Voltage-Dependent Anion Channel Protein 2 (VDAC2) and Receptor of Activated Protein C Kinase 1 (RACK1) Act as Functional Receptors for Lymphocystis Disease Virus Infection

Lymphocystis disease virus (LCDV) is the causative agent of lymphocystis disease in fish, which has caused huge economic losses to the aquaculture industry worldwide, but the molecular mechanism underlying the LCDV-host interaction remains unclear. Here, the 27.8-kDa putative cellular receptor for LCDV was identified as voltage-dependent anion channel protein 2 (VDAC2) and receptor of activated protein C kinase 1 (RACK1), and our results revealed that VDAC2 and RACK1 expression was sufficient to allow LCDV entry and that they are functional receptors that initiate LCDV infection for the first time, which leads to a better understanding of the molecular mechanism underlying LCDV infection and virus-host interactions.

ABSTRACT

In previous research, a 27.8-kDa protein in flounder Paralichthys olivaceus gill (FG) cells was identified as a putative cellular receptor (27.8R), which mediated lymphocystis disease virus (LCDV) infection via interaction with a 32-kDa viral attachment protein (VAP) of LCDV, and monoclonal antibodies (MAbs) against 27.8R and 32-kDa VAP were developed. In this study, the 27.8R was identified as voltage-dependent anion channel protein 2 (VDAC2) and receptor of activated protein C kinase 1 (RACK1) of flounder. Recombinant VDAC2 (rVDAC2) and RACK1 (rRACK1) were obtained by prokaryotic expression, and rabbit anti-VDAC2/RACK1 polyclonal antibodies were prepared. The rVDAC2, rRACK1, and 27.8-kDa proteins in FG cells were recognized by anti-27.8R MAbs and anti-VDAC2/RACK1 polyclonal antibodies simultaneously. Preincubation of FG cells with anti-VDAC2/RACK1 polyclonal antibodies significantly decreased the percentages of LCDV-infected cells and LCDV copy numbers, blocked virus infection, and delayed the development of cytopathic effect. The mRNA expressions of VDAC2 and RACK1 in FG cells were upregulated to maximum levels 12 h and 48 h after LCDV infection, respectively. VDAC2/RACK1 knockdown through short interfering RNA (siRNA) significantly reduced VDAC2/RACK1 expression and LCDV copy numbers in FG cells compared with negative controls, while VDAC2/RACK1 expression on LCDV-nonpermissive epithelial papillosum cells (EPCs) conferred susceptibility to LCDV infection, indicating the VDAC2 and RACK1 were sufficient to allow LCDV entry and infection. All these results collectively showed that VDAC2 and RACK1 function as receptors for LCDV entry and infection.

IMPORTANCE Lymphocystis disease virus (LCDV) is the causative agent of lymphocystis disease in fish, which has caused huge economic losses to the aquaculture industry worldwide, but the molecular mechanism underlying the LCDV-host interaction remains unclear. Here, the 27.8-kDa putative cellular receptor for LCDV was identified as voltage-dependent anion channel protein 2 (VDAC2) and receptor of activated protein C kinase 1 (RACK1), and our results revealed that VDAC2 and RACK1 expression was sufficient to allow LCDV entry and that they are functional receptors that initiate LCDV infection for the first time, which leads to a better understanding of the molecular mechanism underlying LCDV infection and virus-host interactions.

