Detection of replicative Kashmir Bee Virus and Black Queen Cell Virus in Asian hornet Vespa velutina (Lepelieter 1836) in Italy

Information concerning the pathogenic role of honey bee viruses in invasive species are still scarce. The aim of this investigation was to assess the presence of several honey bee viruses, such as Black Queen Cell Virus (BQCV), Kashmir Bee Virus (KBV), Slow Paralysis Virus (SPV), Sac Brood Virus (SBV), Israeli Acute Paralysis Virus (IAPV), Acute Bee Paralysis Virus (ABPV), Chronic Bee Paralysis Virus (CBPV), in Vespa velutina specimens collected in Italy during 2017. Results of this investigation indicate that among pathogens, replicative form of KBV and BQCV were detected, assessing the spillover effect of both these viruses from managed honey bees to hornets.


Material and Methods
Sampling. V. velutina specimens were sampled in Liguria region (North West Italy) from late April to mid-November 2017. From 30 th April to 30 th May (named early-season) fifteen workers were sampled in Airole area (43°52′26.8″N 7°33′00.3″E). On the 24 th July (named mid-season) fifteen workers were collected in Bordighera (43°46′45.2″N 7°39′50.3″E), Sanremo (43°49′24.8″N 7°44′24.2″E) and Dolceacqua (43°51′25.5″N 7°37′20.3″E) areas. On 15 th November (named late-season), six newly-emerged specimens (three fermales and three males) were sampled in Ventimiglia within the botanic garden "Giardini Hanbury" (43°46′57.9″N 7°33′14.7″E). For the early-season and mid-season sampling, the hornets have been collected in front of the apiaries during their predatory activity. While, the late-season samples were collected directely from their nest. The caste of the three newly-emerged females collected in the late season was determined by performing the wet weight measure 57 . Total RNA extraction. Total RNA was extracted from 6 pools (each composed of 5 individuals), three pools belonging to early-season for a total of 15 specimens (Airole) and three (5 specimens each) from mid-season sampling (Sanremo, Bordighera and Dolceacqua respectively). RNA from 6 late-season specimens (Ventimiglia) was extracted from each individual. This different extraction approach was used in order to differentiate the viral presence between the only three male and three female specimens and to evaluate the potential infection in the reproductive caste (gynes and drones).
Total RNA extraction procedure was performed as previously described 56 . Briefly, total RNA was obtained by RNeasy Mini Kit (Qiagen, Hilden, Germany) following tissue homogenization using a TissueLyser II (Qiagen) carried for 3 minutes at 25 Hz. Samples were eluted in 30 µl RNase-free water, quantified by Qubit using the RNA HS assay kit (Life-Technologies, Stafford, USA) and stored in aliquots at −80 °C until use. As a negative control, RNA obtained from a fly (Musca domestica) was used. PCR assays. All extracted RNAs were retro-transcribed by M-MLV Reverse Transcriptase (Invitrogen, Carlsbad, USA), using a blend of oligo-d (T) primers and random hexamers following manufacturer's instruction. Five microliters of the obtained cDNAs were used as a template for the PCR reactions, which were carried out with HotStarTaqPlus Polymerase Mix (Qiagen).
Primers to amplify viral genomes of the honey bee viruses here investigated are reported in Table 1. Samples giving positive results to PCR were sequenced (BMR genomics, Padova) and results analysed by BLAST 58 . Phylogenetic analysis was performed by the Maximum Likelihood method based on the Tamura-Nei model using MEGA software 59 .
Strand-specific RT-PCR. The replication activities of detected viruses were evaluated by strand specific RT-PCRs using specific primers as previously described 60 . All cDNAs were amplified by PCR for the related viral target. Amplicons were visualised on a 2% agarose gel and confirmed by sequence analysis (BMR Genomics, Padova).
Negative staining electron microscopy (nsEM). The three females sampled in the late season were used for nsEM analyses. The extracts were prepared and treated using the method commonly applied for honey bees 61 . Each individual was placed in 2.4 ml 0.001 M potassium buffer (PB), pH 6.7, containing 0,2% sodium diethyldithiocarbamate (DIECA) to prevent melanization and then mechanically homogenized (Ultraturrax -Ika Werk, Staufen, Germany). The extracts were emulsified with 0.3 ml chloroform and 0.3 ml diethyl ether, and then cleared by low-speed centrifugation at 4500 g for 30 min. The supernatants were separated and again centrifuged at 9500 g for 30 min. Next, 100 μl of each supernatant was ultracentrifuged with an Airfuge (Beckman, Indianapolis, USA), operating at 21 psi 82000 g for 15 min, and fitted with an A100 rotor holding six 175 μl test tubes equipped with specific plastic adapters which permit to directly pelleting virions onto 3 mm carbon-coated Formvar copper grids. The grids were negatively stained with 2% sodium phosphotungstate (NaPT) at pH 6.8 for 90 seconds and observed with a FEI Tecnai G2 Biotwin transmission electron microscope (FEI Company, Hillsboro, USA) operating at 85 kV at 16500-43000 magnifications.

