A comparison of biological characteristics of three strains of Chinese sacbrood virus in Apis cerana

We selected and sequenced the entire genomes of three strains of Chinese sacbrood virus (CSBV): LNQY-2008 (isolated in Qingyuan, Liaoning Province), SXYL-2015 (isolated in Yulin, Shanxi Province), and JLCBS-2014 (isolated in Changbaishan, Jilin Province), by VP1 amino acid (aa) analysis. These strains are endemic in China and infect Apis cerana. Nucleotide sequences, deduced amino acid sequences, genetic backgrounds, and other molecular biological characteristics were analysed. We also examined sensitivity of these virus strains to temperature, pH, and organic solvents, as well as to other physicochemical properties. On the basis of these observations, we compared pathogenicity and tested cross-immunogenicity and protective immunity, using antisera raised against each of the three strains. Our results showed that compared with SXYL-2015, LNQY-2008 has a 10-aa deletion and 3-aa deletion (positions 282–291 and 299–301, respectively), whereas JLCBS-2014 has a 17-aa deletion (positions 284–300). However, the three strains showed no obvious differences in physicochemical properties or pathogenicity. Moreover, there was immune cross-reactivity among the antisera raised against the different strains, implying good protective effects of such antisera. The present study should significantly advance the understanding of the pathogenesis of Chinese sacbrood disease, and offers insights into comprehensive prevention and treatment of, as well as possible protection from, the disease by means of an antiserum.


Screening of representative strains by multiple sequence alignments of VP1. Sixteen CSBV VP1
genes of isolates were sequenced (Table 1) Fig. 1). Group I did not have amino acid deletions and was less mutated; therefore, we used the latest isolate SXYL-2015 as a representative strain; the other strains have not been isolated in the same area since 2012 (Table 1). In group II, the amino acid sequence of strain FZ-2012 at positions 87, 192, 195, 202, 242, and 277 was mutated from G, Y, A, Q, K, and T to C, N, V, H, Q, and A, respectively, whereas the other strains are highly conserved. Thus, we used the first isolated and less mutated LNQY-2008 as a representative strain, whose VP1 homology with that of other strains is more than 99.4%, and which reappeared in Qingyuan Liaoning in 2009 and 2010, and in Tieling, Liaoning; Siping, Jilin in 2010 (Table 1). In group III, the strains showed less variation; accordingly, we used JLCBS-2014 (first isolated by our laboratory), which shows less variation as a representative strain, whose VP1 homology with other strains is more than 98.1%, and which reappeared in Changbaishan, Jilin in 2015 (Table 1). The deduced amino acid sequences of mammalian picornaviruses and insect picornalike viruses were then aligned and compared. The results revealed that the structural proteins are located at the 5′ end and the non-structural proteins at the 3′ end 16 . The helicase domains A, B, and C 17 are located between amino acid positions 1353 and 1490 in LNQY-2008, SXYL-2015, and JLCBS-2014. This region includes highly conserved amino acids within the first two domains, GxxGxGKS and Qx5DD in domains A and B; however, the C domain appears to be the least conserved, containing only three of the six residues potentially associated with this site (Fig. 2). The equivalent of the conserved cysteine protease motif GxCG and the putative substrate-binding residues in the GxHxxG domains were identified within the protease domains in the deduced amino acid sequences of the viruses 18 Table 2. Nucleotide and deduced amino acid sequences homology (%) among the three CSBV representative strains and the reference sequences. Homology (%) of the deduced amino acid sequences for the coding regions among the three CSBV representative strains and the reference sequences. Next, we determined the amino acid sequence homology among the SBV strains. Our results (  (Fig. 4).

Analysis of molecular biological characteristics.
