Genomics and proteomics of Apis mellifera filamentous virus isolated from honeybees in China

Abstract

Apis mellifera filamentous virus (AmFV) is a large DNA virus that is endemic in honeybee colonies. The genome sequence of the AmFV Swiss isolate (AmFV CH–C05) has been reported, but so far very few molecular studies have been conducted on this virus. In this study, we isolated and purified AmFV (AmFV CN) from Chinese honeybee (Apis mellifera) colonies and elucidated its genomics and proteomics. Electron microscopy showed ovoid purified virions with dimensions of 300–500 ​× ​210–285 ​nm, wrapping a 3165 ​× ​40 ​nm filamentous nucleocapsid in three figure-eight loops. Unlike AmFV CH–C05, which was reported to have a circular genome, our data suggest that AmFV CN has a linear genome of approximately 493 ​kb. A total of 197 ORFs were identified, among which 36 putative genes including 18 baculoviral homologs were annotated. The overall nucleotide similarity between the CN and CH–C05 isolates was 96.9%. Several ORFs were newly annotated in AmFV CN, including homologs of per os infectivity factor 4 (PIF4) and a putative integrase. Phylogenomic analysis placed AmFVs on a separate branch within the newly proposed virus class Naldaviricetes. Proteomic analysis revealed 47 AmFV virion-associated proteins, of which 14 had over 50% sequence coverage, suggesting that they are likely to be main structural proteins. In addition, all six of the annotated PIFs (PIF-0–5) were identified by proteomics, suggesting that they may function as entry factors in AmFV infection. This study provides fundamental information regarding the molecular biology of AmFV.

Highlights

• •The AmFV CN contains a 493 ​kb linear genome encoding 197 ORFs.
• •Proteomics revealed 14 putative major structural proteins.
• •AmFV belongs to the class Naldaviricetes but not the order Lefavirales.

1. Introduction

Honeybees are indispensable pollinators in natural ecosystems and contribute to the production of approximately 70% of the crops used for human consumption (Klein et al., 2007). However, multiple factors threaten honeybees, such as pesticides, and parasites including viruses (Brosi et al., 2017). To date, more than 30 honeybee viruses have been reported, most of which are RNA viruses (McMenamin and Genersch, 2015; Remnant et al., 2017; Beaurepaire et al., 2020). Apis mellifera filamentous virus (AmFV) is one of the few DNA viruses identified in honeybees and has been studied less than the RNA viruses.

AmFV was initially reported as a honeybee pathogen in the United States in 1978 (Clark, 1978). Acutely infected bees become weak and gather at the hive entrance, while severely infected honeybees exhibit milky-white hemolymph due to tissue degradation (Clark, 1978). In general, AmFV appears to be a weak pathogen, but is endemic in honeybee colonies. For example, it is considered to be the most common and least harmful bee virus in Britain (Bailey, 1982). While AmFV appears to be a weak pathogen, it is not unreasonable to suggest that it may weaken the bee to an extent that makes it more susceptible to other pathogens. Initially, the presence of AmFV was diagnosed using electron microscopy. AmFV was first reported to be an ellipsoidal (400 ​× ​100 ​nm), enveloped virus with a long filamentous nucleocapsid (3060 ​× ​60 ​nm) (Clark, 1978). Later, it was characterized as a DNA virus of slightly different size (450 ​× ​150 ​nm and 3000 ​× ​40 ​nm for virion and nucleocapsid, respectively) (Bailey et al., 1981). The nucleocapsid morphology of AmFV is unique in that it forms three figure-eight loops inside the envelope (Sitaropoulouab et al., 1989).

Partial sequences of AmFV were first derived from genome sequencing of Varroa destructor mites, but at the time they were referred to as baculovirus-related (Cornman et al., 2010). A breakthrough in AmFV research was made in 2015, when the complete viral genome was sequenced from infected worker honeybees collected in Switzerland (Gauthier et al., 2015); this led to the molecular detection of AmFV in practice.

The prevalence of AmFV has been surveyed in the USA, Switzerland, France, Sweden, China, Syria, Czech Republic, and Argentina, showing that it is commonly found worldwide (Gauthier et al., 2015; Hartmann et al., 2015; Hou et al., 2016; Abou Kubaa, 2018; Prodelalova et al., 2019; Quintana, 2019). AmFV is detectable year-round, with higher viral copy numbers found in the Spring (Hartmann et al., 2015). Apart from honeybees, AmFVs have been detected in a wide range of solitary bee species (Ravoet et al., 2014) and even in honey (Bovo et al., 2018, 2020).

