De novo assembly of CvBV virions in the calyx cell nucleus of C. vestalis

BV replication begins with amplification of the proviral genome, followed by virion assembly within calyx cell nuclei [11,23]. In C. vestalis, pupal development lasts ~4.5 days [32] and is accompanied by ovarian development and CvBV virion assembly (Fig 1A and S1A-I Fig). We previously identified 30 dsDNA circles in the CvBV genome and characterized wasp integration motifs [31]. Absolute qPCR targeting the integration sites showed that the abundance of circularized CvBV DNAs was zero at Ov2d (S1 Table), first appeared at Ov3d, and increased sharply by Ov4d, peaking at OvFA1d (Fig 1B; raw CvBV copy numbers and p-values are shown in S1 Table). The relative abundance of individual CvBV circles appears to be established as early as the Ov3d stage, with circles of high abundance remaining dominant and low-abundance circles staying consistently low throughout ovarian development. For example, CvBV circle 14 shows the highest abundance at Ov3d and continues to be the most abundant circle during both Ov4d and OvFA1d. In addition, the other five circles in the top-six group (8, 29, 7, 18, and 19) exhibit a similar pattern. DAPI staining and TEM images showed progressive enlargement of calyx cells and nuclei from Ov3d to OvFA1d (S1J-L Fig), with the cytoplasm still visible at Ov3d but completely absent at the late stages Ov4d and OvFA1d (Fig 1C-E). TEM image showed CvBV virion formation was first observed at Ov3d, where capsids and envelopes began to co-assemble (Fig 1F). At Ov4d, some assembled virions are visible, consisting of multiple nucleocapsids enclosed by a single-layer envelope (Fig 1G), while many unassembled nucleocapsids are present at the periphery of the virogenic stroma, awaiting assembly (S1M Fig). At OvFA1d, the calyx cell nucleus was filled with mature virions (Fig 1H). As calyx cells lysed, these mature virions mixed with the wasp egg (S1N and S1O Fig), and were then injected together into the host larvae during parasitization. The concordance between qPCR and TEM results supports the proposed temporal sequence of CvBV replication and virion morphogenesis.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. Production of CvBV virions in C. vestalis calyx cells during pupal development.

C. vestalis pupal-to-adult development timeline with corresponding ovarian images. Yellow regions indicate maturing or fully mature eggs at Ov4d and OvFA1d. P2d, 2-day-old pupa; P3d, 3-day-old pupa; P4d, 4-day-old pupa; FA1d, 1-day-old female adult; Ov, Ovary. (B) Heatmap showing the abundance of CvBV circles 1–30 across ovarian development from Ov3d to OvFA1d. Rows represent ovary developmental stages, with three biological replicates per stage (n = 3), and columns correspond to CvBV circles 1–30. Color intensity reflects log₁₀-transformed CvBV copy numbers (raw CvBV copy numbers and corresponding raw and adjusted p-values are provided in S1 Table), with darker colors indicating higher abundance. Statistical significance was assessed using two-sided t-tests with Benjamini–Hochberg correction. (C-E) TEM overview of virion morphogenesis at Ov3d, Ov4d, and OvFA1d, respectively. The white rectangle in (D) marks the virogenic stroma, and the magnified view is shown in S1M Fig. The white arrow in (E) indicates mature virions present in the calyx lumen. (F-H) Enlarged views showing virion assembly within a calyx cell nucleus at Ov3d, Ov4d, OvFA1d respectively. In (F), black arrows indicate three representative structures: (i) an unenveloped nucleocapsid (a small electron-dense dot); (ii) an empty envelope without a nucleocapsid (a rounded membrane structure); and (iii) an assembling virion containing a nucleocapsid enclosed by a unit-membrane envelope (an electron-dense dot within a membrane-bound circle). Black arrows in (G) indicate unenveloped nucleocapsids (small electron-dense dots) and mature virions containing multiple nucleocapsids within a single envelope (multiple dots inside one circle). Black arrows in (H) indicate mature virions. N, nucleus. Cyt., cytoplasm. VS, virogenic stroma. CL, calyx lumen. Scale bars: (C–E), 5 μm; (F–H), 0.2 μm.

https://doi.org/10.1371/journal.ppat.1014492.g001

Dynamic expression of CvBV nudivirus-like genes during pupal ovary development of C. vestalis

To investigate gene expression programs underlying CvBV virion formation, we performed RNA-seq on ovaries from five developmental stages using the Illumina HiSeq 2500 platform. Approximately 707 million raw reads were generated from all samples, with clean reads exhibiting mapping ratios ranging from 94% to 97% (S2 Table). These high-quality RNA-seq data enabled us to explore gene expression dynamics across developmental stages. We next conducted a genome-wide analysis of gene expression levels in C. vestalis based on our newly assembled genome and annotation (GenBank accession No.: GCA_054085095.1). K-means clustering (k = 8) identified eight clusters with distinct temporal expression dynamics across the five ovarian developmental stages in C. vestalis (S2A Fig).