A phylogenetic tree was constructed on the basis of the high sequence variability among the partial amino acid sequences of the VP1 region obtained from China, Korea, Vietnam, India, Astralia and the United Kingdom to illustrate the probable genetic relations among the selected SBV strains. Phylogenetic analysis showed that group III and the strains isolated in Korea (AcSBV-Kor and AmSBV-Kor19, GenBank: JQ390592) and Vietnam  (AcSBV-Viet1, and AcSBV-Viet2, GenBank: KM884991) could be classified into a clade. Group II was clustered into a separate subgroup except for FZ-2012. SXYL-2015, SXnor1-2012, and BJ-2012 were clustered into a subgroup, and HBQHD-2012, LNSZ-2011, GZ-2000, and GZ-2002 were also clustered into a separate subgroup, but the latter formed a closely related cluster with groups II and III (Fig. 5).
Histopathological analysis (hematoxylin and eosin [H&E] staining; Fig. 7) revealed that infection of larvae by one of the three CSBV strains caused lesions in the internal organs and tissues of the larvae. The lesions caused by each of these three strains appeared similar. Normal tissue cells after H&E staining were intact, with small round nuclei; the clearance between the epidermis and dermis was small, with few signs of granular liquid. Three days after the inoculation, histopathological analysis showed increased clearance between the epidermis and dermis, disappearance of a portion of the dermis, and deformation of the cells and nuclei. Six days after the inoculation, the gap between the epidermis and dermis increased further and was filled with a watery fluid, and became increasingly hollow. Additionally, the dermis gradually disappeared. The shapes of the cells and nuclei become irregular, and they even disintegrated in some instances. Various organelles disintegrated, resulting in cell lysis. Moreover, all the larvae inoculated with the same copy numbers of one of the three strains developed the same signs of the disease (Fig. 7). Larvae infected with one of the three CSBV strains failed to pupate, and ecdysial fluid accumulated beneath their unshed skin. Larvae changed in colour from white to pale, or even dark yellow, and died. Shortly afterwards, they dried out, forming dark brown gondola-shaped scales.
Comparative analysis of physicochemical properties. As shown in Fig. 8, the mortality rates of larvae infected by each CSBV strain incubated at 50 °C, 60 °C, or 70 °C were not significantly different (P > 0.05), whereas infected larvae incubated at 75 °C and 80 °C and the virus-free control showed significantly lower mortality (P < 0.01), indicating that CSBVs can be inactivated by incubation at 75 °C for 1 h. By contrast, pH 3, ethyl ether, and chloroform seemed to have no effect on viral activity because larvae infected with the viruses exposed to these conditions showed 100% mortality. Moreover, there were no significant differences in the resistance of the three CSBV strains when exposed to high temperatures, pH 3, ethyl ether, or chloroform (P > 0.05).
Analysis of immunogenicity. The three strains of purified CSBV have four major proteins, with estimated molecular weights of 30.5, 31.5, 37.8, and 44.2 kDa (Fig. 9). The results of agar gel immunodiffusion (AGID) assays ( Fig. 10) revealed that there were three kinds of antigens and antisera, each with a clean lane. This result indicated cross-immunogenicity among the three representative strains and cross-reactivity among the three antisera, whereas the immunoprecipitation band was not observed for non-immune serum or saline. In the virus neutralisation assay (Fig. 11), the three strains of the virus were incubated with the three types of antisera and fed to healthy larvae. The larvae showed normal pupation after 4 days. No significant differences were observed among the groups (P > 0.05). Larvae inoculated with the viruses that were neutralised with non-immune serum did not show normal pupation and eventually died. Therefore, the immunisation with different CSBV strains seems to offer cross-protection.