The double-stranded DNA genome of the Swiss strain AmFV (AmFV CH–C05) is approximately 498 ​kb, encoding 247 open reading frames (ORFs). However, only a small portion (∼16%) of the predicted ORFs have been annotated, including 13 homologs of baculovirus genes, including per os infectivity factors (PIFs) and baculovirus repeated ORFs (BROs) (Gauthier et al., 2015). Apart from genome sequencing, very few molecular studies have been conducted on AmFV. An early biochemistry study revealed that there were 12 AmFV structural proteins (Bailey et al., 1981), but to our knowledge, no further relevant investigation has been performed since then. Molecular characterization of AmFV is indispensable to better understand its pathogenicity and its interaction with its host. In this study, we isolated and purified AmFV virions from honeybee colonies in China and conducted next-generation genomic sequencing and compared it to that of the Swiss strain. We performed proteomics to identify virion-associated proteins.

2. Materials and methods

2.1. Virus purification and viral DNA extraction

A. mellifera naturally infected with AmFV were collected from Henan Province, China. The AmFV infection was detected by PCR. Briefly, total DNA was extracted from bee workers by phenol-chloroform and ethanol precipitation. PCR amplification was performed using AmFV specific primers (5′-CAGAGAATTCGGTTTTTGTGAGTG-3′ and 5′-CATGGTGGCCAAGTCTTGCT-3′) (Gauthier et al., 2015). The identity of the PCR products was further confirmed by Sanger sequencing. Virions (AmFV CN) were purified from approximately 40 honeybee adults as previously described (Bailey et al., 1981; Laughton and Siva-Jothy, 2011) with slight modifications. Briefly, bees were homogenized in extraction buffer (0.01 ​mol/L ammonium acetate, 0.02% diethyldithiocarbamate, and 0.01% Triton X-100). The homogenate was filtered using 4-layer gauze, then centrifuged at 1000 ​g for 30 ​min. The supernatant was centrifuged on a sucrose density gradient (20%–60%) at 40000 ​g for 1 ​h. The band at 50% sucrose was collected and purified to remove sucrose. The purified virions were imaged by transmission electron microscopy (TEM) using a 100 ​kV Hitachi H-7000FA microscope. DNA extraction was performed as previously described (Gauthier et al., 2015).

2.2. Genome sequencing and bioinformatic analysis

Virion DNA was sequenced using the Illumina Hiseq 3000 System with shotgun strategy at the Sequencing Platform of the National Key Laboratory of Crop Genetic Improvement at Huazhong Agricultural University (Wuhan, China). The reads were quality controlled and preprocessed using Trimmomatic (version 0.32), then assembled with Trinity (version 2.5.1). Gaps and unreadable sequences were amplified by PCR and confirmed by Sanger sequencing. ORFs were identified using the FGENESV program (http://linux1.softberry.com/berry.phtml) and ORFfinder (http://www.ncbi.nlm.nih.gov/orffinder/), adopting the criteria of polypeptide length >100, standard ATG start codon, and minimal overlap. A genome map was constructed using an in-house Python script. Gene annotation and function prediction were performed using the NCBI BLASTP algorithm (https://blast.ncbi.nlm.nih.gov/Blast.cgi) against the nr protein and UniRef90 (https://www.uniprot.org/blast/) databases. Conserved domains were determined using RPS-BLAST with the Conserved Domain Database (https://www.ncbi.nlm.nih.gov/Structure/cdd/wrpsb.cgi) and the hmmscan to Pfam database, with a minimum e-value of 1.0e^-3 in both cases. The annotated genome sequence data have been submitted to GenBank (http://www.ncbi.nlm.nih.gov/genbank) under accession number OK392616. The genome sequence of AmFV CH–C05 was used as a reference for comparison (GenBank accession number NC_027925.1, Gauthier et al., 2015).

2.3. Phylogenetic analysis

All PIF sequences were aligned using ClustalW with MEGA6 by using default settings. Phylogenetic analysis was conducted using the concatenated PIF amino-acid sequences. The phylogenetic tree was constructed using MEGA6 with the substitution model (LG ​+ ​G ​+ ​I) by using the maximum-likelihood method with 1000 bootstrap replicates.

2.4. Proteomics

Shotgun proteomics was used to identify AmFV virion-associated proteins. Briefly, purified AmFV virions were suspended, reduced, alkylated, and subjected to in-solution trypsin digestion. Digested peptides were subjected to liquid chromatography-tandem mass spectrometry (LC-MS/MS) using Q-Exactive Plus coupled to an Easy nLC 1200 (Thermo Fisher Scientific). Proteomics were performed by Bioprofile (Shanghai, China). Peptide sequences were analyzed using the UniProt Protein Database (https://www.uniprot.org/) and Conserved Domain Database (https://www.ncbi.nlm.nih.gov/Structure/cdd/cdd.shtml; Lu et al., 2000) with an E-value < 0.1.

3. Results

3.1. AmFV purification and morphology

AmFVs were purified from naturally infected honeybees collected in Henan Province, China. Transmission electron micrographs showed the virions were ovoid in shape with a size of approximately 300–500 ​× ​210–285 ​nm (Fig. 1A), wrapping a long filamentous nucleoprotein into three superimposed figure-eight loops (Fig. 1B). The size of the filamentous nucleoprotein was approximately 3165 ​× ​40 ​nm. The morphology we observed was similar to those reported in previous studies (Clark, 1978; Bailey et al., 1981; Sitaropoulouab et al., 1989).