We first identified 51 nudivirus-like genes from the C. vestalis genome annotation based on homology to nudivirus-like genes from C. congregata. We then performed an additional BLASTP search using RNA-seq de novo–assembled transcripts to recover genes missing from the genome annotation, resulting in the identification of 20 additional nudivirus-like genes. Together, this integrated genome- and transcriptome-based approach resulted in a total of 71 nudivirus-like genes, ensuring a more complete identification of the nudiviral gene set (S3 Table). Notably, all are single-exon genes and share 66–99% amino acid identity with C. congregata homologs [22], supporting the accuracy of our annotation. To better understand the genomic context of these nudivirus-like genes, we examined their chromosomal distribution and organization. These genes are distributed across nine chromosomes and organized into three major clusters: cluster 1 on chromosome 7 (23 genes, ~ 82 kb), cluster 2 on chromosome 10 (6 genes, ~ 11 kb), and cluster 3 on chromosome 2 (3 genes, ~ 3.3 kb), with clusters 2 and 3 composed of odv-e66 family genes (S2B Fig). Interestingly, most clusters except cluster 6 identified in C. vestalis were also present in C. congregata [22], where they similarly exhibit clustered genomic organization.

RNA-seq analysis revealed dynamic expression of all 71 nudivirus-like genes in C. vestalis ovaries during pupal-to-adult development (S4 Table). Some genes showed peak expression at Ov3d and declined thereafter (early-expressed; corresponding to K-means cluster 2), while others peaked at Ov4d (late-expressed; corresponding to K-means cluster 3) (Fig 2A). Notably, functional predictions based on homology suggest that early-expressed genes were mainly associated with DNA replication and transcriptional regulation, whereas late-expressed genes may contribute to structural components and virion assembly factors. qPCR validation of five representative genes (38k, HzNVorf140–1, HzNVorf140–2, vlf-1, and vp91) confirmed the transcriptome patterns (Fig 2B). These results provide the first comprehensive catalog of nudivirus-like genes in C. vestalis and reveal their temporally coordinated expression during ovary development.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. Conserved nudivirus-like genes identified in C. vestalis.

(A) Left panel: Heatmap showing scaled expression levels (z-score) of 71 nudivirus-like genes in C. vestalis ovaries across different developmental stages. Expression values were calculated as TPM from three biological replicates for each developmental stage. Differential expression analysis was performed using DESeq2, with Ov1d as the reference condition. Log₂ fold changes and adjusted p-values are provided in S4 Table. Blue and red indicate lower and higher expression, respectively. Colored squares along the clustering tree represent predicted viral functions based on experimental validation and baculovirus homology. Right panel: Heatmap depicting mass spectrometry-based peptide detection for the 71 genes. Numbers within the boxes represent unique peptides and PSMs identified by mass spectrometry (only values ≥ 2 are shown). Darker blue shades indicate higher peptide counts, with values above 30 shown at the maximum blue intensity. Box colors indicate genes with high, medium, or low expression levels, consistent with validation in (B). The 24 genes marked with red asterisks were subjected to functional validation by RNAi in this study. K-means analysis (see S2A Fig) grouped the genes into two clusters, with cluster labels shown to the right of each gene name. The gene fen-1-2 was excluded from the K-means clustering due to its extremely low expression. (B) qPCR analysis (n = 3 biological replicates) of five nudivirus-like genes in C. vestalis shows expression patterns consistent with those observed in heatmap (A). (C) Mass spectrometry-based proteomic analyses of calyx fluid (R1) and purified CvBV virions (R2). R1 identified 56 proteins and R2 19 proteins, each supported by ≥ 2 unique peptides.

https://doi.org/10.1371/journal.ppat.1014492.g002

CvBV virions contain conserved nudivirus-like proteins

To distinguish structural proteins of CvBV virions from other proteins, we performed proteomic analyses on both calyx fluid (R1) and purified CvBV virions (R2, including two technical replicates: R2-1 and R2-2), which allowed us to compare their protein compositions (S3A Fig). To identify nudivirus-like proteins, mass spectra were searched against a custom protein database translated from the C. vestalis ovary transcriptome (S5 and S6 Tables). After summing the two technical replicates of R2 for both unique peptides and peptide-spectrum matches (PSMs), we assessed their consistency with R1. This analysis showed strong correlations (r = 0.79 for unique peptides, r = 0.60 for PSMs), and FPKM (Fragments Per Kilobase of transcript per Million mapped reads) showed a reasonable correlation with proteomic measurements in both samples (R1: r = 0.49–0.65; R2: r = 0.61–0.75; S3B Fig).

We hypothesized that sample R1 likely contains nearly all CvBV nudivirus-like proteins and partial wasp proteins, while sample R2 is expected to be enriched for structural components of CvBV virions. Of the 71 nudivirus-like proteins, 56 highly expressed proteins in R1 were identified with at least two unique peptides, while 19 were identified in R2 with at least two unique peptides (Fig 2C; S5 and S6 Tables). Proteins in R2 were predominantly highly expressed capsid or envelope components, e.g., capsid protein vp39 (21 unique peptides in R1, 24 in R2), the most highly expressed viral transcript (Fig 2A; S4 Table). As expected, several early-expressed genes with homology to nudivirus genes were exclusively identified in sample R1, including helicase (7 unique peptides), p47 (4), lef-8 (6), lef-9 (5), lef-4 (3), and lef-5 (3). Taken together, our proteomic analysis allows us to tentatively distinguish major structural proteins of CvBV from auxiliary factors involved in viral replication and virion maturation.

