Reply to “Classification of Emergent U.S. Strains of Porcine Epidemic Diarrhea Virus by Phylogenetic Analysis of Nucleocapsid and ORF3 Genes”

REPLY

Porcine epidemic diarrhea virus (PEDV) was first detected in U.S. swine in April 2013 (1). We subsequently published a paper, “Isolation and Characterization of Porcine Epidemic Diarrhea Viruses Associated with the 2013 Disease Outbreak among Swine in the United States,” in January 2014 in the Journal of Clinical Microbiology (2), where we described isolation and characterization of U.S. PEDVs. In addition, based on comparative sequence analyses and phylogenetic analyses of the entire genomes of 33 PEDV strains and nucleotide sequences of full-length spike (S), the S1 portion of the gene (the first 2.2 kb of the S gene), ORF3, envelope (E), membrane (M), and nucleocapsid (N), we suggested in that paper that the full-length S gene or the S1 portion is appropriate for sequencing to determine the genetic relatedness of U.S. PEDVs if only one gene is to be selected for this purpose. However, Wang et al. wrote a letter to the editor (Classification of Emergent U.S. Strains of Porcine Epidemic Diarrhea Virus by Phylogenetic Analysis of Nucleocapsid and ORF3 Genes [3]) to dispute our viewpoint and concluded “In summary, the phylogenetic reconstructions may indicate that the N and ORF3 genes should be given significant focus in the evaluation of genetic diversity of PEDV strains emerging in the United States.” Upon request from the Journal of Clinical Microbiology, we provide the following responses to the letter of Wang et al.

PEDV is an enveloped, single-stranded, positive-sense RNA virus with a genome of approximately 28 kb. The genome includes ORF1a and ORF1b, encoding the replicase polyproteins, and other ORFs, encoding four structural proteins (S, E, M, and N) and one nonstructural protein, NS3B (encoded by ORF3). PEDV S, M, and E proteins are three membrane proteins and are important components of viral envelope. The S protein is a glycoprotein on the virion surface, and it functions as the virus attachment protein, interacting with the cell receptor (4). In addition, the PEDV S protein is postulated to harbor epitopes to induce neutralizing antibodies (5–7) and is associated with growth adaptation in vitro and attenuation in vivo (8, 9). The coronavirus M protein is a glycoprotein that plays a pivotal role in the viral assembly process together with the E protein (10, 11). The coronavirus N protein is an unglycosylated structural protein that binds to virion RNA to form the nucleocapsid, which is further incorporated into a viral particle by budding into the compartments of the exocytic pathway, thereby acquiring the viral envelope (12). The PEDV accessory protein encoded by ORF3 is thought to be associated with cell culture adaptation and virus attenuation, though its association with virus attenuation remains to be experimentally confirmed by the reverse-genetics approach (13, 14). Generally speaking, PEDV S, E, and ORF3 genes have higher genetic diversity than M and N genes (15, 16).

