HSPA5 Promotes Attachment and Internalization of Porcine Epidemic Diarrhea Virus through Interaction with the Spike Protein and the Endo-/Lysosomal Pathway

ABSTRACT

Porcine epidemic diarrhea virus (PEDV) has caused huge economic losses to the global pig industry. The swine enteric coronavirus spike (S) protein recognizes various cell surface molecules to regulate viral infection. In this study, we identified 211 host membrane proteins related to the S1 protein by pulldown combined with liquid-chromatography tandem mass spectrometry (LC-MS/MS) analysis. Among these, heat shock protein family A member 5 (HSPA5) was identified through screening as having a specific interaction with the PEDV S protein, and positive regulation of PEDV infection was validated by knockdown and overexpression tests. Further studies verified the role of HSPA5 in viral attachment and internalization. In addition, we found that HSPA5 interacts with S proteins through its nucleotide-binding structural domain (NBD) and that polyclonal antibodies can block viral infection. In detail, HSPA5 was found to be involved in viral trafficking via the endo-/lysosomal pathway. Inhibition of HSPA5 activity during internalization would reduce the subcellular colocalization of PEDV with lysosomes in the endo-/lysosomal pathway. Together, these findings show that HSPA5 is a novel PEDV potential target for the creation of therapeutic drugs.

IMPORTANCE PEDV infection causes severe piglet mortality and threatens the global pig industry. However, the complex invasion mechanism of PEDV makes its prevention and control difficult. Here, we determined that HSPA5 is a novel target for PEDV which interacts with its S protein and is involved in viral attachment and internalization, influencing its transport via the endo-/lysosomal pathway. Our work extends knowledge about the relationship between the PEDV S and host proteins and provides a new therapeutic target against PEDV infection.

INTRODUCTION

Porcine epidemic diarrhea (PED) is an infectious and highly contagious viral disease characterized by anorexia, vomiting, diarrhea, dehydration, and weight loss. Newborn piglets are particularly prone to PED, with mortality rates as high as 80% to 100%, which has caused huge economic losses to the pig industry around the world (1, 2). PED virus (PEDV) is a single-strand, positive-stranded RNA virus belonging to the genus Alphacoronavirus of the family Coronaviridae, with a 28.5-kb genome consisting of ORF1a, ORF1b, and ORF2-6, encoding the structural proteins spike protein (S), envelope protein (E), membrane protein (M), and nucleocapsid protein (N) (3).

The virus life cycle begins with the attachment of the virus particle to the target cell (4, 5). Subsequently, viral internalization introduces viral genomic material into the host cell. These first steps are known as viral entry (6–8). During the PEDV entry process, the highly glycosylated trimeric S protein is responsible for virus fusion into the host cell after engaging with a receptor on the host cell surface (9, 10). The amino-terminal S1 subunit is required for host receptor recognition, while the carboxy-terminal S2 subunit is required for viral fusion (11). There is evidence that several host membrane proteins may be critical entry cofactors, except for the determined receptor for productive infection by coronaviruses (12–14). Therefore, exploring the mechanism of interaction between host cell membrane protein and S1 protein can further reveal the pathogenic and immune-escape mechanisms of PEDV.

Heat shock protein family A member 5 (HSPA5), also known as 78-kDa glucose-regulated protein (GRP78) or heavy chain-binding protein (BiP), belongs to the heat shock protein 70-kDa family, which are known as molecular chaperones for protein folding (15, 16). HSPA5 is widely expressed in the endoplasmic reticulum (ER) and plays a role in maintaining protein folding and quality control of processing proteins (17–19). HSPA5 has been reported as a peripheral membrane protein present on the cell surface, with its N-terminal and C-terminal domains partially existing outside the cell membrane as a multifunctional receptor to trigger different signaling pathways (20–23). The protein is also involved with cell proliferation, cell survival, apoptosis, and metabolism (24–27). Furthermore, HSPA5 has been reported as a coreceptor for several viruses, such as coxsackievirus A9, Borna disease virus, and Zika virus (28–32). HSPA5 has been reported to be involved in the entry and infection of β-coronaviruses in mammalian cells (33–36), but not in α-coronaviruses such as PEDV.

In this study, we screened membrane proteins of Vero cells and identified HSPA5 as an important cofactor in PEDV entry into Vero and LLC-PK1 cells. We systematically proved, in depth, that HSPA5 is involved in the PEDV attachment and internalization processes, as well as in virus transport via the endo-/lysosomal pathway.

RESULTS

Membrane protein HSPA5 affects PEDV infection.

To obtain PEDV S1-related host cell surface proteins, we performed a pulldown assay. Initially, we acquired plasma membrane proteins from Vero cells (Fig. 1A) and used recombinant PEDV S1 protein as prey to capture these proteins (Fig. 1B). Next, the captured proteins were detected using liquid-chromatography tandem mass spectrometry (LC-MS/MS). We compared the data from mass spectrometry to the control and detected 211 S1-related membrane proteins (data not shown). To identify the prominent proteins among these, we first used the Betweenness Centrality (BC) algorithm (37) to analyze and score the correlations between them (Fig. 1C). Subsequently, we utilized pathway-process enrichment analysis to locate the processes and pathways linked with these protein enrichments, using Molecular Complex Detection (MCODE) analysis (38) to further identify the critical proteins in these processes and pathways. These results were presented as interaction network diagrams to help us find correlations (Fig. 1D and E). In the results, some HSP family proteins received high BC scores and were enriched in both the linkage processes “protein processing in endoplasmic reticulum” and “SARS-CoV infection,” suggesting that these proteins may be related to coronavirus infection and have potential research value. Furthermore, among the HSP family proteins, HSPA5 and HSPD1 also had the highest correlation in the enrichment of MCODE1, so we used them as candidate proteins. Given the comprehensive analysis and their similar characteristics, 9 proteins were selected for further screening experiments.

FIG 1.

Membrane protein heat shock protein family A member 5 (HSPA5) affects porcine epidemic diarrhea virus (PEDV) infection. (A) Western blot (WB) of Vero cell plasma membrane protein extract samples. Sodium/potassium-transporting ATPase subunit alpha-1 (ATP1A1) was shown as a n plasma membrane control. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as an intracellular control. (B) Silver stain and WB of the eluted interacting S1-His of the indicated bait proteins. Red arrows denote PEDV S1-His protein in each lane. (C) Host membrane protein interaction networks and Betweenness Centrality (BC) of protein scores were graded and generated using Cytoscape software; top 50 scored proteins are shown. Node size indicates BC score: the larger the circle, the higher the score. Light gray lines indicate protein-protein interaction (PPI) curated in the publicly available STRING database. (D) Pathway-process enrichment analysis network of enriched terms. Left: terms colored by cluster ID, where nodes sharing the same ID are usually close to each other. Right: terms colored by P value, where terms containing more genes tend to be more significant. (E) Molecular Complex Detection (MCODE) analysis. Left: relationships between the highest rated proteins in each of the top 9 MCODE components. Light purple lines indicate PPI curated in the publicly available STRING database. Right: descriptions of the top 6 MCODE components.