CvBV nudivirus-like genes are efficiently knocked down by RNAi

To functionally verify CvBV nudivirus-like genes, we selected 24 highly expressed genes for RNAi knockdown based on transcriptomic and proteomic data (Fig 2A and S4A Fig). These genes were grouped into four functional categories:

1. (1). DNA replication-related: helicase, integrase-1, and integrase-2;
2. (2). transcription-related: p47, lef-5, and lef-9;
3. (3). a functionally validated homolog or paralog in baculoviruses, nudiviruses, or other bracoviruses—many of which are implicated in virion structure or assembly—although whether these genes are conserved functionally in CvBV remains unclear: vp39, 38k, HzNVorf9–1, HzNVorf9–2, HzNVorf106, 27b, K425_459, 11k, 17a-1, 35a-1, 35a-2, vlf-1, HzNVorf140–1, HzNVorf140–2, pif-0, and vp91;
4. (4). highly expressed genes of unknown function: PmV and K425_461.

Double-stranded RNAs (dsRNAs) were injected into one-day-old female pupae to optimize gene silencing. Using the late-expressed 38k as a test case, we found that 200–400 ng dsRNA achieved >99% knockdown as measured by qPCR (S4B Fig). RNAi of all 24 genes yielded knockdown efficiencies of 74–98% (S4B Fig). To assess potential off-target effects of RNAi, we first performed in silico analyses of sequence similarity among CvBV nudivirus-like genes using BLASTN, identifying a limited number of short regions (≥ 21 nt) of potential homology between dsRNA target regions and non-target genes (S4C Fig). These included shared homologous fragments between HzNVorf9–1/HzNVorf9–2, vp91, and PmNVorf87, as well as between HzNVorf106 and helicase. In addition, off-target siRNA generation was evaluated using DSIR (http://biodev.extra.cea.fr/DSIR/DSIR.html) predictions. Importantly, only homologous regions overlapping dsRNA target sequences were considered potential off-target sites, whereas matches outside target regions were excluded. To experimentally evaluate RNAi specificity, we further quantified the expression of target genes, predicted off-target genes, and two C. vestalis housekeeping genes by qPCR following dsRNA treatment. Primer information for all off-target analyses is provided in S13 Table. Except for two C. vestalis genes (dnapolδ and ef1α), all primers were designed to span the predicted off-target regions. QPCR analyses showed no significant off-target effects were detected under the experimental conditions used (S4D and S4E Fig). Together, these results confirm efficient and specific RNAi-mediated gene silencing, supporting subsequent phenotypic analyses.

Early-expressed genes coordinate CvBV replication and late-expressed gene transcription

P143 helicase binds DNA in baculoviruses, and integrase-1 is required for proviral DNA excision in M. demolitor [20,27]. In C. vestalis, early-expressed genes linked to DNA replication and transcriptional regulation (Fig 2A) were detected at high levels by LC-MS/MS, including helicase and integrase-1/-2. RNAi knockdown of ds-helicase, ds-integrase-1, and ds-integrase-2 each led to a significant reduction (>50%) in the abundance of CvBV circles, as determined by the average remaining CvBV abundance ratio across three biological replicates (Fig 3A; raw CvBV copy numbers and p-values are shown in S7 Table). Specifically, ds-helicase and ds-integrase-2 treatments reduced the abundance of 26 circles, while ds-integrase-1 treatment decreased all 30 circles below 50%. Collectively, these results suggest that helicase, integrase-1, and integrase-2 may play conserved roles and broadly influence the abundance of CvBV circles 1–30.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 3. RNAi of early-expressed genes reduces CvBV abundance and downregulates late-expressed gene expression. qPCR analysis showing relative abundance of CvBV circles 1-30 in FA1d female wasps pretreated with ds-helicase, ds-integrase-1, or ds-integrase-2, compared to ds-GFP-injected controls (n = 3 biological replicates).

The y-axis represents the relative CvBV ratio for each sample, calculated as the CvBV copy number in each sample divided by the mean copy number of the three ds-GFP-injected replicates (raw CvBV copy numbers and corresponding raw and adjusted p-values are provided in S7 Table). (B) qPCR analysis of lef-5, lef-9, and p47 expression levels in FA1d female wasps pretreated with ds-GFP or ds-lef-5 (n = 3 biological replicates). (C) qPCR analysis of relative expression levels of eight randomly selected late-expressed viral genes and wasp genes in FA1d female wasps pretreated with ds-GFP, ds-p47, ds-lef-9, or ds-lef-5 (n = 3 biological replicates). Wasp genes dnapolδ and ef1α were used as negative controls. Adjusted p-values (padj) are shown. (D) qPCR analysis showing relative abundance of CvBV circles 1-30 in FA1d female wasps pretreated with ds-p47, ds-lef-9, or ds-lef-5, compared to ds-GFP-injected controls (n = 3 biological replicates). The y-axis represents the ratio of CvBV copy numbers relative to ds-GFP controls, as defined in (A) (raw CvBV copy numbers and corresponding raw and adjusted p-values are provided in S8 Table). (E) TEM images of virion morphogenesis in FA1d female wasps pretreated with ds-GFP or ds-lef-5. White arrows with black outlines indicate mature virions in the lower-left insets of ds-GFP (ii) and ds-lef-5 (ii). (F) Virion density (number per μm²) in the calyx lumen of FA1d female wasps pretreated with ds-GFP or ds-lef-5, as quantified from TEM images (n = 4 biological replicates, measured from a 1 μm² central region selected in each of four independent images). (G) Cross-sectional area (μm²) of individual virions in the calyx lumen of FA1d female wasps pretreated with ds-GFP or ds-lef-5, as quantified from TEM images. A total of n = 50 randomly selected, non-overlapping virions without obvious aggregation were measured in each group. Median values are indicated on the plot. N, nucleus; E, wasp egg; CL, calyx lumen. Scale bar: 5 μm. (A, C, and D) Statistical significance was assessed using two-sided t-tests with Benjamini–Hochberg correction. *padj< 0.05; ** padj< 0.01; *** padj< 0.001; ns, not significant. (B, F, and G) Statistical significance was assessed using two-sided t-tests.

