Discovery of novel astrovirus genotype species in small ruminants

Samples

Fecal samples submitted for parasitological diagnostics were kindly donated by the Institute of Parasitology, Vetsuisse Faculty, University of Bern (Bern, Switzerland). In total 164 fecal samples, derived from 56 sheep, 56 goats, 30 alpacas, 12 deer, four llamas, two bisons, two chamois, one giraffe, and one ibex were investigated. Additional information on the animals and their health status was not available. Fecal samples were suspended 1:10 (w/v) in sterile PBS (137 mM NaCl, 10 mM Phosphate, 2.7 mM KCl; pH 7.4) and stored at −80 °C until further analysis. BoAstV-CH13 and OvAstV-CH16 positive and negative brain tissues were used as positive and negative controls for RT-PCRs and were available from the archive of the Division of Neurological Sciences, Vetsuisse Faculty, University of Bern (Bern, Switzerland). The summary of the laboratory workflow is shown in Fig. 1.

RNA extraction

All RNA extractions from fecal suspensions for AstV-screening by RT-PCR were done with the QIAamp Viral RNA Mini Kit (Qiagen, Hilden, Germany) and RNA isolation from brain tissue was performed with TRI Reagent (Sigma-Aldrich, St. Louis, MO, USA) according to the manufacturers’ protocols. To prepare samples for NGS, 500 μL of fecal suspensions were centrifuged at 16,000×g for 3 min. The supernatants were then centrifuged through Vivaclear MINI Clarifying filters with 0.8 μm PES (Sartorius, Göttingen, Germany) at 2,000×g for 5 min. Next, 280 μL of the filtrates were treated with 2 μL Benzonase (1U/μL) (Merck, Kenilworth, NJ, USA) for 2 h at 37 °C. The benzonase was inactivated by adding ethylenediaminetetraacetic acid (EDTA) to a final concentration of 5 mM. Finally, RNA extraction was performed with the QIAamp Viral RNA Mini Kit (Qiagen, Hilden, Germany) with the modification that the carrier RNA was omitted. RNA extracts were stored at –80 °C until further analysis.

Detection of neurotropic astroviruses BoAstV-CH13 and BoAstV-CH15/OvAstV-CH16 by RT-qPCR

To detect BoAstV-CH13, the RNA extracts were analyzed by either of two different probe based RT-qPCR assays (CH13-A or CH13-B) as described previously. While the CH13-A assay targets the 5′ part of the viral genome in ORF1a, the CH13-B assay detects the center of the viral genome in ORF2. Both protocols revealed a very similar excellent accuracy, precision, and analytical sensitivity (Lüthi et al., 2018).

For the detection of BoAstV-CH15 as well as OvAstV-CH16 a primer pair and a probe were designed based on full-genome sequence alignments of the following three strains from Switzerland and Germany; BoAstV-CH15, OvAstV-CH16 and BoAstV-BH89/14, targeting the 3′ end of ORF2 using the Geneious software package (Version 11.1.4; Biomatters, Auckland, New Zealand). The RT-qPCR was performed as described previously (Kuchler et al., 2019). After each cycle, the fluorescence was measured with the 6-carboxyfluorescein (FAM) channel and data was analyzed using the Sequence Detection Software (Version 1.4; Applied Biosystems, Foster City, CA, USA) with automatic baseline detection and a manual threshold of 0.2. The positive-negative cut-off was set at the cycle threshold (ct) value of 35. Each RT-qPCR run was performed using a positive and negative brain tissue control as well as a water control. All primers and probes are provided in Table S1.

Pan-astrovirus RT-PCR

For the detection of other AstV, a previously described heminested RT-PCR protocol using five degenerated primers; four forward and one reverse primer, which target a 450 nt long sequence at the 3′ end of ORF1b of a broad panel of AstV was used (Chu et al., 2008). Primer sequences are provided in Table S1. First-Strand cDNA synthesis was performed using the GoScript reverse transcriptase (Promega, Madison, WI, USA) and the gene specific reverse primer. The heminested PCR was done with a GoTaq Green Master Mix System (Promega, Madison, WI, USA) in two reaction rounds. For the first round, the PCR was set up in 25 μL reactions containing 12.5 μL of 2× GoTaq Green Master Mix, 4.5 μL of cDNA and a mixture of two forward primers PanAstV_forward 1 and −2 and the PanAstV_reverse primer, each in final concentration of 10 μM (Chu et al., 2008). The PCR was carried out with the following setting: 2 min, 95 °C and 30 cycles each of 30 s, 95 °C; 30 s, 50 °C; 30 s, 72 °C and final elongation 7 min, 72 °C. For the second round, the same reverse primer PanAstV-reverse was used as well as a mixture of PanAstV_forward_nested 1 and −2 (final concentration of 10 μM) and 1 μL of the first-round PCR-product as a template. Temperature settings were the same as in the first PCR, but with 40 cycles instead of 30 cycles.

