A new gene family (BAPs) of Cotesia bracovirus induces apoptosis of host hemocytes

ABSTRACT Polydnaviruses (PDVs), obligatory symbionts with parasitoid wasps, function as host immune suppressors and growth and development regulator. PDVs can induce host haemocyte apoptosis, but the underlying mechanism remains largely unknown. Here, we provided evidence that, during the early stages of parasitism, the activated Cotesia vestalis bracovirus (CvBV) reduced the overall number of host haemocytes by inducing apoptosis. We found that one haemocyte-highly expressed CvBV gene, CvBV-26-4, could induce haemocyte apoptosis. Further analyses showed that CvBV-26-4 has four homologs from other Cotesia bracoviruses and BV from wasps in the genus Glyptapanteles, and all four of them possessed a similar structure containing 3 copies of a well-conserved motif (Gly-Tyr-Pro-Tyr, GYPY). Mass spectrometry analysis revealed that CvBV-26-4 was secreted into plasma by haemocytes and then degraded into peptides that induced the apoptosis of haemocytes. Moreover, ectopic expression of CvBV-26-4 caused fly haemocyte apoptosis and increased the susceptibility of flies to bacteria. Based on this research, a new family of bracovirus genes, Bracovirus apoptosis-inducing proteins (BAPs), was proposed. Furthermore, it was discovered that the development of wasp larvae was affected when the function of CvBV BAP was obstructed in the parasitized hosts. The results of our study indicate that the BAP gene family from the bracoviruses group is crucial for immunosuppression during the early stages of parasitism.


Introduction
The endoparasitoid wasps have adapted various methods to adjust their host's physiology and development to guarantee the survival of wasp offspring [1,2]. Haemocytes play important roles in the immune responses of insects against parasites [3,4]. Therefore, parasitoid wasps induce a series of changes in host haemocytes to avoid haemocyte-mediated immune responses by polydnavirus (PDV) and venom, which are injected into host larvae along with parasitoid eggs during oviposition. For example, Cotesia vestalis parasitization inhibits recognition and encapsulation of the host haemocytes [5]; Diadegma semiclausum parasitization affects haematopoietic regulation and haemocytemediated immune responses in host larvae [6]; Cotesia kariyai parasitization reduces the total haemocyte count (THC) of the hosts [7]; The parasitization of Pimpla turionellae and C. vestalis alter the haemocyte population of their hosts and even the functions of the host haemocytes [8,9]. Among the changes of haemocyte-mediated immune responses post parasitization, apoptosis of haemocytes plays an important role, which was induced by PDVs [10][11][12].
This paper aims to reveal the functional PDV gene-(s) responsible for host haemocyte apoptosis. Herein, it is demonstrated that C. vestalis reduced the host THC by apoptosis induced by C. vestalis BV (CvBV). The global transcriptome statistics of CvBV genes in host haemocytes showed that CvBV-26-4 is secreted by haemocytes and degraded into functional peptides that induced haemocyte apoptosis. Further analysis displayed that CvBV-26-4 and its analogous proteins from other BVs all contain 3 well-conserved motifs, Gly-Tyr-Pro-Tyr and function in inducing apoptosis. Hence, a new PDV gene family, Bracovirus Apoptosisinducing Protein (BAP), was suggested for these proteins. Ectopic expression of CvBV BAP induced apoptosis of fly haemocytes and increased susceptibility of male flies to Staphylococcus aureus. Knockdown of BAP had a deleterious impact on the development of C. vestalis progeny. Thus, we inferred that the BV BAP family is important for host immunosuppression during parasitization.

Materials and methods
Insects P. xylostella and C. vestalis were maintained in our lab as described in a previous study [43]. Drosophila melanogaster stocks were raised on standard cornmeal/ yeast/agar medium at 18°C. In this study, W 1118 was used as the wild-type stock and Bloomington stock Hml-GAL4 (BS#8700) was also used.

Haemocyte viability and counts
About 30 parasitized or non-parasitized P. xylostella haemocytes were collected. Cell viability and number counts were determined by trypan blue staining using Cell Counter (Countstar, Shanghai, China). Hemocyte counting was done 4 h post parasitization (pp). Cell viability was determined at 6 h, 2 h, and 4 h pp. 16 repeats (collections of 30 P. xylostella haemocytes) of each group were determined.

