Phylogenomic assessment of 23 equid alphaherpesvirus 1 isolates obtained from USA-based equids

Equid alphaherpesvirus 1 (EHV-1), a varicellovirus of the Herpesviridae family, is a widespread pathogen of equids which is capable of causing respiratory, ocular, reproductive (i.e., abortion), neurological disease (equine herpes myeloencephalopathy, EHM) and neonatal death [1‐5]. First isolated from abortion material in 1933 [6], EHV-1 is now recognized as a major threat to the global equine industry [7]. In particular, neurological disease in host equids represents a serious welfare and economic problem and has therefore been the focus of numerous prior investigations [8‐11].

EHV-1 is a double stranded DNA virus with an approximately 150 kilobase genome consisting of 76 unique open reading frames (ORF) and 4 duplicated repeat regions [12]. Previous studies have indicated that ORF 30 (DNA polymerase) substitution A2254G (leading to amino acid variation N752D) is associated with neurological disease in host animals [9, 13‐15]. Although the underlying mechanism for this is unclear, it has been suggested that EHV-1 with the A2254G substitution replicates at a higher level and induces a longer-lasting viraemia than isolates without this substitution [16]. Notably, isolates recovered from host animals with neurological disease do not consistently possess this substitution [17]. Although prior work has found that certain isolates are more capable of inducing abortion in experimental studies [18, 19], EHV-1 genetic determinants for abortion have not been identified. The ORF 68 region (which encodes a non-essential membrane associated component) has previously been utilized as a genetic marker for grouping EHV-1 isolates [20].

Although several phylogenomic assessments of naturally occurring EHV-1 have been previously performed, limited numbers of viral isolates obtained from USA-based host animals have been included in these analyses [17, 21, 22]. Ongoing phylogenomic investigation of existing isolates is likely to improve global surveillance and subsequent control of this virus. The purpose of this study was to perform phylogenomic analysis of EHV-1 isolates acquired from USA-based hosts, and to compare these isolates to previously sequenced global isolates. Based on previous work, we hypothesized that isolates would form clades based on geographic origin [17] and that the previously described ORF 30 A2254G substitution would be present only in isolates obtained from host animals with neurological disease [20].

Phylogenomic and recombinational assessment of the 23 newly sequenced EHV-1 isolates obtained from USA-based hosts revealed many similarities with previously sequenced global isolates. Phylogenomic assessment of previously sequenced EHV-1 isolates and related alphaherpesviruses have been reported, facilitating comparison with the results presented herein.

As previously described [17], the genome of EHV-1 isolates obtained from multiple hosts involved in the same disease outbreak were found to share a high degree of similarity. This finding is not unexpected and has previously been reported in studies of a similar nature for related alphaherpesviruses such as canid alphaherpesvirus 1 (CHV-1) [24] and felid alphaherpesvirus 1 (FHV-1) [25]. Within our sample set, multiple isolates from three notable outbreaks of reproductive disease were included: Iowa in 2011, Indiana in 2014 and 2016. In all 3 cases, most isolates from the same outbreak (with the exception of LS113812, Iowa 2011) were found to be near-identical. While isolates obtained from hosts in the same geographic region (USA, Asia, Australia or Europe) clustered together, there were numerous exceptions to this pattern which was observed. For example, newly sequenced USA isolates clustered with varying combinations of isolates from distant geographic locations, including Europe, Australia and Asia. While small variations in pairwise interregional genomic distances were detected, it is clear that EHV-1 isolate geographic origin cannot be determined solely on sequence data. In contrast, FHV-1 isolates obtained from widespread geographic locations mostly do form clades based on geographic origin [38]. Although the definitive reason for this discrepancy is unknown, possible reasons include sample size, degree of viral intraspecies conservation and host species differences in global animal (or animal product) movements. Although previous assessments [20, 39, 40] have sought to utilize EHV-1 ORF 68 as the primary method to classify isolates into clade structures, we chose to perform viral full genome sequencing to identify nucleotide substitutions throughout the genome, as has been previously performed [17] and described.

