Genetic diversity and detection of atypical porcine pestivirus infections

Abstract

Atypical porcine pestivirus (APPV), an RNA virus member of the Flaviviridae family, has been associated with congenital tremor in newborn piglets. Previously reported quantitative polymerase chain reaction (qPCR)–based assays were unable to detect APPV in novel cases of congenital tremor originated from multiple farms from U.S. Midwest (MW). These assays targeted the viral polyprotein coding genes, which were shown to display substantial variation, with sequence identity ranging from 58.2% to 70.7% among 15 global APPV strains. In contrast, the 5′-untranslated region (5′ UTR) was found to have a much higher degree of sequence conservation. In order to obtain the complete 5′ UTR of the APPV strains originated from MW, the 5′ end of the viral cDNA was obtained by using template switching approach followed by amplification and dideoxy sequencing. Eighty one percent of the 5′ UTR was identical across 14 global and 5 MW strains with complete or relatively complete 5′ UTR. Notably, some of the most highly conserved 5′ UTR segments overlapped with potentially important regions of an internal ribosome entry site (IRES), suggesting their functional role in viral protein translation. A newly designed single qPCR assay, targeting 100% conserved 5′ UTR regions across 19 strains, was able to detect APPV in samples of well documented cases of congenital tremor which originated from five MW farm sites (1–18 samples/site). As these fully conserved 5′ UTR sequences may have functional importance, we expect that assays targeting this region would broadly detect APPV strains that are diverse in space and time.

Introduction

Pestivirus is a genus within the Flaviviridae family, currently represented by 11 different species (A–K) (https://talk.ictvonline.org/taxonomy/). One of these species, atypical porcine pestivirus (APPV), was discovered by metagenomic sequencing of swine clinical samples (Hause et al., 2015) and is currently the sole species represented in Pestivirus K. Subsequent research associated APPV with hypomyelination in brain and spinal cord and congenital tremor in newborn piglets, known also as myoclonia congenita, “shaker pigs” or “dancing pigs” (Arruda et al., 2016; Schwarz et al., 2017; Mósena et al., 2018). Severe tremor can impact the ability of the piglets to nurse, leading to high incidence of pre-weaning mortality (Schwarz et al., 2017; Sutton et al., 2019). It was suggested that trans-placental transmission of virus to fetuses during gestation led to the infection and expression of congenital tremors in newborn piglets (de Groof et al., 2016). Our recent research suggested that the substantial diversity at genomic level among APPV strains (Sutton et al., 2019) could prevent identification of novel strains when using quantitative polymerase chain reaction (qPCR)–based diagnostic tools. Although the polyprotein sequences are characterized by high genetic diversity, the 5′-untranslated region (5′ UTR) of the viral genomes is conserved across pestivirus species (Deng and Brock, 1993; Hellen and de Breyne, 2007). The viral 5′ UTR is involved in modulating viral RNA replication and translation (Grassman et al., 2005). Pestiviruses utilize an internal ribosome entry site (IRES) instead of a 5′ cap to facilitate the promotion of translation. Specifically, the IRES recruits host ribosomal units to the initiation codon of the viral genome for translation (Kolupaeva et al., 2000). The 5′ UTR is commonly used in pestiviruses for detection (e.g., APPV; de Groof et al., 2016; Kaufmann et al., 2019; Folgueiras-González et al., 2020.) or classification of species/genotypes (e.g., BVDV; Neil et al., 2019). The first objective of this research was to systematically assess sequence diversity of the viral 5′ UTR and polyprotein sequences, taking into account our recent detection of multiple APPV strains across U.S. Midwest and current global sequencing data. The next objective was to identify and evaluate conserved targets in these regions for the development of qPCR assays allowing identification and quantification of diverse APPV strains. The long-term goal of this research is to develop robust quantification assays to assess the role of swine host-genetics in APPV susceptibility.

Materials and Methods

All procedures were approved by the Institutional Animal Care and Use Committee of the University of Nebraska–Lincoln.