Next, we designed small interference RNAs (siRNAs, Table 1) and used them to knock down select proteins in Vero cells to confirm their involvement in PEDV infection (Fig. 2A). The activity of Vero cells was not reduced by transfection with these siRNAs (cell viability > 90%, Fig. 2B). We then exposed the Vero cells to PEDV at a multiplicity of infection (MOI) of 0.1 for 12 h after treatment for 48 h with siRNA or si-negative control (siNC). We used flow cytometry to determine the relative infection efficiency of PEDV in these cells. As shown in Fig. 2C, HSPA5-knockdown Vero cells significantly inhibited PEDV infection, whereas STIM1-, HSPD1-, and ITGA3-knockdown cells had a modest inhibitory effect and SLC19A1-, PTK7-, HSP90B1-, and ITGB3-knockdown cells had no detectable impact. In addition, the knockdown of SLC3A2 resulted in a significant boost in PEDV infection. In conclusion, of the 9 candidate proteins, HSPA5 knockdown significantly reduced the infection efficiency of PEDV, suggesting a correlation between HSPA5 and PEDV infection.

TABLE 1.

siRNAs used in this study

Target gene	Sequence (5′–3′)
	Sense	Antisense
Monkey-HSPA5	GGCAGCUGCUAUUGCUUAUTT	AUAAGCAAUAGCAGCUGCCTT
Monkey-SLC3A2	GUGCAGUGGUCAUAAUCGUTT	ACGAUUAUGACCACUGCACTT
Monkey-SLC19A1	GCAAGCAGUUCCAGUUAUATT	UAUAACUGGAACUGCUUGCTT
Monkey-ITGA3	GUGGGACUUAUCUGAGUAUTT	AUACUCAGAUAAGUCCCACTT
Monkey-ITGB3	GCCCAUGUUUGGCUACAAATT	UUUGUAGCCAAACAUGGGCTT
Monkey-STIM1	GGAUGAUGUAGAUCAUAAATT	UUUAUGAUCUACAUCAUCCTT
Monkey-HSPD1	GGCUAUAUUUCUCCAUACUTT	AGUAUGGAGAAAUAUAGCCTT
Monkey-PTK7	GCCUCCUUCAACAUCAAAUTT	AUUUGAUGUUGAAGGAGGCTT
Monkey-HSP90B1	GGGACUGGGAACUUAUGAATT	UUCAUAAGUUCCCAGUCCCTT

FIG 2.

Preliminary screening of host membrane proteins which are related to PEDV S1 and affect PEDV infection. (A) Mass spectrometric data of selected proteins. “Gene,” “Description,” and “Accession” values are cited from Uniprot (https://www.uniprot.org/). “Score” is expressed as the credibility of the protein-matching degree. “%Cov(95)” indicates the coverage of the peptide sequence. (B) Cytotoxicity test of the indicated small interfering RNAs (siRNAs). Vero cells were plated in 96-well plates, transfected with the indicated siRNAs for 48 h, and supplemented with 10 μL of cell counting kit-8 (CCK-8) solution per well for cell proliferation assay. The viability of si-negative control (siNC)-transfected cells was set to 1.0. (C) PEDV infection of Vero cells. Cells transfected with siRNA were inoculated with PEDV (multiplicity of infection [MOI] = 0.1) for 12 h. Flow cytometry with anti-mouse monoclonal antibody (MAb) against PEDV intracellular nucleocapsid (N) protein on mock-infected or siRNA-transfected cells after infection. Top row: flow cytometry histograms of the percentage of cells infected with PEDV. Dotted lines: non-infected cells (<0.5%, negative infection) serve as a blank controls to exclude fluorescent background signal and determine whether cells were infected with PEDV. Bottom row: mean fluorescence intensity (MFI) of intracellular PEDV N protein-FITC (fluorescein isothiocyanate) illustrates abundance of viral protein in cells. *, P < 0.05; ***, P < 0.001; ns, not significant.

HSPA5 interacts with PEDV S protein.

To demonstrate the interaction between HSPA5 and PEDV, we co-transfected plasmids with the capacity to express Myc-tagged HSPA5 and Fc-tagged PEDV S proteins into 293T cells for co-immunoprecipitation (co-IP) assays. The recombinant HSPA5 and PEDV S proteins were both successfully expressed as soluble proteins and detected with mouse anti-Myc or mouse anti-PEDV S monoclonal antibody (MAb) (Fig. 3A). Subsequently, anti-Myc or protein A/G magnetic beads were exploited for exogenous co-IP analysis. As shown in Fig. 3A, exogenous proteins Fc-tagged spike or Myc-tagged HSPA5 were used as bait proteins, and subsequently successfully captured each other, which was not observed in the control groups. Further, the interaction between endogenous HSPA5 and PEDV S protein was confirmed by a co-IP analysis in the PEDV-infected Vero cells (Fig. 3B). In addition, the interaction between endogenous HSPA5 and PEDV S protein was further confirmed by confocal microscopy and supported by the Manders correlation coefficient (Fig. 3C). These results provide evidence that HSPA5 specifically interacts with PEDV S protein.

FIG 3.

HSPA5 interacts with PEDV spike (S) protein. (A) Exogenous co-immunoprecipitation (co-IP) analysis to identify HSPA5 interaction with PEDV S. HEK-293T cells were co-expressed with Myc-tagged HSPA5 (Myc-HSAP5) and Fc-tagged PEDV S (Fc-Spike) at 37°C for 48 h. Cell lysates were separately immunoprecipitated with anti-Myc or protein A/G magnetic beads, and the precipitated proteins were immunoblotted (IB) with mouse anti-Myc or mouse anti-PEDV S MAb, respectively. Endogenous co-IP analysis to identify HSPA5 interaction with PEDV S. Vero cells were seeded with PEDV (MOI = 1) at 37°C and then lysed with NP-40 lysate supplemented with protease phosphatase inhibitor at 18 hours post-infection (hpi). Using HSPA5 and PEDV S as bait proteins, rabbit anti-HSPA5 protein pAB or mouse anti-PEDV S protein MAb were used to identify the precipitated proteins in cell lysates by IB. Isotype immunoglobulin antibody was used as a negative control. IB, immunoblot. (B) Colocalization analysis of endogenous HSPA5 (red) and S protein (green) in the different susceptible infected cells at 12 hpi. Nuclei were stained with DAPI (4′,6-diamidino-2-phenylindole) (blue). Images were acquired at ×63 magnification. Representative images are from three independent experiments. Scale bars = 10 μm. Colocalization between HSPA5 and PEDV S was assessed by analyzing the Manders colocalization coefficient using ZEN software. Each value is the mean ± standard error of the mean (SEM) of three replicates.

HSPA5 is involved in PEDV infection.

Next, we utilized an siRNA targeting HSPA5 to knock down endogenous HSPA5 expression in Vero cells and verify its biological significance during PEDV infection. The expression level (~40%, Fig. 4A) and mRNA abundance (~55%, Fig. 4B) of endogenous HSPA5 were significantly reduced in the treated cells. Next, the Vero cells were inoculated with PEDV (MOI = 0.5) for 12 h to detect HSPA5 function. HSPA5 knockdown significantly decreased the PEDV progeny viral titers, as shown by 50% tissue culture infective dose (TCID₅₀) (~0.958 log₁₀ TCID₅₀/mL; Fig. 4C, left panel) and PEDV N gene copy number detected by absolute quantitative PCR (qPCR; ~1.26 × 10⁷ copies/reaction; Fig. 4C, right panel). The virus infectivity (~644.6 arbitrary units of fluorescence intensity detected by immunofluorescence assay [IFA], Fig. 4D) was also decreased.

FIG 4.