Discussion
The incidence of CSBV infection has increased considerably in the past few years, and the virus is seriously threatening apiculture. Currently, CSBV research is focused on genetic characterisation, cell culture, immunisation with structural proteins, and treatment 14,[19][20][21] . Studies have shown that cross-species transmission is more frequent for RNA viruses than for other pathogens of the honeybee [22][23][24][25] . Since 2008, we have monitored the prevalence of CSBV in China and obtained 16 strains of CSBV from different regions and time points (Table 1). Sequence analyses of the viral VP1 genes indicate that there are three kinds of the VP1 gene. In this study, we compared three strains from China (LNQY-2008, SXYL-2015, and JLCBS-2014) in terms of molecular biological characteristics,   Analysis of temperature resistance shows that the mortality of larvae infected by each CSBV strain preincubated at 50 °C, 60 °C, and 70 °C was not significantly different (P > 0.05), whereas larvae infected with the virus pre-incubated at 75 °C or 80 °C showed significantly lower mortality (P < 0.01), indicating that CSBVs can be inactivated by incubation at 75 °C for 1 h.  Genetic exchange (by either recombination or reassortment) plays an important role in evolution by rapidly increasing variation and was suggested to have evolved to offset fitness losses 26 . Some studies have shown that most of the genomic sequences diverged considerably in the VP1 region 27 . In this study, we compared some of the SBV genome reported by Reddy et al. 27 and that published in GenBank for our strains. We found deletions or insertions near the VP1 gene region in structural and non-structural proteins. Our analysis revealed that amino acid deletions or insertions are common phenomena in SBV and may be associated with regional differences and host species.
The phylogenetic tree of VP1 revealed that strains in group III and the strains isolated in Korea and Vietnam tend to be grouped together, suggesting that strains in group III might have originated in Korea in 2010 and then spread to China and Vietnam. Amino acid sequence analysis also showed less variation among group III strains. The strains in group II independently form a clade except for FZ-2012, which is closely related to group III because the mutated amino acids in FZ-2012 VP1 (C, N, V, H, Q, and A) are the same as those in strains of group III. In group I, although SXYL-2015, SXnor1-2012, and BJ-2012 were isolated from the Chinese honeybee A. cerana, those strains formed a closely related cluster with the strains originating overseas, such as AcSBV-IndK1A, AcSBV-IndII-2 (GenBank: JX270795), AmSBV-Australia (GenBank: KJ629183), AmSBV-Kor21, and SBV-UK. We deduced that SXYL-2015, SXnor1-2012, and BJ-2012 probably originated from SBV infecting Apis mellifera, and then infected A. cerana, indicating that SBV can cause interspecies infections, and these data are consistent with Gong's results 28 . By contrast, HBQHD-2012, LNSZ-2011, GZ-2000, and GZ-2002 form a closely related cluster with groups II and III; this finding shows that these isolates originated from GZ-2000, which was first isolated from the Chinese honeybee A. cerana in Guangzhou in 2000. VP1 variability may affect the biological characteristics of CSBV.
The high pathogenicity of CSBV towards A. cerana has been the focus of intensive research. By comparing different genotypes of three CSBV strains in terms of pathogenicity and the pathological damage to A. cerana, we found that the mortality rate of 2-day-old A. cerana inoculated with one of the three CSBV strains rises with the increasing CSBV copy number. When the number of copies per larva reached 1.25 × 10 7 , the mortality rate was 100%. To compare the three viral strains in terms of characteristic clinical signs and pathological changes in bee larvae, we chose this 100% lethal minimal gradient. The three CSBV strains, when inoculated into 2-day-old A. cerana at the same copy number, showed no significant differences in lethality, clinical signs, or pathological changes. It should be noted that artificial breeding of bee larvae cannot ensure 100% survival, and we used only a small number of selected samples, individual differences among larvae and other possible reasons may explain why the larval mortality was not entirely consistent in the three repeated experiments. Nonetheless, all of the results showed a positive correlation within a certain range between the mortality of infected larvae and the inoculated virus copy number. In addition, the characteristic clinical signs and pathological changes in the larvae were the same. Infected larvae were examined microscopically after H&E staining and showed irregularities in the shapes of cells and nuclei after infection by one of the three strains. A liquid-filled cavity was observed between the epidermis and dermis of diseased larvae. Compared to that of uninfected larvae, the body surface of infected larvae was swollen 3-4 days after the inoculation. This result may be attributed to the large gap between the epidermis and dermis. When the infection progressed, we observed disintegrated cells and various broken organelles. This finding is consistent with the typical clinical signs observed, including swelling of the larval surface and increased accumulation of cyst fluid. Furthermore, because all honeybee viruses, including CSBV, have no suitable culture system, we could not accurately measure the 50% lethal dose of CSBV in this study. Therefore, we did not carry out more in-depth and meticulous research on the entire course of the disease and conducted only a preliminary study on CSBV pathogenicity, using the number of virus copies as a quantitation standard. In the future, we will further explore the methods for determination of the 50% lethal dose, to carry out a more in-depth comprehensive comparative study on the pathogenesis of CSBV infection.
