1. Introduction
Avian orthoreoviruses (avian reoviruses, ARVs) are a persistent challenge to poultry producers in the United States and globally. Infection with ARVs has been associated with a variety of symptoms and syndromes in commercial poultry, including tenosynovitis/viral arthritis, enteric symptoms such as watery diarrhea, respiratory symptoms, myocarditis, viral hepatitis, runting-stunting syndrome (RSS), poor feed conversion with compromised weight gain and high rates of morbidity, and occasionally, mortality [
1,
2,
3,
4]. ARVs are ubiquitous among poultry flocks, and despite vaccination efforts, the frequency of ARV outbreaks affecting broiler chickens in the United States, Canada and China has increased in the last decade [
2,
3,
5,
6]. ARVs belong to the
Orthoreovirus genus of the
Reoviridae family, and their genome consists of ten double-stranded RNA segments classified into three large (L1, L2 and L3), three medium (M1, M2 and M3) and four small (S1, S2, S3 and S4) segments with sizes that vary between 1100 and 4000 bp [
7]. Traditionally, ARV genomic classification into genogroups has been based exclusively upon a single gene encoded by the S1 chromosomal segment, the SigmaC gene. The SigmaC (σC) gene product is the primary antigenic viral protein to which neutralizing antibodies bind and is the best characterized gene and gene product of the ARVs [
2,
4,
8,
9]. However, there is not a strong correlation between the SigmaC sequence (or genogroups based upon SigmaC sequence) and the observed pathogenicity and clinical presentation of the strain in the field [
9].
With the arrival of next-generation sequencing (NGS) technology, whole-genome sequencing (WGS) has become an important tool to determine the genetic background of ARV strains that escape neutralizing vaccine responses [
2,
10]. Although the workflow for the production of NGS data appears to be straightforward, the NGS of dsRNA viral genomes introduces some complications and is a relatively new practice for which “gold standards” and widely-accepted best practices have not yet been established. Viruses are obligated intracellular pathogens. If no enrichment of the viral genetic material or depletion of host genetic material is done, the majority of the reads recovered will map to the host genome, while only a very small percentage will correspond to the targeted virus [
2]. Even by disregarding the economic losses associated with up to >90% of the sequencing cartridge being occupied by non-target sequences, without a highly pure viral genomic sample, the sequencing run may not yield enough targeted reads to have sufficiently complete genome coverage, and therefore, a reliable genome assembly [
11].
Currently, there are several publications that report the use of WGS to determine the genetic relatedness of ARV strains isolated from poultry, most of which use short-read sequencing technology and different enrichment strategies to overcome host genomic contamination. Generally researchers opt for sequencing a small number of ARV isolates per sequencing run, hoping to obtain enough viral reads for good coverage [
12]. Egaña-Labrin and collaborators introduced a host rRNA depletion step with a Terminator
TM 5′-Phosphate-Dependent Exonuclease degradation step (the degradation of RNAs without a 5′ triphosphate cap) before converting the ARV RNA genome into cDNA [
2], while Chrzastek et al. [
13] used a reovirus single primer amplification (R-SPA) approach to enrich ARV cDNA. There are no reports showing if these intermediate steps significantly reduced host contamination and/or enriched for ARV-mapping reads after sequencing. James and colleagues first developed a protocol for the purification of ARV virions from crude cell lysates [
14] using Capto Core 700 kDa resin, which we have integrated into our purification/isolation scheme.
Herein, we have performed a systematic comparison of four different approaches (see
Figure 1) for ARV purification and enrichment at different levels (virion, RNA and cDNA) to determine which workflow will generate the higher number of ARV-reads after NGS, with the aim of setting up a gold standard for ARV WGS. Our protocol has been entered into protocols.io (DOI:
dx.doi.org/10.17504/protocols.io.14egn38z6l5d/v1, accessed on 29 June 2023) and is available to the public.