HSPA5 is involved in PEDV infection. (A to D) Knockdown of HSPA5 inhibits PEDV infection. Vero cells were inoculated in 24-well plates and transfected with 50 pM siRNA targeting HSPA5 or siNC for 48 h, and knockdown efficiency was determined. Next, PEDV infection was determined by inoculation with PEDV (MOI = 0.5) for 12 h after transfection. (A and B) Knockdown efficiency of HSPA5 from the protein (A) and gene levels (B) in Vero cells, respectively. (C and D) Effect of HSPA5 knockdown on viral infection was detected by 50% tissue culture infective dose (TCID50) to determine progeny virus titer (panel C, left) and absolute quantitative PCR (qPCR) (panel C, right) and indirect immunofluorescence assay (IFA) (D). (E to G) Overexpression of HSPA5 promotes PEDV infection. Two HSPA5-overexpressing Vero cell lines were screened by lentivirus and subcloning and inoculated with PEDV (MOI = 0.1) for 24 h in 24-well plates. (E) Protein levels of exogenous Myc-HSPA5 for pLV-HSPA5-1 Vero and pLV-HSPA5-2 Vero, with pLV-NC Vero as a blank control. TCID₅₀ (panel F, left) was used to detect PEDV progeny virus titers. Absolute qPCR (panel F, right) and IFA (G) were used to detect PEDV infection. Relative intensity of HSPA5 was quantified by analyzing the gray value of different bands using ImageJ software and normalized to GAPDH. Representative IFA images in infected cells stained with PEDV N (green) and obtained by a ×20 confocal microscope. Fluorescence intensity of PEDV N (green) was obtained by ImageJ analysis. Scale bars = 50 μm. Data represent means ± SEM from three independent experiments. Statistical analysis was performed using Student’s t test. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001.

To further demonstrate the biological significance of HSPA5 in PEDV infection, we screened two Vero cell lines overexpressing Myc-HSPA5 established by the lentiviral system and subcloning. As shown in Fig. 4E, exogenous HSPA5 was overexpressed compared to the pLV-Nc Vero cell line control group. After 24 h of PEDV infection (MOI = 0.1), HSPA5 overexpression significantly enhanced virus infectivity (~0.5 × 10⁷ and ~1.23 × 10⁷ PEDV N gene copies/reaction detected by absolute qPCR, Fig. 4F, right panel; ~1,466 and ~1,892 arbitrary units of fluorescence intensity detected by IFA, Fig. 4G) and increased the titer of progeny virus (~0.270 and ~0.375 log₁₀ TCID₅₀/mL; Fig. 4F, left panel) compared with the control group. Collectively, these results provide evidence that HSPA5 is involved in PEDV infection in Vero cells.

HSPA5 is involved in the PEDV virion’s attachment to the cell membrane.

To reveal the role of HSPA5 in PEDV virion attachment, we first focused on its cellular localization. Laser confocal microscopy was carried out with permeabilized or nonpermeable Vero and LLC-PK1 cells. Because the intracellular HSPA5 signal was intense, we used non-permeabilized cells to demonstrate the cell surface HSPA5 signal to reduce interference from the intracellular fluorescence signal. As shown in Fig. 5A, the colocalization pixels between the red fluorescently labeled HSPA5 and the green fluorescent cell membrane probe DiO indicated the presence of HSPA5 on the membrane. The experimental evidence from flow cytometry (Fig. 5B) and immunoblotting on extracted membrane protein (Fig. 5C) further confirmed the presence of HSPA5 on the cell membranes of Vero and LLC-PK1 cells.

FIG 5.

HSPA5 localizes to the plasma membrane of PEDV-susceptible cells. (A) Detection of endogenous HSPA5 on cell membranes in Vero and LLC-PK1 cells by laser confocal imaging. Left column images show permeabilized and non-permeabilized cells stained with HSPA5 rabbit pAb as the mock group. Right column images shows non-permeabilized cells stained with HSPA5-Alexa Fluor 647 (red) and cell membrane probe DiO (green). Nuclei were stained with DAPI (blue). White arrows denote colocalized pixels. Scale bars = 5 μM. (B) Detection of HSPA5 signal on PEDV-susceptible cell membrane by flow cytometry. Vero cells and LLC-PK1 cells were cryo-digested, harvested, fixed, incubated with rabbit HSPA5, and stained with Alexa Fluor 647 anti-rabbit Ab for flow cytometry. The mock group had no incubation antibody, while the control group used rabbit IgG as the first antibody and Alexa Fluor 647 anti-rabbit Ab as the second antibody. The positive signal range of HSPA5 is marked with a solid line in the superimposed histogram. Solid line (<0.5%, negative): group stained with rabbit IgG serves as an isotype-matched control to exclude fluorescent background signal. (C) Western blot detection of total cell lysates and purified plasma membrane proteins of Vero cells and LLC-PK1 cells with HSPA5 antibody. Western blotting with GAPDH antibody was used to verify extraction and purification of the plasma membrane.

To investigate whether PEDV S and HSPA5 can interact at the cell membrane during viral attachment, we exposed Vero cells to PEDV at MOI = 5 for 2 h at 4°C, fixed the cells without permeabilizing them, and used laser confocal imaging to detect colocalization of the S protein with HSPA5 during the PEDV attachment phase. As shown in Fig. 6A, we observed colocalization of the S protein with HSPA5, confirming that PEDV S can bind to HSPA5 during attachment. We subsequently measured the effect of HSPA5 knockdown during the PEDV attachment phase. Vero cells were transiently transfected with siRNA against HSPA5 (siHSPA5) or siNC for 48 h, followed by incubation with PEDV at MOI = 5 for 2 h at 4°C, and observed by indirect IFA and flow cytometry, respectively. Knockdown of HSPA5 significantly reduced the MFI (mean fluorescence intensity) of PEDV S protein on the Vero cell membrane (~109.1 arbitrary units, Fig. 6A). As shown in Fig. 6B (left), the siNC-treated Vero cells (gray) showed a stronger PEDV S protein signal than the siHSPA5-treated cells (red). Subsequent analysis of the PEDV S positive signal (~4.253 × 10³ arbitrary units of MFI; Fig. 6B, middle), percentage of PEDV-infected cells with signal (~7.83% of cells positive; Fig. 6B, right), and PEDV titer on the cell membrane (~0.5 log₁₀ TCID₅₀/mL, Fig. 6C) demonstrated a significant decrease, illustrating that knockdown of HSPA5 could significantly reduce the attachment of PEDV virions.

FIG 6.

Knockdown of HSPA5 reduces attachment of PEDV virus. (A) Laser confocal detection of PEDV S interaction with HSPA5 during the attachment phase. (B) IFA detection of PEDV S signal on the surface of siHSPA5-treated or siNC-treated Vero cells. Vero cells were transfected with siRNA or siNC for 48 h, seeded with PEDV (MOI = 5) for 2 h at 4°C in the confocal dish, washed three times with pre-cooled phosphate-buffered saline (PBS), fixed with 4% paraformaldehyde, blocked overnight with 5% bovine serum albumin (BSA), incubated with mouse anti-PEDV S Ab as the primary antibody for 2 h at room temperature (RT), and incubated with Alexa Fluor 488 anti-rabbit Ab as the secondary antibody for 1 h at room temperature, followed by detection. The three exhibition images were randomly selected from each group from independent experiments. Nuclei were stained with DAPI (blue). Scale bars = 20 μM. The fluorescence intensity of the images was obtained using Image J software. (C) Flow cytometry detection of PEDV S signal on the surface of siHSPA5-treated or siNC-treated Vero cells. Vero cells were seeded in 6-well plates and the process of transfection and PEDV inoculation was the same as described in Fig. 7A. Treated cells were detached with 10 mM EDTA in pre-cooling PBS, collected, incubated with mouse anti-PEDV spike Ab for 1 h at RT, and stained with Alexa Fluor 488 anti-rabbit Ab for 0.5 h at RT, followed by flow cytometry. Overlay histograms and average MFI relative to S protein positive percentage are shown. According to the negative fluorescent signal of the control group incubated with mouse IgG, the positive signal range of S protein is marked with a solid line (<0.5%, negative) in the overlay histogram. Published data came from three separate experiments, each containing biological repeats in which no less than 10,000 cells were detected by flow cytometry. (D) Progeny virus titer was detected by TCID₅₀. All data are means ± SEM of three independent experiments. **, P < 0.01; ***, P < 0.001.

