Full-length genomic sequencing and characterization of Borna disease virus 1 isolates: Lessons in epidemiology

2.1 Virus strains and cell culture

Human OL cells and virus stocks of the natural human strain BoDV-1 Hu-H2 and equine strain Equ-Cres, the nonnatural vaccine strain DessVac and strain V propagated in OL cells were kindly supplied by Hanns Ludwig, Professor of Virology (Free University of Berlin, Berlin, Germany) in 2012 and 2010, respectively. These BoDV-1 strains were distinct strains of completely different origins, the histories of which are described in full detail in Tables 1 and 2.

For each virus strain, strictly separated lamina air flow facilities of our cell laboratory at Chongqing Medical University were used. OL cells were maintained in Dulbecco's Modified Eagle's Medium (DMEM; Invitrogen Life Technologies) with high glucose (4.5 g/L, Gibco), containing 1% penicillin, 1% streptomycin (Sigma-Aldrich), and 5% heat-inactivated fetal bovine serum (Gibco). Cells were incubated at 37°C in a 5% CO₂ incubator. Cells were harvested after reaching 70% to 80% confluence.

2.2 Amplification of BoDV-1 genomes

For sequencing complete genomes of strains BoDV-1 Hu-H2, Equ-Cres, and DessVac, viral RNA was extracted from OL cells using Trizol (Invitrogen Life Technologies) and quantified by a NanoDrop spectrophotometer. RNA (500 ng) was reverse-transcribed by SuperScript I reverse transcriptase (Takara, Japan) using random hexamers. Eight pairs of primers were designed to amplify nucleotides (nt) 52 to 8788 of the BoDV-1 genome by overlapping polymerase chain reaction (PCR) (Table 3). To sequence the 5′ and 3′ termini, viral RNA was circularized by T4 RNA ligase (Promega, USA). The ligated RNA was reverse-transcribed using primer P9R, and amplified with primers P9F/P9R. Thereafter, 3′-RACE was performed to determine the 3′-terminus using the 3′-Full RACE Core Set (Takara, Japan), according to the manufacturer's instructions. Briefly, after RNA was reverse-transcribed into complementary DNA (cDNA) by the 3′-RACE adapter primer in this kit, nested PCR was performed using primers P10F/P10R and P11F/P11R for the inner and outer PCRs, respectively. All primers are listed in Table 3 and synthesized by the Shanghai Songon Biological Engineering Biotechnology Company.

PCR reactions were performed using 2 μL of cDNA with high fidelity Pfu DNA polymerase in a 25 μL reaction mixture applied to a Corbett Research Rotor-Gene 6000 Thermocycler (Corbett Research, Australia). Cycling conditions consisted of an initial hold for 10 minutes at 94°C, followed by 35 amplification cycles at 94°C for 30 seconds, at 56 to 60°C for 30 seconds, and at 72°C for 2 to 4 minutes, with a final extension step at 72°C for 10 minutes. Each PCR was independently repeated at least three times. All PCR products were directly sequenced after purification or cloned into the pGM-T Easy vector system (Promega, USA) supplied by the Shanghai Songon Biological Engineering Biotechnology Company, using an ABI3730xl DNA Analyser (Macrogen Inc, Amsterdam, The Netherlands). For quality control of sequences, firstly, the PCR products were analyzed by 1.5% agarose gel electrophoresis. Only PCR products in the expected position, demonstrating one bright product band without any nonspecific amplification, were selected for sequencing (Figure 1). Secondly, PCR products of each virus strain were repeatedly analyzed, at least three times. Based on the Overlap-Layout-Consensus algorithm, DNAMAN 8.0 (Lynnon Biosoft, USA) was used for sequence assembly.

2.3 Sequence alignment and phylogenetic analysis

Before phylogenetic analysis, multiple sequence alignments were conducted using ClustalW based on progressive alignment. The highly variable regions in alignments did not contain gaps which were manually verified.

Considering what model of nucleotide evolution might be most appropriate for analyzing our data, several phylogenetic trees were created by the MEGA 6.0 program using both the neighbor-joining (NJ) and the maximum-likelihood (ML) methods with different algorithms, thereby employing 1000 replicates of bootstrap resampling analysis for each tree. The tree with the highest supporting bootstrap values was chosen. The trees were rooted with the sequence of BoDV-2 No/98 (AJ311524).