INTRODUCTION

Fish lymphocystis disease (LCD) is a chronic viral disease with typical symptoms of clusters of enlarged hypertrophied dermal cells on the skin, fins, and internal organs of affected fish (1). Lymphocystis disease virus (LCDV) is the causative agent of LCD and belongs to the genus Lymphocystivirus within the family Iridoviridae (2). LCDV has infected more than 140 wild and cultured species of marine, brackish water, and freshwater fish worldwide, causing huge economic losses to the aquaculture industry (3, 4). In China, LCD is often observed in some cultured fish, especially the flounder Paralichthys olivaceus; even though the disease rarely causes death, it may lead to secondary infection by other microorganisms, resulting in high mortality rates (5–7). LCDV is a macromolecular DNA virus with an envelope (8, 9), and the complete genome sequences of two LCDV isolates, LCDV-1 from the flounder Platichthys flesus in Europe and LCDV-C from the flounder P. olivaceus in China, have been determined (10, 11). Although great advances in LCDV studies have been made, the molecular mechanism of LCDV infection, replication, and pathogenesis remains elusive (12, 13), and research on cellular receptor proteins is of great significance to elucidate the molecular mechanism underlying LCDV-host interactions.

In our previous studies, two putative cellular receptors for LCDV infection with molecular weights of 27.8 kDa (27.8R) and 37.6 kDa (37.6R) were identified from flounder gill (FG) cells (14, 15). Monoclonal antibodies (MAbs) against 27.8R (3D9 and 2G11), which were able to inhibit LCDV infection of FG cells, were developed (16). Besides, LCDV infection could induce the upregulation of 27.8R expression, which in turn increased the susceptibility and availability of FG cells for LCDV entry (17), and 27.8R was widely detected on tissues and leukocytes of flounder and turbot (Scophthalmus maximus) (7, 18). In the following research, a 32-kDa envelope protein of LCDV, encoded by the open reading frame (ORF) 038 gene of LCDV-C, was found to function as a viral attachment protein (VAP) and initiate LCDV infection via interaction with 27.8R (19), and anti-32-kDa VAP MAbs were prepared (20). However, further studies are required to determine the biological function of 27.8R and the specific mechanism of 27.8R-mediated LCDV infection.

Voltage-dependent anion channel proteins (VDACs) were originally characterized as abundant proteins found in the outer mitochondrial membrane. However, evidence that VDACs can also be expressed in plasma membranes has begun to accumulate (21). VDAC2 is one of the isoforms of VDACs involved in apoptosis, and viral proteins have been reported to be involved in the VDAC pathway to regulate host cells (22, 23). Grouper VDAC2 is required for nervous necrosis virus (NNV) infection for maintaining the cellular ATP level and has a positive impact on virus-induced apoptosis (24). In flounder, it has been reported that VDAC may mediate antiviral immune responses through induction of apoptosis (25). As for receptor of activated protein C kinase 1 (RACK1), it has homology to the β-subunit of G proteins and interacts with several cell surface receptors and proteins in the nucleus. Moreover, RACK1 is a key mediator of various pathways and contributes to numerous aspects of cellular function, including cell adhesion, apoptosis, the immune response, and signaling pathways (26, 27). VDAC2 and RACK1 have been reported to form a complex with VP5 of infectious bursal disease virus (IBDV) through inhibition of cell apoptosis to promote viral replication (28–30). However, there has been no report that the two proteins serve as viral cellular receptors. In this study, 27.8R was identified to be VDAC2 and RACK1 through two-dimensional (2D) Western blotting, 2D far-Western blotting, a 2D virus overlay protein binding assay (VOPBA), and mass spectrometry (MS). Afterwards, recombinant VDAC2 (rVDAC2) and RACK1 (rRACK1) and their rabbit polyclonal antibodies were prepared; the abilities of 27.8R and rVDAC2/rRACK1 to react with anti-27.8R MAbs and anti-VDAC2/RACK1 polyclonal antibodies were analyzed by Western blotting; and the ability of anti-VDAC2/RACK1 polyclonal antibodies to inhibit LCDV infection was elucidated by fluorescence-activated cell sorter (FACS) analysis, real-time quantitative PCR (qPCR), and an indirect immunofluorescence assay (IFA). In addition, the mRNA expression of VDAC2/RACK1 after LCDV infection and LCDV copy numbers in VDAC2/RACK1 knockdown FG cells were investigated by qPCR, and VDAC2/RACK1 protein expression in FG cells was detected by confocal microscopy. Moreover, LCDV infection of nonpermissive epithelial papillosum cells (EPCs) from the carp Cyprinus carpio, which were transiently transfected with pcDNA3.1-VDAC2/pCIneo-GFP-RACK1 (pcDVDAC2/pcGRACK1) of flounder, was determined by qPCR, and VDAC2/RACK1 and LCDV were doubly stained in EPCs by immunofluorescence.

