Introduction

Porcine epidemic diarrhea virus (PEDV), a member of the genus Alphacoronavirus, is a pleomorphic enveloped virus that contains a large (approximately 28 kb), positive-sense, single-stranded RNA genome. It is a causative agent of PED, a highly contagious gastrointestinal disease in pigs with mortality reaching 100% in suckling piglets. While PEDV replicates efficiently in porcine enterocytes in vivo, it grows poorly in cultured cells unless adapted through multiple passages [6]. Like those of other coronaviruses (CoVs), the PEDV spike protein (PEDV-S) is responsible for both receptor recognition and fusion between the virus and host cells [10, 15]. Notably, the identity of the receptor used by PEDV is still a subject of controversy. While earlier studies have suggested that PEDV uses porcine aminopeptidase N (pAPN) as a cellular receptor [9, 12, 22, 24], more-recent studies have supported a contrary view based on several lines of evidence [17, 29].

CoV-S proteins, including PEDV-S, are members of the class I viral membrane fusion proteins, which also include hemagglutinin (HA) from influenza A virus (IAV) [3, 14]. Both CoV-S and HA fold into a metastable prefusion conformation and require proteolytic cleavage (priming) by host proteases to make them fusogenic. This cleavage produces a receptor-binding subunit (HA1 of influenza virus or S1 of CoV-S) and a membrane-fusion subunit (HA2 or S2), which remain associated through non-covalent interactions. While details of the prefusion structure of PEDV-S are lacking, recent studies of the spike proteins of the betacoronaviruses MHV and HKU1 have suggested that the overall prefusion structures of CoV-S and HA are similar [14] and that they might share a common mechanism of membrane fusion. However, a few fundamental differences between the two proteins should also be considered. While the fusion peptide of HA is located at the N-terminus of HA2, CoV-S has an internal fusion peptide located downstream of the N-terminus of S2. Consequently, two proteolysis sites are usually required for the conformational change of S2: one at the S1/S2 junction and the other at the so-called S2’ site located at the N-terminus of the fusion peptide [1, 2]. In addition, unlike HA, many CoV-S, proteins probably including that of PEDV, are not sufficiently primed by proteolysis during viral packaging release and thus require further proteolysis by extracellular proteases, cell-surface proteases, or lysosomal proteases [14]. Considering these common and unique properties of the two proteins, it would be interesting to investigate whether they could functionally replace each other in a virus infection context.

Recombinant retrovirus pseudotyped with PEDV-S has been generated and employed to investigate the PEDV entry mechanism [20, 33]. However, cells transduced by pseudotyped retrovirus usually display characteristics that are distinct from those of cells that are infected with PEDV. For example, cell–cell fusion is rarely observed in transduced cells. Notably, with the successful development of a system to construct single-cycle influenza A virus (scIAV) [23, 36], it is now possible to assess whether scIAV can be generated when PEDV-S is provided in trans instead of HA. The resulting pseudotyped virus (scIAV-S) will be beneficial especially for studying entry mechanism mediated by PEDV-S. The present study was undertaken to test the hypothesis that PEDV-S can functionally replace HA to drive replication of IAV by constructing scIAV-S and examining its ability to infect PEDV-permissive cells. We also showed, using our pseudotyped virus system, that sialic acid is critical for mediating PEDV entry. In addition, we demonstrated that IAV carrying PEDV-S derived from cell-adapted and field-isolated strains exhibited distinct characteristics in infected cells, suggesting different modes of entry among PEDV strains.

Materials and methods

Cells and viruses

Human embryonic kidney (HEK) 293T, VeroE6-APN, and Madin-Darby canine kidney (MDCK) cells were maintained at 37 °C in Opti-MEM (Thermo Scientific) supplemented with 10% heat-inactivated fetal bovine serum (FBS) and 100 IU of penicillin and 100 mg of streptomycin per ml in humidified 5% CO2 incubators. All scIAV-PEDV-S used in this study were generated in the HEK293T cells and stored at −80 °C until use. Recombinant PEDVAVCT12 was generated by reverse genetics and propagated in VeroE6-APN cells as described previously [8, 37]. Influenza A virus (A/PR/8/34) was generated and titrated as described previously [34].