Next generation sequencing

Prior to NGS, libraries were prepared using TruSeq DNA Nano Kit (Illumina, San Diego, CA, USA). For cDNA synthesis and polyA-selection the SMARTer Ultra Low Input RNA Kit (Takara Bio Inc., Kusatsu, Japan) was applied. NGS was performed on an Illumina HiSeq 3000 or NovaSeq 6000 in paired-end mode (2 × 150 bp).

Bioinformatics analysis

Reads were quality-trimmed with trimmomatic (Version 0.36) and mapped to their respective host genomes (alpaca: BioProject PRJNA30567, assembly Vicugna_pacos-2.0.1, deer: BioProject PRJNA324173, assembly CerEla1.0, goat: BioProject PRJNA340281, assembly ARS1, sheep: BioProject PRJNA179263, assembly Oar_v4.0) using STAR (Version 2.5.3a). Quality-trimmed and unmapped reads were assembled via SPAdes (Version 3.11.1).

Resulting scaffolds were then aligned to virus databases (Genbank and RefSeq viral nucleotide sequences downloaded on 12th of April 2018, UniProt viral amino acid sequences downloaded on 12th of March 2018) using BLASTn (Version 2.7.1+, default settings) and DIAMOND (Version 0.9.10, default settings). In order to exclude false positives, the scaffolds with a virus hit were aligned to an in-house non-viral database consisting of archaeal, bacterial, fungal, mammal, and protozoal sequences. Scaffolds were considered false positive if they had a longer hit on a sequence of the in-house database compared to the virus databases or if they had a nucleotide hit of more than 10% of their own length to any sequence of the non-viral database. Scaffolds with hits to AstV on nucleotide and/or amino acid level were considered valid AstV hits. For further analysis, scaffolds with a length over 5,550 nt and the three AstV-typical ORF including the ribosomal frameshift sequence between ORF1a and ORF1b were selected.

Rapid amplification of cDNA ends

In order to complete the inchoate 5′ ends of the five almost full-length scaffolds, a rapid amplification of cDNA ends (RACE) was performed. This 5′ RACE was carried out with a RACE-System (Invitrogen™, Carlsbad, CA, USA) according to the manufacturer’s instructions except of the usage of SuperScript™ III Reverse Transcriptase (Invitrogen™, Carlsbad, CA, USA) instead of SuperScript™ II Reverse Transcriptase (Invitrogen™, Carlsbad, CA, USA). Nested PCR was performed with gene specific primers (Table S1) and Taq DNA Polymerase with Standard Taq Buffer (New England BioLabs, Inc., Ipswich, MA, USA) as described by the manufacturer and an annealing temperature of 55 °C.

After visualization on a 1% agarose gel, PCR-products were excised of the gel and purified using NucleoSpin Gel and PCR Clean-up system (Macherey-Nagel, Düren, Germany) according to the manufacturer’s protocol.

Each 3 μL of the purified PCR products was Sanger-sequenced using the BigDye^® Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems, Foster City, CA, USA) in a 3730 DNA Analyzer (Applied Biosystems, Foster City, CA, USA) according to the manufacturer’s protocol. Sequence data were analyzed using the Geneious software package (Biomatters, Auckland, New Zealand, Version 10.2.6), trimmed with an error probability limit of 0.05 on both ends and aligned to the respective scaffolds.

Phylogenetic analysis

Capsid precursor sequences of 36 AstV strains were used for phylogenetic analysis and included AstV strains of various ruminants and our newly discovered sequences. Based on these sequences, a phylogenetic tree was constructed using the maximum likelihood method with 1,000 bootstrap replicates based on the Le_Gascuel_2008 with Freqs. model (Le & Gascuel, 2008) in MEGA 7 (Version 7.0.26). The model was chosen using the Find Best DNA/Protein Models option in MEGA 7.

Recombination analysis

Putative recombination events were assessed using the Recombination Detection Program (RDP4, Version 4.94) (Martin et al., 2015), following the RDP4 Instruction Manual (http://web.cbio.uct.ac.za/∼darren/RDP4Manual.pdf) and with the highest acceptable p-value set to 0.01. Recombination events were considered only when involving at least one of our newly generated sequences and having a highest acceptable p-value of 0.01 with all of the following methods: RDP (Martin & Rybicki, 2000), GENECONV (Padidam, Sawyer & Fauquet, 1999), Bootscan (Martin et al., 2005), Maxchi (Smith, 1992), Chimaera (Posada & Crandall, 2001), SiScan (Gibbs, Armstrong & Gibbs, 2000) and 3Seq (Lam, Ratmann & Boni, 2018).