In accord with the full-length BoDV-1 sequences (8910 nucleotides each; N = 23), an overview on accession numbers at GenBank, original hosts of isolated viruses and hosts’ diseases, biological sources of sequences, and the corresponding references are provided in Table 4. Sequences of published strains/isolates were retrieved from the NCBI nucleotide database. Additional analysis of 28 sequences based on P-protein, and 31 sequences based on N-protein acquired from the GenBank database by using BLAST are shown in Supporting Information Figure S1 and Table S1; Figure S2 and Table S2, respectively.

2.4 Accession numbers of new full-length genomes

The new full-length genome sequences provided in this study have been submitted to GenBank (National Center for Biotechnology Information, USA; NCBI).

The genomic sequence of Isolate_Hu-H2_DE_1994_CN_2019 is accessible through accession number MT375542. The genomic sequence of Isolate_DessVac_DE_1960_CN_2019 is accessible through accession number MT375543. The genomic sequence of Isolate_Equ-Cres_DE_1995_CN_2019 is accessible through accession number MT375544 at GenBank.

2.5 Cell cycle analysis

Cell cycle analysis was performed as described previously.⁴⁰ In brief: strains Hu-H2, Equ-Cres, DessVac, and strain V propagated in OL cells as well as uninfected OL cells were cultured without serum for 24 hours to induce a G1 arrest. After addition of 5% serum to finish arrest, the cells were harvested over a 24-hour period. Then, OL cells (1 × 10⁶ cells) were washed with cold phosphate-buffered saline (PBS), fixed in cold 70% ethanol at 4°C for at least 2 hours, and stained with 0.5 mL propidium iodide (50 µg propidium iodide/mL, 100 U of RNase A/mL, PBS) solution for 30 minutes in the dark. Measurement of DNA content and analysis of cell cycle phases was performed by flow cytometry (BD Biosciences, USA). Flow cytometric analysis for DNA content cannot distinguish between G2 and M phases. Therefore, the percentage of cells in the G1, S, and G2/M phases was determined at different time points after release from G1 arrest.

Cell cycle analysis was used to determine the proliferation index (PI) which is defined as the percentage of cells in the S-phase + G2/M-phase. The PI was applied to assess proliferation rates.

2.6 Cell apoptosis assay

Apoptosis of OL cells infected with strain Hu-H2, Equ-Cres, DessVac, and strain V, as well as of uninfected control cells was measured using an Annexin V-FITC Apoptosis Detection Kit (Beyotime, China), according to the manufacturer's instructions as described previously.⁴⁰ Briefly, cells were washed twice with PBS and re-suspended with binding buffer at a concentration of 1 × 10⁶/mL, followed by the addition of 5 μL Annexin V-FITC and 5 μL propidium iodide. After gentle mixing, the cells were incubated for 15 minutes at room temperature in the dark. The cells were analyzed by flow cytometry within 1 hour. Experiments were repeated in triplicate.

2.7 Statistical analysis

Data were expressed as the means ± standard deviations. SPSS 19.0 software was used to analyze the experimental data. The cells were compared using ANOVA. Two-tailed Student t tests were applied to identify significant differences at a 95% confidence level. P < .05 was deemed to be statistically significant.

3.1 Profile and taxonomical classification of BoDV-1 strains

This study not only aimed to provide new sequence and phylogenetic data but also to profile three unique BoDV-1 strains. Two of them, BoDV-1 Hu-H2 and BoDV-1 Equ-Cres were isolated from PBMCs by long-term cocultivation with further passaging in human OL cells. Isolate Hu-H2 came from a patient with obsessive compulsive disorder (OCD), located in Berlin, Germany,²⁴ and isolate Equ-Cres from a horse with nonfatal Borna disease, located in Munich, Germany. These viruses belong to the very few which could have been retrieved from blood instead of brain cells. In contrast to these natural isolates, BoDV-1 DessVac, and strain V, originally isolated from brains of horses with fatal Borna disease (BD) underwent multiple in vivo passaging in rabbits to become either an adapted vaccine strain for horses, or a laboratory strain by further passaging in rats and finally in vitro in cell culture.^{42, 43} Strain V was the first laboratory strain to become completely sequenced²⁶ and served taxonomically as a reference strain of the new family Bornaviridae.¹ A detailed description of the history of isolate Hu-H2, isolate Equ-Cres, strain DessVac, and strain V⁴¹ is provided in Tables 1 and 2, respectively.

3.2 Full-length genomic sequences of BoDV-1 strains and phylogenetic analysis

Multiple sequence alignments of full genomes of Hu-H2, Equ-Cres, and DessVac compared with 20 other BoDV-1 strains confirmed strikingly low divergence levels not exceeding a maximum of 5.55%.