Plasmid construction

The full-length PEDV S derived from PEDVAVCT12 (GenBank LC053455.1), PEDVYN144 (GenBank KT021232.1) and PEDVG2 (field isolate) were codon-optimized for expression in mammalian cells and synthesized (Genscript and Synbio Tech). Subsequently, each construct was subcloned into the pCAGGS expression plasmid. To ensure optimal surface expression, ER retention signals at the C-terminal end were removed from all constructs.

The pHW-HA-mCherry and pHW-ΔM2 plasmids were constructed as described previously [36]. To construct pHW-ΔNA, pHW2000 encoding the NA gene of A/PR/8/34 was subjected to site-directed mutagenesis to introduce two consecutive stop codons at amino acids 164 and 165. All plasmids were subjected to nucleotide sequencing to ensure that no unwanted mutations were inadvertently introduced.

Recovery of scIAV -S

scIAV-S expressing the mCherry protein were rescued using a strategy similar to one described in a previous study [36]. Briefly, HEK293T cells in a six-well plate were transfected with 0.5 μg each of pHW2000 plasmids encoding the seven segments (PB2, PB1, PA, NP, NA, M, and NS) from A/PR/8/34 and pHW-HA-mCherry together with 2 μg of the pCAGGS plasmid encoding each construct of PEDV-S, using Fugene HD (Promega). To construct scIAV-S lacking NA, pHW-ΔNA was used instead of pHW2000-NA. Likewise, scIAV-S lacking M2 was constructed by using pHW-ΔM2 instead of pHW2000-M. At 72 h after transfection, cell supernatants were harvested and adsorbed directly onto VeroE6-APN cells for further analysis.

Western blot assay

Western blot assays were carried out according to the published procedure with some modifications [35]. Transfected cells were collected and lysed in 200 μl of mammalian cell lysis buffer (50 mM Tris [pH 8.0], 5 mM EDTA, 100 mM NaCl, 1% NP-40 and protease inhibitor mixture) for 30 min on ice. After centrifugation at 10,000 × g for 5 min, lysates were separated on 10% SDS-PAGE gels and subsequently transferred to nitrocellulose membranes (Bio-Rad), followed by blocking with 5% non-fat milk in TBS-T for 1 h. Membranes were probed with one of the primary antibodies, including anti-PEDV-S mouse polyclonal antibody (a kind gift from Dr. Qigai He), and anti-β-actin mouse monoclonal antibody clone C-4 (Santa Cruz Biotechnology) followed by goat anti-mouse antibodies conjugated to HRP (Biolegend). The signals were visualized with western blotting detection reagent (Bio-Rad).

Immunofluorescence assay

VeroE6-APN cells were grown on a Lab-Tek II chamber slide (Thermo Scientific) in Opti-MEM supplemented with 10% FBS for 12 h at 37 °C. The adherent cells were inoculated with scIAV-S for 1 h at 37 °C. After three washes with PBS, cells were cultured in serum-free Opti-MEM in the presence of trypsin (2 μg/ml) for 24 h. Cells were fixed in 80% chilled acetone for 10 min and blocked in 10% FBS/1%BSA/PBS for 30 min. The slides were subsequently incubated with mouse anti-influenza-A-virus NP antibodies (Clone 2C9; Southern Biotech) for 1 h, washed three times with PBS, and incubated with FITC-conjugated anti-mouse IgG antibodies. After additional washes, the slides were mounted with Antifade Mounting Medium with DAPI (Vector Laboratories). Cells were examined using an Olympus IX51 fluorescence microscope.

Statistical analysis

Data were expressed as mean ± SD. Statistical tests were performed using the Student t-test in GraphPad Prism 5.0 (GraphPad Software).