Virus and venom collection
From the female wasps, CvBV virions were collected and incapacitated using UV light as described in a previous study [44]. C. vestalis venom gland was dissected and the stored venom protein was extracted. We defined the quantity of CvBV or venom from one female as one female equivalent (FE).

Injection
The third instar larvae of P. xylostella (2 h post-ecdysis) and male flies (7 days post-eclosion) were used for injection of different fluids as previously described [44].

Apoptosis detection
Haemocytes collected from 30 P. xylostella after different treatments or Drosophila larvae were suspended in 00 μL phosphate buffered saline (PBS, pH 7.4) and plated into the Lab-Tek II Chambered Cover glass (Thermo Fisher, MA, USA) for 30 min at room temperature. Apoptotic haemocytes were visualized using a terminal deoxynucleotidyl transferase dUTP nick end labelling (TUNEL) assay kit (Vazyme, Nanjing, China) according to guidelines. And 4, 6-diamidino-2-phenylindole (DAPI) (Sigma-Aldrich) was used for nuclei counterstaining. The images were captured using a confocal microscope, Zeiss LSM 800. To quantify the percentage of apoptotic haemocytes from P. xylostella and D. melanogaster, the labelled nuclei of > 3 images were counted using the Image J software.

Caspase activity assay
After 4 h post treatment, the haemocytes were collected from P. xylostella. And the Caspase-Glo® 3/7 assay Kit (Promega, USA) was used to perform the Caspase activity assay. Each group contains 15 biological replicates.

RNA extraction
To analyse gene expression at different times pp, we extracted the total RNA of P. xylostella at 0. 5

mRNA sequencing and data analysis
The haemocytes of P. xylostella after being parasitized for 6 h, 2 h, and 4 h were collected and subject to transcriptome sequencing. In detail, each group contained 3 biological replicates in the form of 200 parasitized P. xylostella. For each library preparation, a total of 1 μg RNA was used as input material. And the VAHTS mRNA-seq v2 Library Prep Kit (Vazyme, Nanjing, China) was used for library generation. Clustering of the index-coded samples was performed on a cBot Cluster Generation System (Illumina, USA) according to guidelines. Then, the sequencing was performed on an Illumina HiSeq X Ten platform using the 150 bp paired-end module.

Quantitative PCR
Quantitative PCR (qPCR) was used to validate the transcriptome data and study the expression profiles of selected genes in varying developmental stages and tissues. qPCR was performed as previously described [48]. β-actin gene (GenBank accession No. AB282645) and β-tubulin gene (GenBank accession No. EU127912) of P. xylostella were used as internal controls. The relative expression levels were calculated using the 2 −ΔΔCt method [48].

Sequence analysis
A homolog search was carried out against other species using BLASTP (http://www.ncbi.nlm.nih.gov/), with CvBV-26-4 as the seed sequence. The signal peptide of CvBV-26-4 was predicted by SIGNALP 6.0 (http://www. cbs.dtu.dk/services/SignalP/). MEGA 7.0 software was used to perform the alignment analyses. The figures of the sequence analysis were created using Jalview [49].

RNA silencing
The T7 RiboMAXTM Express kit (Promega, USA) was used to synthesize double-stranded RNA (dsRNA) of the CvBV-26-4 gene and GFP gene. Primers used for dsRNA synthesis are displayed in Table S7. Before parasitization, 500 ng of dsCvBV-26-4 was injected into the middle 3 rd instar larvae of P. xylostella, and dsGFP was used as the negative control. The efficacy of the RNA interference was determined by qPCR at 4 h pp.

Western blot
A polyclonal antiserum against CvBV-26-4 was generated by the Shanghai Sangon Biotechnology Company using a peptide antigen for antiserum development. Epitopes are predicted by the GenScript Optimum Antigen design tool. The peptide sequence synthesized for the CvBV-26-4 antiserum development was YPYPSSDGSSGYFS.
Haemocytes and plasma of P. xylostella were separated by centrifugation. After adding SDS PAGE Loading Buffer, the samples were boiled for 0 min to denature the proteins. The western blot was performed as previously described [21]. In this study, polyclonal antiserum against CvBV BAP (1:500) and actin (1:3000), and secondary antibody against rabbit IgG (1:3000) were used. The ECL western blotting substrate (Promega, USA) was used to visualize the bands.