To confirm which domain of HSPA5 on the cell membrane surface interacts with PEDV virions, plasmids expressing truncated HSPA5 fused with Myc tag were constructed as shown in Fig. 7A. Plasmids expressing Myc-HSPA5 truncated domain (NBD, intermediate domain of nucleotide- and substrate-binding domain [NSD], and substrate-binding domain [SBD]) and Fc-Spike were co-transfected into HEK293T cells for immunoprecipitation analysis (Fig. 7B). As shown in Fig. 7C, Myc-NBD (amino acids 1 to 410) was found to interact with PEDV S protein, whereas Myc-NSD (amino acids 212 to 531) and Myc-SBD (amino acids 410 to 651) did not. To further verify this result, specific polyclonal antibodies were utilized to block the N and C termini of HSPA5, respectively. qPCR was performed to measure the amount of PEDV RNA in each group. As shown in Fig. 7D, intracellular PEDV RNA and progeny virus titers were significantly reduced in the N terminus-specific antibody-blocked Vero and LLC-PK1 cells (~10.26 × 10⁸ copies/reaction in Vero cells, ~6.24 × 10⁸ copies/reaction in LLC-PK1 cells by qPCR; ~0.917 log₁₀ TCID₅₀/mL in Vero and ~0.958 log₁₀ TCID₅₀/mL in LLC-PK1 by TCID₅₀), while no significant difference was detected in the C terminus-specific antibody-treated cells. These results demonstrated that the N terminus of HSPA5 is involved in PEDV virion attachment to the cell membrane.

FIG 7.

HSPA5 N-terminal antibody blocks PEDV infection. (A to C) A co-IP assay was performed to determine which domain of HSPA5 binds to the PEDV S protein. Fc-spike was separately co-transfected with Myc-NBD, Myc-NSD, and Myc-SBD into 293T cells for 48 h. The transfected 293T cells were lysed with NP-40 lysis buffer, then immunoprecipitated with anti-Myc or protein A/G magnetic beads, and the precipitated proteins were IB with mouse anti-Myc or mouse anti-PEDV S MAb, respectively. (A) Structure of HSPA5 and construction of truncated structural plasmids. (B) WB detection of correct expression of exogenous proteins. GAPDH was used as an internal control. (C) Myc-NBD and Fc-S can both capture each other by co-IP. (D) Viable progeny titers and PEDV RNA copy numbers were used to detect PEDV infection in Vero and LLC-PK1 cells incubated with anti-HSPA5 antibody by TCID₅₀ and absolute qPCR. Vero and LLC-PK1 cells were seeded in 24-well plates and incubated with 12.5 μg/mL rabbit IgG antibodies or HSPA5 C and N terminus-specific rabbit antibodies at 37°C for 1 h, exposed to PEDV (MOI = 5), and incubated at 4°C for 2 h. Next, cells were washed three times with pre-cooled PBS and incubated at 37°C for 8 h. Treated cells were collected, and the total mRNA was extracted with TRIzol. After reverse transcription, absolute qPCR was performed to determine the N gene copy number. After repeated freezing and thawing 3 times, the treated cells were collected into tubes and centrifuged at 12,000 rpm for 10 min to collect progeny virus for TCID₅₀ determination. Data presented are from three independent experiments, each containing biological replicates. Each value is the mean ± SEM of three replicates. **, P < 0.01; ***, P < 0.001; ns, not significant.

Role of HSPA5 in PEDV internalization.

Because the lead compound HA15 immediately binds to HSPA5 and inhibits its activity (39), we used it to explore the role of HSPA5 in the internalization of PEDV virions. We first established the concentration range for HA15 and the internalization time of PEDV virions. Vero cells were adsorbed by PEDV virion (MOI = 5) at 4°C for 2 h, followed by incubation at 37°C for internalization. As determined by IFA, most virions were internalized within 1 h (Fig. 8A). The cytotoxicity of HA15 was examined using a cell counting kit-8 (CCK-8) in Vero cells. As shown in Fig. 8B, concentrations of 25 μM or lower will not reduce the activity of treated Vero cells (cell viability > 90%). Next, we examined the role of HSPA5 in PEDV internalization.

FIG 8.

Role of HSPA5 in PEDV internalization. (A) Most virions were internalized within 1 h of attachment. Vero cells were incubated with PEDV (MOI = 5) at 4°C for 2 h and then kept at 37°C for internalization. IFA was used to detect the internalization of PEDV at different time points. PEDV S protein was stained with Alexa Fluor 488 (green) and the nuclei were stained with DAPI (blue). Scale bars = 10 μm. (B) Viability of Vero cells treated with HA15. Vero cells grown in 96-well plates were separately treated with gradient doses of HA15 (1.5625, 3.125, 6.25, 12.5, 25, 100 μM) or dimethyl sulfoxide (DMSO; 100 μM) for 24 h, and cell viability was evaluated by a CCK-8 solution cell proliferation assay. Cell viability in DMSO-treated cells was set to 1.0. (C) Schematic of HA15 treatment of PEDV-infected Vero cells. (D) HA15 treated at the beginning of internalization. Vero cells were seeded on the cell culture plate and incubated with PEDV (MOI = 5) at 4°C for 2 h, then unbound PEDV virions were washed off with pre-cooled PBS, HA15-diluted medium was added, and cells were incubated at 37°C for 6 h and finally collected for testing. (E) HA15 treated after internalization. PEDV (MOI = 5) was cultured in Vero cells for 2 h at 4°C. The unbound PEDV virus was removed, the cells were incubated at 37 C for an hour, HA15 diluted medium was added, and cells were incubated for an additional 6 h. Finally, cells were gathered for examination. The TRIzol method was used to extract cellular mRNA from the experimental (treated with 3.125, 6.25, 12.5 μM HA15) and control groups (treated with 12.5 μM DMSO), and absolute qPCR was used to detect the PEDV N copy number after reverse transcription. The collected cells were repeatedly frozen and thawed three times and centrifuged to obtain progeny viruses for TCID₅₀ determination. Harvested cells were fixed and permeabilized, stained with mouse anti-PEDV N MAb for 1 h at RT, washed, and incubated with Alexa Fluor 488- goat anti-mouse pAb as the secondary antibody at RT. Finally, the PEDV N protein signal was detected by flow cytometry. Non-infected cells served as a blank control to exclude fluorescent background signal. Negative background signal, the positive signal range of PEDV N, is marked with a solid line (<0.5%) in the superimposed histogram. All data represent the means ± SEM from three independent experiments. *, P < 0.05; **, P < 0.01; ****, P < 0.0001.