Selection of target gene(s) for analysis really depends on the purposes of the study. For developing a diagnostic reverse-transcription PCR (RT-PCR) assay to detect various PEDV strains, the more conserved genes, such as M and N, could be chosen (15, 16). For molecular epidemiology study, if the purpose is to have a conservative estimate of virus evolution without influence of immune pressure, the N gene (the encoded N protein is located inside the viral envelope) could be used; if the purpose is to determine the relatedness and genetic diversity of viruses, the S, E, M, and ORF3 genes could be used (8, 13, 15–19). After PEDV emergence in the United States, swine practitioners and diagnosticians frequently asked which gene(s) of PEDV is appropriate for sequencing to study the relatedness and genetic diversity of PEDVs in U.S. swine. Ideally the whole-genome sequences should be determined to truly reflect the genetic profiles of PEDVs, but it is too expensive and time-consuming to perform whole-genome sequencing as a routine tool to serve the swine industry's need, even though advances in technology, such as next-generation sequencing, made it feasible. After comparing phylogenetic trees based on individual genes to the tree based on the whole-genome sequences as well as considering the fact that the PEDV S protein harbors the postulated neutralization epitopes, we suggested in our paper (2) utilization of the full-length S gene or S1 portion for sequencing and molecular analysis to reflect the relatedness and genetic diversity of different U.S. PEDVs. Wang et al. (3) performed phylogenetic analyses of 72 PEDV N sequences and 54 PEDV ORF3 sequences in their article to support their conclusion that the N and ORF3 genes should be considered to evaluate the genetic diversity of PEDV strains emerging in the United States. In our opinion, in order to determine which gene(s) is more appropriate to reflect the genetic diversity of PEDVs, phylogenetic analyses based on individual genes of the same viruses should be compared. Thus, we retrieved sequences of 47 PEDVs with whole-genome sequences available in GenBank as of 30 June 2014. We also included one PEDV (USA/Illinois [IL]/2013/59573-5) whose whole-genome sequences were recently determined by us for analysis. Phylogenetic trees based on full-length S, the S1 portion (the first 2.2 kb of the S gene), N, and ORF3 nucleotide sequences of 48 PEDVs are shown in Fig. 1A to D, respectively. The structures of phylogenetic trees based on full-length S or the S1 portion were similar. Among 20 U.S. PEDV sequences, 18 U.S. PEDVs clustered together, whereas 2 U.S. PEDV variants (USA/Ohio [OH]/2014/OH851 and USA/IL/2013/59573-5) formed a distinct cluster (Fig. 1A and B), indicating that U.S. PEDVs include at least two genotypes based on the S gene sequence analysis. However, in the phylogenetic trees of N (Fig. 1C) and ORF3 (Fig. 1D), all 20 U.S. PEDVs formed one cluster, indicating that the N and ORF3 sequences of the USA/OH/2014/OH851 and USA/IL/2013/59573-5 PEDV variants did not have significant differences from those of the remaining 18 U.S. PEDVs. Further comparative sequence analyses demonstrated that the 18 U.S. PEDVs and 2 U.S. PEDV variants had nucleotide identities of 96.2 to 96.6%, 93.1 to 93.6%, 99.7 to 99.9%, and 99.4 to 100% based on full-length S, the S1 portion, N, and ORF3, respectively. This again confirmed that the N and ORF3 sequences did not reflect the genetic diversity of U.S. PEDVs as accurately as the spike gene sequences, at least at the current stage of PEDVs circulating in U.S. swine.

FIG 1.

Phylogenetic analyses of the full-length spike (A), S1 portion of spike (B), nucleocapsid (C), and ORF3 (D) gene nucleotide sequences of 48 PEDVs (20 U.S. PEDVs and 28 non-U.S. PEDVs). The trees were constructed using the distance-based neighbor-joining method of the software MEGA5.2. Bootstrap analysis was carried out on 1,000 replicate data sets, and values are indicated adjacent to the branching points. The two U.S. PEDV variant strains are indicated by bullet points. Bar, 0.002 nucleotide substitution per site. IA, Iowa; IN, Indiana; MN, Minnesota; CO, Colorado; NPL, Newport Laboratories.

In summary, (i) ideally, the whole-genome sequences should be determined to truly reflect the genetic profiles of PEDVs; (ii) to a lesser extent, several genes, such as S, E, M, and ORF3, can be used to determine the genetic relatedness and diversity of PEDVs; and (iii) if one gene is selected, the spike gene would be more appropriate than the N and/or ORF3 gene to reflect the relatedness and genetic diversity of U.S. PEDVs, at least for current PEDVs in U.S. swine. This is in alignment with a common practice of using a viral gene encoding a surface protein, such as GP5 and hemagglutinin, for molecular epidemiology of porcine reproductive and respiratory syndrome virus and swine influenza virus, respectively, to assess genetic relatedness/diversity among field viruses in conjunction with pig flow management, biosecurity, and even vaccine selection (20–25).