Sample collection and processing

Blood samples from piglets experiencing congenital tremor were collected from five different farms from the U.S. Midwest (MW1 to MW5). The farm operations varied in size and purpose from niche small family farms to research and commercial swine operations. The genetic source of affected piglets varied from purebred lines to commercial maternal and terminal crossbreds. The number of tissue samples from each site ranged from 2 to over 400 samples. Serum was obtained by centrifuging whole blood samples at 2,400 × g at 4 °C for 30 min. Serum samples were stored at −80 °C. Viral RNA was isolated using MagMAX-96 Viral RNA Isolation kit (Applied Biosystems) and MagMAX Express Magnetic Particle Processor (Applied Biosystems) following the manufacturer’s protocol. The viral cDNA was generated using random hexamers and the First strand cDNA Synthesis Kit (GE Healthcare Bio-Sciences).

Sequencing and genome assembly of a novel APPV strain

A set of 22 pigs expressing congenital tremor and originating from Midwest Site-2 (MW2) tested negative for APPV using qPCR assays described in Arruda et al. (2016) and Sutton et al. (2019). To evaluate the role of sequence diversity in potential lack of detection, sequencing of APPV from MW2 samples was performed based on genome preamplification of the viral genome and Oxford Nanopore MinION sequencing as described (Sutton et al., 2019). Briefly, 39 pairs of overlapping primers covering the whole viral genome were designed using Primal Zibra (https://primal.zibraproject.org) and two APPV genome sequences described as references (MK728876 and MF167291.1), including one originated from Midwest (MK728876–MW1; Sutton et al., 2019). Following polymerase chain reaction (PCR) amplification, equal amounts of amplicons were pooled and purified using AMPure XP beads. Library preparation and Nanopore sequencing were based on manufacture’s protocols using Ligation Sequencing kit 1D (SQK-LSK108) and the MinION (Oxford Nanopore).

Bioinformatic analyses and genome assembly of the FAST5 sequencing reads were based on tools initially designed for the Zika virus (https://github.com/zibraproject/zika-pipeline). First, adapters were removed from the FAST5 reads with Porechop v0.2.3 (Nanopore). The reads were then mapped to the reference genome (MF167291.1) with minimap2 (https://github.com/lh3/minimap2) and converted to BAM format with “samtools view.” The preamplification primers were trimmed and read coverage was normalized using the align_trim.py script. Sequence variants between the reference sequence and MW2 APPV genomes were called using “nanopolish variants” (https://github.com/jts/nanopolish). Finally, the consensus sequence of the MW2 genome was generated with the margin_cons.py script.

Complete sequence of the 5′ UTR APPV genome

Sequencing of the complete 5′ UTR of the APPV was accomplished through 5′ Rapid Amplification of cDNA Ends (RACE) using the Template Switching RT Enzyme Mix (New England BioLabs) and available polyprotein DNA sequences, according to the protocol. Briefly, a template switching oligo (TSO) (5′—GCT AAT CAT TGC AAG CAG TGG TAT CAA CGC AGA GTA CAT rGrGrG—3′) was incorporated at the 3′ end of the cDNA following reverse transcription initiated with a N^pro gene specific primer (5′—TGT AAG CTG TTG GGG ACT AGG T—3′), the first gene of the APPV genome. The complete 5′ UTR of the MW2 strain was amplified via nested PCR using the Q5 High-Fidelity DNA polymerase (New England Biolabs), a primer complementary to TSO (5′ CAT TGC AAG CAG TGG TAT CAA C 3′) and APPV specific reverse primers targeting N^pro gene (outer, 5′—TTT TGT ATG ACT TTC TCC TAC CAC CA—3′) and the 5′ UTR (inner, 5′—GTG ACG TCT GCC CCG TAC TCG—3′). Following dideoxy sequencing of the 5′ RACE amplicons, APPV specific primers (forward, 5′ GCA TAA TGC TTG GAT TGG CT 3′; reverse, 5′—GAC TTT CTC CTA CCA CCA GTT C—3′) were used for amplification and dideoxy sequencing the 5′ UTR in APPV-infected samples from multiple MW sites (MW1-5).