Fig. 1.

Transmission electron micrographs of purified AmFV. A. Whole virions. B. Representative filamentous nucleocapsid with three figure-eight loops. Scale bar, 200 ​nm.

3.2. Genome overview

The AmFV CN genome was assembled from 37,330,706 high-quality reads with an average coverage of 7576 ​×. There were a few gaps (total of ∼2843 bp) that could not be filled by PCR and Sanger sequencing, likely due to the complexity of the DNA sequence and/or structure in those regions. Likewise, gaps of approximately 2500 bp also exist in the reported AmFV CH–C05 genome (Gauthier et al., 2015). The size of the assembled genome was 492,752 bp, which was 3644 bp shorter than that of CH–C05 (∼496,396 bp). The overall identity between AmFV CN and CH–C05 was 96.9%. The G ​+ ​C content of AmFV CN was 50.6%, similar to the 50.9% of CH-05 (Gauthier et al., 2015). However, unlike the CH–C05 genome, which was reported to be circular (Gauthier et al., 2015), our data indicate that the AmFV CN has a linear genome, with no evidence of overlap between the two ends during genome assembly.

Initially, 157 methionine-initiated ORFs with a minimum length of 100 residues were predicted in AmFV CN. When the genome of AmFV CH-05 (Gauthier et al., 2015) was used as the reference, an additional 40 ORFs with lengths shorter than 100 residues were identified in AmFV CN, increasing the total ORF number to 197 (Fig. 2). Coding regions constituted 63% of the AmFV CN genome, similar to that observed in the CH-05 genome (65%). In comparison to CH–C05 (a total of 241 ORFs), 45 ORFs were missing in AmFV CN, of which 32 were shorter than 100 residues (Table 1). For consistency, the ORFs in AmFV CN were named from their homologs in CH-05 (NC_027925.1). Notably, ORF183 of CH-C05 was split into two separate ORFs in AmFV CN, which were designated AmFV_183 and AmFV_183a. The genome organization between CN and CH-C05 is highly conserved.

Fig. 2.

Diagram of the AmFV CN genome. The linear genome of AmFV is shown with marked length. The arrows indicate predicted ORFs and direction of transcription. ORFs predicted to be related to DNA replication/metabolism, PIFs, and BROs are shown in red, green, and brown, respectively. ORFs identified by proteomics are displayed in a blue font.

Table 1.

AmFV CH-05 ORFs not present in AmFV CN.

No.	ORF	Protein	Length (aa)
1	AmFV_014	hypothetical protein	57
2	AmFV_026	hypothetical protein	105
3	AmFV_032	hypothetical protein	61
4	AmFV_033	hypothetical protein	52
5	AmFV_035	hypothetical protein	68
6	AmFV_036	hypothetical protein	72
7	AmFV_038	hypothetical protein	59
8	AmFV_039	hypothetical protein	93
9	AmFV_040	hypothetical protein	105
10	AmFV_044	hypothetical protein	56
11	AmFV_046	hypothetical protein	57
12	AmFV_047	hypothetical protein	33
13	AmFV_061	hypothetical protein	66
14	AmFV_090	hypothetical protein	114
15	AmFV_094	hypothetical protein	55
16	AmFV_121	hypothetical protein	50
17	AmFV_131	hypothetical protein	37
18	AmFV_153	hypothetical protein	58
19	AmFV_155	hypothetical protein	83
20	AmFV_160	hypothetical protein	52
21	AmFV_163	hypothetical protein	58
22	AmFV_167	hypothetical protein	69
23	AmFV_176	hypothetical protein	35
24	AmFV_179	hypothetical protein	26
25	AmFV_186	hypothetical protein	98
26	AmFV_190	hypothetical protein	108
27	AmFV_191	hypothetical protein	56
28	AmFV_192	hypothetical protein	50
29	AmFV_194	hypothetical protein	84
30	AmFV_196	hypothetical protein	173
31	AmFV_197	hypothetical protein	116
32	AmFV_199	hypothetical protein	49
33	AmFV_202	hypothetical protein	346
34	AmFV_204	hypothetical protein	28
35	AmFV_205	hypothetical protein	44
36	AmFV_208	hypothetical protein	58
37	AmFV_209	hypothetical protein	136
38	AmFV_217	hypothetical protein	143
39	AmFV_222	hypothetical protein	75
40	AmFV_227	hypothetical protein	37
41	AmFV_236	hypothetical protein	648
42	AmFV_238	hypothetical protein	80
43	AmFV_239	hypothetical protein	206
44	AmFV_240	hypothetical protein	278
45	AmFV_241	hypothetical protein	118

The ORFs were annotated on the basis of homology. Thirty-six ORFs had homologs in the genomes of other species (from viruses, eukaryotes, and bacteria) in public sequence databases (Table 2). Based on predicted functions from databases, these ORFs were categorized as: 6 potential DNA replication and nucleotide metabolism, 6 PIFs, 9 BROs, and 15 with putative other or unknown functions. All 36 annotated ORFs in AmFV CN were present in AmFV CH-C05 with high protein sequence identities (>90%, Table 2).