Comparative analysis of the physicochemical properties of the three strains of CSBV yielded no significant differences in sensitivity to temperature, low pH (3.0), ethyl ether, and chloroform (P > 0.05). This finding indicates that CSBVs are non-enveloped viruses that are resistant to ether and chloroform. Larvae infected by CSBV that was exposed to ethyl ether and chloroform were unable to pupate. The same results were obtained upon exposure of the viruses to pH 3, indicating that these virus strains are not inactivated by an acid. After virus inactivation at 75 °C for 1 h, the viral titre and infectivity decreased, and the inoculated larvae could then morph into pupae. In this study, we also found that a few larvae could morph into pupae after inoculation with a lethal dose of CSBVs (1.25 × 10 7 copies/larva) that were incubated at 50 °C, 60 °C, or 70 °C. The mortality rate did not reach 100%. We presumed that biological activity of CSBV probably was affected with the increase in temperature from 50 °C to 75 °C. In addition, because of the difficulty with rearing of larvae, we could not avoid deaths of the larvae caused by differences in individual, indoor, and outdoor environments and in nutrient composition, machine operation, or other reasons; thus, ~25% of the larvae are expected to die of natural causes. To detect the death of the larvae caused by infection with CSBV preincubated at 75 °C or 80 °C, we performed a regression test on dead larvae, and it showed that there is no significant difference in the mortality rate between the reseeded group and virus-free control group (P > 0.05). These results can serve as a reference for further research on physicochemical characteristics of this virus.
SDS-PAGE analysis of CSBVs purified by caesium chloride gradient centrifugation revealed four proteins with estimated molecular weights of 30.5, 31.5, 37.8, and 44.2 kDa, in agreement with reports that the CSBV structural protein is composed of four proteins with molecular weights 30.5, 31.5, 37.8, and 44.2 kDa 12,29 . This finding indicates that we obtained purified viral proteins, and thus avoided contamination with bee proteins in the immunological experiments. The AGID assay is a rapid method that can be performed in a veterinary diagnostic laboratory 30 . Clear precipitation bands were obtained for combinations of each of the three viral antigens and three kinds of antisera, whereas immunoprecipitation was not detected with non-immune serum and saline control. These data provide preliminary evidence of immune cross-reactivity among these three CSBV strains and the three kinds of antisera. The virus neutralisation assay supported this hypothesis. We carried out the neutralisation reaction with a custom-made antiserum and three strains of the virus, respectively. The larvae pupated normally after inoculation by feeding. Conversely, after inoculation with the virus incubated with non-immune serum, the bee larvae did not pupate. These results further corroborate the cross-protection after immunisation with different strains of CSBV and provide experimental evidence for the production of vaccines using any of these three CSBV strains because all three can confer protective immunity against CSBV.
In this study, we found that although there are sequence-specific features in each of the three genotypes, there are no significant differences in pathogenicity, physicochemical properties, or immunogenicity. Our results lay the foundation for more in-depth research on the properties of CSBV, including epidemic patterns, mechanisms of infection, and preventive measures. The specific function of the high-variability protein structure is a good topic for future studies.