https://doi.org/10.1371/journal.ppat.1014492.g003

Other early-expressed genes include nudivirus-like transcriptional regulators p47, lef-4, lef-9, and lef-5. Given its homology to the baculovirus Autographa californica multiple nucleopolyhedrovirus (AcMNPV) and other systems, lef-4 and lef-9 act as RNA polymerase subunits, while lef-5 initiates late-expressed gene transcription [21,27,33]. RNAi showed that ds-lef-5 strongly suppressed the expression of two other early-expressed genes, p47 and lef-9, and most late-expressed genes (vp39, vlf-1, 38k, pif-0, HzNVorf140–1, HzNVorf140–2, HzNVorf9–1, and PmV) (Fig 3B and 3C). Similarly, ds-p47 and ds-lef-9 treatments also significantly reduced the expression of late-expressed genes, consistent with their roles in transcriptional regulation (Fig 3C). Given the putative role of late-expressed genes as structural components of CvBV virions (including but not limited to vp39) [27], we hypothesized that their reduced expression would result in decreased virion production. As expected, ds-p47 and ds-lef-9 treatments significantly reduced the abundance of most CvBV circles to below 30% (p47: 27 circles; lef-9: 28 circles), whereas ds-lef-5 treatment decreased the abundance of all CvBV circles to below 6% (Fig 3D; raw CvBV copy numbers and p-values are shown in S8 Table). TEM further confirmed severe impairment of virion production after ds-lef-5 injection (Fig 3E and 3F). Compared with ds-GFP controls, virions were rarely observed in either the calyx cell nuclei or the calyx lumen in ds-lef-5–treated wasps. Quantitative analysis of TEM cross-sectional areas further suggested that CvBV virions in ds-lef-5–treated wasps were noticeably smaller than those in the ds-GFP control group (Fig 3G). Together, these results suggest that these early-expressed genes play important roles in CvBV genome abundance and virion formation.

Functional characterization of capsid and envelope genes in CvBV

In M. demolitor, the late-expressed genes are mainly associated with structural proteins and viral assembly-related components [23,27]. In C. vestalis, the number of genes encoding capsid and envelope proteins remains unclear. LC-MS/MS analysis identified vp39 as the most abundant protein, with the highest number of unique peptides in both calyx fluid (R1: 21) and purified virions (R2: 24) (Fig 2A; S3 Table). Twelve additional proteins (PmV, 38k, HzNVorf9–1, HzNVorf9–2, HzNVorf106, 27b, K425_459, 11k, 17a-1, 35a-1, 35a-2, and K425_461) also showed high unique peptide counts and peptide-spectrum matches (PSM ≥ 15) in the R2 sample, supporting their roles as CvBV structural components, though their specific localization within the virion remains unclear.

Previous studies have shown that knockdown of capsid proteins disrupts bracovirus integrity and decreases viral DNA abundance [27,28]. For example, vp39, which encodes a nucleocapsid structural protein in baculoviruses, showed no apparent effect on envelope formation upon knockdown. However, TEM analysis revealed the presence of irregularly assembled virions, including enveloped particles containing aberrant or incomplete nucleocapsids, as well as empty envelopes completely lacking nucleocapsids, which accumulated and filled the nucleus. We hypothesize that knockdown of capsid genes in C. vestalis results in decreased CvBV abundance and aberrant virion formation, which may help distinguish capsid proteins from envelope proteins based on phenotypic differences observed. As expected, TEM analysis revealed severe structural defects in CvBV virions following vp39 knockdown (S5A-C Fig). In the calyx cell nucleus, envelope formation appeared largely unaffected, with envelopes still being normally produced. However, these envelopes rarely contained nucleocapsids. No nucleocapsids were detected near the virogenic stroma, and correspondingly, few to no mature virions were observed in the calyx lumen. Similarly, knockdown of PmV, HzNVorf9–1, HzNVorf9–2, and HzNVorf106 caused comparably severe phenotypes. Envelopes were still present but contained aberrantly assembled nucleocapsids, and the number of mature virions in the calyx lumen was markedly reduced compared with ds-GFP controls (Fig 4A-F and S5D-L Fig). In addition, knockdown of the other three genes (38k, 27b, and K425_459) also affected viral capsid protein accumulation and nucleocapsid formation (Fig 4G-I and S5M-R Fig). However, depletion of the three genes was not sufficient to completely abolish nucleocapsid assembly, as a subset of nucleocapsids was still observed within the nucleus. Consistent with this partial defect, a small number of virions were still observed in the calyx lumen. Furthermore, all capsid gene knockdowns led to significantly reduced virion density in 1 μm² calyx lumen areas, significantly decreased virion cross-sectional area, and a marked global decrease in CvBV DNA abundance (Fig 4J, 4K, and S6 Fig). Specifically, injection of capsid genes dsRNA significantly reduced the average levels of CvBV to 1.9–55.5% of control levels, with vp39 knockdown causing the most severe reduction (raw CvBV copy numbers and p-values are shown in S9 Table). Together, these results indicate that knockdown of capsid genes primarily disrupts nucleocapsid assembly and encapsidation rather than envelope biogenesis, leading to defective virion assembly and a pronounced reduction in CvBV abundance.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 4. CvBV structural proteins are critical for virion integrity.