2. Materials and Methods
2.1. Experimental Design
This section details the overall experimental procedure, while the subsections (below) provide details for each procedure. Sixteen different combinations involving four purification steps were tested in these experiments, each of which is detailed in
Figure 1. Briefly, five T175 flasks of LMH cell (ATCC CRL-2117) monolayers at 95% confluency were infected with approximately 20 μL of ARV-infected cell culture supernatant and placed in an incubator (38 °C, humidified, with 5% CO
2) for 5 days. After that time, the infected LMH cells and supernatant were harvested, centrifuged at 3000×
g for 10 min at room temperature (RT), and the pellet was resuspended in 1600 µL of Virus Dilution Buffer (VDB) [
14]. Cells were lysed via sonication on ice (3 pulses at 30% amplitude, 10 s on and 30 s off, using a Branson Digital Sonifier 450 (Branson Ultrasonics Corporation, Brookfield, CT, USA). The initial 1600 µL of infected cell lysate was split into two aliquots of 800 µL each that were subjected to virion enrichment using Capto Core 700 resin (Cytiva, Marlborough, MA, USA, catalog number GE17-5481-01), or ultracentrifugation on a sucrose gradient. This resulted in two vials of enriched virions at a volume of 375 µL each. Then, each of the vials were divided into two, and 187.5 µL of each vial (~1.25 T175 flask of infected cells) was treated with the nuclease Benzonase (Sigma-Aldrich, St. Louis, Missouri, catalogue number 9025-65-4), while the other 187.5 µL remained untreated. After this second purification step, there were four 187.5 µL vials containing purified virions (two of which were treated with Benzonase, and two were untreated) from which RNA was extracted. Once again, each of the four virion preparations was divided into two, and half of each extract was subjected to host ribosomal RNA (rRNA) depletion, while the other half remained untreated. After this third purification step, there were eight vials of RNA (four whose rRNA was depleted, and four were untreated) that were converted to cDNA via reverse transcription. Finally, each of the eight cDNA samples were split into two, and half of the sample received ARV genome amplification using the R-SPA method, and the other half did not. The sixteen DNA samples resulting from this experiment were sequenced in a single short-read sequencing run.
For subsequent experiments in which the number of treatment combinations was reduced from sixteen to eight or to four, we reduced the number of infected cells (for testing eight treatment conditions, two T175 flasks were used, and one T175 flask was used when testing four) and volumes for purification steps one and two, accordingly. In total, seven ARV strains/isolates were used in this study (
Table S1).
2.2. Cell Lines and Culture Conditions
The LMH (chicken epithelial liver from ATCC, catalog number CRL-2117) cell line was propagated in Dulbecco’s Modified Eagle Medium (DMEM, Gibco, Waltham, MA, USA, catalog number 1195-065) supplemented with 8% heat-inactivated fetal bovine serum (FBS, R&D Systems, catalog number S11150) and 0.1% gentamycin antibiotic (Lonza, Walkerville, MD, USA, catalog number 17-518L). The cells were maintained at 38 °C and 5% CO2 with humidification.
2.3. Purification of ARV Virions Using Capto Core 700 Resin
ARV virion preparations were purified using the Capto700 slurry protocol previously described by James et al. [
14]. Briefly, 350 µL of infected lysed cells in VDB were centrifuged at 800×
g for 10 min to remove the nuclei and cell debris. The supernatant was collected in a fresh tube containing 100 µL of 50% Capto700 slurry. The samples were mixed in an end-over-end tumbler for 45 min at room temperature and subsequently centrifuged at 800×
g for 10 min. The top phase was transferred to a new tube containing 100 µL of 50% Capto700 slurry for another purification round in the end-over-end tumbler (45 min at room temperature). To clear all the Capto Core 700 resin, the samples were passed through an Illustra MicroSpin column (GE Healthcare, catalog number GE27-3565-01) at 800×
g for 5 min. Viruses were kept at 4 °C until processed for RNA.
