of NcVgR (GenBank accession no. KX022098.1) by microinjecting double-stranded RNAs (dsRNAs) targeting this gene (dsNcVgR) into the newly eclosed adult females. Our preliminary experiments showed that microinjection of 3.2 ng dsNcVgR into the bodies of individual female N. cincticeps caused an approximately 50% reduction of NcVgR accumulation in the ovaries at the transcript and protein levels ( Fig. 3G and H) but only slightly affected leafhopper survival until 11 days (Fig. 3I) and egg hatching rate (Fig. 3J). Although 3.2 ng dsNcVgR treatment delayed the ovary development and reduced fecundity, the ovary still remained intact ( Fig. 3K and Fig. S4). In the dsNcVgR-treated females, the accumulation of NcVg decreased by approximately 30% in the ovaries but increased in other tissues at the transcript and protein levels ( Fig. 3H and L).
Porin is a class of outer membrane proteins (OMPs) of Gram-negative bacteria (14). Previously, we have prepared antibodies against Nasuia porin and Sulcia OMP (13,14). Although the oocyte entry process of Nasuia and Sulcia was delayed owing to the retarded ovary development following the dsNcVgR treatment (Fig. S4), their abundance remained constant in the ovaries and other tissues, as revealed by Western blot assay with antibodies against Nasuia porin and Sulcia OMP (Fig. 3H). Confocal microscopy confirmed that Nasuia-NcVg colocalization in the ovaries remained congruent after the knockdown of NcVgR expression ( Fig. 3M and N). All these results thus provide strong evidence that the entry of the Nasuia-NcVg complex into the oocytes is independent of the NcVgR trafficking system. N. cincticeps exploits Nasuia for NcVg transport into the oocytes by establishing a specific interaction between Nasuia porin and NcVg. We then investigated how NcVg can reside in the cytoplasm of Nasuia. We reasoned that porins, the abundant OMPs in bacterial membranes that form channels to facilitate the passage of some cellular molecules (20), might mediate the entry of NcVg into the Nasuia cytoplasm. Yeast two-hybrid assay showed that the NcVg2 domain of Vg precursor can directly interact with Nasuia porin (Fig. 4A). Furthermore, NcVg did not interact with the Sulcia OMPs (13, 14) (Fig. 4A). A glutathione S-transferase (GST) pulldown assay also confirmed that GST-fused Nasuia porin specifically bound to His-fused NcVg2 (Fig. 4B). These results suggested that NcVg-porin interaction may activate the porin channel on Nasuia envelope to open, allowing NcVg to pass through the envelope into the Nasuia cytoplasm.
To further characterize the role of Nasuia in the oocyte entry of NcVg, we knocked down the expression of a gene encoding an insect proline-rich protein (N. cincticeps proline-rich protein [NcPRP]) (GenBank accession no. MK722101), which was reported to determine the abundance of Nasuia rather than Sulcia in N. cincticeps by a yet unknown mechanism (21). RT-PCR and Western blot assays revealed an enriched NcPRP expression in the bacteriome but not in other tissues ( Fig. 4C and Fig. S5A). Microinjecting synthesized dsRNAs targeting the NcPRP gene (dsNcPRP) into the newly emerged insects resulted in a significant reduction in the abundance of Nasuia in both rather than with Sulcia in the hemocytes. Bar, 25 m. (C) Confocal micrographs showed that NcVg, Nasuia, and Sulcia were absent in the epithelial plug of the pre-vitellogenic-stage ovary. Bar, 100 m. (D) Confocal micrographs showed that NcVg accompanied Nasuia, rather than Sulcia from the epithelial plug (panels I) into the oocyte to form a "symbiont ball" (panels II), wherein NcVg was released into the oocyte (panels III). Bars, 50 m. Red arrows marked the colocalization of NcVg and Nasuia. The female bacteriocytes (A), hemocytes (B), and ovarioles (C and D) were stained with Sulcia-cy5 (blue), Nasuia-cy3 (red), and NcVg-FITC (green). (E and F) Immunoelectron micrographs showed that NcVg was absent within the bacteriocytes. Bars, 2 m (E) and 500 nm (F). (G and H) Immunoelectron micrographs showed the presence of NcVg within the cytoplasm of Nasuia in the epithelial plug. Bars, 500 nm. (I and J) Immunoelectron micrographs showed the distribution of NcVg along the envelope invaginations of Nasuia. Bars, 500 nm. (K) Immunoelectron micrograph showed the presence of NcVg within the developing yolk granules in the oocyte cytoplasm. Bar, 500 nm. The female bacteriocytes (E and F) and ovaries (G to K) were immunolabeled with NcVg-specific IgG as the primary antibody, followed by treatment with 15-nm gold particle-conjugated goat antibodies against rabbit IgG as the secondary antibody. Panels F, H, and J are the enlargements of boxed areas in panels E, G, and I, respectively. Red arrows mark gold particles. Ep, epithelial plug; Ev, envelope; Fc, follicular cell; Sb, symbiont ball; O, oocyte; N, Nasuia; S, Sulcia. All images are representative of at least three replicates. the ovaries and the remaining tissues, while the Sulcia abundance remained constant (Fig. 4D to F). The abundance of NcVg in the ovary was decreased in the insects microinjected with dsNcPRP (Fig. 4H), indicating that knocking down NcPRP expression inhibited the oocyte entry of NcVg. Consistently, confocal microscopy indicated that the dsNcPRP treatment almost completely inhibited the coentry of NcVg and Nasuia into the epithelial plug and oocyte but did not significantly affect the efficiency of Sulcia oocyte entry (Fig. 4H and I). Nevertheless, the dsNcPRP treatment of N. cincticeps reduced the survival rate and fecundity of the females, indicating that the Nasuia titer was critical for egg maturation (Fig. S5B and C). Together, these results reveal a previously undescribed function for insect obligate bacterial symbionts, wherein Nasuia is involved in the process of NcVg oocyte entry.

Insect Bacterial Symbionts Mediate Vitellogenin Uptake
Nasuia-carried NcVg supports efficient egg development. We next determined how this newly discovered Nasuia-mediated oocyte entry pathway for leafhopper Vg influenced the reproduction of the insect hosts. The in vivo interaction of NcVg and Nasuia porin in the hemocytes was determined by microinjecting antibodies against Nasuia porin, Sulcia OMP, or NcVg into the insect hemocoel. It was expected that treatment with NcVg2 antibody, but not with NcVg3 or NcVg4 antibodies, caused a significant reduction of the colocalization of Nasuia and NcVg in the hemocytes (Fig. 5A). Similarly, the frequency of colocalization of Nasuia and NcVg in the hemocytes was significantly decreased after treatment with Nasuia porin antibody but remained constant after treatment with Sulcia OMP antibody (Fig. 5B). Most importantly, treatment with porin antibody caused an approximately 20% reduction of NcVg deposition in insect eggs without affecting Sulcia or Nasuia abundance (Fig. 5C). Interestingly, porin antibody treatment significantly reduced insect egg hatching and prolonged the egg hatching period (Fig. 5D to F). Together, these results reveal that Nasuia-carried Vgs provided at least 20% of the total Vgs in the developing eggs for efficient embryonic development.