Table 2.

Putative genes in AmFV CN Genome.

Putative function	ORF	size (aa)	Putative protein	Best match with Pfam-A database			Best match with BLASTP					AmFV CH–C05 identity
				Pfam code dmain	E-value	Pfam no	Species/Virus	Score	E-value	Similarity	aacession no
DNA Replication and nucleotide metabolism	AmFV_027	610	Thymidylate synthase	Thymidylat_synt	1.5E-85	PF00303.19	Malassezia pachydermatis	284	3.0E-84	45.5%	XP_017993541.1	98.2%
	AmFV_042	556	Integrase	phage_intergrase	2.5E-05	PF00589.22	Fibrobacter sp.	55.1	2.0E-04	27.2%	NLD99342.1	98.5%
	AmFV_074	1955	DNA Pol	DNA_pol_B	4.5E-14	PF00136.21	–	–	–	–	–	99.1%
	AmFV_095	1603	DNA ligase	DNA_ligase_A_N	8.3E-09	PF04675.14	Athalia rosae	99	9.0E-17	24.0%	XP_025602269.1	99.6%
	AmFV_114	879	RR1	Ribonuc_red_IgC	0.0E+00	PF02867.15	Cytophagaceae bacterium	730	0.0E+00	44.4%	ODS80019.1	99.7%
	AmFV_216	2469	RR2	Ribonuc_red_sm	2.1E-101	PF00268.21	Uncultured virus	391	2.0E-121	100.0%	ADD74390.1	100.0%
PIFs	AmFV_057	334	PIF-5	–	–	–	Tipula oleracea nudivirus	170	4.8E-06	20.0%	YP_009116743.1	99.1%
	AmFV_060	830	PIF-1	PIF	7.0E-16	PF05092.12	Cyclophragma undans nucleopolyhedrovirus	91	4.0E-15	30.0%	YP_010086634.1	98.6%
	AmFV_077	1188	PIF-0	Baculo_p74		PF04583.12	Hyphantria cunea nucleopolyhedrovirus	156	6.4E-07	29.4%	YP_473207.1	98.6%
	AmFV_088	280	PIF-3	PIF3	4.4E-04	PF05006.12	Malacosoma sp. alphabaculovirus	54	1.0E-04	20.7%	ANW12283.1	100.0%
	AmFV_100	402	PIF-2	PIF2	6.0E-18	PF04631.12	Mauternbach virus	92	2.0E-16	31.2%	AYP97928.1	99.8%
	AmFV_157	204	PIF-4	Baculo_19	3.2E-09	PF04798.12	–	–	–	–	–	100.0%
BROs	AmFV_008	182	BRO-1	–	–	–	Chrysodeixis chalcites nucleopolyhedrovirus	53	1.0E-04	33.3%	YP_249718.1	100.0%
	AmFV_016	1306	BRO-2	Bro-N	2.9E-09	PF02498.17	Helicoverpa armigera nucleopolyhedrovirus	179	1.60E-6	29.1%	AMN15974.2	94.0%
	AmFV_069	262	BRO-3	–	–	–	AmFV	1309	0.0E+00	98.9%	YP_009165820.1	98.9%
	AmFV_075	434	BRO-4	–	–	–	Spodoptera litura granulovirus	52	9.0E-05	28.0%	YP_001257066.1	98.9%
	AmFV_106	667	BRO-5	Bro-N	3.2E-07	PF02498.17	Chrysodeixis includens nucleopolyhedrovirus	77	3.0E-12	25.0%	AOL57177.1	95.8%
	AmFV_108	627	BRO-6	Bro-N	8.6E-04	PF02498.17	Spodoptera frugiperda ascovirus 1a	66	6.0E-09	29.0%	YP_762434.1	92.3%
	AmFV_110	499	BRO-7	Bro-N	6.6E-29	PF02498.17	Chrysodeixis chalcites nucleopolyhedrovirus	87	2.0E-15	29.8%	AGE61478.1	97.0%
	AmFV_111	437	BRO-8	–	–	–	AmFV	201	1.9E-14	25.2%	YP_009165857.1	98.4%
	AmFV_133	157	BRO-9	Bro-N	4.1E-03	PF02498.17	–	–	–	–	–	98.7%
Others	AmFV_006	1699	Protein kinase (PK)	–	–	–	Acanthamoeba polyphaga mimivirus	111	6.7E-03	26.0%	AKI80069	99.0%
	AmFV_009	442	PARP	Trypan_PARP	3.1E-04	PF05887.11	Paramecium bursaria Chlorella virus CviKI	179	2.4E-09	28.4%	AGE51657.1	99.8%
	AmFV_023	648	ATPase	AAA domain	4.0E-26	PF00004.29	Lasius niger	124	1.0E-27	34.0%	KMQ86848.1	99.2%
	AmFV_034	501	Serpin-like	Pacifastin inhibitor (LCMII)	4.6E-10	PF05375.13	Blattella germanica	269	6.5E-24	34.9%	PSN39366.1	97.0%
	AmFV_043	1633	Myristoylated membrane	–	–	–	Mimivirus sp. SH	186	6.2E-12	46.7%	AZL89416.1	96.0%
	AmFV_068	326	RING finger protein 413R	zf-C3HC4_3	2.1E-03	PF13920.6	–	–	–	–	–	100.0%
	AmFV_080	572	RING finger protein	–	–	–	Collichthys lucidus	52	1.0E-03	33.3%	TKS65457.1	98.6%
	AmFV_082	510	HZV 115-like	DUF4580	2.3E-05	PF15162.6	Oryctes rhinoceros nudivirus	85	3.0E-15	31.1%	YP_002321369.1	99.8%
	AmFV_091	534	hypothetical protein	–	–	–	Hirundo rustica rustica	144	5.0E-04	40.2%	RMB88007.1	98.2%
	AmFV_101	1982	Gamma-glutamyltranspeptidase	G_glu_transpept	1.6E-14	PF01019.21	Diachasmimorpha longicaudata entomopoxvirus	147	5.9E-07	34.6%	AKS26328.1	97.5%
	AmFV_113	583	hypothetical protein	–	–	–	Thalassiosira oceanica	145	2.4E-08	32.5%	EJK57244.1	90.5%
	AmFV_123	948	hypothetical protein	–	–	–	Harpegnathos saltator	122	1.0E-24	29.4%	EFN83926.1	99.8%
	AmFV_168	1236	MdSGHV 070	–	–	–	Musca hytrosavirus	166	2.3E-04	29.6%	YP_001883398.1	97.3%
	AmFV_193	969	Chitin-binding	LOMP_10	8.6E-36	PF03067.15	Apis mellifera	136	2.0E-31	42.4%	WP_180560000.1	96.0%
	AmFV_235	334	PLC	PI-PLC-X	3.6E-11	PF00388.19	Taibaiella koreensis	72	6.0E-10	33.6%	WP_118975952.1	100.0%