2.4. Purification of ARV Virions Using a Sucrose Gradient
LMH-ARV crude lysate was made via the ultracentrifugation of cell culture media and debris at 32,000 rpm for 2 h at 4 °C. The LMH-ARV crude pellet was resuspended in 800μL DMEM (supplemented as described above) after centrifugation and layered on a 4-step sucrose density gradient. The sucrose density gradient was prepared from a stock 3 M sucrose working solution that was further diluted with nuclease-free H2O at 48%, 30%, and 15% of the original stock. The following ratios of sucrose and H2O from the bottom to top layers were as follows: 700 μL 3 M sucrose (1.59 g/cm3); 494 μL and 205 μL (48%) (1.23 g/cm3); 308 μL and 391 μL (30%) (1.13 g/g/cm3); and 154 μL and 545 μL (15%) (1.05 g/cm3), respectively. The crude lysate was layered gently over the gradient and centrifuged in an Optima L- 90 K ultracentrifuge (Beckman Coulter, Brea, CA, USA) at 36,000 rpm using a SW60 rotor for 4 hrs at 4 °C. The lower band of the gradient (enriched virions) was collected and resuspended in a 30% sucrose solution. The resulting suspension was ultracentrifuged again for 2 h under the same conditions (36,000 rpm at 4 °C in a SW60 rotor (Beckman Coulter, catalogue number 335650) to pellet the virions out of the sucrose solution. Finally, the virion pellet was resuspended in nuclease buffer (see below) for downstream procedures.
2.5. Nuclease Treatment
Purified virions were subjected to treatment with Benzonase. Purified virions from 2.5 T175 flasks of cells were pelleted (36,000 rpm for 4 h at 4 °C) and resuspended in 187.5 µL nuclease buffer (10 mM Tris HCl pH 7.5; 2 mM MgCl2; 10% sucrose) and mixed with 2.25 µL of Benzonase (~562 U) and incubated at 37 °C for 30 min. The reaction was stopped with 9 µL of 0.5 M EDTA.
2.6. RNA Extraction and Host/Bacteria rRNA Depletion
RNA was extracted using MagMAX™ Viral RNA Isolation Kit (Applied Biosystems, Waltham, MA, USA, catalog number AM1939) from an initial volume of 350 µL following the manufacturer’s protocol. Bacterial and host rRNA depletion was performed using the protocol from Parris et al. [
15]. Briefly, 1 µL of ssDNA probes (10 µM, IDT) was hybridized with 12 µL of total RNA in the presence of 2 µL NEB Probe Hybridization Buffer (New England Biolabs, Ipswich, MA, USA). Hybridization was conducted in a thermal cycler (BioRad, Hercules, CA, USA, model T100) at 95 °C for 2 min, followed by a decreasing temperature cycle from 95 °C to 22 °C at 0.1 °C/second and a final incubation at 22 °C for 5 min. The probe-bound rRNAs present in the sample were degraded immediately after hybridization with 2 µL of RNase H enzyme (New England Biolabs, catalogue number M0297S), 2 µL RNase H reaction buffer, and 1 µL of nuclease-free water with 5 µL of the hybridization mix (incubation at 37 °C for 30 min). Finally, the unbound ssDNA probes were digested at 37 °C for 30 min from each sample (10 µL) using 2.5 µL DNase I enzyme (New England Biolabs, catalog number M0303S) in the presence of DNase I reaction buffer (5 µL) and nuclease-free water (22.5 µL). The resulting rRNA-depleted RNA sample was further purified using 2.2X RNAClean XP beads (Beckman Coulter, catalog number A63987) following the manufacturer’s recommended protocol. RNA concentrations and the percentage of rRNA contamination were measured with a 2100 Bioanalyzer system (Agilent, Santa Clara, CA, USA) and the RNA 6000 Pico Kit (Agilent, catalog number 5067-1513).