Mass spectrometry of plasma peptides
To confirm whether CvBV-26-4 protein was secreted outside haemocytes and degraded into peptides, we collected peptides from the plasma of P. xylostella.~0 μl of haemolymph from 4 h pp P. xylostella larvae was added to 00 μl anticoagulant buffer [44]. The diluted haemolymph was centrifuged at 000 g at 4 for 5 min to obtain the plasma in the supernatant. The peptides from the plasma were isolated with a 3 kDa Millipore ultrafiltration device. The peptide mixtures were subjected to mass spectrometry analysis using a Thermo Scientific Easy nanoLC 1000 (Thermo Fisher Scientific, MA, USA). Peptidespectrum matches (PSMs) were set to 10.

Ectopic expression of CvBV-26-4 in drosophila melanogaster
TheORF sequence of CvBV-26-4 was cloned into the pUAST-attb vector [51] to obtain the CvBV BAP transgenic lines. Integrase-mediated insertion by phiC31 into the attP2 landing-site locus on the 3 rd chromosome was used to obtain the transgenic Drosophila line carrying the UAS-BAP gene.
The survival assay was conducted as described [52]. Briefly, each male fly at 7 days post-eclosion was injected with 40 nL of the Staphylococcus aureus resuspension using an Eppendorf Femtojet (Eppendorf, Germany). The death was recorded every 2 h; flies that died within the first 6 h were removed. Flies were kept at 25°C and were transferred to new food every day.

Determination of C. vestalis development
To detect the effects of CvBV-26-4 on C. vestalis offspring, 0.1 μg each of CvBV BAP antiserum and rabbit IgG (negative control) were injected into the middle 3 rd instar P. xylostella, which were then parasitized by C. vestalis. The body length of wasp pupae was measured one day after pupation. The rate of wasp pupa formation was recorded. The eclosion of C. vestalis was also noted following wasp larva pupation.

Statistical analysis
All statistical analyses were performed using SPSS 20.0 software. Data were presented as means ± SD and were examined by the One-way ANOVA and Tukey's test, with a p-value of 0.05 as the significance threshold. Different survival curves were compared using log rank tests. Chi-squared tests were used to assess the rate of wasp pupa formation and eclosion.

Parasitization reduced the total count of host haemocytes through apoptosis
We analysed the total number of haemocytes from parasitized host larvae at 4 h pp by trypan blue staining using Cell Counter. The results showed that the overall number of haemocytes of parasitized host larvae was less than that of the control group (Figure 1(a)). As the number of haemocytes decreased within 4 h pp, the survival ratio of haemocytes should be affected earlier after parasitization. As expected, the survival ratio of haemocytes started to be affected at 2 h pp (Figure 1(b)). The survival ratio of haemocytes from parasitized host larvae at 4 h pp was below 70% (Figure 1(b)). Additionally, the detection of apoptosis at 4 h pp by the TUNEL assay revealed many positive haemocytes from parasitized host larvae, whereas most haemocytes from non-parasitized host larvae were negative (Figure 1(c) and S1a). Furthermore, we detected the caspase activity of haemocytes from P. xylostella at 4 h pp and the result showed that parasitization induced the caspase activity of host haemocytes ( Figure S2a).

The activated CvBV-induced apoptosis of host haemocytes
The injection of CvBV or venom into host larvae during wasp oviposition might induce apoptosis of host haemocytes. To evaluate the effects of CvBV and venom on the apoptosis of host haemocytes pp, we detected the apoptotic haemocytes of P. xylostella following CvBV and venom injection by a TUNEL assay. The results showed that 0.05 FE CvBV, a dosage close to that administered during real parasitism [44], induced apoptosis of haemocytes while 0.05 FE venom or PBS could not (Figure 2 and S1b). Further analyses indicated that apoptosis of host haemocytes was not induced by 0.05 FE inactivated CvBV ( Figure 2 and S1b), which suggested that apoptosis was induced by virulent CvBV instead of its capsid proteins.