Results

Recovery of scIAV pseudotyped with PEDVAVCT12 spike (scIAV-SAVCT12)

Recently, we established a reverse genetics system to construct scIAV in which the HA gene was replaced by the mCherry gene. The virus was recovered in HEK293T cells and subsequently propagated in MDCK cells stably expressing the HA protein [36]. Using a similar strategy, we attempted to rescue scIAV-SAVCT12 by transfecting HEK293T cells with plasmids necessary for IAV reverse genetics together with codon-optimized SAVCT12 as depicted in Fig. 1A. Notably, we did not add trypsin to transfected cells during virus rescue to avoid the detachment of HEK293T cells. Moreover, since cells expressing SAVCT12 underwent rapid cell–cell fusion, cells stably expressing SAVCT12 could not be obtained for the propagation of scIAV-SAVCT12. Alternatively, the trypsin-free viruses released from HEK293T cells were directly inoculated onto VeroE6-APN cells treated with trypsin 12 h after infection. At 24 after infection, mCherry expression was assessed by fluorescence microscopy. Unexpectedly, not only did we observe a strong mCherry signal, but we also detected extensive syncytia in scIAV-SAVCT12-infected cells (Fig. 1B). In the absence of trypsin treatment, we could still detect a few cells with an mCherry signal but no syncytium formation (Fig. 1B), suggesting that scIAV-S AVCT12 might be able to enter cells but trypsin is essential for induction of cell–cell fusion. It should also be noted that control supernatants from HEK293T cells transfected with pCAGGS expressing SAVCT12 alone did not yield detectable syncytium formation in VeroE6-APN cells, suggesting that the detected syncytia were not due to PEDV-S expressed from residual plasmids (data not shown).

Fig. 1
figure 1

Construction and infectivity of scIAV-SAVCT12. A. Schematic representation of the IAV reverse genetics platform used to generate scIAV-SAVCT12. B. scIAV-SAVCT12 (undiluted) was adsorbed onto VeroE6-APN cells for 1 h. The infected cells were further incubated in the absence or presence of trypsin (2 μg/ml) at 37°C for 24 h. The expression of mCherry and nuclear staining (using Hoechst 33342 fluorescent stain) were examined using a fluorescence microscope. Scale bars, 20 μm and 100 μm (for no trypsin treatment) C. VeroE6-APN cells on chamber slides were infected with scIAV-SAVCT12 (undiluted). At 24 h after infection, infected cells were fixed and subjected to immunofluorescence staining for ANP. Cells were also stained with DAPI to show nuclei. D. scIAV-SAVCT12 was constructed in the presence or absence of pHW2000 encoding PB2. The resulting viruses (undiluted) were adsorbed onto VeroE6-APN cells for 1 h. Infected cells were further incubated in the presence of trypsin (2 μg/ml) at 37°C for 24 h. The expression of mCherry and syncytium formation were examined by fluorescence microscopy. Scale bars, 20 μm

Full size image

To characterize scIAV-SAVCT12, we initially tried to purify rescued viruses from the supernatants and perform western blot analysis to detect co-expression of PEDV-S and IAV nucleoprotein (ANP). Unfortunately, despite several attempts, we could not obtain sufficient purified virus for the assay (data not shown). Alternatively, we performed immunofluorescence staining to examine the expression of ANP in the syncytium of scIAV-SAVCT12-infected VeroE6-APN cells and detected strong expression of ANP (Fig. 1C). To further test whether the mCherry expression in infected cells required active influenza A virus polymerase activity, we attempted to construct scIAV-SAVCT12 by omitting the plasmid encoding the PB2 polymerase and examined its infectivity in VeroE6-APN cells. As expected, in the absence of PB2, neither an mCherry signal nor syncytium formation could be detected in infected cells (Fig. 1D). These results suggest that scIAV can be pseudotyped with PEDV-S and that the pseudovirus can efficiently infect PEDV-permissive cells. It is also notable that the massive syncytium formation in scIAV-infected cells was unexpected, and this observation has never been reported elsewhere. This points to the possibility that scIAV-S might enter cells by a mechanism similar to that of PEDV.