Interaction of Nasuia and Vg in different leafhopper species. The above results demonstrate that NcVg hitchhikes with the obligate symbiont Nasuia for transport into the ovaries of N. cincticeps; however, it is unknown whether this phenomenon is unique to this host-symbiont relationship or whether other cases exist. We therefore investigated whether the interaction between Nasuia and Vg also occurred in other rice leafhopper species, such as Nephotettix nigropictus, Nephotettix virescens, and Recilia dorsalis (16,(22)(23)(24). Phylogenetic analysis of the Nasuia 16S rRNA sequences of N. cincticeps, N. nigropictus, R. dorsalis, and N. virescens using Bayesian methods revealed that they had formed a close relationship (Fig. S6A). The multiple alignment of the Nasuia porin amino acid sequences revealed that the sequences of N. cincticeps, N. nigropictus, and N. virescens were conserved with a high sequence similarity (ϳ67% identity) but shared a relatively lower similarity with Nasuia porin in R. dorsalis (ϳ41% identity) (Fig. S6B). A sequence alignment of the Vg2 domain in these leafhopper species indicated that they shared high amino acid sequence similarities, ranging from 91% to 95% (Fig. S6C). Furthermore, we examined the localization of Sulcia, Nasuia, and Vg in the ovaries of these insects, revealing that Vg colocalized with Nasuia rather than Sulcia in all tested leafhopper species (Fig. 6A). Yeast two-hybrid assay also revealed that Nasuia porins of these insect species interacted specifically with their respective Vg2 domains (Fig. 6B). Taken together, these conserved mechanisms indicate that the leafhoppers have evolved to exploit Nasuia for transporting Vg into their oocytes by establishing the specific interaction between Vg and the Nasuia porin (Fig. 6C).

DISCUSSION
To facilitate embryonic development, large quantities of the yolk protein precursor Vgs are transported into insect oocytes through receptor-mediated endocytosis (8,9). In this study, we report a novel receptor-independent Vg uptake system, in which Vg directly hitchhikes the oocyte entry path of endosymbiont to ensure insect egg development success (Fig. 6C). A rice leafhopper female can oviposit ϳ200 eggs at the vitellogenic stage, and thus, massive Vgs are required for egg maturation (25,26). Rice leafhopper Vg precursor is biosynthesized by the fat body, secreted into the hemolymph for cleavage into the 35-and 178-kDa subunits, whereas only the 178-kDa subunit is taken up by the ovary germarium through the VgR-dependent system (10). Simultaneously, a large amount of Nasuia bacteria (5 ϫ 10 8 copies) leave the bacteriocytes, move into hemolymph, and then move to the posterior poles of the terminal oocytes (14). In the hemolymph, the large cleaved Vg subunits can recognize and bind the porin on Nasuia envelope, which potentially induces the opening of porin channels for Vg to pass into the cytoplasm of Nasuia. The Vg-containing Nasuia then moves into The numbers of eggs laid by adult females treated with porin antibody or preimmune serum were comparable. (E and F) Nasuia porin antibody treatment significantly reduced egg hatching rate (E) and prolonged the egg hatching duration (F). Data in panels A, B, and D to F are presented as means Ϯ SD of three independent experiments. The significance of any differences was tested using an independent t test. *, P Ͻ 0.05; **, P Ͻ 0.01; ****, P Ͻ 0.0001; NS, not significant.
Insect Bacterial Symbionts Mediate Vitellogenin Uptake ® the posterior poles of terminal oocytes. Finally, Vg is potentially released through the porin channels on Nasuia envelope invaginations or through the broken Nasuia envelopes. The released Vg is ultimately internalized within the developing yolk granules in the oocyte. We calculate that Nasuia-carried Vgs provide at least 20% of the total Vg contents in the developing eggs. The supplementation of Nasuia-carried Vgs complements the egg development, which guarantees optimal insect reproduction. Notably, the Nasuia-mediated oocyte entry of Vg represents a conserved mechanism employed by different leafhopper species, suggesting that this conserved mechanism could play an important role in maintaining insect reproduction.