Genetic diversity of APPV genomes

Fourteen APPV strains from various locations around the world, with complete or relatively complete genome sequence, including 5′ UTR, were identified in the NCBI database. Although the number of submissions claiming “complete APPV genome sequences” is larger, our analysis of available APPV sequences including our own work suggested that the number of actual complete genomes is relatively limited. MultAlin software (Corpet, 1988) was used to align and assess sequence diversity across the 5′ UTR and polyprotein sequences, when available, between MW and the worldwide APPV strains. Conservation of the 5′ UTR was assessed at each nucleotide site across 5 MW strains and 14 global APPV strains. The same approach was performed for each of the twelve viral genes, at the nucleotide and amino acid sequences; due to availability of the relatively complete genome sequence, the only MW strain used in this analysis was MW1 (MK728876). The alignments of the 5′ UTR and polyprotein sequences performed by MUSCLE (Edgar, 2004) and implemented in MEGA X (Kumar et al., 2018) were used to generate the phylogenetic trees by using the Maximum Likelihood method and Tamura-Nei model (Tamura and Nei, 1993). The secondary structure of the APPV 5′ UTR, strain MW2, was obtained using RNAfold (Lorenz et al., 2011) and mfold software (Zuker, 2003).

Quantitative PCR assay development

A real-time quantitative (qPCR) TaqMan assay (Forward: 5′ AAC CTG AGA GAG AGG TAC CGA 3′; Reverse: 5′ TCA GTA GAC CCT ACA CCT A 3′; Probe: 5′ 6-FAM/AG ACG TCA C/Zen/C GAG TAG TAC ACC CA/IABkFQ 3′) was developed targeting the conserved 5′ UTR regions across Midwest and a subset of global APPV strains. These sequences originated from samples collected from Asia, Europe, and North-America. This qPCR assay was evaluated testing samples obtained from piglets affected by congenital tremor and unaffected controls from five U.S. Midwest sites. The qPCR was performed using the TaqMan Universal Master Mix II no UNG, 2X (Life Technologies), 0.25/0.5 μM of probe/primers, cDNA template (1.5 μL) in a 10 μL final volume. The thermal cycling profile included an initial denaturation at 95 °C for 10 min followed by 40 cycles of denaturation at 95 °C for 15 s and annealing and extension at 60 °C for 60 s using a CFX384 Real-Time PCR (BioRad) instrument. All the samples were tested in technical triplicates; outlier replicates with a difference in Cq > 0.5 were discarded. Samples with an average Cq < 35 were considered positive for APPV. The number of piglets affected by congenital tremor and piglets from unaffected litters (as controls) tested for APPV presence varied between sites, including the following: MW1 (n = 18 positive, 11 unaffected), MW2 (n = 6 positive), MW3 (n = 5 positive), MW4 (n = 2 positive), and MW5 (n = 1 positive, 1 unaffected). The positive piglets exhibited clinical signs consistent with congenital tremor including different degrees of body trembling and shaking and were identified by trained technicians or veterinarians. Unaffected controls were randomly selected from litters not affected by tremor. The same qPCR assay was tested on samples collected from piglets affected by congenital tremor across multiple years from the same site (MW1), including 2017 (n = 8), 2018 (n = 2 positive), 2020 (n = 2), and 2021 (n = 6). Assay specificity was also assessed in silico by aligning the complete APPV 5′ UTR with reference sequences of common swine viral pathogens including porcine respiratory and reproductive syndrome virus (PRRSV), porcine epidemic diarrhea virus (PEDV), porcine circovirus 2 (PCV2), senecavirus A (SVA), classical swine fever virus (CSFV), and foot and mouth disease virus (FMDV).

The analytical sensitivity of the qPCR assay was evaluated using standard dilution curves and 10-fold serial dilutions, APPV amplicons overlapping the assay (Forward: 5′ TTA ACC AGG CCT CTA GTA CCA C 3′; Reverse: 5′ GAC TTT CTC CTA CCA CCA GTT C 3′) and clinical samples (MW1 strain) assayed in triplicates. APPV copy number was calculated based on the fragment size concentration of APPV amplicon.