Phylogenetic analysis revealed that human isolate Hu-H2 and horse isolate Equ-Cres were more distant from DessVac, but could be assigned to the strain V cluster. Other clusters could be assigned to the laboratory strain He/80 and USA derivatives of He/80, Japanese strains, Austrian horse strains, and German encephalitis cases (Figure 2).

As for human-derived isolates, the Hu-H2 genome isolated in 1994 from blood cells of a German OCD patient,²⁴ displayed the largest distance to BoDV-1 sequences retrieved from German encephalitis cases of transplant recipients reported in 2018⁶ (5.20%). Much lower levels were found compared to hP2br, isolated in 2000 from the brain of a Japanese schizophrenic patient^{25, 29} (1.89%), as well as compared to the two other German encephalitis cases (112-6 and ER2)^{7, 8} reported in 2018/2019 (1.88% and 2.31%), respectively.

Isolate Equ-Cres, originating from PBMCs of a Bavarian horse with nonfatal BD in 1995, displaying a very low divergence to Hu-H2, could also be assigned to the strain V cluster. Compared with other natural horse-derived strains like recent Austrian strains originating from the brain (8072/15, 1407/15, 272/16, 271/16),²⁸ moderate divergence levels (between 2.18% and 2.28%) were found. The vaccine strain DessVac from 1960 displayed the largest genetic distance (5.22%) to the laboratory strain He/80 (C6BV and related derivatives)^{27, 30, 31} and a moderate distance (1.63%) to laboratory strain V^{26, 30} originating from 1927.⁴³ A detailed divergence matrix is provided in Supporting Information Table S3.

3.3 Phylogenetic analysis based on nucleotide sequences of major BoDV-1 proteins

In contrast to full-length genomes, a higher number of P-gene and N-gene sequences are available, many of which originating from human patients differing by source, time of isolation, and country. Their phylogenetic analyses elicited valuable data in addition to complete genomes, as evaluated below. The corresponding figures and tables can be found in the Supporting Information, as detailed below.

3.4 Effects of BoDV-1 strains on OL cell proliferation

To address the different effects of BoDV-1 strains on OL cell proliferation, we measured cell cycles by flow cytometry. In the cell cycle, the primary phases of cell proliferation are the S and G2 phases. We determined the cell proliferation index (PI), defined as the percentage of cells in the S-phase + G2/M-phase. Thus, cell proliferation rates could be assessed by estimating the relative cell numbers in the S + G2/M-phase fraction.

The percentages of S + G2/M-phase cells in natural viruses, namely OL/Hu-H2 and OL/Equ-Cres, were not significantly different from each other (t = 0.840, P = .448), but were significantly lower than those found in the nonnatural laboratory viruses OL/strain V and OL/DessVac, as well as in uninfected control OL cells (P > .05). In addition, the proliferation rate of laboratory OL/strain V was comparable to that of laboratory OL/strain DessVac (t = 0.261, P = .807) (Figure 3A).

3.5 Effects of BoDV-1 strains on OL cell apoptosis

Next, we assessed the effects of BoDV-1 strains on OL cell apoptosis by using Annexin V-FITC staining. The apoptosis rates of OL/Hu-H2 and OL/Equ-Cres were significantly increased compared to uninfected OL cells (t = 4.220, P = .013; t = 3.317, P = .029, respectively). In contrast, significantly decreased apoptosis rates of OL/strain V and OL/DessVac were measured compared to that of uninfected OL cells (t = −3.399, P = .027; t = −4.229, P = .013, respectively). Furthermore, the apoptosis rates did not significantly differ between either natural viruses Hu-H2 and Equ-Cres (t = 0.456, P = .672) or laboratory strains V and DessVac (t = 0.485, P = .653) (Figure 3B). Moreover, we could show here that these opposite biological signatures were independent of the strains’ host background.

4.1 New full-length genomes

Full-length genomic sequencing of isolates Hu-H2 and Equ-Cres, and strain DessVac confirmed that they are authentic BoDV-1 viruses, displaying similarly high levels of sequence conservation (≥95%) characteristic of BoDV-1 viruses.^{6-8, 25-31}

4.2 Phylogenetic trees’ analyses

Phylogenetic trees altogether revealed a maximum divergence of 5.55%, 5.34%, and 4.94% at the genome, P-gene, and N-gene levels, respectively. Notably, the genetic distance could be at the same level between different batches of the same strain (DessVac, 1960 and 1990) as between different strains, shown for the P-gene.