HSPA5 is a multifunctional host protein (40), and inhibition of its activity may affect multiple physiological processes and thus influence the results of PEDV infection detection. We designed two sets of experiments to confirm the effect of HSPA5 inhibition on viral internalization, as shown in Fig. 8C. After 2 h of inoculation with 5 MOI PEDV at 4°C, the cells in group A were incubated in medium containing HA15 or dimethyl sulfoxide (DMSO) at 37°C for 6 h. Group B cells continued to incubate for 1 h before being incubated in the medium containing HA15 or DMSO for 6 h. As shown in Fig. 8D and E, treatment with HA15 significantly reduced the amount of intracellular PEDV RNA in the cells (~4.76 × 10⁷ copies/reaction of PEDV N gene in 12.5-μM HA15-treated group A; ~0.97 × 10⁷ copies/reaction of PEDV N gene 12.5-μM HA15-treated group B). In Vero cells treated with 12.5-μM HA15, the titers of intracellular progeny virus (~1.875 log₁₀ TCID₅₀/mL in group A; ~0.507 log₁₀ TCID₅₀/mL in group B) and the percentage of infected cells (~70.93% cells positive in group A; ~19.60% cells positive in group B) were decreased. Furthermore, it was found that inhibiting HSPA5 activity during PEDV internalization had a significantly greater impact on virus infection than inhibiting it after internalization, indicating the significance of HSPA5 during internalization. In conclusion, inhibition of HSPA5 activity affects the process of PEDV internalization and reduces the efficiency of virus infection.

HSPA5 participates in the endocytosis and transport pathway of the PEDV virion.

During internalization, the endocytosed viruses are regarded as cargo, transported to the early endosomes and then to the fusion sites (41). According to previous studies, PEDV enters cells through the endocytic pathway (42) and transfers virions via the endo-/lysosomal pathway (43–45). Since we showed in this study that HSPA5 is involved in the attachment and the internalization of PEDV, we further investigated whether HSPA5 is involved in the endocytosis and transport pathways of the virus. After 1 h of PEDV (MOI = 5) infection, the flow cytometry results revealed that the level of HSPA5 on the cell surface had significantly dropped (~6.195 × 10³ arbitrary units of MFI, Fig. 9A). This finding raises the possibility that HSPA5 was internalized along with the PEDV virion. Next, we used subcellular localization to detect HSPA5 to further confirm whether it participates in the endo/lysosomal pathway by binding with the PEDV S. As shown in Fig. 9B, the subcellular localization of PEDV in Vero cells transfected with BFP-HSPA5 and the subcellular colocalization of PEDV-HSPA5-BFP complex and early endosome antigen 1 (EEA1, labeled early endosome), Ras-related protein Rab-7 (RAB7, labeled late endosome), and lysosome-associated membrane glycoprotein 1 (LAMP1, labeled lysosome) is indicated as white spots marked by an arrow in the merged image; PEDV-HSPA5 complex was observed in early/late endosomes and lysosomes. In general, this suggests that HSPA5 binds with PEDV S and is involved in endosomal-lysosomal transport through viral internalization.

FIG 9.

HSPA5 accompanies PEDV into cells through endocytosis and traffics via the endo-/lysosome pathway. (A) Flow cytometry determined that HSPA5 on the host cell membrane was significantly reduced after PEDV infection. The process of virus internalization was the same as described above. At 2 h post-internalization, treated cells were harvested after dissociation with 10 mM EDTA in pre-cooled PBS, fixed in 4% PFA for 20 min without permeabilizing, stained with rabbit anti-HSPA5 MAb or rabbit IgG Ab for 1 h at RT, then stained with Alexa Fluor 647 donkey anti-rabbit pAb as the secondary antibody for 30 min at RT, followed by flow cytometry. Cells which were not incubated with PEDV were regarded as negative controls. Overlay of the histogram and MFI is shown. The positive signal area of HSPA5 on the superimposed histogram is marked with a solid line according to the negative background signal. Published data are from three independent experiments, each containing biological replicates with no fewer than 10,000 cells detected by flow cytometry. All data are means ± SEM of three independent experiments. (B) PEDV (green) and HSPA5 (blue) co-located in endo-/lysosome (red). Vero cells were planted in confocal dishes and transfected to express BFP-HSPA5 fusion protein for 24 h and treated with PEDV (MOI = 5) for 2 h at 4°C and then transferred to 37°C for internalization. Confocal dishes were collected at different time points (30 min, 40 min, 50 min), fixed, permeabilized and blocked with 5% BSA for 2 h. According to different collection time points, the blocked cells were co-incubated with mouse anti-PEDV S MAb and separated with rabbit anti-EEA1 (early endosome marker), RAB7 (late endosome marker), and LAMP1 (lysosome marker) pAb at 4°C overnight, then stained with Alexa Fluor 488 goat anti-mouse pAb and Alexa Fluor 647 donkey anti-rabbit pAb as the secondary antibodies for 1 h at RT. Because the BFP-HSPA5 fusion protein is a blue fluorescent protein (excitation wavelength: 405 nm), the nuclei were not stained with DAPI. Finally, the treated cells were observed with a super-resolution laser confocal microscope. Images were acquired at ×63 magnification. Representative images are from three independent experiments, and boxed regions are magnified in panels to the right. Scale bars = 5 μm. ***, P < 0.001.

Inhibition of HSPA5 activity affected PEDV colocalization in lysosomes during internalization.

Next, we further investigated the relevance of HSPA5 activity during PEDV internalization using the subcellular localization of PEDV. After 2 h of incubation with PEDV at 4°C, the Vero cells were treated with preheated HA15 or DMSO medium, followed by culturing at 37°C. After fixation, permeabilization and co-staining with MAb, Vero cells were observed by laser confocal microscopy. As shown in Fig. 10, the colocalization pixels of PEDV S protein with the early endosomal marker EEA1 and the late endosomal marker RAB7 were not affected by inhibition of HSPA5 in individually infected cells, whereas the lysosomal marker LAMP1 was significantly decreased. A weighted colocalization coefficient (WCC) analysis also showed that the inhibition of HSPA5 significantly reduced colocalization between S proteins and lysosomes (from 0.7786 to 0.4557), while colocalization between S proteins and early or late endosomes was not detected. Therefore, we conclude that inhibition of HSPA5 activity affects the subcellular localization of PEDV in lysosomes via the endo-/lysosomal pathway.

FIG 10.

Inhibition of HSPA5 activity affected PEDV colocalization in lysosomes during internalization. Vero cells seeded in confocal dishes were incubated with PEDV (MOI = 5) for 2 h at 4°C, then washed three times with pre-cooled PBS to remove unattached virions. The Vero cells were treated with preheated diluted HA15 or DMSO medium and cultured at 37°C for different time periods. After fixed permeabilization and blocking, the cells were co-stained with mouse anti-PEDV S MAb (green), rabbit anti-EEA1 MAb (red), rabbit anti-RAB7 MAb (red), or rabbit anti-LAMP1 MAb (red). Alexa Fluor 488 goat anti-mouse IgG and Alexa Fluor 647 goat anti-rabbit IgG were employed as secondary antibodies. DAPI (blue) was used to stain nuclei. White arrows indicate colocalized pixels. Images were taken at ×63 magnification and are represented as a single slice of a stack from three independent experiments; boxed regions are magnified in panels to the right. Scale bars = 10 μm. All data represent means ± SEM from three independent experiments. ****, P < 0.0001; ns, not significant.

HSPA5 is important for PEDV infection in LLC-PK1 cells.