Abrogation of M2 ion channel activity does not affect scIAV-SAVCT12 infectivity

Given that scIAV-SAVCT12-infected cells formed large syncytia in the presence of trypsin (Fig. 1B) and that SAVCT12 was not synthesized de novo in scIAV-SAVCT12-infected cells, we speculated that the observed syncytium formation might have been mediated by SAVCT12 remaining at the cell surface following internalization of scIAV-SAVCT12. In other words, replacing HA with SAVCT12 might have switched the mode of cell entry of scIAV-SAVCT12 from endocytosis to direct fusion at the plasma membrane. Moreover, since the M2 ion channel is essential for the uncoating of IAV in host cell endosomes, entry of scIAV-SAVCT12 by direct fusion with the plasma membrane would bypass viral-endosomal fusion, rendering the proton channel function of M2 unnecessary. To test this hypothesis, we constructed scIAV-SAVCT12 lacking a functional M2 ion channel (scIAVΔM2-SAVCT12) and assessed its infectivity in VeroE6-APN cells. As shown in Fig. 2, we detected undisturbed mCherry expression with large syncytium formation in scIAVΔM2-SAVCT12-infected cells when infected cells were treated with trypsin. Notably, we have shown previously that scIAVΔM2 carrying IAV HA showed no mCherry signal in MDCK cells [36].

Fig. 2
figure 2

M2 is not required for scIAV-SAVCT12 infection. scIAV-SAVCT12 was constructed using the pHW2000 plasmid encoding M with a disrupted M2 ORF. The resulting virus was inoculated onto VeroE6-APN cells for 1 h. Infected cells were further incubated for 24 h in the presence of trypsin (2 μg/ml) before being examined under a fluorescence microscope for mCherry expression and syncytium formation. Scale bars, 100 μm

Full size image

Infectivity of scIAV-SAVCT12 and recombinant PEDVAVCT12 was not affected by hypertonic sucrose and NH4Cl treatment

Previously, it was reported that PEDV strain KPEDV-9 is likely to enter host cells via endocytosis, as infectivity was remarkably impaired when cells were treated with endocytosis inhibitors [26]. It is still not known however, whether PEDV can also fuse directly with the plasma membrane without being internalized by endocytosis, as reported previously for other CoVs. Given that scIAV-SAVCT12 lacking the functional M2 ion channel could infect VeroE6-APN cells and form syncytia without de novo synthesis of SAVCT12, we thus hypothesized that scIAV-SAVCT12 and PEDVAVCT12 could enter host cells by direct fusion at the plasma membrane. To this end, VeroE6-APN cells were pretreated with hypertonic sucrose (0.45 M) to impair the formation of both coated and uncoated pits at the plasma membrane [4] before being infected with scIAV-SAVCT12 and recombinant PEDVAVCT12. In contrast to what has been reported previously [26], we detected no changes in either syncytium formation or mCherry expression in infected cells regardless of hypertonic sucrose treatment (Fig. 3A). Likewise, VeroE6-APN cells treated with NH4Cl (50 mM for 2 h) to neutralize the intracellular pH also displayed strong mCherry expression and extensive syncytium formation when infected with scIAV-SAVCT12 (Fig. 3A). In line with the observation in scIAV-SAVCT12-infected cells, infection of VeroE6-APN cells with PEDVAVCT12 was also hardly affected by treatment with hypertonic sucrose or NH4Cl solution (Fig. 3B). It is also notable that pretreatment of VeroE6-APN or MDCK cells with hypertonic sucrose or NH4Cl could substantially inhibit replication of influenza A virus (Fig. 3C). These results collectively suggest that at least PEDVAVCT12 can enter host cells via a mechanism that is independent of the clathrin-mediated endocytosis pathway.