2.7. cDNA Production and Amplification using R-SPA
The single primer amplification protocol (R-SPA) described by Chrzastek et al. [
13] was used to produce ARV cDNA for whole-genome sequencing. Briefly, SuperScript IV reverse transcriptase (ThermoFisher Scientific, Waltham, MA, USA, catalog number 18090010) was used to convert ARV RNA into cDNA. First, 10 µL of RNA was mixed with 1 µL of 100 µM primer R8N (produced using our specifications by Integrated DNA Technologies), 1 µL of 10 mM dNTP mix (New England BioLabs, catalog number N0447S), and 1 µL of nuclease free water and incubated at 95 °C for 4 min, followed by it being left for 1 min on ice. The mix was then added to 7 µL of SuperScript IV reaction mix consisting of 4 µL of 5X buffer, 1 µL of 100 µM DTT (ThermoFisher Scientific, included in the SuperScript IV mix), 1 uL recombinant RNaseOUT Recombinant Ribonuclease Inhibitor (40 U/μL, ThermoFisher Scientific, catalog number 10777019) and 1 μL of SuperScript IV reverse transcriptase (200 U/ μL). The 20 μL reaction was incubated in a T100 Thermal Cycler (BioRad) under the following conditions: 10 min at 23 °C, and 10 min at 55 °C, followed by 10 min at 80 °C. cDNA second strand synthesis was performed immediately using Klenow polymerase (New England Biolabs, product number M0210S). First, 1 μL of primer 10 μM R8N (Integrated DNA Technologies (IDT)), 1 µL of 10 mM dNTP mix (New England BioLabs, catalog number N0447S) and 2 µL of 10X Reaction Buffer were added to the 20 µL of the previous step, and the mix was incubated at 94 °C for 3 min and cooled to 4 °C in a thermocycler. Then, 1 uL of Klenow polymerase was added to the reaction (final volume 25 uL), and the mix was incubated at 37 °C for 60 min. cDNA was purified using X1.8 AMPure XP beads (Beckman Coulter, product number A63880) following the manufacturer’s protocol and subsequently amplified using the HiFi PCR Phusion kit (New England Biolabs, catalog number M0530S). The PCR reaction mix conditions were as follows: 31 μL of nuclease free water, 10 μL of 1X Phusion HF buffer, 1 μL of 10 mM dNTP mix, 2.5 μL of primer R (10 μM) and 0.5μL of Phusion DNA polymerase. The 45 μL mix was incubated in a thermal cycler (BioRad, model T100) under the following conditions: 30 s of denaturation at 98 °C; 35 amplification cycles of 30 s at 98 °C; 30 s at 50 °C; 1 min at 72 °C; and a final 10 min extension at 72 °C. The PCR products were again purified using X1.8 AMPure XP beads following the manufacturer’s protocol.
2.8. Whole-Genome Sequencing and Bioinformatic Analysis
Genomic libraries were generated with the Nextera XT DNA Library Preparation Kit (Illumina, San Diego, CA, USA catalog number FC-131-1024) and IDT for Illumina DNA/RNA UD Indexes Set A (Illumina, catalog number 20027213). Samples were run using a MiSeq Reagent Nano Kit v2 500 cycles cartridge (Illumina, catalog number MS-103-1003) on an Illumina Miseq instrument (Illumina). Raw reads were trimmed, and quality-filtered (Phred score > 30) using Trimmomatic [
16]. High-quality reads were mapped against the chicken genome (NCBI accession no. GCF_016699485.2) using BWA aligner [
17]. Reads that did not map against the chicken genome (potential ARV reads) were extracted using SAMtools v1.16.1 [
18] and used for de novo assembly using SPAdes v3.15.3 [
19] and reference-guided assembly with MIRA v3.4 in GalaxyTrakr [
20]. The generated contigs were mapped to the ARV strain S1133 genome (NCBI accession number KF741756–KF741765) and extracted using Geneious mapper with the highest sensitivity and set to five iterations (Geneious Prime 2022.1.1,
https://www.geneious.com). ARV-mapping contigs were used to produce complete genomes that were polished using Pilon v1.24 [
21]. The quality of the assemblies and estimated genome length was assessed using QUAST [
22] Galaxy v5.2.0+galaxy1 was also used on extracted ARV-mapping contigs in GalaxyTrakr [
20]. Genome coverage was estimated using SAMtools after mapping the filtered reads (with BWA) to each isolate’s complete genome.