As for human viruses, apart from the level of divergence, phylogenetic trees suggested clustering trends by country and/or country region rather than by clinical diagnosis or year of isolation. At the genome or single-gene level, clusters could be assigned to countries (Japan, USA, Austria, Australia, and Germany) and German regions. In contrast, different diseases of patients (MDD, BP, OCD, CFS, and encephalitis) or time of isolation (1993 to 2019) turned out to be unrelated to genetic differences of virus isolates/strains.

4.3 BoDV-1 epidemiology

What our phylogenetic analyses suggested controverts the currently promoted hypothesis. The bottom line of the latter boils down to three points: (a) human disease is always fatal, (b) BoDV-1 infection is mainly restricted to single endemic clusters in Germany, and (c) transmission occurs through a reservoir animal, the bi-colored white-toothed shrew.¹⁰ In contrast and apart from global healthy carriers and psychiatric patients independently arguing against zoonosis,^{3, 17-22} this study here suggested that sequence similarities alone are likely inappropriate to establish circumscribed endemic areas, given the high genetic conservation of BoDV-1 (Table 5).

Regional clustering were determined earlier for animal viruses and shrews.^{11, 33} However, independently evaluated phylogenetic trees assigning Japanese macaque #33 strain⁴⁴ to other GenBank sequences (N- and P-gene), revealed that the degree of homology was not only independent of host species and time of sampling, but also unrelated to countries.⁴⁵

4.4 Overlapping genetic signatures

Furthermore, phylogenetic analysis unraveled that instead of distinct host-specific genotypes, human and animal BoDV-1 strains were differing by highly conserved individual genetic signatures. Single varying mutations were shared with other strains from either human or animal origin, but each signature's profile was unique. The view of characteristic signatures of human isolates assumed earlier^{24, 35} was supported by this study.

With respect to animal natural and nonnatural strains, we found a coincidental distribution among different human strains at more or less separate branches on either the P- or N-gene level. German psychiatric human and equine strains were more close to strain V than Australian viruses. Notably, partial sequencing of German isolates²⁴ were done in an expert laboratory in the United States, which never dealt with strain V.³⁵ At the N-gene level, the human encephalitis cases (transplant recipients) were more close to the American brain autopsy sequences, and at the P-gene level, also more close to those and the Austrian CFS patient as well as to laboratory strain He/80. Japanese human and bovine isolates were more close to strain H1766, another horse-derived laboratory strain³⁰ than to DessVac (batch in 1960). Japanese isolates²⁵ were, however, sequenced in collaboration with the above expert laboratory in the United States, which had sequenced laboratory strain C6BV (He/80)²⁷ but not strain H1766. Thus, at various sequence levels, this study could rule out previous⁴⁶ and currently repeated misconception¹⁰ which claimed that human isolates^{24, 25} are contaminants.

4.5 Phenotypes of BoDV-1 strains

Earlier findings had revealed that even very low genetic variability below 0.05% could be associated with significant differences in disease phenotype. He/80 variants differing by only two mutations each in the glycoprotein and the L polymerase induced a more severe neurologic disease in Lewis rats than nonmutated strains.³¹

Recently, our team could show that natural human strain Hu-H1 and nonnatural strain V were differently impacting host cell reproduction⁴⁰ and metabolic signatures.⁴¹ This study finally indicated that different phenotypes are universally applicable to any natural vs nonnatural strain. We found significantly decreased proliferation and enhanced apoptosis by two other natural strains of either human or equine origin (Hu-H2 and Equ-Cres), respectively, compared to the opposite effects by nonnatural (laboratory) strains V and DessVac. Opposite phenotypes could clearly distinguish between natural and laboratory strains which are otherwise genetically closely related, and thus confirm their authenticity. Given the still existing contamination debate,¹⁰ this is an important finding. Nevertheless, the question of what mechanisms may cause these differences, remained unsolved. At least, the opposite phenotypes appeared to be dependent upon significantly different levels of virus strain adaptation, namely different manipulation of cell metabolism, rather than to be influenced by different host species. According to a previous study of our group, human strain Hu-H1 was shown to promote cell apoptosis via upregulating expression levels of proapoptotic protein Bax and downregulating antiapoptotic Bcl-2 protein, whereas laboratory str. V did the opposite, as shown by Western blot analysis.⁴⁰ Another remarkable difference concerned the sensitivity to antiviral drugs. Although the in vitroreplication of human strain Hu-H1 could be inhibited by amantadine sulfate in a dose-dependent manner (0.2 to 1.2 µg/mL), no effects were observed using laboratory str. V.⁴⁷ Notably, the high sensitivity of human BoDV-1 strains to amantadine held also true in infected depressed patients who benefitted significantly from low dose well-tolerated amantadine treatment vs placebo.⁴⁷