Results
Caste determination of late-season newly-emerged females. Among the six late-season specimens (three females and three males) sampled in Ventimiglia, the wet weight of females resulted in 420 mg, 487 mg and 517 mg respectively, heavier than the 386.4 mg recorded as highest wet weight in V. velutina workers collected in July. Such results indicate these individuals as gynes 57 . Negative staining electron microscopy (nsEM). The nsEM analysis highlighted the presence of few scattered particles, morphologically closely resembling virions in all the three specimens examined. Observed particles were roundish, around 35 nm in diameter (Fig. 3) and mainly empty revealing a sharp rim; thus, their shape and size were compatible with those described for Dicistroviruses. Since these morphological characters   www.nature.com/scientificreports www.nature.com/scientificreports/ are the same for KBV and BQCV but no specific antisera were available to perform immunoelectron microscopy (IEM), it was not possible to confirm the nature of the particles observed as virus and to exactly identify them.

Discussion
In view of a global control strategy for the recent spread of the invasive V. velutina in European countries 15,62,63 , studies on the relationship among pathogens and the predator assume great significance. In this study, the presence of viruses previously detected (1993 and 2010) by Lavazza and colleagues 64 54,56 .
Data obtained by NsEM analysis, indicate that the particles observed could be effectively referable to dicistrovirus (KBV and BQCV), as indicated by sequence analysis. However, due to their overlapping morphology, it was not possible to exactly define which dicistrovirus was present. In fact, only by using IEM with specific antisera it might be possible to immune-aggregate particles, thus obtaining both an enrichment of the sample and a true viral species identification within the same genus. Nevertheless, since molecular methods can only indicate the presence of genomic material but do not prove the existence of mature "complete" virions, the detection by nsEM of particles with morphological pattern typical of either viruses could be considered a further indication of the   www.nature.com/scientificreports www.nature.com/scientificreports/ presence of replicative forms. This hypothesis is enforced by the evidence that such nsEM analysis was performed on newly-emerged specimens that cannot have had yet the possibility to be contaminated with bee viruses during predation, but only by being fed during their larval stage with infected honey bees foragers captured by hornet workers.
Sequence analysis of KBV and BQCV from V. velutina indicates high identity rates to viral sequences identified in A. mellifera. Considering the predatory activity of V. velutina versus A. mellifera, this genetic similarity suggests a possible horizontal transmission route of these pathogens by ingestion of infected honey bees 18,23,[26][27][28] . Moreover, the molecular phylogenetic analysis of KBV and BQCV from V. velutina identifies a close relationship to recent European A. mellifera virus sequences, therefore excluding the involvement of viruses of Asiatic origin.
Moreover, the relatively low number of particles observed, that, according to the established detection limits of the nsEM Airfuge method here applied, should be around 10 3 -10 4 particles/ml, is suggestive of subclinical infection, i.e. the situation normally detected also in normoreactive/healthy honey bees 81 . In fact, by using the www.nature.com/scientificreports www.nature.com/scientificreports/  www.nature.com/scientificreports www.nature.com/scientificreports/ same nsEM Airfuge method in comparison with quantitative RT-qPCR and a sandwich Mab-based ELISA for the detection of DWV it was previously shown that the viral load in clinically affected A. mellifera is usually considerably higher i.e. >10 6 viral copies 82 .
In Italy, only three KBV infected apiaries have been previously detected: one in Tuscany region nearby Liguria, and two in Lazio region 38 . At the light of the detection of KBV within this study, a retrospective analysis performed on 2015 archived V. velutina workers maintained at −80 °C, collected in Liguria region in Taggia area (43°51′53.1″N; 7°50′57.2″E), indicated a previous circulation of KBV (GenBank -MK238796). A possible explanation for KBV presence in this area could be the high rate of migratory beekeeping activity from other Italian sites to Liguria. It is likely that these "introduced" migratory colonies were asymptomatically infected by KBV, that could be transmitted between colonies during migration and foraging activity related to honey flow 83 , thus increasing KBV presence and making it accessible to predators such as V. velutina.
The positivity of only newly-emerged females collected in November could be related to the higher incidence of KBV in honey bees in the late autumn season 84 . Similarly, the presence of BQCV in Asian hornets both in mid-season and in late-season samples could be related to the high frequency of this virus in honey bee during summer 84 .
The BQCV was detected exclusively in gynes, while KBV in both male and female individuals. The small dimension of the late-season samples does not allow formulating a conclusive hypothesis on the sex-distribution of KBV and BQCV infection. Finally, variation of incidence of KBV and BQCV detected in V. velutina throughout the season is compatible with the increasing trend of infection usually found in the honey bee colonies [84][85][86][87][88] . Therefore, in late summer/early autumn there is a higher infection rate of V. velutina in larvae following their feeding with infected honey bee thoraxes.
In conclusion, the results of this investigation indicate that the honey bee pathogens KBV and BQCV could successfully infect V. velutina, although in an asymptomatic form. Additional studies should be performed in vitro to clarify if KBV and BQCV infection could have clinical evidence in V. velutina, as well as the possibility that these viruses could be transmitted vertically, in order to discuss the hypothetical role of honey bee viruses in invasive species.