(A–I) TEM images of virion morphogenesis in FA1d female wasps after RNAi knockdown of capsid genes. (A, D, and G) Overview of calyx cell nuclei in FA1d female wasps pretreated with ds-GFP, ds-PmV, or ds-38k. (B, E, and H) Magnified views of the boxed areas in (A, D, and G), respectively. Insets (black boxes) show regions outlined by white boxes, illustrating the effects of RNAi on virion assembly. (E) Black arrows indicate four types of aberrantly assembled virions: (i) an incomplete nucleocapsid lacking an envelope (a small hollow circle); (ii) a single large, irregularly shaped nucleocapsid enclosed by an envelope (a large hollow circle containing a slightly smaller hollow circle); (iii) multiple incomplete nucleocapsids enclosed by an envelope (a large hollow circle containing several small hollow circles); and (iv) an empty envelope lacking nucleocapsids (a large hollow circle), which represents the most frequently observed phenotype. (H) Black arrows indicate four types of aberrantly assembled virions: (i) a nucleocapsid lacking an envelope (a small electron-dense dot; different from E(i)); (ii) a single large, irregularly shaped nucleocapsid enclosed by an envelope (a large hollow circle containing a slightly smaller hollow circle; same as E(ii)); (iii) a single small incomplete nucleocapsid enclosed by an envelope (a large hollow circle containing a small hollow circle; similar to E(iii)); (iv) a partially assembled virion containing only a nucleocapsid enclosed by a single-membrane envelope (an electron-dense dot within a membrane-bound circle; different from E(iv)). (C, F, and I) TEM images of the calyx lumen of FA1d female wasps pretreated with ds-GFP, ds-PmV, or ds-38k. Black arrows indicate mature virions. (J) Virion density (number per μm²) in the calyx lumen of FA1d female wasps pretreated with ds-GFP, ds-vp39, ds-PmV, ds-HzNVorf9-1, ds-HzNVorf9-2, ds-HzNVorf106, ds-38k, ds-27b, or ds-K425_459, as quantified from TEM images (n = 4 biological replicates, measured from a 1 μm² central region selected in each of four independent images). padj values are shown. (K) Heatmap showing CvBV relative abundance in ovaries of FA1d female wasps pretreated with nine dsRNAs shown in (J) (n = 3 biological replicates). In the heatmap, the color of each cell represents the mean ratio of CvBV copy numbers from three biological replicates relative to the ds-GFP controls (raw CvBV copy numbers and corresponding raw and adjusted p-values are provided in S9 Table). The x-axis represents individual CvBV circles (circles 1–30). Lighter colors indicate a greater reduction in CvBV abundance following knockdown. (L–N) TEM images of virion morphogenesis in FA1d female wasps after RNAi knockdown of envelope gene 11k. (L) Overview of the calyx cell nuclei in FA1d female wasps pretreated with ds-11k. (M) Magnified view of the boxed area shown in (L). White dashed circles highlight regions where multiple nucleocapsids, lacking an envelope, are aggregated together. (N) TEM image of calyx lumen in FA1d female wasps pretreated with ds-11k. White arrows indicate nucleocapsids not enclosed by an envelope. N, nucleus; VS, virogenic stroma; E, wasp egg; CL, calyx lumen. Scale bars: (A, D, G, and L), 5 μm; (B, C, E, F, H, I, M, and N), 0.5 μm; insets of (B, E, and H), 0.25 μm. Statistical significance was assessed using two-sided t-tests with Benjamini–Hochberg correction.