Results and Discussion

Variation in APPV RNA sequence led to strain-specific qPCR assays to detect APPV in piglets with congenital tremor

In a metagenomic analysis (Sutton et al., 2019), we reported a novel APPV strain (MK728876) associated with congenital tremor in a population of pigs from Midwest (MW1). Initially, a veterinary diagnostic laboratory did not detect the presence of APPV in these samples. We were also unable to detect APPV using a previously reported qPCR assay (Arruda et al., 2016). Genome sequencing and assembly of the MW1 strain revealed that local sequence variation overlapping the reported qPCR assay (Arruda et al., 2016) prevented APPV detection. More recently, using two available qPCR assays (Arruda et al., 2016; Sutton et al., 2019) APPV was not detected in piglets from another U.S. Midwest site (MW2) that exhibited clinical signs of congenital tremor. DNA sequence analysis of the putatively novel APPV strain (MW2) revealed three and two sequence nucleotide polymorphisms (SNPs) overlapping the qPCR probe and reverse primer, respectively, of the assay previously reported in Sutton et al. (2019). Specifically, this assay was previously designed based on MW1 APPV strain (MK728876). These sequence differences most likely affected the ability of the qPCR assay to detect the APPV genome in MW2 samples, leading to false-negative results. Genome sequences of the MW2 strain were utilized to create a strain/site-specific qPCR assay which was tested on all 22 MW2 samples. This novel qPCR assay detected APPV in 19 of the 22 MW2 samples. Inconsistent qPCR detection of APPV across farm sites emphasized the need for the development of a robust assay that can detect a diverse array of APPV strains, despite their level of genetic variation.

Substantial variation across viral polyprotein genes could lead to challenges in accurate genetic diagnostic of APPV in piglets with congenital tremor

Using the same approach described in Sutton et al. (2019), Oxford Nanopore sequencing was employed to sequence a novel APPV strain detected in MW2 samples. There were 1,637 DNA variations between this APPV genome (MW2) and MK728876 (MW1). There were also 811 DNA variations between MW2 and an APPV strain observed in Europe (MF167291.1). The MW farm sites from where these samples originated were approximately 160 km apart and yet had more differences between their sequences compared to the reference APPV genome isolated from Europe (MF167291.1). This finding emphasizes that the high degree of sequence variation among APPV strains could result in underdiagnosis of congenital tremor induced by APPV. In order to demonstrate this, sequence diversity across the polyprotein sequences was assessed in APPV strains with complete or relatively complete genomes (with 5′ UTR), including MW1 and 14 other strains sampled across the globe. The level of DNA sequence conservation in viral polyprotein genes varied from 58.9% in N^pro to 70.7% in NS3. A similar average sequence conservation was observed between structural (64.5%) and non-structural genes (65%) (Table 1). The less conserved viral gene, N^pro, coding for an autoprotease, cleaving itself, and creating the amino terminal end of the C protein (Stark et al., 1993) is nonessential for viral replication. Specifically, mutant classical swine fever virus (CSFV) with N^pro removed was found viable but replicating less efficient (Tratschin et al., 1998). Also, following infection with CSFV, N^pro appears to inhibit host-type I interferon and apoptosis and as a result promoting viral replication (Ruggli et al., 2003). In contrast, the most conserved gene, NS3, also a non-structural gene, has multiple functions harboring a chymotrypsin-like serine protease domain, a helicase, and a NTPase domain (Wiskerchen and Collett, 1991; Tamura et al., 1993; Warrener and Collett, 1995).

Table 1.

DNA and amino acid sequence similarity between 14 global and five U.S. Midwest APPV strains across 5′ UTR and polyprotein genes.