Materials and Methods
Sample collection. A total of 185 A. cerana larval samples (each collected sample included five larvae from a single colony) were obtained in China in 2008-2015 (Table 1). In all larvae, CSBV infection was detected by reverse-transcriptase PCR (RT-PCR) 31 with VP1 primers (F: 5′ -GCGGATCCATGGATAAACCGAAGGATATAAG-3′ , R: 5′-GCAAGCTTTTATTGTACGCGCGGTAAATA-3′ ). The infection rate in a test-positive sample was calculated by means of the following formula: infection rate (%) = (total number of infected larvae ÷ total number of larvae in test-positive sample) × 100.
We selected one test-positive larva randomly for sequencing from the same colony at the same time.
Screening representative strains by multiple sequence alignments of the VP1 gene. Using VP1 as a target gene, we carried out multiple gene sequence comparisons for all the CSBV isolates and reference strains in GenBank using the Clustal W method in the MegAlign software (DNA STAR, Inc., Madison, WI, USA). We named the samples after the strain that originated in the same region and showed 100% homology, then submitted the data to GenBank (CSBV uniformly renamed, first isolated geographic location + years). According to the sequence comparison results, we selected LNQY-2008, SXYL-2015, and JLCBS-2014 as representative strains.
Isolation and purification of the representative strains. For each of the three strains (LNQY-2008: ~70.0% proportion of the infected larvae, SXYL-2015: ~80.0% proportion of the infected larvae, and JLCBS-2014: ~73.3% proportion of the infected larvae), 50 infected A. cerana larvae were collected. After weighing, larvae were completely homogenised in sterile water (1.5-fold amount, by weight) using a pestle and mortar. CSBV purification was performed by cesium chloride gradient centrifugation, according to Ma's method 12,32 . The supernatant was then passed through a 0.45-μ m cell filter first and then through a 0.22-μ m cell filter. The filtrate was then fed to the second instar of A. cerana larvae for virus passage, and 50 diseased larvae were taken from each culture plate after 8 days. The virus was isolated and purified again by the above method. Next, virus suspension was analyzed by the RT-PCR method 31 for the following viruses: black queen cell virus (BQCV) 33 , acute bee paralysis virus (ABPV) 33 , chronic bee paralysis virus (CBPV) 34  Israeli acute paralysis virus (IAPV) 36 , and CSBV 31 . Three virus suspensions, after we proved that they did not contain other viruses in addition to CSBV, were stored at − 80 °C until use.
As previously described 12 , we designed the primers in this work on the basis of the nucleotide sequences of GZ-2002 and SBV-UK. These primers were used to prepare full-length, single-stranded cDNA of LNQY-2008, SXYL-2015, and JLCBS-2014 genomes. The cDNAs were amplified by PCR (30 cycles, annealing at 50-55 °C for 45 s, and elongation at 72 °C for 50-60 s). The 3′ end was cloned by the 3′ -RACE method (Clontech). The PCR-amplified cDNAs were cloned into the pMD 18-T vector (Takara Biotechnology Co., Ltd., Dalian, China). The plasmids were then used to transform Escherichia coli DH5α cells (Takara Biotechnology Co., Ltd.). The plasmids were extracted using the Plasmid Extraction Kit (Axygen Biotechnology Co., Ltd.).
Nucleotide sequencing was performed by Sangon Biotech Co., Ltd. The nucleotide sequences from all of the fragments were assembled to build a continuous complete sequence by means of the DNASTAR software. Sequence analysis was performed by the Clustal W method in the MegAlign software. Phylogenetic trees were constructed by the neighbour-joining method (p = distances), and up to 1000 bootstrapping replicates were used in the MEGA 5.0 software for VP1 region amino acid sequences obtained from China, Korea, Vietnam, India, Australia, and the United Kingdom.
Each larva was fed 20 μ L of a virus suspension mixed with an equal amount of basic larval diet 38 (BLD; 50% royal jelly, 37% sterile water, 6% glucose, 6% fructose, 1% yeast extract) at 95% relative humidity and 34 °C. The virus-free control was fed 20 μ L of sterile water with an equal amount of BLD. BLD was used subsequently for daily feeding. The clinical signs in each group of larvae were examined and recorded every day until death of larvae. All the larvae that were orally inoculated in this study were analysed by the RT-PCR method 31 for BQCV 33 , ABPV 33 , CBPV 34 , DWV 34 , KBV 35 , IAPV 36 , and CSBV 31 .
