4. Discussion
Porcine epidemic diarrhea (PED) to date remains a colossal burden to the global swine industries owing to the lack of successful vaccine in the field [12,13]. Moreover, other emerging and re-emerging swine enteric coronaviruses, including porcine deltacoronavirus (PDCoV) [24], and swine acute diarrhea syndrome coronavirus (SADS-CoV) [25], further complicate the field condition by affecting diagnostic accuracy and increasing the risk of viral recombination [26]. Hence, there is still a pressing need to develop a safe and effective vaccine, particularly the LAV, to mount this disastrous disease. The reverse genetics system is a powerful and widely used tool to study viral pathogenesis and novel LAV design by active modification of genes of interests. Previous studies using reverse genetics on other coronaviruses have identified many virulent/attenuating determinants that might be shared among different genus of coronaviruses [12]. Nevertheless, direct evidences of attenuation due to harboring those mutated determinant(s) in pigs are still limited in PEDV [16,23,27]. In the present study, to investigate the role of the S gene in the attenuation process of the PEDVPT 52 strain, we generated four infectious cDNA clones of G2b PEDV, including the parental virulent iPEDVPT-P5, attenuated iPEDVPT-P96, as well as two chimeric viruses (iPEDVPT-P5-96S and iPEDVPT-P96-5S) with exchanged S gene. We found that the iPEDVPT-P96 virus fully regained the virulence after the exchange of S gene derived from the highly virulent iPEDVPT-P5 virus, showing comparable patterns of viral shedding, diarrhea and histopathological changes. Our data confirmed that the S gene is the primary attenuating determinant of the iPEDVPT-P96 virus and the genetic backbone other than the S gene involving the attenuation of the iPEDVPT-P96 might be of less importance. However, the iPEDVPT-P5-96S still exhibited partial virulence resulting in severer villous attenuation and higher mortality rate compared to those caused by the iPEDVPT-P96 virus. This result supports that the virulence of PEDV might be a multigenic event.
In the present study, the first pair of cDNA clones of a virulent G2b PEDV and its derived attenuated strain were generated. Since these two viruses were closely related to each other, this established platform would allow for easier manipulate in studying the effect of single nucleotide polymorphism for identifying potential virulent/attenuating determinate(s). In the present study, we demonstrated that reversion of virulence or attenuation occurred after the S gene was reciprocally replaced. However, this finding is in disagreement with previously published results by Wang et al. [17], indicating that there was an absence of virulence reversion after S gene substitution between the highly virulent G2b strain, BJ2011C, and the avirulent G1a strain, CHM2013. There could be a few reasons for the discrepancy between our observations that of previous studies. First, the autonomy of the viruses is different. Viral attenuation in serial cell-culture or animal passage is a progressive process that involves a series of gene mutations and the subsequent alternated cooperative interplay between gene products and mechanisms influencing cell-virus interaction. Although the patterns of attenuating mutation are similar among G2b PEDV strains, the asynchronous mutations between different PEDV strains could lead to loss of the cooperative or complementary function(s) for other gene products. In the previous study, the used viral strains belonged to different genogroups [17,28]. Efficient viral assembly requires proper signaling and interaction between each structural protein. Since the BJ2011C virus belonged to genogroup 2b, we speculated that the reason a singular S gene exchange failed to reverse the virulence of the avirulent G1a CHM2013 strain might be due to, at least in part, the suboptimal cooperation between structural proteins in terms of PEDV morphogenesis. Second, the degree of attenuation differed between the CHM2013 virus and PEDVPT-P96 virus. Comparing to the avirulent CHM2013 virus that induced no detectable viral shedding in two-day-old piglets, iPEDVPT-P96 virus retained some levels of virulence as it caused viral shedding, observable clinical symptoms, and even mortality in the seven-day-old conventional piglets. Besides the S gene [17,28], the abolishment of the function of non-structural proteins (NSPs) alone has been demonstrated to attenuate highly virulent PEDVs by disrupting the antagonistic ability of host interferons (IFN) [16,27]. That means, when the function of certain NSP was abated, PEDVs lose its virulence, regardless of the S gene they carry. Therefore, variations of NSPs in the CHM2013 virus might have a critical effect on its virulence, and the effect might be stronger than that contributed by the S gene derived from the highly virulent BJ2011C virus. Similarly, we also speculated that the IFN-suppressive function provided by NSP(s) might also contribute to virulent differences between the iPEDVPT-P5-96S and iPEDVPT-P96 since nine mutations were identified in NSPs in the iPEDVPT-P96 as compared with that of PEDVPT-P5 [18]. Regardless, further studies are needed to clarify the hypotheses mentioned above.
Sequence comparison between the autologous highly virulent PEDVPT-P5 and attenuated PEDVPT-P96 revealed several amino acid substitutions [18], especially in the S gene (see Table 1). Consistent with other studies, these mutations chiefly accumulated in the S2 domain, presumably because of adaptation to Vero cells [12,29]. In the present study, the enhanced fusogenic ability in vitro was distinctly ascribed to the S gene, more specifically the S2 domain, derived from iPEDVPT-P96 and presumably represented the evolutionary process to increase viral progeny since Vero cells are not of swine-origin but readily susceptible to PEDV infection. Animal experiments with seven-day-old piglets revealed that the exchange of the S gene from the iPEDVPT-P96 resulted in a delayed onset of peak viral shedding, diarrhea, and milder villous atrophy when examined histologically at 3 d post-inoculation. Additionally, in contrast to the findings published by Wang et al. [17], Suzuki et al. [30] reported that replacement of the entire S gene or S1 sequence from a highly virulent OKN-1/JPN/2013 American type PEDV strain enabled the attenuated rPEDVGFP-CV777 to acquire virulence in piglets. In iPEDVPT-P96 virus, however, only two amino acid substitutions were identified in the S10 (C144T) and S1B (T554C) domains compared to the iPEDVPT-P5 virus. For PEDV, S10 and S1B domains are known to contain sialic acid-binding and receptor-binding domains, respectively; thus, are crucial for viral entry [31]. Although Hou et al. [23] previously showed that deletion of 197 amino acids in S10 domain of PEDV resulted in attenuation in piglets, a more specific epitope has yet to be identified. Therefore, we are curious about whether only these two mutations in iPEDVPT-P96 can attenuate the highly virulent iPEDVPT-P5 virus or if the S2 domain must be primarily accounted for the attenuation. Future studies will be conducted to elucidate these questions.
In this study, the first pair of cDNA clones for a virulent G2b PEDV, and its derived attenuated strain were generated, allowing us to use both gain-of-function and lose-of-function approaches to studying the role of the S gene in PEDV pathogenesis. We confirmed that the S gene is a crucial virulent/attenuating determinant for the iPEDVPT-P96, but its importance varies among different PEDV strains. Thus, other studies’ results alongside our results will provide valuable information for the future generation of novel chimeric or multivalent vaccines.