2.9. Statistical Analysis
For the data generated in the first sequencing run, two-way ANOVA with Tukey’s multiple comparisons post hoc test was performed to elucidate potential significant differences in the number of raw reads, quality-filtered reads and ARV-mapping reads between the samples that went under virion purification using Capto Core 700 resin and sucrose gradient with and without subsequent R-SPA amplification. For the analysis of subsequent experiments, one-way ANOVA with Tukey’s multiple comparisons post hoc test was performed to elucidate potential significant differences in the number of raw reads, quality-filtered reads, proportion of chicken genome-mapping reads, proportion of ARV-mapping reads, number of contigs and ARV-mapping contigs between the samples that were subjected to different purification methods. All statistical analysis were carried out using statistical software GraphPad Prism version 9.3.1 (GraphPad Software, San Diego, CA, USA).
4. Discussion
This study aimed to determine the most efficient and cost-effective ARV genome enrichment strategy for WGS purposes. Most of the non-target reads produced via ARV-WGS come from the host (in our case, the chicken), which in some cases, reaches > 99% of the sequenced material [
2]. Theoretically, there are three options to enrich the ARV genetic content in a sample before sequencing: (i) reducing the chicken genetic content, (ii) increasing the ARV genetic content, or (iii) a combination of both. A reduction of the chicken genetic content in a sample can be accomplished before extracting the viral genetic material (while the virions are still intact) via virion purification, or after, by depleting the chicken DNA/RNA present in the sample. We included these two approaches in this study. We compared the performance of ultracentrifugation in a sucrose density gradient (a commonly used technique for virion purification [
1,
23]) to that of a protocol developed by James et al. that uses Capto Core 700 resin to purify ARV virions [
14]. Additionally, we tested a treatment with Benzonase after virion purification to further purify the intact virions before opening them for RNA extraction. Benzonase is an engineered endonuclease that degrades all forms of DNA and RNA, while having no proteolytic activity. Hence, we reasoned that Benzonase could help degrade residual host nucleic acids without impacting dsRNA protected within intact virions. The last purification step performed in this study to reduce the chicken genomic content was a host rRNA depletion treatment using a set of sixty-nine ~120 bp non-overlapping DNA probes covering the entire length of the chicken 18S and 28S ribosomal RNAs (rRNAs) and mitochondrial RNAs (mtRNAs) [
15]. These probes hybridize with target chicken RNAs, forming RNA:DNA complexes that are degraded via subsequent treatments with RNaseH (endonuclease that catalyzes the cleavage of RNA bound to DNA) first and DNaseI (endonuclease that nonspecifically cleaves DNA) after. The amplification of the ARV genetic content in a sample can be conducted via PCR amplification of the viral genome before NGS [
24,
25] or via target capture NGS [
26]. While we are unaware of an available target capture NGS scheme with ARV-specific probes, early this year, Chrzastek and collaborators reported a single primer amplification assay (R-SPA) that significantly improved the recovery of ARV mapping reads after sequencing [
13]. This approach was evaluated in the experiments documented herein, alone and in combination with the other ARV purification methods mentioned above.