Comparative analysis of physicochemical properties. According to the conventional method 39 , equal amounts of each of the three virus suspensions and sterile water were individually incubated at varying temperatures (50 °C, 60 °C, 70 °C, 75 °C, or 80 °C) in a water bath for 1 h to carry out the temperature resistance experiment, and each group contained 15 larvae. Purified virus suspensions (pH adjusted to 3.0 using an HCl solution, 0.1 M, pH 3.0) were incubated at a constant temperature of 37 °C in a water tank for 12 h, followed by adjusting the pH of the solution to 7.2 with a 5.6% NaHCO 3 solution. Next, sensitivity to pH 3.0 was measured. To analyse sensitivity to chloroform, pure chloroform was added to purified virus (at a final concentration of 4.8%); the mixture was rocked gently at 4 °C for 10 min and centrifuged at 500 rpm for 5 min. The supernatant was extracted to analyse the viral activity. To analyse sensitivity to ethyl ether, each virus suspension was mixed with ethyl ether, and placed on an oscillator shocking for 60 min, and centrifuged at 2000 rpm for 20 min. We used a capillary with appropriate percussion and absorbed the virus for ether sensitivity test. Then, according to the above method, a 2-day-old A. cerana larva was fed 20 μ L of the virus suspension (1.25 × 10 7 copies) along with an equal amount of BLD, and was examined after 6 days to determine whether the larvae developed the signs of CSD (to detect CSBV infection). The virus-free control larvae were fed 20 μ L of sterile water with an equal amount of BLD. A total of four treatments (replicated thrice) were carried out.

Analysis of immunogenicity.
Four-to 6-week-old female pathogen-free BALB/c mice were randomly subdivided into four groups with five mice in each group: negative control, LNQY-2008, SXYL-2015, and JLCBS-2014. All animal procedures were approved by the Ethics Committee of Liaoning Medical University. Immunisations were performed via three intraperitoneal injections at 15-day intervals. All antigens were emulsified with complete Freund's adjuvant for the first immunisation and with incomplete Freund's adjuvant for subsequent immunisations; PBS served as a negative control. The amount of each antigen used for each immunisation was 20 μ g/mouse. The specific antibodies in the serum samples were detected by an enzyme-linked immunosorbent assay (ELISA) and the absorbance at 490 nm was measured on an ELISA plate reader 40 . Serum samples were separated from blood by centrifugation at 3600 rpm for 15 min and then stored frozen until use. Three kinds of CSBV proteins were separated by SDS-PAGE. The proteins were resolved on 12% SDS-polyacrylamide gels using standard protocols 41 .
One gram of agarose and 8 g of NaCl were added to 100 mL of phosphate buffer (0.01 M, pH 7.2); the mixture was shaken well, and microwaved for 2 min to prepare an agar solution, which was slightly cooled and poured into Petri dishes (90 mm in diameter; 20-22 mL of agar per plate) and were allowed to solidify. Seven wells were made in the agar plates for each experimental group; the central hole was loaded with one of the three CSBV strains, and the surrounding wells were loaded with antisera against LNQY-2008, SXYL-2015, or JLCBS-2014 or non-immune serum. The AGID assay was performed for 24-48 h at 37 °C in a humidified chamber, and the results were examined thereafter.
Statistical Analysis. Data on the mortality rates expressed as percentages were normalised using arcsine square root transformation and were subjected to repeated-measures analysis of variance with the Bonferroni adjusted post hoc pair test, in the SPSS 17.0 software (SPSS Inc., Chicago, IL, USA). If Mauchly's test of sphericity showed no significant differences in the repeated-measures data (P > 0.05), then normal one-way analysis of variance was performed. Differences were considered significant at P < 0.05.