Next, used LLC-PK1 cells to study the role of HSPA5 in pig-derived cells rather than IPEC-J2 because of the low efficiency of infection (43, 46). We first determined by CCK-8 that 100 μM or lower concentrations of HA15 had no significant affect on the activity of treated LLC-PK1 cells (cell viability>90%, Fig. 11A). Inhibiting HSPA5 activity significantly reduced the virus copy number (Fig. 11B), progeny virus titer (Fig. 11C), and intracellular PEDV N protein signal (Fig. 11D) in LLC-PK1 cells. Meanwhile, inhibition of HSPA5 activity affected the colocalization of PEDV S protein with lysosomes in the endo-/lysosomal pathway, but not with early or late endosomes during viral internalization (Fig. 11E). WCC analysis results also showed that inhibition of HSPA5 significantly reduced the colocalization between S protein and lysosomes (mean: 0.4651 to 0.2348). In conclusion, HSPA5 has a positive effect on PEDV infection in both Vero cells and pig-derived LLC-PK1 cells.

FIG 11.

HSPA5 is important for PEDV infection in LLC-PK1 cells. (A) Viability of LLC-PK1 cells treated with HA15. Cells grown in 96-well plates were separately treated with gradient doses of HA15 or DMSO for 24 h and evaluated by a CCK-8 solution cell proliferation assay. (B to D) Role of HSPA5 in PEDV infection of LLC-PK1 cells. LLC-PK1 cells were grown and incubated with PEDV (MOI = 0.5) for 12 h at 37°C, then measured by TCID₅₀ and absolute qPCR and flow cytometry, respectively. (B) Intracellular virus copy number was detected by absolute qPCR. Total RNA was extracted from the treated cells by TRIzol and reverse-transcribed into cDNA, followed by absolute qPCR to detect the copy number of the N gene. (C) The treated cells were repeatedly frozen and thawed three times to fully release the intracellular virus, and after centrifugation, TCID₅₀ was used to determine the progeny virus titer. (D) Treated cells were digested, collected, and incubated with mouse anti-PEDV N MAb as the primary antibody for 1 h after fixation and permeabilization at RT, then incubated with Alexa Fluor 488 goat anti-mouse pAb as a secondary antibody for 0.5 h at RT and evaluated by flow cytometry. (E) Inhibition of HSPA5 activity affects the endo-/lysosomal pathway in LLC-PK1 during PEDV infection internalization. LLC-PK1 cells were seeded in confocal dishes and incubated with PEDV (MOI = 5) at 4°C for 2 h, then washed 3 times with pre-cooled PBS. The medium was replaced with a preheated medium of diluted HA15 or DMSO and used at 37°C for different time points. After immobilization and permeation, the cells were co-stained with mouse anti-PEDV S MAb (green), rabbit anti-EEA1 MAb (red), rabbit anti-RAB7 MAb (red), or rabbit anti-LAMP1 MAb (red) overnight at 4°C. AlexaFluor488 goat anti-mouse IgG and AlexaFluor647 goat anti-rabbit IgG were used as secondary antibodies for 1 h at RT. DAPI (blue) was used to stain the nuclei. White arrows indicate co-positioning pixels. WCC was used for co-location analysis. The image was taken at ×63 magnification and represents a single slice stacked from three independent experiments; the boxed area is magnified in the panel on the right. Scale bar = 10 μm. All data represent the means ± SEM from three independent experiments. *, P < 0.05; ****, P < 0.0001; ns, not significant.

DISCUSSION

PED is a widespread viral disease that has caused enormous losses to the global pig industry. The PEDV S protein determines host tropism by interacting with receptors and can also widely recognize host membrane proteins to enhance virus adhesion and invasion, which is crucial in the process of coronavirus infection (47). Although the host factors of PEDV infection have been gradually described, there have been few reports on the interaction between the PEDV S protein and host membrane proteins and a lack of diverse information on the PEDV invasion process (48, 49). Here, we first performed a pulldown-LC-MS/MS to identify 211 plasma membrane proteins which potentially interact with the PEDV S1. Among these proteins, transferrin receptor 1 was found, which has been shown to interact with the PEDV S, providing support for the identification results. Aminopeptidase N, a contentious PEDV receptor (50–54), was not found in this study, possibly due to this protein’s abundance or lack of interaction. Given that viruses utilize viral proteins to hijack host proteins and initiate a large number of host biochemical processes during infection to guarantee effective multiplication (55), we investigated the correlation between S1-related proteins and associated biochemical processes or pathways by BC, pathway-process enrichment, and MOCDE analysis to further explore the diverse interactions between S proteins and hosts (Fig. 1). Notably, some proteins were enriched in known viral processes such as “viral processes,” “maturation of spike protein,” and “translation of structural protein,” among others. It is possible that these proteins, or proteins associated with them in the network, also participate in the PEDV-related infection process and have yet to be identified. Next, we selected 9 proteins for siRNA knockdown screening after comprehensive analysis of the diverse information to find the proteins involved in PEDV infection. Finally, we discovered that HSPA5 most significantly reduced PEDV infection (Fig. 2).

We established the interaction between the exo-/endogenous HSPA5 and S protein by co-IP, as shown in Fig. 3A. Laser confocal imaging also revealed that Vero cells and LLC-PK1 cells colocalized endogenous HSPA5 with S proteins (Fig. 3B). These results provide evidence that HSPA5 specifically interacts with PEDV S protein. In addition, PEDV infection experiments using HSPA5 knockdown cells and overexpression cell lines also demonstrated that HSPA5 is involved in PEDV viral infection (Fig. 4). Notably, HSPA5 has been reported to bind to some viral structural proteins and to be involved in the degradation of viral misfolded proteins during protein production as well as in virus assembly (56–58). Although we have demonstrated that the S protein binds to HSPA5 and affects PEDV infection, further research is required to determine whether this interaction is involved in viral invasion.

We then attempted to determine during the first life cycle of PEDV whether HSPA5 binds to S proteins during the invasion phase. To reveal its role in PEDV attachment, we first confirmed that HSPA5 is present on the cell membrane of Vero and LLC-PK1 cells by laser confocal microscopy and flow cytometry, WB (Fig. 5), and laser confocal microscopy revealed that HSPA5 colocalizes with S protein during PEDV attachment (Fig. 6A). HSPA5-knockdown Vero cells were used to detect PEDV attachment, and significant reduction was observed by flow cytometry, IFA, and TCID₅₀ analysis (Fig. 6). This suggests that HSPA5 on the membrane of Vero cells contributes to PEDV attachment. Furthermore, we used an antibody-blocking assay to corroborate this conclusion. A co-IP assay was used to prove that the N-terminal domain of HSPA5 binds with the PEDV S protein. The polyclonal antibody (pAb) against the N-terminal domain of HSPA5 also significantly reduced PEDV RNA abundance and progeny virus titers in PEDV-infected Vero cells (Fig. 7). Interestingly, the C-terminal domain of HSPA5, not the N-terminal domain, has been reported to interact with the S protein of the β-coronavirus SARS-CoV-2 (severe acute respiratory syndrome coronavirus 2) and can form a complex with the receptor ACE2 to participate in SARS-CoV-2 attachment (36), suggesting the variability of S proteins of different coronavirus species. This also provides clues to finding the receptors of PEDV. Taken together, these results show that the N terminus of HSPA5 is a functional domain for PEDV attachment.