We anticipate that this newly discovered model for the uptake and storage of Vg by endosymbionts to provide a supplementary source of Vg for egg development may be a common pattern shared by many insects. The posterior pole of the terminal oocyte is the initial oocyte entry site for many obligate bacterial symbionts such as Buchnera and Serratia in aphids, bacteriocytes in whiteflies, Sulcia and Nasuia in leafhoppers, and bacteroids in Bradysia tritici (6,14,27). However, the leading end of oocytes, the germarium, is the initial oocyte entry site for the facultative bacterial symbionts Spiroplasma and Wolbachia by hitchhiking the yolk or Vg uptake machinery in a receptor-dependent manner in their respective Drosophila or planthopper hosts (28,29). Here, we find that VgR is absent in the posterior poles of the terminal oocytes of female N. cincticeps, from there, Sulcia and Nasuia directly squeeze and break through the distinct microvilli into the terminal oocyte cytoplasm. Because Vg resides inside the cytoplasm of Nasuia, it is reasonable that the transport of Vg-containing Nasuia into the terminal oocytes is independent of the Vg/VgR transport and internalization machinery. Many viral pathogens have evolved to hitchhike the existing pathways for Vg and bacterial symbiont to enter insect oocytes, facilitating the long-term viral epidemic and persistence in nature (2). For example, Vg can carry rice stripe virus or tomato yellow leaf curl virus into the germarium via the VgR-dependent system, from which the virus can spread into the oocyte cytoplasm along the nutritive cords in their respective planthopper or whitefly vectors (19,30). Our recent finding reveals that rice dwarf virus can recognize and bind Nasuia porin to enable virions to pass through the bacterial envelope into the periplasmic space, enabling Nasuia to carry virions into the posterior poles of rice leafhopper oocytes (14). We hypothesize that the long-term coexistence of bacterial symbionts, Vg, and viral pathogens on their pathways into insect oocytes may have led to the formation of evolutionary cross-kingdom interactions among them in nature.
Our study demonstrates a novel oocyte entry path for Vg by hitchhiking on Nasuia to ensure the success of leafhopper egg development, which guarantees optimal insect reproduction. One female insect generally oviposits dozens or hundreds of eggs at the vitellogenic stage, and thus numerous Vgs are required for egg development and maturation. During insect ovary maturation, the apical and lateral regions of the terminal oocyte are covered by egg envelopes in late vitellogenesis, and consequently, the oocyte entry route for Vg via the nutritive cord may be abolished (7). At this time point, however, the posterior pole of the terminal oocyte is still open for the endosymbionts to enter, and thus, Vg can exploit the terminal oocyte entry paths used by the maternally inherited symbionts. Many insect species, such as aphids, leafhoppers, planthoppers, and whiteflies, harbor obligate bacterial symbionts that are vertically transmitted to the next generation of insects through the posterior poles of the terminal oocytes (1,2). Given the widespread distribution of these insect species in nature, they may also have gained an evolutionary advantage by hitchhiking with obligate bacterial symbionts for carrying Vg into the posterior poles of the terminal oocytes for ensuing efficient egg development. Specifically, the synchronization of intensive endosymbiont absorption into the ovaries of insect hosts with a massive requirement for Vg during egg maturation minimizes insect investment in Vg uptake and maximizes symbiont transmission. targeted the 16S rRNA sequences of Sulcia and Nasuia, as described previously with slight modifications (13,14). After 5-h incubation at 50°C, the samples were thoroughly washed in washing buffer (0.15 M NaCl, 0.015 M sodium citrate) and then observed with a Leica TCS SP5 confocal microscope.
To observe the localization of NcVgR, the ovaries were immunolabeled with NcVgR antibody conjugated to rhodamine (NcVgR-rhodamine) and examined following the above method.