Fig. 3
figure 3

Effect of hypertonic sucrose or NH4Cl treatment on entry of scIAV-SAVCT12 and recombinant PEDV. A. VeroE6-APN cells were pretreated with 0.5 M sucrose for 1 h or 50 mM NH4Cl for 2 h before being infected with scIAV-SAVCT12. Untreated cells were used as a control. At 24 hpi, infected cells were examined for syncytium formation and mCherry expression under a fluorescence microscope. Scale bars, 100 μm. B. VeroE6-APN cells were mock treated or treated with 0.5 M sucrose or 50 mM NH4Cl prior to infection with PEDVAVCT12 (MOI = 0.1). Infected cells were maintained in Opti-MEM supplemented with trypsin (2 μg/ml). At the indicated time points after infection, supernatants were collected and evaluated for TCID50 in VeroE6-APN cells. The mean ± SD from three independent experiments is presented. n.s., not statistically significant. C. VeroE6-APN or MDCK cells were mock-treated or treated with 0.5 M sucrose or 50 mM NH4Cl prior to infection with influenza A virus (A/PR/8/34; MOI = 0.01). Infected cells were maintained in Opti-MEM supplemented with TPCK-trypsin (1 μg/ml). At the indicated time points after infection, supernatants were collected and the amount of infectious virus (TCID50) was determined in MDCK cells. The mean ± SD from three independent experiments is presented. n.s., not statistically significant; **, p < 0.01

Full size image

Infectivity of scIAV-S depends on the degree of cell adaptation of the spike protein

Several studies have shown that naturally isolated PEDV strains, mostly in the G2 genogroup, replicate poorly in Vero cells [25]. Accumulated mutations in the S gene are strongly associated with cell adaptation [11, 18, 31]. It was therefore of interest to assess the infectivity of scIAV pseudotyped with S derived from PEDV with various degrees of cell adaptation. SAVCT12 is closely related to the classical PEDV strain SM98-1, which is considered a highly cell-adapted strain. In addition, a variant (YN144) derived from a non-adapted strain after the 144th passage in Vero cells [5] is deemed to be an intermediately adapted strain. We therefore synthesized a codon-optimized S gene (SG2) based on a PEDV isolated from intestinal tissues of a piglet experiencing severe watery diarrhea during the outbreak in Thailand in 2014. BLAST analysis of the SG2 protein revealed that the amino acid sequence of our protein was 99.12% identical to that of CH/GXNN/2012 (GenBank AFO42860.1). Moreover, a codon-optimized S gene of strain YN144 was synthesized. As mentioned previously, both synthetic constructs were designed to exclude C-terminal ER retention motifs (YxxΦ and KxHxx). When plasmids expressing each PEDV-S were used to transfect VeroE6-APN cells in the presence of trypsin, only those expressing SAVCT12 and SYN144 displayed massive syncytium formation (Fig. 4A). The failure to form syncytia in PEDV-SG2-transfected cells was not due to defects in protein expression in transfected cells (Fig. 4A). We next attempted to construct scIAV pseudotyped with SG2 (scIAV-SG2) and with SYN144 (scIAV-SYN144) and evaluated their infectivity in VeroE6-APN cells. As a control, scIAV-SAVCT12 was also generated and examined. While abundant mCherry-expressing cells and large syncytia were detected in the cell cultures infected with scIAV-SAVCT12 and scIAV-SYN144, those infected with scIAV-SG2 displayed considerably fewer positive cells with punctate expression of mCherry and no apparent syncytium formation (Fig. 4B). The absence of syncytia in scIAV-SG2-infected cells may be due to an intrinsic property of PEDV-SG2, as cells transfected with pCAGGS-SG2 also displayed less syncytium formation than those transfected with pCAGGS-SAVCT12 and –SYN144 (Fig. 4A). Taken together, these results suggest that PEDV strains with different degrees of adaptation might utilize distinct pathways for internalization into host cells.