Gene	No. of strains	Gene length	% Conserved (DNA)	% Conserved (Protein)
5′ UTR (Global+MW)	19	385	81.0	N/A
5′ UTR (Global)	14	385	83.9	N/A
5′ UTR (MW)	5	378	91.5	N/A
Npro	15	540	58.2	67.8
Core Protein C	15	333	63.1	73.9
Erns	15	630	65.9	81
E1	15	597	66.8	86.4
E2	15	723	64.7	80.5
p7	15	192	66.2	79.7
NS2	15	942	64.2	83.1
NS3	15	2061	70.7	92.4
NS4A	15	201	65.2	88.1
NS4B	15	1017	68.6	92.6
NS5A	15	1416	60.7	72.9
NS5B	15	2256	65.8	83.6

Across the 12 APPV genes, the level of conservation was higher between amino acid sequences compared to RNA sequences. The largest level of amino acid sequence similarity was observed for NS4B (92.6%) and NS3 (92.4%), whereas N^pro was characterized by the largest sequence variation (67.8%). NS4B is thought to play a role in viral replication as a part of the replication complex and is also an integral membrane protein (Weiskircher et al., 2009).

The 5′ UTR is characterized by extensive sequence conservation across APPV strains

Like in other Pestiviruses, the vast majority of the APPV 5′ UTR harbors a predicted internal ribosome entry site (IRES). Although the function of the IRES was characterized in other members of Flaviridae and Pestiviruses in particular, there is no information about its role in APPV. The complete sequence of the 5′ UTR was obtained by amplification and sequencing of the 5′ end of the viral cDNA (MW2) using template switching approach. The size of the complete 5′ UTR of the MW2 strain was 378 nucleotides; this information was used to obtain 5′ UTR sequences from strains sampled from other MW locations (MW1, MW3-5). Sequence comparison revealed that the 5′ UTR is the most conserved region of the APPV genome, with identical ribonucleotide sequences ranging from 81.0% across global and MW sites/strains to 91.5% between MW strains (Table 1; Supplementary Figure S1). Across MW-observed strains, there were 32 polymorphic sites across the 378 nucleotide of the 5′ UTR. The MW2 and MW5 were found to be identical and had only 2 polymorphic nucleotides present when compared to MW3. Interestingly, these samples originated from niche small family operations. There were 22 polymorphic sites between the APPV detected in MW1 and MW4 which are commercial or research facilities that use modern swine genetic lines. Across global APPV strains, the 5′ UTR was identical among certain strains originated from Asia (MK216752.1 and MK216753.1; MK216749.1 and MK216751.1).

Our prediction of the secondary structure of the APPV 5′ UTR (MW2 strain), using RNAfold (Lorenz et al. 2011), showed the same multiple stem-loop and side stem-loop structures (Figure 1), as observed in other Pestiviruses such as in CSFV (Fletcher and Jackson, 2002). Some of the largest conserved segments of the APPV 5′ UTR were found to correspond with IRES structural domains that have been previously shown to be important for translation in viruses with similar IRESs (Chon et al., 1998; Sizova et al., 1998; Psaridi et al., 1999; Kolupaeva et al., 2000; Fletcher and Jackson, 2002). In addition, although the ribonucleotide sequence conservation of the 5′ UTR (most of it harboring the IRES) is 81.0% across the 19 APPV strains, sequence conservation pre-IRES region is only ~ 66%.

Figure 1.

Secondary structure of the 5′ UTR of the MW2 APPV strain generated via mfold software, with consecutively conserved nucleotide regions of 20 nucleotides or larger highlighted in green. The analysis of 5′ UTR in MW2 strain predicts that IRES starts at the position 61, whereas stem 2 results of base pairing between nucleotides 338-341 and 362-365.