Electron microscopy. To observe the subcellular entry process of Sulcia and Nasuia in ovaries of N. cincticeps, 40 ovaries from leafhoppers at different days postemergence were dissected and fixed with 0.5% (vol/vol) glutaraldehyde and 3% (vol/vol) paraformaldehyde in phosphate-buffered saline (PBS) (0.1 M, pH 7.2) for 12 h at 4°C. The fixed tissues were dehydrated in a graded series of ethanol from 50% to 100% at Ϫ20°C and embedded in LR Gold resin (SPI Ltd.). For immunoelectron microscopy, ultrathin sections from bacteriocytes and ovaries were immunolabeled with NcVg IgGs as the primary antibody (0.351 mg/ml, diluted with blocking buffer (50 mM PBS [pH 7.0] containing 1% bovine serum albumin [BSA], 0.02% polyethylene glycol 20000 [PEG 20000], and 100 mM NaCl) in 1:200 dilution), followed by treatment with goat anti-rabbit IgG conjugated with 15-nm-diameter gold particles as the secondary antibody (Abcam) (diluted with blocking buffer in 1:100 dilution). The sections were treated only with secondary antibody as a control. Ultrathin sections were examined with an H-7650 Hitachi transmission electron microscope. To determine the distribution relationship of NcVg and the two bacterial symbionts, 23 ovary samples were observed and the distribution was calculated.
Yeast two-hybrid assay. To test the interaction between Nasuia porin and Vg, a yeast two-hybrid assay was performed using the Matchmaker Gal4 Two-Hybrid System 3 (Clontech). The porin gene of Nasuia from N. cincticeps was constructed in the prey plasmid pGADT7. Different domains of Vg (Vg1 to Vg4) were constructed in the bait plasmid pGBKT7 (see Table S3 in the supplemental material). The bait and prey plasmids were cotransformed to the yeast strain AH109, and ␤-galactosidase activity was detected on SD/-Leu/-Trp/-His/-Ade/X-␣-Gal culture medium (100 l of X-␣-Gal [4 mg/ml] [catalog no. 630462; Clontech, Mountain View, CA, USA] spread onto a 10-cm SD medium lacking adenine, histidine, leucine, and tryptophan [catalog no. 630323; Clontech, Mountain View, CA, USA], plated using glass beads). The positive-control pGBKT7-53/pGADT7-T and negative-control pGBKT7-Lam/pGADT7-T were transformed in the same way. Since both the gene sequences of Nasuia porin and Vg2 from the leafhopper N. nigropictus, N. virescens, and R. dorsalis shared high similarity with those of N. cincticeps ( Fig. S5B and C), the same primers were used for gene cloning, and the interactions between Vg2 and Nasuia porin of the three leafhopper species were monitored in a yeast two-hybrid system.
GST pulldown assay. The cDNA fragments of Nasuia porin were amplified and cloned into PGEX-3X for fusion with GST. Domain 2 of vitellogenin (NcVg2) was cloned into pDEST17 for fusion with the His tag (Table S3). All recombinant proteins were expressed in Escherichia coli strain Rosetta and purified. GST-Nasuia porin was first bound to GST-Sepharose 4B beads (GE) for 3 h at 4°C, then the mixture was centrifuged for 5 min at 100 ϫ g, and the supernatant was discarded. Histidine-tagged NcVg2 (His-NcVg2), NcVg3, or NcVg4 was added to the beads and incubated for 2 h at 4°C. After being centrifuged and washed five times with washing buffer (300 mM NaCl, 10 mM Na 2 HPO 3 , 2.7 mM KCl, and 1.7 M KH 2 PO 4 ), the bead-bound proteins were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and detected by Western bloting with His-tagged antibody and GST-tagged antibody (Sigma).
Neutralizing Vg-Nasuia binding in vivo. To examine the binding affinity of NcVg to Nasuia porin in hemocytes of N. cincticeps, we microinjected the antibodies against Nasuia porin, Sulcia OMP, or NcVg into insect hemocoel to neutralize the interaction. The abdomens of 30 adult female leafhoppers at 1 day after emergence were microinjected with 64 nl of antibody solution (0.5 g/l) or preimmune serum (0.5 g/l) as a control, and the leafhoppers were allowed to feed on healthy rice seedlings. The hemolymph collected at 3 days after microinjection was immunolabeled with NcVg-FITC and hybridized with the Sulcia-cy5 or Nasuia-cy3 probe and observed by confocal microscopy. To calculate the colocalization of NcVg and Nasuia-cy3, 30 insect hemolymph samples for each treatment were prepared.