Global transcriptome analysis of CvBV genes in P. xylostella haemocytes
RNA-seq analyses of host haemocytes were performed to elucidate genes or gene families of CvBV that induce apoptosis of host haemocytes. The target gene or genes should be abundantly expressed before 2 h pp, at which time the survival ratio of haemocytes is affected (Figure 1(b)). The Illumina HiSeq X Ten platform was used to sequence the mRNA of haemocytes from host larvae at 6 h, 2 h, and 4 h pp to obtain the gene expression profiles. A total of 537,693,406 clean reads were obtained from 9 cDNA libraries (Table S8). The lowest number of clean reads obtained from sample 6 h-3 nearly reached 55 million, and the highest number of reads obtained from sample 4 h-1 was more than 67 million. A total of 16,810,694 reads (3.13%) corresponding to the CvBV genome were obtained from 9 libraries (Table S1).
The dynamic changes of CvBV gene expression profiles at different times pp were analysed. Among 157 predicted CvBV genes [26], 140 were detected in haemocytes at 6 h, 2 h, and 4 h pp. Their FPKMs are shown in Tables S9, S10, and S11. Global transcriptome statistics of CvBV gene transcripts showed that the relative expression levels of various CvBV genes differed at different times pp (Figure 3(a-c)). Based on the differences between gene expression levels in haemocytes, CvBV genes were classified into 4 categories, high (FPKM≥1000), medium (1000>FPKM≥100), low (100>FPKM≥10), and marginal (10>FPKM>0). To validate the results analysed from transcriptomes, 16 CvBV genes were randomly selected across the observed range of expression for qPCR analysis of the cDNA samples from haemocytes used for sequencing. These qPCR analyses confirmed the trends observed from transcriptome analyses (Figure 3(d-F)). The specific number of CvBV genes expressed in host haemocytes at 6 h, 2 h, and 4 h pp was 129, 124, and 133, respectively ( Figure S3a). There were 14 CvBV genes expressed in haemocytes at a high level within the observed 4 h pp ( Figure S3b).
Furthermore, the expression levels of CvBV-26-4 in parasitized host larvae at different times were determined with the qPCR analysis. Our analyses showed that CvBV-  was expressed abundantly at 4 h pp and rapidly decreased by 8 h pp (Figure 5(a)). However, six more sample time points within 4 h pp were refined to determine the transcript profiles of CvBV-26-4 at the earlier stage. CvBV-26-4 expression level increased rapidly at 4 h pp ( Figure 5(b)). qPCR analysis showed that the transcript level of CvBV-26-4 was much higher in the haemocytes at 4 h pp than that in other tissues ( Figure 5(c)), suggesting that CvBV-26-4 plays an important role in the apoptosis of host haemocytes during the early stages of parasitism. To further test whether the apoptosis was caused by CvBV-26-4, gene silencing was conducted via dsRNA injection, while it was discovered that neither the transcript level of CvBV-26-4 nor its protein content was affected ( Figure 6  (a,b)and S5). Then, prior to parasitization, we interfered with CvBV-26-4 via injecting the CvBV-26-4 antiserum, an injection of rabbit IgG that served as a negative control. We detected fewer TUNEL positive haemocytes after injection of the CvBV-26-4 antiserum compared with negative control ( Figure 6(c,d)), which suggests that CvBV-26-4 indeed caused the apoptosis of host haemocytes.

CvBV-26-4 homologs induced apoptosis of P. xylostella haemocytes
When using BLASTP to conduct the homolog search, four homologs of CvBV-26-4 were identified: two homologs from Cotesia BV (GenBank accession No. CCQ19291.1 and CCQ71098.1) and two homologs from the genus Glyptapanteles (GenBank accession No. ACE75128.1 and ABK57046.1), BV-carrying wasps. We aligned the amino acid sequences of these homologs with CvBV-26-4 to clarify whether there are common structural features. The alignment showed that the selected homologs possessed the same signal peptide and 3 well-conserved Gly-Tyr-Pro-Tyr (GYPY) motifs (Figure 7(a)). To assess whether the two homologous proteins from Cotesia BV have a similar function to CvBV-26-4, 10 4 pfu baculoviruses modified with the insertion of GFP, CvBV-26-4, CcBV_CCQ71098.1, and CsKBV_CCQ19291.1 were injected into middle 3 rd instar P. xylostella, respectively. The expression of CvBV-26-4 was confirmed by western blotting ( Figure S6). Host haemocytes at 4 h pi were observed for apoptosis detection and the results showed that CvBV-26-4, CcBV_CCQ71098.1, and CsKBV_CCQ19291.1 can induce apoptosis of haemocytes while GFP cannot (Figure 7(b) and S1c). With common structural features and a similar function, these proteins were therefore named Bracovirus Apoptosis-inducing Proteins (BAPs).