As a large DNA virus, AmFV CN has several key factors that facilitate DNA replication. AmFV_074 encodes a hypothetical type-B DNA polymerase with a predicted length of 1960 residues. AmFV_042 and AmFV_095 are a putative integrase and DNA ligase, respectively, which are likely involved in viral DNA replication. Like many DNA viruses, AmFV also encodes enzymes involved in nucleotide metabolism. AmFV_027 is a homolog of thymidylate synthase, which catalyzes the conversion of deoxyuridine monophosphate (dUMP) to deoxythymidine monophosphate (dTMP). AmFV_114 and AmFV_216 are homologs of ribonucleoside reductase subunits 1 and 2 (RR1 and RR2), respectively, which catalyzes the formation of deoxyribonucleotides from ribonucleotides and provides precursors for viral DNA synthesis. Among these six ORFs, the putative integrase (AmFV_042) was newly annotated in AmFV CN.

There are six ORFs in AmFV that are homologs of baculoviral PIFs: PIF0 (AmFV_077), PIF1 (AmFV_060), PIF2 (AmFV_100), PIF3 (AmFV_088), PIF4 (AmFV_157), and PIF5 (AmFV_057). In comparison to the reported AmFV CH–C05 genome sequence (Gauthier et al., 2015), PIF4 was newly annotated in this study. PIFs are essential for oral infection by baculoviruses; ten (PIF0–9) have been identified in baculoviruses, with all but PIF5 forming a PIF complex of ∼520 ​kDa to initiate midgut infection (Wang et al., 2019). PIFs are also found in other large arthropod DNA viruses, including nudiviruses, white spot syndrome virus (WSSV), and salivary-gland hypertrophy viruses (Escobedo-Bonilla et al., 2008; Wang and Jehle, 2009; Lietze et al., 2011; Bezier et al., 2015). It is not yet known whether AmFV also produces a functional PIF complex or it has other unidentified PIF homologs.