Fig. 4
figure 4

Infectivity of scIAV bearing PEDV-S derived from viruses with different levels of cell adaptation. A. VeroE6-APN cells were transfected with pCAGGS encoding SAVCT12, SYN144 or SG2 (2 μg each). Transfected cells were maintained in Opti-MEM supplemented with trypsin (2 μg/ml). Transfected cells were examined 24 h later for syncytium formation. Arrows denote syncytia. Likewise, HEK293T cells were transfected with pCAGGS encoding PEDV-S using the same conditions. At 48 h after transfection, cells were harvested, lysed and subjected to western blot analysis using anti-PEDV-S1 antibodies. Beta actin was used as a loading control. B. scIAV psedotyped with SAVCT12, SYN144 or SG2 was constructed, and its infectivity in VeroE6-APN cells was evaluated. At 24 hpi, infected cells were examined for syncytium formation and mCherry expression under a fluorescent microscope. Scale bars, 50 μm (20 μm for scIAV-SG2-infected cells)

Full size image

Neuraminidase (NA) hinders the infectivity of scIAV-S

Several lines of evidence have indicated that PEDV utilizes sialic acid, possibly Neu5Ac, as a co-receptor for attachment and entry [13, 19]. Moreover, pretreatment of host cells with recombinant NA has been shown to substantially reduce PEDV infectivity, and the N-terminal domain (NTD) of PEDV-S (residues 1–249) has been proposed to be critical for sialic acid binding [9]. Unlike other previously described pseudotyped viruses, scIAV-S, produced in this study, contains active NA on its particles. It is therefore ideal for investigating the role of NA in infection by scIAV-S. Since scIAV-S does not require NA for release, we were able to construct an scIAV-S variant bearing a defective NA by introducing two stop codons in the NA gene (ΔNA). Of note, the NA deletion did not compromise the packaging signal of the NA gene, and therefore, packaging of the NA segment should be undisturbed. As expected, in the absence of NA, both scIAV-SAVCT12 and –YN144 could infect VeroE6-APN cells with substantially higher infectivity (Fig. 5A). While scIAV-SG2 with ΔNA did not trigger syncytium formation in infected cells, we observed significantly more cells displaying an mCherry signal when compared to cells infected with the virus with intact NA (Fig. 5B). Taken together, these results indicate that the NA activity of scIAV-S might hinder the attachment of the PEDV-S on the scIAV-S particles with sialic acid on the cell surface.

Fig. 5
figure 5

Effect of neuraminidase on the infectivity of scIAV-S. A. scIAV pseudotyped with SAVCT12 or SYN144 in the presence or absence of a functional NA gene in segment 6 was constructed. Syncytium formation was assessed in VeroE6-APN cells (in a six-well plate) infected with each scIAV-S construct. The mean + SD from three independent experiments is presented. **, p < 0.01. B. scIAV pseudotyped with SG2 in the presence or absence of a functional NA gene in segment 6 was constructed. The expression of mCherry was assessed in infected VeroE6-APN cells at 24 hpi. Scale bars, 20 μm

Full size image

Discussion

Constraints associated with in vitro culture of natural isolates of PEDV have considerably hindered the study of their replication, and, consequently, they have not yet been well characterized. In particular, the mechanism underlying how PEDV enters a host cell is not well understood. We demonstrate in this study that IAV particles could be efficiently pseudotyped with PEDV-S, and such pseudoviruses could be employed to investigate the receptor binding and cellular entry of PEDV. Using a strategy similar to the one used with our previously described scIAV system [36], we successfully rescued scIAV carrying the mCherry gene in HEK293T cells co-expressing PEDV-S derived from various strains and demonstrated that scIAV-SAVCT12 could infect VeroE6-APN cells and produce a strong mCherry signal. Unfortunately, due to the high fusogenicity of SAVCT12, we could not generate cells stably expressing the protein as a platform to propagate scIAV-SAVCT12 in the current study. However, SAVCT12 is now being modified to be less fusogenic without compromising its receptor-binding function.