CvBV BAP peptide induced apoptosis of P. xylostella haemocytes
As CvBV BAP (CvBV-26-4) contains a signal peptide, it was hypothesized that CvBV BAP may be secreted and function outside of haemocytes. In the host haemocytes and plasma, CvBV BAP protein content was detected at 6 h, 2 h, and 4 h pp. The CvBV BAP proteins in the host haemocytes and plasma were detected at 6 h pp, accumulated rapidly by 2 h pp, and were maintained at a high level at 4 h pp (Figure 8(a) and S7). We wonder if the CvBV BAP proteins could be degraded into peptides in host plasma but not in host haemocytes. We collected peptides of host plasma at 4 h pp for mass spectrometry. One peptide of CvBV BAP (peptide-spectrum matches, PSMs = 14) was identified (Figure 8(b)). The CvBV BAP peptide contained 18 amino acids, SSGYSSSPLNKGIGFGYP, which is at the C terminus of the CvBV BAP protein. Apoptosis detection of host haemocytes at 4 h pi of BAP peptide showed that CvBV BAP peptide can induce apoptosis of haemocytes (Figure 8(c) and S1d). And the caspase activity of P. xylostella also increased post injection of the BAP peptide ( Figure S2b). Injection of the CvBV BAP antiserum can rescue the apoptosis haemocytes from parasitized P. xylostella ( Figure 6(d)), but injection of CvBV BAP antiserum can not rescue the apoptosis of haemocytes from parasitized P. xylostella when co-injected with the BAP short active peptide (Figure 8(d) and S1d).

CvBV BAP induced apoptosis of D. melanogaster haemocytes
To confirm the function of CvBV BAP proteins in nonadapted hosts, D. melanogaster, we used the GAL4/UAS binary expression system [53]. A D. melanogaster transgenic cell line carrying a UAS transgene encoding the CvBV BAP protein was produced. Then, we used a haemocyte specific expression driver, Hml-GAL4, to express UAS-BAP in D. melanogaster haemocytes [54]. Morphological observation showed that haemocytes from the 3 rd instar Hml>BAP larvae rarely keep normal forms (Figure 9(a)). The aberrant morphology phenotype was caused by apoptosis (Figure 9(b)). In addition, a survival assay was used to determine the effect of CvBV BAP on D. melanogaster haemocytes. The susceptibility to S. aureus of Hml>BAP flies was increased compared to controls (Figure 9(c) and S1e).

Effects of CvBV BAP on parasitization
For detecting the effect of CvBV BAP on parasitization, the CvBV BAP antiserum was injected into the middle 3 rd instar larvae of P. xylostella which were then parasitized by C. vestalis. The body lengths of the wasp pupae, the wasp pupa formation, and the eclosion of C. vestalis were analysed to evaluate the wasp development. The results showed that the body lengths of wasp pupae were significantly decreased post-injection of BAP antiserum compared to the control group because of the blocked function of CvBV BAPs (Figure 10(a)). Injection of CvBV BAP antiserum reduced the percentage of wasp pupa formation (Figure 10(b)) as well as the eclosion rate of C. vestalis (Figure 10(c)).