Across the 19 APPV strains analyzed, there are five conserved 5′ UTR segments with a size larger than 20 nucleotides and located in stem 1 and 2, and domains II, IIIa, and IIIb. In the alignment of the 19 strains, the largest conserved region was found towards the 5′ end of the 5′ UTR, covering most of the domain II, starting at position 68 and continuing for 43 consecutively conserved nucleotides (Figure 1). Partial or complete deletions of the domain II in a CSFV reduced IRES function (Kolupaeva et al., 2000; Fletcher and Jackson, 2002), defined as a measure of translation initiation. The second largest conserved 5′ UTR segment includes 37 nucleotides and was located in the domain IIIa. Mutations which altered the conformation of domain IIIa in CSFV (Kolupaeva et al., 2000), or an entire deletion of the stem-loop structure replacing it with a single C residue in a CSFV transcription vector (Fletcher and Jackson, 2002), severely reduced IRES function. Deletion of domain IIIb in hepatitis C virus (HCV), a member of the Flaviridae with an IRES structure highly similar to Pestiviruses, was shown to inhibit stable binding of the IRES to eukaryotic initiation factor eIF3, which is essential for translation of viral proteins in both HCV and CSFV (Sizova et al., 1998).

Another highly conserved segment (36 nucleotides) is located at the distal end of the APPV 5′ UTR and overlapping the stem 2/domain IIIe. We predicted that in APPV stem 2 forms as a result of base pairing between nucleotides 345-348 and 369-372 (Figure 1). It was demonstrated that substitutions in stem 2 could reduce IRES function to as little as 15% in a CSFV transcription vector compared to wild type. Interestingly, introduction of novel substitutions that reestablished normal base pairing, while altering the original nucleotide sequence, also restored IRES function to wild-type levels (Fletcher and Jackson, 2002). Similarly, introduction of two nucleotide mutation in the stem 1b in CSFV was shown to reduce the role of the IRES in translation drastically, indicating that this stem is also important for IRES function (Kolupaeva et al., 2000). In addition, deletion of domain IIIe in bovine viral diarrhea virus (BVDV), a pestivirus, had been associated with a reduction in IRES function to 67% compared to wild type (Chon et al., 1998). Introduction of deletions and substitutions into domain IIIe in CSFV resulted in a reduction of IRES function from 47% to 71% (Kolupaeva et al., 2000). In HCV, it was also shown that mutations affecting the domain IIIe loop reduced IRES function to 0%–30% (Psaridi et al., 1999).

Phylogenetic analysis based on APPV 5′ UTR and polyprotein sequences

Phylogenetic analysis of the APPV 5′ UTR across different strains from Asia, Europe, and North-America, including five MW strains (Figure 2a), showed similar clustering at the 5′ UTR and at the polyprotein genes, both at the DNA and amino acid level (Figure 2b and c). In one of the clusters, the European and North-American strains are predominantly represented, whereas the other distinct cluster includes only Asian strains. There were some signs of possible evolution at a regional level, with examples seen in a small cluster of strains originating from Switzerland (MN099165.1, MN099167.1—MN099169.1) or in a cluster originating from China (MK216749.1—MK216753.1). Despite the relatively short geographical distances between farm sites, the MW strains did not cluster together when using 5′ UTR sequences. The small cluster of MW strains (MW2, 3, and 5), originated from small family farms with niche host genetics, share more sequence similarity with two Asian strains than MW1 or MW4. In contrast, MW1 and MW4 originated from modern commercial lines, medium to large swine operations, clustered with APPV strains detected in Europe. This level of genetic diversity between APPV strains sampled in U.S. Midwest underlines the importance of a robust diagnostic approach that could detect any strain.

Figure 2.

Phylogenetic analysis of 14 global and five U.S. Midwest APPV strains using the RNA sequences representing 5′ UTR (a) polyprotein genes (b) and the predicted amino acid sequence (c). The phylogenetic trees of the RNA (a and b) were obtained by using the Maximum Likelihood method and Tamura-Nei model (Tamura and Nei, 1993). The phylogenetic tree of the predicted amino acid sequences (c) was generated by using the Maximum Likelihood method and JTT matrix-based model. These analyses included 19 RNA sequences and 385 positions for the 5′ UTR (a), 15 RNA sequences and 10,908 positions for the polyprotein genes (b), and 15 amino acid sequences and 3,635 positions for the predicted viral proteins (c). The trees with the highest log likelihood are shown.

5′ UTR qPCR-based assays detected cases of APPV infections across different farm sites

Discovery of new cases of congenital tremor across U.S. Midwest highlighted the need of a robust assay that could broadly detect APPV strains, despite their level of genetic diversity.