Confirmation of newly discovered astrovirus sequences

To confirm novel AstV sequences, discovered by NGS, a RT-qPCR using specific primer-probe-combinations for each putative virus-candidate was applied. Specific primers and probes were designed for each of the almost full-length scaffolds using the Geneious software package (Biomatters, Auckland, New Zealand, Version 11.1.4). The primer and probe sequences with the respective nucleotide positions of the target are provided in Table S1.

All RT-qPCR reactions were performed using the TagMan™ Fast Virus 1-Step Master Mix (Applied Biosystems, Foster City, CA, USA) in 10 μL reactions according to the manufacturer’s instructions with a final concentration of 500 nM for each primer and 125 nM for each probe. For each reaction 2 μL RNA were added to 8 μL Master-Mix. The RT-qPCR was performed using a CFX96™ Real Time System on C1000 Touch™ Thermal Cycler (BioRad, Hercules, CA, USA), with the following cycle settings: 10 min, 45 °C; 10 min, 95 °C, and 40 cycles (15 s, 95 °C; 20 s, 61 °C; 30 s, 60 °C). After each elongation step, fluorescence was measured and analyzed with the CFX Maestro software (Version 4.1.2433.1219, BioRad, Hercules, CA, USA) with an auto calculated baseline threshold. Samples with a ct-value <35 were defined as positive.

Retrospective screening of samples for newly discovered astrovirus-candidates

After confirmation of all newly discovered AstV-candidates, all remaining 159 samples were re-tested using the RT-qPCR protocols as described above. For each run, a non-template control was used as negative control. The respective confirmed samples served as target-specific positive controls.

Detection of neurotropic astroviruses BoAstV-CH13 and OvAstV-CH16 by RT-qPCR

Quantitative RT-PCRs were performed for the detection of neurotropic AstV. Positive controls showed ct-values on average of 22.73 (22.62–22.86) for BoAstV-CH13 and 23.22 (21.74–24.61) for OvAstV-CH16, respectively. All negative controls were scored negative. All 164 fecal samples had ct-values >35 or were reported as “undetectable” and were therefore diagnosed as negative for BoAstV-CH13 and OvAstV-CH16.

Detection of various astroviruses using Pan-astrovirus RT-PCR

To ensure the effective implementation and to establish positive controls, the protocol was first applied to brain-extracts of BoAstV-CH13 and OvAstV-CH16 positive animals. These samples showed a strong and clear band with the expected size of 450 bp. Nineteen of the 164 tested fecal samples derived from sheep (n = 6: S1–S6), goats (n = 8: G1–G8), alpaca (n = 3: A1, A3, A4) and deer (n = 2: D2, D5) showed a band at 450 bp and were therefore defined as Pan-AstV RT-PCR-positive.

All 19 samples, which were positive tested in Pan-AstV RT-PCR were further processed for NGS. Therefore, RNA was de novo extracted, capsid-bound viral genomes were relatively enriched and free nucleic acids were depleted using benzonase treatment.

NGS and bioinformatics analysis

Sequencing in paired-end mode (2 × 150bp) generated 45′455′879–233′701′511 reads per sample and raw reads were deposited in the European Nucleotide Archive (Accessions ERR3143214–ERR3143223). In 12 samples, a total of 29 scaffolds ranging from 521 to 6255 bp in size, showed AstV hits in the bioinformatics analysis on amino acid level, with sequence identity to known AstV sequences ranging from 39.3% to 96.4% (Table S2). The vast majority of these hits was to AstV strains identified in bovine fecal samples. Strikingly, in one sample of a sheep (sample S3), a 967 nt scaffold had a hit on the neurotropic BoAstV-CH13, isolate 42,799 (Bouzalas et al., 2016) on nucleotide (identity 83.4%, hit length 296 nucleotides) as well as on amino acid (identity 57.2%, hit length 306 amino acids) level. The k-mer coverage of this scaffold was <×4, which indicates relatively low RNA concentrations. The sequence mapped to the 3′ half of ORF2, which encodes the hypervariable part of the capsid precursor protein (Babkin et al., 2012).

For the remaining AstV hits, we further analyzed scaffolds with a minimum length of 5,500 nt and covering parts of all three AstV-typical ORFs. One such scaffold was identified in two sheep and in three goats (Table S2). These tentative AstV scaffolds were designated CapAstV-G2.1, -G3.1 and -G5.1 and OvAstV-S5.1 and -S6.1. Each of these scaffolds showed the ribosomal slippery sequence (5′-AAAAAAC-3′) at the ORF1a/1b junction and an overlap between ORF1b and ORF2. The translated nsp1ab sequences had an identity between 71.7% and 75.4% to the next best hit while the capsid protein precursor ranged from 57.5% to 73.7% (Table 1). Best hits were AstV strains previously described in cattle (Nagai et al., 2015; Tse et al., 2011), deer (Smits et al., 2010) and takin (Guan et al., 2018). Taken together, these findings clearly support the notion that these five scaffolds represent complete or almost complete viral genomes of novel MAstV strains.