To determine the role of Nasuia-mediated Vg transfer manner in insect reproduction, the adult female leafhoppers at 5 days after emergence were microinjected with Nasuia porin antibody or preimmune serum and were allowed to feed on rice seedlings for oviposition. At 7 days postoviposition, the number of eggs laid by the female was recorded and the duration and hatching rate were measured. The protein levels of NcVg, Nasuia porin, and Sulcia OMP in 100 eggs from each treatment group were detected by Western blot assay.
Knocking down the in vivo expression of NcVgR in N. cincticeps. In order to examine whether Nasuia hitchhiked NcVg transport system, RNA interference (RNAi) was performed to knock down expression of the NcVgR gene. The abdomens of newly emerged adult females of N. cincticeps were microinjected with dsRNAs targeting NcVgR gene (dsNcVgR) or the green fluorescent protein (gfp) gene (dsgfp), and the treated insects were then allowed to feed on healthy rice seedlings. At 6 days after microinjection, 50 ovaries from each treatment group were immunolabeled with Vg-FITC, hybridized with the Sulcia-cy5 or Nasuia-cy3 probe, and then were observed with a Leica TCS SP5 confocal microscope. Thirty ovaries were observed with a Leica M165C stereomicroscope to calculate the number of ovarioles at different developmental stages. The accumulation levels of NcVgR, NcVg, Sulcia OMP, and Nasuia porin in 30 insects were detected by RT-qPCR and Western blot assays after treatment of dsNcVgR or dsgfp, respectively. Relative levels of gene expression were normalized to a housekeeping gene elongation factor 1alpha gene (EF1, GenBank accession no. AB836665) and estimated by the 2 ϪooCt (cycle threshold).
After dsRNA microinjection, 30 adult female leafhoppers were picked out for mating one to one with male leafhoppers in glass tubes containing one rice seedling. The total number of eggs laid by each mated female were recorded to evaluate the female fecundity. The egg hatching rates were evaluated according to the number of neonates/total number of eggs. The longevity of females was also measured. The entire experiment was repeated three times.
Knocking down the in vivo expression of Ncprp in N. cincticeps. We cloned the full-length Ncprp cDNA based on the genomic database of Nasuia. The role of Nasuia abundance in the oocyte entry process of NcVg was examined by knocking down expression of the Ncprp gene. The abdomens of newly emerged adult females were microinjected with dsRNAs targeting the Ncprp gene (dsNcprp) or dsgfp. The ovaries dissected from 50 insects 6 days after microinjection were immunolabeled with NcVg-FITC, hybridized with the Sulcia-cy5 and Nasuia-cy3 probes, and then were viewed by confocal microscopy. The proportion of Nasuia in the epithelial plugs of the ovaries was calculated depending on fluorescence signal. The levels of Ncprp, Nasuia, and Sulcia in 30 insect ovaries were measured by RT-qPCR assay (Table S3). The relative levels of gene expression were normalized to the level of the EF1 gene and estimated by the 2 ϪooCt (cycle threshold) method. The protein expression levels of Ncprp, Nasuia, and Sulcia in 30 insect ovaries were detected using a Western blot assay. In addition, the mortality, fecundity, and egg hatching rates of N. cincticeps treated with dsgfp or dsNcprp were monitored following the same protocol for dsNcVgR treatment. The experiments were repeated three times.
Data analyses. All data were analyzed with SPSS (version 17.0; SPSS, USA). Percentage data were first transformed by the arcsine square root before analysis. Data sets were examined for normality before comparison. Multiple comparisons of the means were conducted using a one-way analysis of variance (ANOVA) followed by Tukey's honestly significant difference (HSD) test at the P Ͻ 0.05 significance level. Comparisons between two means were conducted using an independent t test.
Data availability. Sequence data were deposited in GenBank under accession no. MK722101.

SUPPLEMENTAL MATERIAL
Supplemental material is available online only. TEXT S1, DOCX file, 0.03 MB.