Discussion
The total number of haemocytes including granular cells and plasmatocytes decreased significantly in 4 th instar host larvae parasitized by C. vestalis [9]. This study found that parasitization affected the THC of P. xylostella much earlier in development. The total number of haemocytes from the late 3 rd host larvae at 4 h pp decreased and the survival ratio of haemocytes was affected at 2 h pp. It was similar to the reports indicating that the number of host haemocytes decreased during the early stages of parasitism by C. melanoscela [55], Campoletis sonorensis [56], C. kariyai [7], and Pimpla turionellae [8]. The THC of P. xylostella larvae was reduced because of apoptosis caused by parasitization, which was consistent with that found in Pseudoplusia includes [11] and Spodoptera litura [12]. Inducing apoptosis of host haemocytes is an important strategy for immunosuppression.
C. vestalis has three important wasp-associated factors which influence the physiology and development of host larvae, including CvBV, venom, and teratocytes [57]. Teratocytes are released into the host when the wasp egg hatches [58], while venom and PDV function earlier. Our results of different injections showed that the activated   Error bars indicate ± SD. Differences among samples were tested with tukey-test (**: significant difference between samples p<0.01). Rate of wasp pupa formation (b) and eclosion rate of C. vestalis (c) in 3 rd instar P. xylostella larvae injected with CvBV-26-4 antibody or rabbit IgG before parasitization. Differences among samples were tested with the chi-squared test. (*: significant difference between samples p <.
CvBV induced apoptosis of host haemocytes. In the CvBV genome, 157 ORFs were identified [26]. RNA-seq analyses of the host haemocyte transcriptomes showed that CvBV-26-4 might be involved in the induction of apoptosis of host haemocytes. It is not easy to silence CvBV-26-4 via injection of dsRNA, so we injected its antiserum to block functional CvBV-26-4. As expected, apoptotic haemocytes were reduced by nearly 75% after injection of the CvBV antiserum, which indicated that CvBV-26-4 was one of the key genes inducing apoptosis during the early stage of parasitism.
CvBV-26-4 homologs were found in two Cotesia BVs and two Glyptapanteles wasps. The homologs from CcBV and CsKBV also induced apoptosis of P. xylostella haemocytes. Therefore, we named these proteins as Bracovirus Apoptosis-inducing Proteins (BAPs). These BAP genes may belong to a new gene family of BVs. It is not surprising that BAPs were found only in Glyptapanteles and Cotesia wasps because Glyptapanteles is the sister genus to Cotesia ( Figure S8) [59], which suggests that other variants may exist given the species diversity in these two genera. We proposed that BAPs might have originated between 54 Mya and 17 Mya based on the phylogeny of the microgsterid lineage ( Figure S8).
The transcript level of CvBV BAP in the haemocytes of parasitized P. xylostella at 6 h pp was higher than that at 2 h pp, but BAP protein content in the parasitized host haemocytes was most abundant at 2 h pp when the survival ratio of haemocytes started to be affected, which suggests that BAP proteins were accumulated. An unexpected finding was that CvBV BAP degraded into a functional peptide in host plasma. PDV proteins secreted by haemocytes into plasma could affect a greater number of haemocytes and possibly other tissues. For example, CvBV BAP peptides may be a ligand for some receptors, which can induce apoptosis. Since CvBV BAP peptide was detected in the plasma (Figure 8(b)), but not in the haemocytes (mass spectrometry analysis, data not shown), and the precursor of BAP has a predicted signal peptide, we hypothesized that the precursor of BAP was secreted into plasma and degraded into the active forms. The antiserum made against a different part of the BAP protein may bind BAP secreted into the haemolymph to prevent the processing of BAP into the active forms, which would explain why the antiserum is able to inhibit the protein's function (Figure 6(c,d)). In our study, the level of baculovirus infection is not high, but Bac-CvBV-26-4-induced apoptosis was observed in a large number of host haemocytes, which confirms our hypothesis. It is worth noting that a fraction of CvBV BAP may get into plasma due to cell lysis for more apoptotic haemocytes post parasitization. Further work is required to confirm that the degradation happens in the haemolymph and is necessary for function, to determine which proteases cleave CvBV BAP, and to identify the receptor of the CvBV BAP peptide.
CvBV BAP can also induce apoptosis of Drosophila haemocytes, which suggests that CvBV BAP could affect some conserved factors in the apoptosis pathway. The development of wasp larvae in the host was affected when the function of CvBV BAP was blocked. The host immunity response increased at the early stage of parasitism for the blocking of CvBV BAP, which results in shorter body lengths of wasp pupae, a lower percentage of wasp pupation, and a reduced rate of pupa to adult eclosion.
In summary, a new gene family from BVs was identified: the BAP family. Genes from the BAP family encoded proteins containing GYPY motifs, as well as proteins that induced apoptosis of P. xylostella haemocytes. CvBV BAP was secreted into host haemocoel and degraded into functional peptides to induce apoptosis. The BAP gene family is important for the successful parasitization of Cotesia wasps for its function in immunosuppression.

Data availability statement
The raw data for mRNA sequencing of P. xylostella haemocytes are deposited at SRA database of NCBI at https://www. ncbi.nlm.nih.gov, with accession number SRR10863199~SRR10863201.