Bees are a species-rich group, with more than 20,000 species worldwide [1, 2]. Social bees such as honey bees and bumble bees are well researched and documented, but most species of bees are solitary or only primitively eusocial [1]. Diverse bee communities are important pollinators of flowering plants and provide ecosystem services for both wild and cultivated plants [3]. However, pathogens and parasites drive the decline of many bee populations, together with other factors including pesticides, changes in land use and management intensity, climate change, and invasive alien species [4]. In addition, the introduction of pathogens and parasites from commercial bees into wild and managed bee populations contributes to their decline [5, 6]. This study aimed to discover the pathogens from wild bees in the genus of Osmia, an important spring and summer pollinator, and several Osmia species have been developed as manageable orchard, berry, and seed crop pollinators [7].
To address this issue, 200 samples including wild bees or pupa collected in hollow reeds from ten prefectures were enrolled in this study. Among those, total of 26 newly eclosed Osmia bees from Nagano (6 bees), Fukushima (5 bees) and Ibaraki (15 bees) collected from 2019 to 2021 were used for further analyses. Due to the difficulty in identifying species by morphology, we confirmed the species of the collected wild bees through amplification of the mitochondrial gene cytochrome c oxidase subunit I (COI) following the method in previous report [8]. The nucleotide sequences of 6 bees from Nagano, 1 bee from Fukushima and 15 bees from Ibaraki showed 99.7-100% identity with the sequence from Osmia cornifrons (O. cornifrons) isolate OC0656 (GenBank accession no. EU726549), while the remaining 4 bees from Fukushima showed 99.85-100% identity with the sequence from O. taurus AKBW18-47 (GenBank accession no. LC500084).
Subsequently, attempt was made to unveil previously unknown sequences by high throughput sequencing approaches established in our laboratory with a few modifications [9]. In brief, after depletion of ribosomal RNA, cDNA library of each individual bee was constructed followed by high throughput sequencing on a MiSeq bench-top sequencer (Illumina) with MiSeq v3 Reagent Kit (150 cycles). The sequence reads were trimmed by using the default settings on a CLC Genomics Workbench 12.0.3 (CLC Bio) and were assembled using de novo assembly commands in the CLC software. The generated contigs were analyzed using the BLAST program (National Center for Biotechnology Information). Among them, novel contig sequences of approximately 14,200 nucleotides (nt) were obtained from three out of four O. taurus bees collected in Fukushima prefecture. The sequences were similar to the genomic sequence of an unsegmented chuvirus, Scaldis River bee virus (SRBV), previously detected in O. cornuta [10]. Hence, we named the novel virus as Osmia-associated bee chuvirus (OABV), with extensions of 49, 50, and 52, to indicate the source of the respective contigs.
To obtain the terminal sequences of OABV, RNA from bee no. 52 was reverse-transcribed into Rapid amplification of cDNA ends (RACE)-ready first-strand cDNA with SMARTer RACE 5′/3′ Kit (TaKaRa Bio). For 5′-RACE, the gene-specific primers (GSPs) used for primary and nested PCR were 5′-GGTATTTGCTCCATCTGGGTTTGACCTCCGACT-3′ and 5′-GAGCCTTGTACCCTATACTCAGTTCCTTCATGAGCT-3′, respectively. For 3′-RACE, a polyadenylated tail was added to tested RNA using A-Plus Poly(A) polymerase (CELLSCRIPT) prior to first-strand 3′-cDNA synthesis. The GSPs used was 5′- TGCATGCAAGTTACAGGGAGCATCTTGCTGTGT-3′. PCR was performed for 25 cycles of amplification (94 °C for 30 s, 65 °C for 30 s, 72°C for 120 s) using KOD FX Neo (ToYoBo), SeqAmp DNA Polymerase (TaKaRa Bio) for 5′ and 3′-RACE, respectively. The PCR products were subjected to direct sequencing. We obtained additional 22 and 15 nucleotide sequence at 5′ and 3′ terminus, respectively. Ultimately, the whole genome sequence of OABV-52 in the length of 14,236 nucleotides was confirmed and deposited to GenBank along with the nearly complete genome sequences of OABV-49, 50 under accession numbers LC697316 (no. 52), LC697314 (no. 49) and LC697315 (no. 50), respectively.
Genome organization of OABVs was similar to SRBV, a negative-sense RNA virus, composed G, N, and L genes that encode glycoprotein (GP, 3,129 nt), nucleoprotein (NP, 1,980 nt), and RNA-dependent RNA polymerase (RdRp, 7,470 nt), respectively (Fig. 1B). The nucleotide sequences of the GP, NP, and RdRp of OABV-52 were 69%, 70%, and 75% identity to those of SRBV, respectively. On the other hand, the putative amino acid sequences of the GP, NP, and RdRp of OABV-52 shared 67%, 78%, and 83% identity with those of SRBV, respectively. In addition to the three genes aforementioned, the known chuviruses SRBV, Shayang fly virus 1, and Taiyuan leafhopper virus carry a fourth predicted open reading frame between those for the nucleoprotein and glycoprotein [10,11], which was absent in OABV (Fig. 1A).
Next, phylogenetic analysis was conducted by using the neighbor-joining method with bootstrap values (1000 replicates) in MEGA 11 [12]. As illustrated in Fig. 2, the amino acid sequences of OABV GP were closely related to SRBV and to the hymenopteran chu-related virus (OKIAV123) identified from Dioxys cincta [13] (with 67% amino acid identity). The NP tree indicated that the 3 OABVs were related to SRBV, OKIAV122 [13] from Chelostoma florisomne (shares 78% amino acid identity), and OKIAV123 (shares 75% amino acid identity). Of note, in both phylogenetic trees, OABVs formed an independent subcluster.
To confirm the positive rate of OABV among the 26 Osmia bees, conventional PCR was performed using the primer set targeting RdRp gene: F-5′-GAGCTAGTCCAGGTGAAGAAAG-3′ and R-5′-CCTCTGTCGTTGGGAAGATAAG-3′ by PrimeScript One Step RT-PCR Kit Ver.2 following the instruction of manufacturer. As summarized in Table 1, The OABV was detected in three out of four O. taurus bees collected in Fukushima prefecture, while none of the O. cornifrons bees was detected positive.
Given its morphological similarity to O. cornifrons, O. taurus originated in Asia, appears to have been introduced unintentionally into the United States [14], and these two Asian species have recently naturalized in the Mid-Atlantic region. Since 2003, six species of mason bee native to the United States have been declining significantly, while the exotic O. taurus has increased by 800%. It suggests the introduced species might be competing with native species [15]. It has been reported pathogen spillover from honey bees to not only other bees, but also to a broad range of hymenopteran species that live in close contact with these bees [16]. Nevertheless, the studies of geographical distribution of spillover mostly were originated from America, Europe and New Zealand, while the reports from other countries were less frequent [16].
In the current study, a novel chuvirus, designated OABV, was detected from O. taurus, but not from O. cornifrons. Phylogenetically, OABV is closely related to SRBV and other viral strains (OKIAV122, OKIAV123) identified from European counties. Study of bee pathogens in Asian countries is scarce, and hence whether OABV is an endemic pathogen in Japan remains unclear. Of note, since the bees was eclosed in a laboratory incubator, OABV is likely an endogenous pathogen in pupa. However, at present stage, we cannot conclude that O. taurus is the bona fide host for OABV, or whether it can transmit the virus to other bees. Additional investigations including virus isolation and comprehensive epidemiology study are required to understand the biologic impact of OABV on wild bees.