Nine putative bro genes were found in the AmFV CN genome (Table 2), among which AmFV_069, AmFV_111, and AmFV_113 were not initially annotated as bro genes in the report of the AmFV CH–C05 genome (Gauthier et al., 2015), but are noted in the updated genome sequence in GenBank (NC_027925.1). Bros were first identified in the genome of Lymantria dispar multinucleocapsid nucleopolyhedrovirus (LdMNPV), which contains 16 repeated ORFs (Kuzio et al., 1999). The number of BROs in AmFV is in the same range of the numbers present in baculoviruses, 0 to 16 (Li et al., 2021). BROs constitute a superfamily identified in invertebrate dsDNA viruses, bacteriophages, and bacteria that contain a conserved N-terminal predicted DNA-binding motif (Jakob et al., 2001; Bideshi et al., 2003). Their function(s) is not clear, although several studies have shown that some have DNA-binding activities and interact with laminin or translation-associated proteins (Zemskov et al., 2000; Kang et al., 2003; Kotani et al., 2015). The multiple BROs found in AmFV support the hypothesis that BROs play important roles in viral interactions with invertebrates (Bideshi et al., 2003).

In addition to the above genes, 15 putative genes were identified with homology to genes from other species (Table 2).

3.3. Phylogeny analysis based on PIFs

Of the 36 annotated genes of AmFV CN, 18 (including six PIFs, nine BROs, DNA polymerase, RR1, and RR2) have homologs in baculoviruses, suggesting that it is a baculo-like virus. Based on shared PIFs, the new virus class Naldaviricetes was recently proposed by the International Committee on Taxonomy of Viruses (ICTV) (https://ictv.global/ictv/proposals/2020.006D.R.Naldaviricetes.zip) (Walker et al., 2021). This class contains four families of nuclear arthropod large DNA viruses (NALDVs) including Baculoviridae, Nudiviridae, Hytrosaviridae, and Nimavirida, with AmFV as a free member. To place AmFV among the viruses, we conducted phylogeny analysis using concatenated protein sequences of the PIFs. The results showed that AmFV CN and AmFV CH–C05 formed a separate branch within Naldaviricetes (Fig. 3). Unlike the members of Baculoviridae, Nudiviridae, and Hytrosaviridae, which also encode four subunits of a DNA-directed RNA polymerase, the AmFV CN and CH–C05 genomes lack ORFs encoding these late expression factors and thus, do not belong to the proposed order Lefavirales (https://ictv.global/ictv/proposals/2020.006D.R.Naldaviricetes.zip) (Walker et al., 2021). The PIF phylogenetic tree shows that AmFVs are distinct from the members of Lefavirales, although the AmFV PIFs appeared to be closer to those of Hytrodaviridae with a bootstrap value of 69% (Fig. 3).

Fig. 3.

Phylogenetic tree of members of Naldaviricetes derived from concatenated protein sequences of PIFs. The maximum-likelihood (ML) tree on substitution model (LG ​+ ​G ​+ ​I) is present. Numbers on the nodes indicate ML nonparametric bootstrap supports (1000 replicates). The tree is drawn to scale, with branch lengths measured in the number of substitutions per site. The classifications of viruses are indicated. Virus abbreviations and sequence accession numbers are: AcMNPV, Autographa californica multiple nucleopolyhedrovirus, NC_001623; CpGV, Cydia pomonella granulovirus, NC_002816; NeleNPV, Neodiprion lecontei nucleopolyhedrovirus, NC_005906; CuniNPV, Culex nigripalpus nucleopolyhedrovirus, NC_003084; GbNV, Gryllus bimaculatus nudivirus, NC_009240; HzNV, Helicoverpa zea nudivirus-2, NC_004156; MdSGHV, Musca domestica salivary gland hypertrophy virus, NC_010671; WSSV, White spot syndrome virus, NC_003225; AmFV CH–C05, NC_027925; AmFV CN, OK392616.

3.4. The AmFV CN proteome

Since most of the ORFs of AmFV remain hypothetical, proteomics was performed to determine which ORFs are authentic proteins and whether they are structurally related. The proteins consistently identified in independent LC-MS/MS experiments are summarized in Table 3; 47 proteins were found to be associated with the AmFV virion (Table 3, Fig. 2).

Table 3.

The AmFV CN virion associated proteins.