Our results show that Capto Core 700 by itself slightly, but consistently, increases the number of ARV-mapping reads compared to when not virion purification is applied. Additionally, we observed that the sucrose gradient by itself works better than Capto Core 700 by itself at purifying ARV virions, but it also seems to decrease the amount of ARV in the sample quite dramatically. Losing a virion or viral RNA is expected after every purification step, and generally, this small loss is compensated by a high purity. The application of Benzonase and host rRNA depletion increased the yield of ARV reads from virions purified via Capto Core 700, but not via ultracentrifugation in a sucrose gradient. Capto Core 700 resin is made of porous hollow beads with an inactive shell and a ligand-activated core that traps and holds molecules smaller than viruses (under 700 kilodaltons) [
27]. Hence, small DNA/RNA fragments are expected to enter the core of the beads, bind to the internal ligands and be depleted from the sample. This indicates that the majority of the host contamination left after virion purification with Capto Core 700 are long nucleotide (DNA and RNA) sequences that are depleted with the subsequent enzymatic treatments (Benzonase and host rRNA depletion). We did not explore the application of Benzonase prior to virion purification with Capto Core 700, but this may improve the purification process even further. The further purification of virions isolated via ultracentrifugation in a sucrose gradient with host rRNA depletion seemed to negatively affect the ARV-read yield after NGS, and the application of Benzoanse does not improve the purity as drastically as it did for the Capto Core 700-treated samples in our experiments. Ultracentrifugation over a sucrose density gradient uses the high density (and therefore, lower buoyancy) of the viral particles, which travel across the gradient, to separate them from the host cell debris (membranes, organelle fragments and nucleic acids) that become trapped in between the sucrose layers [
14,
23]. We suspect that this technique, while purifying the virions, may reduce their numbers considerably more than the Capto Core 700 resin one can, and therefore, the subsequent purification steps (that always carry a minimal sample loss) more dramatically impact the outcome.
On the other hand, the amplification of ARV cDNA with R-SPA consistently improved the yield of ARV reads in the samples regardless of the purification method(s) used to remove the chicken genomic content. For this reason, only the strategies that included the R-SPA step were investigated in our subsequent experiments. The results obtained by applying only R-SPA to the samples (
Figure 4, average of 14% ARV reads in a sample) were consistent with the results reported by Chrzastek et al. [
13]. However, we observed that R-SPA seemed to work better for the vaccine strain S1133 (~45% ARV reads) than it did for the clinical isolates (<11% ARV reads). We first hypothesized that S1133 (vaccine strain and reference genome) was most probably among the group of strains used to design the R-SPA oligos, and therefore, we observed a better performance with this isolate. To test this hypothesis, we compared the sequence homology of R-SPA primer R8N [
13] with the ARV whole genomes obtained in this study, and we observed that the 5′ end of the ARV fragments (where R8N aligns) is very conserved, and therefore, it is very unlikely that the observed differences in the R-SPA performance were caused by different primer affinities between the ARV strains. Another potential explanation resides in the fact that although all the isolates were expanded in LMH cells for the same period of time, S1133 is cell-culture-adapted, so it could be disproportionately expanded compared to the other clinical isolates, yielding a higher viral titer, and therefore, more viral genomes for PCR amplification to act upon. This second hypothesis could not be tested because the techniques for quantifying ARV titers (plaque forming units, PFU) and/or genomes (qPCR) are not developed or well standardized. Still, when combined with any other ARV purification method (at the virion level or after RNA extraction), R-SPA consistently improved the enrichment of ASV reads in each sample.
The combination of virion purification with Capto Core 700 (with or without Benzonase treatment) followed by host rRNA depletion and cDNA amplification was the most cost-effective approach and the only strategy that allowed the assembly of whole genomes. This was unexpected as the average genome coverage estimated with the ARV reads of the samples purified with this combination was comparable to the average genome coverage of the samples that went through other enrichment strategies, demonstrating that the latter one yielded enough ARV reads to produce complete genomes as well. The main difference between the samples from which we obtained complete genomes and the other samples is the level of host contamination present in the sample, indicating that high levels of non-ARV contamination (>50%) negatively impact the assembly of whole genomes. Additionally, the combination of Capto Core 700 host rRNA depletion and cDNA amplification allowed the detection of a mix of ARVs in a clinical isolate. However, the efficiency of the enrichment strategy was lower in the mixed samples (<57% ARV reads) compared to that of the isolates that only contained one ARV isolate (>82%).