https://doi.org/10.1371/journal.ppat.1014492.g004

In addition to capsid genes, the number and functional roles of envelope genes in CvBV remain largely unexplored. The envelope primarily forms the outer membrane structure that surrounds nucleocapsids. Knockdown of envelope-associated genes may impair the efficiency of nucleocapsid envelopment and, in some cases, also affect viral assembly efficiency and/or the infectivity of nucleocapsids. A previous report showed that envelope gene HzNVorf64 knockdown affects the abundance of a subset of Microplitis demolitor bracovirus (MdBV) circles, rather than causing a global reduction in viral circle abundance [26]. Therefore, we reasoned that CvBV envelope genes can be distinguished from capsid genes by combining TEM-based assessment of virion morphology with absolute qPCR quantification of CvBV circle abundance. Consistent with this expectation, knockdown of 11k resulted in a pronounced accumulation of unenveloped nucleocapsids in the calyx cell nuclei from one-day-old female adult (Fig 4L and 4M). These nucleocapsids frequently appeared in clustered aggregates and were markedly more abundant than those observed in ds-GFP controls. The density of mature virions in the calyx lumen was markedly different from that observed in capsid gene knockdown groups, and showed no obvious difference compared with ds-GFP controls (Fig 4N). Furthermore, RNAi targeting four additional genes (17a-1, 35a-1, 35a-2, and K425_461) produced similar phenotypes in TEM images (S7A-L Fig). Nucleocapsid morphology was largely unchanged, but a small number of nucleocapsids clustered within the nucleus. Some unassembled nucleocapsids were also released into the calyx lumen following cell lysis. Based on TEM images of the calyx lumen, we quantified the number of mature virions and found no obvious difference compared with ds-GFP controls (S7M Fig). Quantification of virion cross-sectional areas further suggested that knockdown of 17a-1 and 35a-1/35a-2 moderately reduced virion size, whereas no obvious size reduction was observed following 11k or K425_461 knockdown (S7N Fig). Similarly, qPCR analysis of CvBV circles 1–30 showed that knockdown of any single gene did not cause a global change in CvBV circle abundance and that only a subset of circles exhibited moderate up- or down-regulation (S7O Fig; raw CvBV copy numbers and p-values are shown in S10 Table). Overall, the TEM phenotypes and qPCR results support a clear distinction between the CvBV envelope and capsid genes. Envelope genes appear to primarily affect nucleocapsid envelopment efficiency, without broadly altering nucleocapsid structure or viral circle abundance.

Functional characterization of CvBV assembly-related genes

Given that vlf-1 is implicated in virion assembly of MdBV [27,28,33], we examined its function and that of its homologs HzNVorf140–1 and HzNVorf140–2 in C. vestalis [31]. RNAi knockdown of vlf-1 caused four major defects in calyx cell nuclei: (i) an unenveloped nucleocapsid in the nucleus; (ii) an envelope enclosing one or more incompletely processed, hollow nucleocapsids; (iii) an incompletely processed, hollow nucleocapsid lacking an envelope; and (iv) an envelope containing a nucleocapsid with an elongated rod-like shape (Fig 5A, S8A, and S8C Fig). These defective virions were also released into the calyx lumen following calyx cell lysis and exhibited variable size (Fig 5B, S8B, and S8D Fig). Similarly, individual knockdown of the two vlf-1 homologs (HzNVorf140–1 and HzNVorf140–2) produced similar phenotypes in the calyx cell nuclei and calyx lumen. Based on TEM images of the calyx lumen, knockdown of each of the three genes caused a significant reduction in virion density, with remaining virions at comparable levels in all knockdowns (Fig 5C). Quantification of virion cross-sectional areas further showed that all three knockdowns significantly reduced virion size (S8E Fig). Similarly, qPCR analysis showed that knockdown of these three genes caused a significant overall reduction in CvBV abundance, without pronounced circle-specific effects (Fig 5D; raw CvBV copy numbers and p-values are shown in S11 Table). These findings indicate that vlf-1 and its homologs are essential for CvBV virion assembly and maturation, with all three genes showing largely comparable effects.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 5. Knockdown of vlf-1, HzNVorf140-1, and HzNVorf140-2 causes significant defects in CvBV virion assembly.

(A) TEM image of calyx cell nuclei in FA1d female wasps pretreated with ds-vlf-1. Black arrows indicate four types of aberrantly assembled virions: (i) an unenveloped nucleocapsid in the nucleus; (ii) an envelope enclosing one or more incompletely processed, hollow nucleocapsids; (iii) an incompletely processed, hollow nucleocapsid lacking an envelope; and (iv) an envelope containing a nucleocapsid with an elongated, rod-like morphology. (B) TEM image of the calyx lumen in FA1d wasps pretreated with ds-vlf-1. Black arrows indicate the same aberrantly assembled virions described in (A). (C) Virion density (number per μm²) in the calyx lumen of FA1d female wasps pretreated with ds-GFP, ds-vlf-1, ds-HzNVorf140-1, or ds-HzNVorf140-2 (n = 4 biological replicates, measured from a 1 μm² central region selected in each of four independent images). The quantified virions included only particles containing both an envelope and a nucleocapsid, including most aberrantly assembled virions but excluding nucleocapsids lacking an envelope. padj values are shown. (D) Heatmap showing CvBV abundance in ovaries of FA1d female wasps pretreated with ds-GFP, ds-vlf-1, ds-HzNVorf140-1, or ds-HzNVorf140-2 (n = 3 biological replicates). In the heatmap, the color of each cell represents the mean ratio of CvBV copy numbers from three biological replicates relative to the ds-GFP controls (raw CvBV copy numbers and corresponding raw and adjusted p-values are provided in S11 Table). The x-axis represents individual CvBV circles (circles 1–30). Lighter colors indicate a greater reduction in CvBV abundance following knockdown. N, nucleus; CL, calyx lumen. Scale bar: 0.5 μm. Statistical significance was assessed using two-sided t-tests with Benjamini–Hochberg correction.