We sought to elucidate whether HSPA5 affects the viral internalization of PEDV. The lead compound HA15, which specifically binds to HSPA5 and inhibits its activity (39), was used to explore the role of HSPA5 in the internalization process. It was found that 25 μM or lower concentrations of HA15 will not reduce the activity of treated Vero cells. Given that inhibition of HSPA5 activity may affect a variety of biological processes, thus affecting the efficiency of PEDV infection, we established separate pre- and post-internalization inhibition of HSPA5 activity to highlight its effect on virus internalization. As expected, inhibition of HSPA5 activity with 12.5 μM or lower concentrations of HA15 during internalization was found to reduce PEDV infection more significantly than inhibition after internalization in Vero cells (Fig. 8). This suggests that HSPA5 plays a more important role during viral internalization.

Although HSPA5 is becoming more widely known as a multifunctional receptor on the cell surface, more information about the viral adhesion components involved in viral attachment has been reported but little is known about the internalization part of the invasion phase (59, 60). Therefore, we attempted to describe the effect of HSPA5 on PEDV internalization. During internalization, most enveloped viruses are endocytosed into intracellular compartments before membrane fusion (61), and then internalized viruses act as cargo via specific transport pathways to the fusion site (62, 63). According to previous research, endocytosed PEDV virion is transported to the fusion site via the endo-/lysosomal pathway (43, 45). Significant reduction in HSPA5 signal on the cell surface was observed along with the completion of cell endocytosis. This suggests the possibility that HSPA5 internalizes together with PEDV virion. Unsurprisingly, subcellular localization of S protein and HSPA5 was discovered in endosomes and lysosomes during PEDV internalization by laser confocal microscopy (Fig. 9). This suggested that HSPA5 binds with S for endo-/lysosomal transport through viral internalization. In addition, it was also found that inhibition of HSPA5 activity significantly reduces the subcellular colocalization of PEDV in lysosomes during Vero or LLC-PK1 cell internalization (Fig. 10 and 11). Notably, these results are not sufficient to prove a direct role of the HSPA5-S protein complex in the endo-/lysosomal pathway, but do demonstrate that HSPA5 is involved in the endosomal-lysosomal transport process of PEDV. It has also been reported that HSPA5 on the cell membrane can be regulated by autophagy, selectively bound to misfolded proteins and targeted to lysosomes for degradation (64). Although there have been few studies of HSPA5 regarding the endosomal-lysosomal pathway (17), we can propose a hypothesis that PEDV facilitates internalization of the virus through S protein hijacking of HSPA5 regarding targeting endosome-lysosome related functions; this direct function of HSPA5-binding S protein requires more studies to be interpreted. In conclusion, this work suggests that HSPA5 promotes intracellular PEDV trafficking to the lysosome via the endo-/lysosomal pathway in Vero and pig-derived cells.

In conclusion, our results demonstrate, for the first time, a correlation between S1-related host membrane proteins. This provides a valuable reference for future studies on PEDV S1 protein. Additionally, we identified the host membrane protein HSPA5 as a new target which interacts with the PEDV S protein through its N-terminal domain, regulating attachment and contributing to the internalization of PEDV. Upon further analysis, it was discovered that HSPA5 is involved in the trafficking of PEDV through the endo-/lysosomal pathway, which aids in viral translocation in both Vero and pig-derived cells. This new information extends our understanding of PEDV infection, and the identification of HSPA5’s involvement in PEDV attachment and internalization provides a new basis for the development of potential therapeutic agents.

MATERIALS AND METHODS

Cells and virus.

African green monkey kidney cells (Vero cells) and human embryonic kidney cells (HEK293T cells) were cultured in Dulbecco modified Eagle medium (DMEM; cat no. 12100; Solarbio, Beijing, China). LLC-PK1 cells were kept in the modified Eagle medium (MEM; cat no. PM150410; Pricella, Wuhan, China). All media were supplemented with 10% heat-inactivated fetal bovine serum (FBS; cat no. FS201-02; Australian origin, TransGen, Beijing, China) containing antibiotics (100 μg/mL penicillin-streptomycin; cat no. P1400; Solarbio) in a humidified 37°C, 5% CO₂ incubator.

The PEDV DR13 strain (GenBank accession no. JQ023162) was stored in our laboratory (65). The propagation and titration of PEDV was performed using Vero cells, and the titers were determined by Karber’s method.

Antibodies.

Rabbit anti-GAPDH MAb (ab1811602), rabbit anti-HSAP5 BiP N terminus pAb (ab32618), rabbit anti-HSPA5 BiP C terminus pAb (ab21685), mouse anti-ATP1A1 MAb (ab7671), Alexa Fluor 488 goat anti-mouse pAb (ab150113), and Alexa Fluor 647 donkey anti-rabbit pAb (ab150075) were purchased from Abcam (Cambridge, United Kingdom). Mouse anti-PEDV S1 MAb (SD-3) and mouse anti-PEDV N MAb (SD-2) were obtained from Medgene Labs (SD, USA). Rabbit anti-HSPA5 pAb (11587-1-AP), mouse anti-c-Myc tag MAb (67447-1-lg), rabbit anti-HA tag pAb (51064-2-AP), horseradish peroxidase (HRP)-conjugated goat anti-mouse IgG antibody (SA00001-1), and HRP-conjugated goat anti-rabbit IgG antibody (SA00001-2) were from Proteintech (Chicago, IL, USA).

Reagents.

Lipofectamine 3000 reagent (L3000001) and Lipofectamine RNAiMAX transfection reagent (13778150) were purchased from Invitrogen (USA). PrimeScript RT master mix kit (RR036A) was purchased from Takara Bio (Japan). TransZol Up Plus RNA kit (ER501-01) was purchased from TransGen (China). A Cell Plasma Membrane Staining kit with DiO (C1993S) was purchased from Beyotime (China). Universal SYBR Green Master (04913914001) was purchased from Roche (Mannheim, Germany).

Expression vector construction and transfection.

The gene fragment of codon-optimized PEDV spike (GenBank: JQ023161.1) was synthesized by GeneWiz (Suzhou, China) and cloned to pFUSE-lgG1-Fc (Invitrogen, CA). The full-length HSPA5 cDNA fragment according to the green monkey HSPA5 gene (GenBank: XM_008006256.2) was synthesized and inserted into the pLVX-IRES-tdTomato and pmTagBFP2-N vector by Sangon (Shanghai, China). The HSPA5 NBD, HSPA5 SBD, and the middle domain between HSPA5 NBD and SBD were cloned into the pCAGGS vector with c-Myc tag for mammalian cell expression.

Pulldown-LC/MS-MS and analysis.

The PEDV S1 protein was expressed and purified in our laboratory. Vero cells were grown to 100% confluence in a T75 cell culture flask and a Minute Plasma Membrane Protein Isolation and Cell Fractionation kit (Invent Biotechnologies, Plymouth, MN, USA) was used to obtain purified membrane protein. After confirmation that the membrane proteins were extracted correctly, the samples were divided and stored at −20°C. Cobalt resin was incubated with purified 200 μg His-labeled S1 protein at room temperature for 2 h and then incubated with purified membrane protein at 4°C overnight. After overnight incubation at 4°C, the enriched cobalt resin was washed 6 times with Tris-NaCl solution buffer (15 mM Tris, 150 mM NaCl [pH 8.0]) and eluted with 500 mM imidazole. The eluates were analyzed through 10% SDS-PAGE and treated with a fast silver stain kit (P0017S, Beyotime, China). The 150- to 30-kDa protein strips were intercepted and applied to LC-MS/MS by Lumingbio (Shanghai, China). All experiments were biologically repeated three times. When the score sequest HT was ≥5 and the number of peptides and unique peptides was >1, the proteins on the protein bands were characterized.

The BC algorithm and the drawing of the interaction network diagram were all completed using Cytoscape software (66). Metascape (http://metascape.org/) was used for pathway-process enrichment analysis and MCODE analysis (67).