Since being described almost 20 years ago [20], the ORF 30 A2254G substitution has been investigated in the context of equine herpes myeloencephalopathy (EHM) [2, 9, 13‐15, 17, 41, 42]. Although the substitution has been shown to be significantly associated with neurological disease [20], notably the substitution is not present in all isolates recovered from hosts with EHM [17]. As the underlying mechanism for this association between this specific substitution and EHM is unknown, the reasons for this apparent inconsistency are presently unknown. Most (22/23) of the newly sequenced USA isolates included in the present study were obtained from host animals with reproductive disease, with only 1 originating from a host equid with neurological disease. As noted above, only 1/23 isolates possessed the ORF 30 A2254G substitution, which originated from a host with reproductive disease. The single isolate obtained from the host animal with neurological disease did not possess the ORF 30 A2254G substitution. Inclusion of a higher number of isolates originating from hosts with neurological disease in future studies of similar design are suggested.

Several other ORF 30 substitutions have been described, in addition to the well-known ORF 30 A2254G substitution. Certain EHV-1 isolates (obtained from animals with reproductive disease) have been noted to contain a non-synonymous substitution in ORF 30 at position 2258 (A2258C) [43]; this substitution was not detected in any of the isolates in our sample. In addition, a recently described non-synonymous ORF 30 substitution at position 2254 (A2254C) has been described and documented in isolates from the USA and France [44‐46]; again, this substitution was not detected in any of the isolates in our sample. In addition to the ORF 30 A2254G substitution, we detected 6 other unique non-synonymous substitutions in the 23 newly sequenced isolates, some of which have been previously described [17, 20]. Surveillance using assays targeting this highly conserved region is likely to be beneficial for use in future outbreaks. The effect of each of these substitutions on virulence, if any, is unknown.

Both intraspecies and interspecies recombination have been shown to be prevalent mechanisms of diversity in EHV-1 and related alphaherpesviruses [17, 24, 25, 38, 47‐51]. We detected evidence of recombination in most (22/23) of the newly sequenced EHV-1 isolates in the present study. It is unknown why one of the newly sequenced isolates (LS050627) did not demonstrate evidence of recombination, although this could represent the limitations of the predictive computational processes utilized for this purpose. Increased understanding of the mechanisms by which EHV-1 substitutions develop is likely to have significant implications for both disease surveillance and control in host equids.

The present study has several limitations, including sample size, incomplete host animal information, lack of host disease type diversity, and viral genome gaps following sequencing. The present study included 23 newly sequenced EHV-1 isolates from USA based host animals. While concerted attempts were made to include a higher number of isolates, sample availability was a limiting factor. Approximately 46 samples (suspected to contain either EHV-1 DNA or viable EHV-1) were initially screened, but only 23 yielded high quality EHV-1 DNA in quantities suitable for Illumina MiSeq sequencing. During sample collection, efforts were made to collect as much host animal data as possible. As many of the samples had been collected and archived many years prior, host information was not consistently readily available. Most (22/23) of the EHV-1 isolates were obtained from hosts with reproductive disease. Inclusion of a higher number of isolates from animals with neurological and/or respiratory disease may have facilitated assessment of relationships between viral genome substitutions and host disease type. Finally, in common with all Illumina platform viral genome sequencing assessments, the reference-based assembly of high GC content regions resulted in sequence gaps with the genome sequences of the isolates from this study. While unavoidable with this approach, this was accounted for during analysis and is therefore not expected to have affected the results or subsequent conclusions.

Overall, the genomes of the 23 newly sequenced EHV-1 isolates obtained from USA-based hosts were similar to previously sequenced global isolates. The previously described ORF30 A2254G substitution was infrequently detected in the newly sequenced USA isolates, most of which were obtained from host animals with reproductive disease. In line with previous findings, recombination was likely to be partially responsible for genomic diversity in the newly sequenced USA isolates.