No.	ORFsa	Protein	Mol. weight (kDa)	Protein length (aa)	Test 1		Test 2
					Unique peptides	Sequence coverage (%)	Unique peptides	Sequence coverage (%)
1	AmFV_002	hypothetical protein	108.0	961	8	7.8	5	5.1
2	AmFV_017	hypothetical protein	46.6	409	10	27.7	5	14.5
3	AmFV_021∗	hypothetical protein	36.8	321	18	66.9	11	31.6
4	AmFV_022∗	hypothetical protein	30.6	271	12	52.2	7	26.7
5	AmFV_023∗	AAA ​+ ​ATPase	69.2	648	35	57.1	29	48.2
6	AmFV_051∗	hypothetical protein	39.1	342	18	52.5	14	45.2
7	AmFV_053	hypothetical protein	15.1	127	2	12.7	1	7.1
8	AmFV_054	hypothetical protein	111.5	1062	14	17.7	17	23
9	AmFV_056	hypothetical protein	32.3	286	3	12.1	2	9.3
10	AmFV_057∗	PIF5	36.1	334	29	68.3	21	58.9
11	AmFV_058∗	hypothetical protein	148.7	1353	64	61.9	45	42.9
12	AmFV_060	PIF1	93.9	830	22	29.6	16	21.4
13	AmFV_062	hypothetical protein	77.6	676	6	7.6	3	4.1
14	AmFV_064	hypothetical protein	38.6	350	7	31.1	6	26.9
15	AmFV_073	hypothetical protein	63.3	629	1	1.4	4	10.2
16	AmFV_077	PIF0	131.9	1188	23	21.7	22	25.8
17	AmFV_078	hypothetical protein	37.2	336	12	29.6	8	27.2
18	AmFV_085	hypothetical protein	13.0	118	2	23.1	2	23.1
19	AmFV_088	PIF3	30.8	280	6	29	3	15.1
20	AmFV_089	hypothetical protein	48.8	450	16	46.3	8	19.8
21	AmFV_092	hypothetical protein	40.0	357	13	44.7	11	36.8
22	AmFV_097	hypothetical protein	70.2	628	24	47.4	18	32.4
23	AmFV_099∗	hypothetical protein	42.6	383	32	90.9	22	66.8
24	AmFV_100	PIF2	45.0	402	11	39.4	9	30.2
25	AmFV_104	hypothetical protein	19.7	174	5	36.1	3	13.9
26	AmFV_117	hypothetical protein	10.0	88	2	18.4	1	9.2
27	AmFV_122	hypothetical protein	56.7	507	4	10.3	2	4.5
28	AmFV_128∗	hypothetical protein	56.6	490	31	61.6	21	43.1
29	AmFV_130∗	hypothetical protein	19.8	173	10	64.8	9	51.1
30	AmFV_138∗	hypothetical protein	53.4	450	26	53	18	40.3
31	AmFV_139	hypothetical protein	18.0	168	5	32.9	2	19.8
32	AmFV_140	hypothetical protein	164.8	1440	55	44.2	33	27.4
33	AmFV_141	hypothetical protein	22.8	196	4	17.9	1	6.2
34	AmFV_146∗	hypothetical protein	24.9	218	16	70	11	51.2
35	AmFV_148	hypothetical protein	38.9	346	4	14.8	5	20.9
36	AmFV_149∗	hypothetical protein	12.6	127	6	52.4	2	21.4
37	AmFV_151∗	hypothetical protein	44.0	396	20	52.2	14	37.7
38	AmFV_154	hypothetical protein	95.3	850	19	28.9	9	12
39	AmFV_156∗	hypothetical protein	29.3	262	16	63.5	17	63.5
40	AmFV_157	PIF4	23.0	204	7	41.9	3	12.3
41	AmFV_161	hypothetical protein	36.3	332	1	1.8	2	5.1
42	AmFV_164	hypothetical protein	19.9	182	4	19.9	2	13.3
43	AmFV_182	hypothetical protein	102.4	895	39	48.9	29	33.9
44	AmFV_200	hypothetical protein	111.4	1019	8	8.6	4	4.2
45	AmFV_215	hypothetical protein	223.4	1965	8	3.1	8	3.9
46	AmFV_233	hypothetical protein	9.2	83	4	42.7	2	31.7
47	AmFV_235	hypothetical protein	38.7	334	17	45.6	13	34.8

^aThe ORFs with sequence coverage over 50% in at least one test were marked with ∗.

All the annotated PIFs, PIF0 (AmFV_077), PIF1 (AmFV_060), PIF2 (AmFV_100), PIF3 (AmFV_088), PIF4 (AmFV_157), and PIF5 (AmFV_057), were detected by proteomics (Table 3), suggesting that they are virion structural proteins that likely function as entry factors. AmFV_023, encoding a putative AAA ​+ ​ATPase, was also identified (Table 2, Table 3). The functions of the remaining detected proteins are largely unknown (Table 3).

Further more, we detected 14 proteins with predicted molecular weights of 12.6–148.7 ​kDa with over 50% sequence coverage in at least one of the LC-MS/MS runs: AmFV_021 (36.8 ​kDa), AmFV_022 (30.6 ​kDa), AmFV_023 (69.2 ​kDa), AmFV_051 (39.1 ​kDa), PIF5 (AmFV_057, 36.1 ​kDa), AmFV_058 (148.7 ​kDa), AmFV_099 (42.6 ​kDa), AmFV_128 (56.6 ​kDa), AmFV_130 (19.8 ​kDa), AmFV_138 (53.4 ​kDa), AmFV_146 (24.9 ​kDa), AmFV_149 (12.6 ​kDa), AmFV_151 (44.0 ​kDa), and AmFV_156 (29.3 ​kDa), suggesting they are likely major structural proteins (Table 3).

4. Discussion

AmFV is an endemic DNA virus in honeybee colonies; however, very little was known about its molecular biology. In China, the prevalence of AmFV in honeybee colonies varies from 10% to 85% (Hou et al., 2016, 2017). In this study, we have produced useful information on this mysterious, large DNA virus.