https://doi.org/10.1371/journal.ppat.1014492.g005

Vp91 is essential for CvBV infectivity but dispensable for virion production

In calyx fluid samples, proteomic analysis identified several enriched proteins in addition to structural components and assembly-related factors. One of these proteins, vp91, is recognized as a per os infectivity factor (PIF-8) in AcMNPV [34–36], but its function in bracoviruses of C. vestalis remains unknown. In this study, we first examined the effect of vp91 knockdown on virion morphology using TEM. We found that nucleocapsids and the surrounding envelope structures in calyx cell nuclei were largely unaffected, and the density of virions in the calyx lumen showed no obvious difference compared with ds-GFP controls (S9A-C Fig). Quantification of virion cross-sectional areas likewise revealed no obvious change in virion size following knockdown (S9D Fig). Moreover, qPCR analysis of CvBV circles 1–30 following vp91 knockdown showed no significant changes in viral circle abundance (S9E Fig; raw CvBV copy numbers and p-values are shown in S12 Table), suggesting that vp91 is unlikely to serve as a major structural component or assembly-related factor. To further evaluate whether vp91 influences CvBV infection in the host, we quantified the absolute abundance of two CvBV circular DNAs, CvBV_19 and CvBV_22, in seven P. xylostella tissues within 1 h post-parasitization by ds-vp91-treated wasps, when most of the CvBV DNAs had not yet integrated into the host genome [31]. The abundance of both viral circles was reduced in most host tissues relative to the ds-GFP control group (Fig 6A), although CvBV_22 abundance in the fat body and testis showed no obvious decrease. Meanwhile, we detected CvBV gene expression in P. xylostella tissues parasitized by control or ds-vp91-treated wasps. Given that pif-0 (p74) is a known per os infectivity factor in AcMNPV [27,37], we used knockdown of CvBV pif-0 as a positive control to compare the effects of vp91 knockdown. Knockdown of vp91 markedly reduced the expression of the CvBV virulence genes CvBVre22–6 and CvBVre19–3 [38], across multiple P. xylostella tissues (Fig 6B). Notably, the extent of CvBV gene downregulation varied among different host tissues following vp91 or pif-0 knockdown. In the midgut, knockdown of vp91 resulted in a marked reduction in CvBV expression, ranging from 8.5 to 38.8-fold, whereas in all other tissues, CvBV expression decreased more significantly following pif-0 knockdown. These results indicate that vp91 knockdown reduces CvBV accumulation across most host tissues. Consequently, the decreased expression of CvBVre22–6 and CvBVre19–3 is likely due to reduced viral infection efficiency and lower numbers of CvBV genomes delivered into host tissues. Together, these findings support a role for vp91 as an infectivity factor required for efficient CvBV infection in host tissues, with no apparent role as a major structural component of CvBV virions.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 6. vp91 functions as a viral infectivity factor.

(A) Absolute quantification qPCR analysis of circular CvBV_19 and CvBV_22 copy numbers per ng of genomic DNA across seven P. xylostella tissues. Tissues were collected within 1 h after parasitization by one-day-old female wasps to minimize potential interference from viral genome integration into the host genome at later stages of parasitization. HEM, hemocytes; MG, midgut; FB, fat body; CNS, central nervous system; TES, testis; MT, Malpighian tubules; EPI, epidermis. n = 3 biological replicates. (B) qPCR analysis of the expression levels of two CvBV virulence genes in various tissues (same as in (A)) of P. xylostella at 24 h after parasitization by wasps pretreated with ds-GFP, ds-pif-0, or ds-vp91 (n = 3 biological replicates). p-values, two-sided t-tests.

https://doi.org/10.1371/journal.ppat.1014492.g006

Insect rearing and sampling

C. vestalis and its host P. xylostella were maintained under standard laboratory conditions [52]. Developmental stages of C. vestalis were identified based on timing and morphology [32]. Ovaries were collected from females at four pupal stages (day 1: Ov1d, day 2: Ov2d, day 3: Ov3d, day 4: Ov4d) and one-day-old adults (OvFA1d), with three biological replicates per stage (30, 20, 6, 6, and 6 females, respectively) for RNA-seq.

CvBV nudivirus-like genes identification

Bracovirus morphogenesis-related genes from Cotesia congregata were retrieved from published sources via ParWaspDB (https://bipaa.genouest.org/is/parwaspdb/) [22]. These served as queries for blastp searches against predicted coding sequences of the C. vestalis genome to identify putative CvBV nudivirus-like genes. De novo transcriptome assembly of all RNA-seq data was performed with Trinity v2.12.0 [53], and coding regions were predicted using TransDecoder v5.5.0. The resulting coding sequences were compared with the bracovirus dataset to refine CvBV gene identification. Genes within 10 kb were considered clusters.

RNA-seq library construction, sequencing, and analyses

Total RNA from 15 ovary samples was used to construct RNA-seq libraries with Illumina kits (Illumina, USA) and sequenced as 150 bp paired-end reads on a HiSeq 2500 platform. Paired-end reads from C. vestalis ovary libraries were mapped to the reference genome (GenBank accession No. GCA_054085095.1) using HISAT2 v2.2.1 [54] with default settings. Gene-level read counts were generated from the mapped reads and used for differential expression analysis with DESeq2 [55]. Expression levels for visualization were quantified and normalized as transcripts per million (TPM). A combined protein dataset, including annotated genome proteins and newly predicted transcriptome proteins, was used for mass spectrometry-based annotation.