Rapid amplification of cDNA ends

In four of the five samples analyzed, the 5′ RACE did work and Sanger sequencing added to the 5′ end of the respective almost full AstV scaffolds (G3.1, G5.1, S5.1, S6.1). This resulted in the completion of these four scaffolds, which, together with CapAstV-G2.1, were deposited on GenBank (Accessions: CapAstV-G2.1 (MK404645.1), CapAstV-G3.1 (MK404646.1), CapAstV-G5.1 (MK404647.1), OvAstV-S5.1 (MK404648.1), and OvAstV-S6.1 (MK404649.1)).

Phylogenetic analysis

To assess the genetic relationship of the newly discovered strains and partial sequences to other known MAstV, a phylogenetic comparison of all generated putative genomes to representative MAstV genotype species and so far unclassified strains with high sequence similarity based on the capsid precursor protein sequences (Fig. 2) was conducted. For the goats, all three novel CapAstV strains clustered in different branches of the phylogenetic tree. CapAstV-G2.1 and CapAstV-G3.1 are both related to two different BoAstV strains: BoAstV/JPN/Hokkaido 11-55/2009 (p-dist. 0.293) and BoAstV-B170/HK (p-dist. 0.361), respectively. CapAstV-G5.1 is very similar to the new OvAstV strains S5.1 (p-dist. 0.236) and S6.1 (p-dist. 0.027), clustering together in one clade of the phylogenetic tree. They cluster with a p-distance of 0.419 to the closest related AstV strain CcAstV-1/DNK/2010, which has been detected in feces from deer (Smits et al., 2010) (Table 1; Fig. 2). Partial sequences generated from goat feces cluster together with the here described CapAstV/OvAstV or with enterotropic strains from other ruminants. In the partial sequences generated from sheep and deer feces, the situation is similar to the one described in goats. As previously indicated, one partial sequence, S3.1, is not clustering close to other enterotropic AstV but rather near the neurotropic BoAstV-CH13/NeuroS1 cluster. In samples with multiple AstV hits, partial sequences cluster at different positions of the phylogenetic tree. OvAstV-S5.1 clusters together with CapAstV-G5.1/OvAstV-S6.1 and S4.1, while scaffold S5.2, detected in the same sample, clusters together with sequences from sample G8 (G8.3/G8.5).

Recombination analysis

Knowing about the high tendency toward recombination in RNA viruses (Matsui et al., 1998) and previous reported indications of interspecies transmission of AstV based on recombination analysis (De Battisti et al., 2012; Lan et al., 2011), it was decided to determine whether such recombination events may have occured in the newly discovered strains. Subsequent to the phylogenetic analysis, the five new AstV sequences were analyzed together with selected AstV strains for plausible recombination events. Three putative recombination events were reported (Table S4). Between the three strains OvAstV-S6.1, CapAstV-G5.1 and OvAstV-S5.1, two recombination events were reported, one starting at position 4,062 and ending at position 6,114 (recombination event 1) and the other starting at position 1,737 and ending at position 2,443 (recombination event 3) (Fig. 3A) in OvAstV-S6.1. All three sequences may be the recombinant. Recombination event 3 was predicted to the overlap between ORF1a and ORF1b, whereas recombination event 1 was predicted for almost the entire ORF2. Another putative recombination event (2) was identified between BoAstV-GX27/CHN/2014 and BoAstV/JPN/Hokkaido11-55/2009 or closely related sequences as potential parental sequences, with the 5′ breakpoint at position 36 and the 3′ breakpoint at position 1,154 resulting in CapAstV-G2.1 as the recombinant (Fig. 3B). Identities on nucleotide and amino acid level of the strains involved in the recombination events can be found in Table S3.

Confirmation of newly discovered astrovirus sequences

Using specific RT-qPCRs, all five newly discovered AstV-candidates could be confirmed with the following ct-values: G2, 27.65; G3, 25.26; G5, 26.99; S5, 25.77; S6, 25.28. The non-template control remained negative for all targets. Sample G5 showed a positive signal (ct-value 27.09) for the primer-probe-pair S6.1, which is not surprising, considering the fact that these two samples share highly similar sequences. In all other samples, no cross-reactivity could be detected between the individual targets.

Retrospective screening of samples for newly discovered astrovirus-candidates

All 159 samples remained negative for the primer-probe combinations G3.1, G5.1, S5.1, and S6.1. Surprisingly one sample originating from a goat, scored positive for G2.1 (ct-value 29). Based on information from the original submission site (Parasitology, Vetsuisse Bern), it could be found that this sample originated from the same holding as the sample G2.