One of the most intriguing findings is our observation that scIAV-SAVCT12-infected cells could form massive syncytia similar to those produced by PEDV-infected cells. Since SAVCT12 is not synthesized de novo in scIAV-SAVCT12-infected cells, the syncytium formation observed at neutral pH is likely to be due to interaction between the receptor and SAVCT12 remaining on the cell surface after internalization of the virus. Our data thus underscore the likelihood that scIAV-SAVCT12 can enter cells by direct fusion at the plasma membrane. It is worth noting that some studies have shown that treatment with trypsin allows some CoVs to bypass the endosome-dependent pathway for entry [1, 21, 30]. However, scIAV-SAVCT12 was recovered in the absence of trypsin, and virus inoculation was carried out using trypsin-free viruses. Notably, unless trypsin was added at 12 hpi, we could not detect any syncytium formation. Indeed, syncytium formation was entirely unexpected, as PEDV has been reported to enter cells by clathrin-mediated endocytosis even in the presence of large amounts of trypsin during infection [26]. We also showed that scIAV-SAVCT12 lacking the M2 proton channel could infect VeroE6-APN cells and trigger formation of large syncytia similar to those observed in scIAV-SAVCT12-infected cells. In contrast to our previous report that scIAV-HA requires functional M2 for infectivity [36], our findings here suggest that scIAV-SAVCT12 can undergo replication without the need for M2-mediated internal virion acidification in the endosome. That PEDV can enter permissive cells by both direct fusion with the plasma membrane and endocytosis is not surprising as similar phenomena have also been observed with other CoVs [7, 30, 32].

In this study, we also speculated that differences in the S proteins of different PEDV strains might play a role in determining the mode of entry. In line with our hypothesis, Park and colleagues have recently demonstrated that the cleavage of the S protein during viral assembly appears to affect the mode of MERS-CoV entry. If S is cleaved at the S1/S2 junction during biogenesis by host proteases such as furin or proprotein convertases in producer cells, the released virions are primed and able to enter target cells by fusing at the plasma membrane. Alternatively, virions released with uncleaved S need to be processed further by endocytic proteases in the target cell and therefore require endocytosis for entry [27]. Although PEDV-S, like other alphacoronaviruses, is believed to be present in an uncleaved form [15], data showing the cleavage of S in different PEDV strains are lacking. S proteins of cell-adapted PEDV such as PEDVAVCT12 or PEDVYN144 may be cleaved during morphogenesis of the virion, whereas those of PEDVG2 may require further processing in the target cells. Further analysis of cleavage of the S protein in cell-adapted and field-isolated strains will be critical for better understanding of the mechanisms underlying PEDV entry.

Another important point worth mentioning is that we showed, using scIAV-S, that sialic acid moieties located on the cell surface plays a role in PEDV attachment. Although it has been demonstrated that the NTD of PEDV-S could bind to sialic acid, most data of the available have been based on assays using recombinant S proteins [13]. Because the construction of recombinant PEDV bearing specific mutations in the S gene can be technically challenging, the available data concerning the binding of PEDV S to sialic-acid-containing glycoproteins in the context of infection have so far been limited to the use of natural isolates such as the PEDV-UU strain [15]. With a well-established IAV reverse genetics system, scIAV-S serves as a robust platform to assess binding of the S protein to sialic acid in the context of infection. Furthermore, since many CoVs, such as MERS-CoV [16], and transmissible gastroenteritis coronavirus (TGEV) [28], have been reported to bind to sialic acids, it is also possible to construct scIAV pseudotyped with CoV-S of interest to investigate this interaction in more detail.

It is noteworthy that cells infected with scIAV-S display massive cell–cell fusion that is very similar to what is observed in cells infected with PEDV but does not occur in cells transduced with other pseudoviruses. Moreover, we showed here that PEDV-S derived from a field isolate, despite its poor growth in vitro, can be cloned into the expression vector and subsequently used to construct scIAV-S for further investigation. scIAV thus serves as a new platform not only for investigating the mechanism by which PEDV-S mediates cell entry but also to gain insights into the entry mechanisms utilized by cell-adapted and naturally isolated PEDV.