Calyx fluid collection and CvBV virions purification

Two CvBV virion samples were isolated from C. vestalis ovaries for proteomic analysis: R1, representing calyx fluid, and R2, consisting of purified CvBV virions. The R2 sample was divided into two technical replicates (R2-1 and R2-2) to assess reproducibility for LC-MS/MS analysis (S3A Fig). Calyx fluid from 1,000 (R1) or 1,200 (R2-1 or R2-2) pairs of one- or two-day-old female wasps was collected and filtered through a 0.22 μm syringe filter (Merck Millipore, USA) to remove debris and eggs. The R1 filtrate was centrifuged (20,000 × g, 5 min, 4 °C) and washed three times with PBS to obtain a crude virion extract. The R2 sample underwent further purification by sucrose gradient centrifugation (25–65%, 100,000 × g, 2 h, 4 °C) as described previously [56]. The virion-containing fraction (45–55% sucrose) was collected, pelleted (20,814 × g, 2 h), and resuspended in ~400 μl PBS.

Proteomic analysis of CvBV virions

Both calyx fluid (R1) and purified CvBV virions (R2) were used for mass spectrometry-based protein identification. In-gel trypsin digestion was performed overnight at 37 °C with 3 µg trypsin (Promega, USA) in 100 mM NH₄HCO₃. Peptides were desalted using C18 cartridges (Sigma-Aldrich, USA), vacuum-dried, and reconstituted in 0.1% formic acid. LC-MS/MS analysis was conducted on an LTQ Orbitrap Elite mass spectrometer coupled to an Easy-nLC II system (Thermo Fisher Scientific, USA). Peptides were loaded onto a trap column (75 µm × 2 cm, C18) and separated on an analytical column (50 µm × 15 cm, C18) using a five-step linear gradient from 3% to 90% acetonitrile (0.1% formic acid) over 150 min at 250 nL/min. Data were acquired in data-dependent mode with the top 20 ions selected for CID MS/MS. Proteins were identified by searching against a custom database of predicted ORFs from the ovary transcriptome datasets, and only proteins supported by at least two unique peptide spectrum matches were retained.

Gene cloning and verification

RT-PCR was performed with KOD polymerase (TOYOBO, Japan) to validate full-length or partial ORFs of CvBV nudivirus-like genes. For selected genes, mRNA expression was validated by qPCR using total RNA extracted from pooled ovaries (6–20 per sample) collected at four developmental stages (Ov2d, Ov3d, Ov4d, and OvFA1d). qPCR primers listed in S13 Table. Ribosomal protein Cv_18s rRNA (GenBank JX399880.1) and Cv_β-tubulin (GenBank MT459787) were used as reference genes for C. vestalis, while Px_β-tubulin (GenBank EU127912) and Px_β-actin (GenBank NM_001309101) served as internal controls for P. xylostella. We used the comparative 2^−ΔΔCT method to calculate the relative gene expression levels [57].

RNAi of selected genes

Gene-specific primers with T7 promoter adaptors (S13 Table) were used to amplify 300–900 bp templates for dsRNA synthesis. dsRNA was synthesized using the MegaScript RNAi Kit (Ambion, USA). A 0.1 μL dsRNA solution (2–5 μg/μL) was injected into one-day-old female pupae using a Fentojet Express microinjector (Eppendorf, Germany) and a Narishige micromanipulator, as previously optimized for RNAi efficiency [58]. Controls were injected with non-specific GFP dsRNA. One-day-old adults were sampled, and ovaries were dissected with ophthalmic scissors for downstream analysis.

Absolute quantification of CvBV circles by qPCR

To quantify 30 CvBV circles, genomic DNA was extracted from the ovaries of pupae and adult females using the Wizard Genomic DNA Purification Kit (Promega, USA) following the manufacturer’s instructions. Primers and absolute quantification curves are described in [31]. For absolute quantification of CvBV abundance in different P. xylostella tissues, genomic DNA was isolated from hemocytes, midgut, fat body, central nervous system, testis, Malpighian tubules, and epidermis collected within 1 h after parasitization by one-day-old female wasps to minimize potential interference from viral genome integration at later stages of parasitization. The copy numbers of circular CvBV_19 and CvBV_22 were quantified by absolute qPCR and normalized to total genomic DNA input, and the abundance of each viral circle was calculated as copy number per ng of genomic DNA.

TEM observation

Ovaries from one- or two-day-old female wasps were dissected and fixed in 2.5% glutaraldehyde in PBS at 4 °C for 8–12 h. Non-essential ovarian tubules were removed, and calyx regions containing CvBV particles were post-fixed in 1% osmium tetroxide at room temperature for 1 h. Samples were dehydrated through graded ethanol, embedded in resin, and ultrathin sections were cut with a LEICA EM UC7 microtome. Sections were double-stained with lead citrate and Uranyless (DeltaMicroscopy, France) for 5–10 min and examined on a Hitachi H7650 electron microscope at 100 kV (Bio-ultrastructure Analysis Lab, Zhejiang University). At least three females were analyzed per dsRNA treatment to ensure consistent observations.

Statistics

Graphs and statistical analyses were performed using GraphPad Prism 8 (GraphPad, USA). Two-sided Student’s t-tests were used to assess differences between groups. Multiple-testing correction using the Benjamini–Hochberg method was applied when multiple hypotheses were tested within the same experimental framework. Adjusted p-values < 0.05 were considered statistically significant. Exact n, p-values and adjusted p-values are provided in figure legends. Data are presented as mean ± s.d. unless stated otherwise. Heatmaps were generated with the pheatmap v1.0.12 in R v4.1.0.