RNA interference.

siRNAs targeting HSPA5 or other proteins and negative control (siRNA-NC) were purchased from GenePharma (Shanghai, China). Vero cells were transfected with the indicated siRNAs at a final concentration of 25 to 50 nM using Lipofectamine RNAiMAX according to the manufacturer’s instructions. The Vero cells were harvested 48 h later and checked using real-time qPCR (RT-qPCR) and WB. The transfected Vero cells were used for subsequent experiments. The siRNAs used in this work are listed in Table 2.

TABLE 2.

Primers used for qPCR

Target gene	Sequence (5′–3′)
	Sense	Antisense
PEDV N probe	FAM-TGTTGCCATTACCACGACTCCTGC-BHQ3	-
PEDV N	CGCAAAGACTGAACCCACTAAC	TTGCCTCTGTTGTTACTTGGAGAT
Monkey-GAPDH	ACATCATCCCTGCCTCTACTG	CCTGCTTCACCACCTTCTTG
Monkey-HSPA5	GCTGGTACAGTAACAACTGCATGGGT	CCGAGAACACGGTCTTTGACGC

Cell viability assay.

Cell viability was determined using a CCK-8 (GlpBio, Montclair, CA, USA) according to the manufacturer’s instructions. Cells were seeded in 96-well plates and treated with the indicated inducer or inhibitor at a gradient concentration at 37°C for 24 h, or transfected with siRNAs for 48 h. The cells were then supplemented with 10 μL CCK-8 solution per well, followed by incubation for another 1 h. The optical density at 450 nm (OD₄₅₀) was recorded with a microplate reader (Fluostar Omega; BMG Labtech, Ortenberg, Germany).

RT-qPCR.

Cells were given the appropriate treatment or infection, total RNAs were extracted using a TransZol Up Plus RNA kit, and reverse transcription was performed using a PrimeScript RT Master Mix kit. A TaqMan RT-qPCR method based on the PEDV N gene was used to determine PEDV RNA copy numbers in infected cells (68). The host cDNAs were amplified by relative quantification RT-qPCR using universal SYBR Green Master Mix, and GAPDH was used as endogenous control. The doubling change in gene mRNA expression level is expressed as the mean and standard deviation (SD) from the results of three independent experiments. The gene primers and probes listed in Table 2 were synthesized by Sangon (Shanghai, China).

Flow cytometry.

Vero cells were detached with pre-cooled PBS containing 10 mM EDTA and fixed at 4% paraformaldehyde (PFA) for 20 min without cell permeabilization. Cell fixing was followed by immunolabeling with rabbit anti-HSPA5 pAb or mouse anti-PEDV spike as the primary antibody for 1 h at room temperature (RT), and Alexa Fluor 647 donkey anti-rabbit pAb or Alexa Fluor 488 goat anti-mouse pAb was used as the secondary antibody for 0.5 h at RT. For experiments with intracellular staining of PEDV N protein, cells were detached with 10 mM EDTA in PBS, fixed in 4% paraformaldehyde, permeabilized with 0.1% TritonX-100 in PBS, and then stained with mouse anti-PEDV N protein pAb as the primary antibody and Alexa Fluor 488 goat anti-mouse pAb as the secondary antibody. Flow cytometry was performed using CytoFLEX (Beckman Coulter, Brea, CA) and data were analyzed using FlowJo 10.6.2 (Tree Star Inc.).

Co-immunoprecipitation assay.

Vero cells were seeded into 6-cm dishes and infected with PEDV. Cells were collected on ice using 0.8 mL NP-40 lysis buffer (Beyotime, China) containing protease inhibitor cocktail (Thermo Fisher, Waltham, MA, USA) and then centrifuged at 12,000 × g for 10 min at 18 h postinfection. The lysate was incubated overnight with 2.5 μg pAB of anti-HSPA5 or rabbit IgG and MAb of anti-PEDV S or mouse IgG at 4°C on a rocker platform. Next, 80 μL of fresh protein A/G magnetic beads (MedChemExpress, Dallas, TX, USA) was added to the mixture and incubated for 0.5 h at RT on a rocker platform. The HSPA5 or spike-enriched magnetic beads were washed with PBS with Tween 20 (PBST), combined with SDS-PAGE loading buffer, and heated for 10 min at 100°C. The proteins were analyzed by Western blotting with anti-HSPA5 pAb or anti-spike MAb.

The plasmids expressing spike-Fc and HSPA5-Myc, HSPA5 NBD-Myc, HSPA5 SBD-Myc, or HSPA5 NSB-Myc were co-transfected in HEK293T cells (in 35-cm dishes). Co-transfection was conducted using the Lipofectamine 3000 reagent following the manufacturer’s protocol. After 48 h, cells were lysed by pre-cooled NP-40 on ice and then centrifuged at 12,000 × g for 10 min. The protein A/G or anti-c-Myc magnetic beads washed by PBST were added to lysate supernatant and incubated on a rocker platform for 2 h at RT. The enriched magnetic beads were combined with SDS-PAGE loading buffer and heated for 10 min at 99°C after washing with PBST. The anti-Myc MAb or anti-spike MAb as the primary antibodies were labeled and analyzed by Western blotting.

IFA and confocal microscopy.

Vero or LLC-PK1 cells were cultured in 35-mm confocal petri dishes or tissue culture plates for 12 h. After being treated, they were washed three times with pre-cooled PBS, fixed with 4% PFA for 15 min, and permeabilized with 0.1% Triton X-100 in PBS at room temperature for 10 min. After three rinses with pre-cooled PBS, the cells were incubated in PBS containing 5% bovine serum albumin (BSA, Beyotime, China) overnight, then incubated with suitable primary antibody at RT for 2 h. After full washing, the cells were incubated with secondary antibodies at RT for 1 h. Finally, the nuclei were either stained with DAPI (4′,6-diamidino-2-phenylindole) for 10 min or not stained. Fluorescence images were obtained by a confocal laser scanning microscope (LSM700; Carl Zeiss AG, Germany). The confocal laser scanning images were taken with a ×20 objective lens and a ×63 oil lens as single slices stacked from three independent experiments.

Instead of calculating Pearson correlation coefficients, we calculated the overlap coefficient according to Manders (69), which allowed more reliable quantification of colocalization coefficients in the images where the fluorescence of one antigen was stronger than that of the other (70). To compare the fluorescence localization of different experimental groups during viral internalization, the colocalization coefficients of EEA1, RAB7, LAMP1, and PEDV S protein channels were calculated by the WCC, which is recommended for analyzing scattergram regions of interest (ROI) (70, 71). According to the final version of the LSM700 brief manual, using the confocal toolbox of Zen software, we first defined the threshold value of single-channel fluorescence on the scattergram through the picture of a single signal, selected the ROI area to obtain colocalization pixels, and finally calculated the WCC using the formula WCC = Σ_if_i,coloc/Σ_if_i,all: “fi,coloc” represents the amount of collocated fluorescence pixels between green (PEDV S) and red (EEA1, RAB7, LAMP1) channels after removing the background value, and “fi,all” represents the total fluorescence of pixels of the green (PEDV S) channel within a cell.

Statistical analysis.

The experimental data were shown as group average ± standard error of the mean (SEM) from at least three independent experiments. Prism 8.0.2 software (GraphPad, San Diego, CA) was used to perform unpaired two-tailed Student’s t tests or one-way analysis of variance (ANOVA) for all statistical analysis and calculation.