The average size of the AmFV CN particle was 300–500 ​× ​200–290 ​nm (Fig. 1), which were similar but slightly wider comparing to those previous reported (Clark, 1978; Bailey et al., 1981; Sitaropoulouab et al., 1989). The initial reported size of the AmFV virion was 450 ​× ​150 ​nm (Bailey et al., 1981), but these authors also mentioned that after purification on sucrose gradients, the particles appeared irregularly ellipsoidal with varying dimensions of 250–500 ​× ​120–200. Therefore, the slight difference we observed was most likely a consequence of different purification methods. The size of the filamentous nucleocapsid we observed is consistent with previous reports (Bailey et al., 1981; Sitaropoulouab et al., 1989). Electron microscopy showed the unique wrapping of the long nucleocapsid within the AmFV virion (Fig. 1), supporting the earlier hypothesis that the nucleocapsid has to form three figure-eight loops to fit into the envelope (Sitaropoulouab et al., 1989).

The overall high nucleotide identity (96.9%) between the CN and CH–C05 strains indicates that these two viruses are closely related. Thirty-two of the 45 ORFs missing in AmFV CN (Table 1) encoded ORFs of less than 100 residues, suggesting that they may not be authentic ORFs. The gene order of AmFV CN was similar to that of AmFV CH–C05. However, unlike circular genome reported for the CH–C05 strain (Gauthier et al., 2015), our sequence data suggested that AmFV CN has a linear genome, as no reads overlapped both ends. In fact, our data support an earlier hypothesis that AmFV contains a linear genome based on electron microscopic observations of AmFV DNA (Bailey et al., 1981). Given that the length of the nucleocapsid is about 3.1 ​μm (Clark, 1978; Bailey et al., 1981; Sitaropoulouab et al., 1989) and the measured DNA length is 5.8 ​± ​0.3 ​μm (Bailey et al., 1981), viral DNA appeared to be only slightly condensed during packaging. How AmFV DNA interacts with proteins to form nucleocapsids, and how the long neucleocapsid is wrapped in the virion envelope into three figure-eight loops, remain open and intriguing questions.

Identifying structural proteins is fundamental for understanding the unique structure of AmFV and will provide useful information for the development of immunological AmFV diagnostic kits. By proteomic analyses, we identified 47 structural proteins that were present in at least two replicate runs (Table 3). As expected, all six PIFs (PIF 0–5) were present, suggesting that they are AmFV structural proteins and likely functional entry factors. In addition, AmFV_023, with homology to the ATPases associated with diverse cellular activities (AAA+) ATPase family, was identified. These ATPases are a large protein family that utilize energy from ATP hydrolysis to participate in multiple cellular functions. The presence of AmFV_023 in the AmFV virion suggests that it may be involved in providing required energy for nucleocapsid packaging.

Among the 47 proteins, 14 with predicted molecular weight of 12.6–148.7 ​kDa appeared to be major structural proteins, as they were detected with over 50% sequence coverage (Table 3). Previous study has shown that 12 proteins, ranging in size from 13 to 70 ​kDa, had been associated with the AmFV virion by polyacrylamide gel electrophoresis (Bailey et al., 1981). Of these 12, P40 and P13 appeared to be major nucleocapsid proteins (although P13 was also present in the envelope), while P37 and P23 appeared to be major envelope proteins (Bailey et al., 1981). We speculate that AmFV_149 (predicted 12.7 ​kDa), AmFV_146 (24.9 ​kDa), PIF5 (36.1 ​kDa), and AmFV_51 (39.1 ​kDa) reflect the previously identified P13, P23, P37, and P40, respectively. However, this speculation is based solely on the rough matches in molecular weight and need verification in the future. One way to do this is to generate specific antibodies for probing western blots or performing immunoelectron microscopy.

In summary, our study showed that AmFV CN contains a linear genome of ∼492,752 bp, encoding 197 ORFs. Forty-seven of the ORFs were associated with virions, including six PIFs, which likely function in viral entry. The results provide fundamental information for future molecular studies on the virus.

Data availability

The genome sequence of AmFV CN has been deposited in GenBank under accession number OK392616.

Ethics statement

This article does not contain any studies with human or animal subjects (except insects) performed by any of the authors.

Author contributions

Dahe Yang: investigation, data curation, conceptualization, formal analysis, writing-original draft, writing-review and editing. Jun Wang: data curation, formal analysis, investigation, writing-original draft, writing-review and editing. Xi Wang: investigation, methodology. Fei Deng: funding acquisition, resources, supervision. Qingyun Diao: funding acquisition, resources, supervision. Manli Wang: funding acquisition, resources, supervision. Zhihong Hu: conceptualization, formal analysis, funding acquisition, resources, supervision, writing-original draft, writing-review and editing. Chunsheng Hou: conceptualization, funding acquisition, resources, supervision, writing-review and editing.

Conflict of interest

The authors declare that they have no conflict of interest.