Sequence alignment of the 5′ UTR obtained from samples collected from different Midwest uncovered conserved segments that could be ideal for detection of multiple APPV strains. The new assay targeted the highly conserved segments located from stem 1 and domain IIIb of the APPV IRES.

The newly designed qPCR was evaluated on a subset of available samples from Midwest sites (MW1–5; 1–18 samples/site). This assay detected the presence of APPV in 100% of well documented cases of congenital tremor across MW sites (Cq < 35). The same assay detected presence of APPV in samples collected from piglets affected by congenital tremor from multiple years (2017, 2018, 2020, and 2021) and originated from the same site (MW1). Control samples obtained from unaffected litters were found negative for APPV. Due to a lack of sample availability, we were unable to validate this qPCR in samples of European or Asian origin. However, since this assay is targeting 100% conserved 5′ UTR regions across 19 global strains, regions that may also have functional importance, we expect that assays would broadly detect APPV strains that are diverse in space and time.

The analytical sensitivity of the qPCR assay was evaluated using standard dilution curves using both standard APPV amplicons overlapping the assay and clinical samples. The limit of detection of APPV in clinical samples was 6.5 APPV copies (Cq ~38.4). Assay specificity was assessed in silico by aligning the complete APPV 5′ UTR with reference sequences of common swine viral pathogens (PRRSV, PEDV, PCV2, SVA, CSFV, and FMDV). No significant similarity was found between the complete APPV 5′ UTR, which was targeted by our assay, and the reference genomes of these viral species.

Conclusion

The high degree of sequence variation observed in the coding genes between various APPV genomes makes it difficult to use single targets in an assay to correctly determine viral presence. Following inconsistent qPCR diagnostic results involving samples collected from piglets affected by congenital tremor and originated from different farms, novel APPV strains from multiple sites across the Midwest were sequenced and analyzed. Substantial sequence variation across the RNA sequence encoding polyproteins prevented accurate detection of APPV by qPCR, leading to false negative results. Although some diagnostic laboratories currently employ the use of degenerate probes and primers to have a greater chance of capturing multiple strains (Peddireddi, 2020), there are still limitations to this approach. Since the viral genome will continue to evolve in time, even different combinations of degenerate primers could lead to spurious results. The RNA changes are more tolerated at the gene level compared to 5′ UTR, since synonymous substitutions will not impact protein sequence but could impact hybridization based detection. Our in silico analyses showed that previously reported assays targeting NS3 (Yuan et al., 2021) or NS5B (Beer et al., 2017; Yuan et al., 2021) genes have limited ability of detection (6.7% to 46.7%) when targeting the subset of 15 APPV strains used in our study. In contrast, as cited literature demonstrated, in vitro changes in conserved regions of 5′ UTR had important impact of translation initiation of viral proteins in Flaviridae. Full-length sequence analysis of multiple APPV genomes determined that the 5′ UTR region is characterized by high level of conservation, making it an ideal candidate region for qPCR diagnostic assays. Our newly designed qPCR assay targeting the 5′ UTR of APPV was able to detect APPV in piglets affected by congenital tremor across five Midwest farms. As the APPV genome continues to evolve, the qPCR assay will potentially need to be reevaluated to ensure the nucleotide sites are still conserved for accurate detection of APPV. For example, our in silico analysis showed that a previously reported assay targeting 5′ UTR (de Groof et al., 2016) has identical sequence alignments with only a subset (42.1%) of the 19 APPV strains with complete or relatively complete 5′ UTR used in our study. In contrast, a 5′ UTR qPCR assay that is reported more recently (Kaufmann et al., 2019) has identical alignment with the 5′ UTR of all 19 APPV strains.

Glossary

Abbreviations

APPV

atypical porcine pestivirus

UTR

untranslated region

IRES

internal ribosomal entry site

MW

midwest sites

PCR

polymerase chain reaction

qPCR

quantitative polymerase chain reaction

TSO

template switching oligo

Conflict of Interest Statement

The authors declare no real or perceived conflicts of interest.