RESULTS

Identification of the 27.8R protein.

FG cell membrane proteins were first desalinated, followed by first-dimension isoelectric focusing (IEF) and second-dimension electrophoresis with silver staining. The horizontal axis represents the protein isoelectric point, with the pH ranging from 3 to 10, and the vertical axis represents the protein molecular weight (Fig. 1A). According to the results of 2D Western blotting of FG cell membrane proteins with MAbs against 27.8R, five protein spots between 25 kDa and 35 kDa were visualized (Fig. 1B) and designated spots a to e after comparison with the FG cell membrane protein gel using PDQuest software. Furthermore, the results of 2D far-Western blotting of FG cell membrane proteins with the LCDV 32-kDa VAP (Fig. 1C) and a 2D VOPBA of FG cell membrane proteins with LCDV (Fig. 1D) presented two positive protein spots between 25 kDa and 35 kDa, and the two spots corresponded to spot c and spot e in 2D Western blots compared with FG cell membrane proteins. Protein identification was achieved by MS analysis for all five immunoreactive spots, and the results were compared with data in the Mascot database. As shown in Table 1, four proteins, VDAC1 (spot b), RACK1 (spot c), uncharacterized protein LOC109641635 (spot d), and VDAC2 (spot e), were identified. A protein score of >35 was considered significant (P < 0.05) in this experiment. Combined, the results of 2D Western blotting, 2D far-Western blotting, and a 2D VOPBA confirmed that VDAC2 (spot e) and RACK1 (spot c) are the putative cellular receptors corresponding to the 27.8R protein. No spot was detected in polyvinylidene difluoride (PVDF) membranes in negative controls (data not shown).

FIG 1.

Identification of LCDV 27.8R. (A) Two-dimensional electrophoresis of FG cell membrane proteins. The gel was visualized by sliver staining and scanned. The five protein spots, which are marked with red arrows and designated spots a to e in an FG cell membrane protein gel, were subjected to MS analysis. (B) 2D Western blotting of FG cell membrane proteins with anti-27.8R MAbs. (C) 2D far-Western blotting of FG cell membrane proteins with the LCDV 32-kDa VAP. (D) 2D VOPBA of FG cell membrane proteins with LCDV particles. Circles, protein spots identified by 2D Western blotting, 2D far-Western blotting, and a 2D VOPBA, respectively. M, marker.

TABLE 1.

List of protein spots identified by matrix-assisted laser desorption ionization–time of flight (MALDI-TOF/TOF) MS

Spot	Protein	GenBank accession no.	Theoretical Mr/pI	Sequence coverage (%)	Protein score	Peptide identified	Species
a	No significant hits
b	Voltage-dependent anion-selective channel protein 1	XP_019952890.1	30,567/6.53	46	364	AVPPTYIDLGKWAEHGLTFTEKWNTDNTLGTEITLEDQLAKMTFDSSFSPNTGKITQSNFAVGYKTDEFQLHTNVNDGTEFGGSIYQKVNDQLETAVNLAWTAGNSNTRYQIDPDASFSAKLTLSALLDGK	Paralichthys olivaceus
c	Receptor of activated protein C kinase 1	XP_019964456.1	35,458/7.06	9	40	DETNYGIPQRDVLSVAFSADNRVW QVTIGTR	Paralichthys olivaceus
d	Uncharacterized protein LOC109641635	XP_019961708.1	25,688/7.63	27	87	ENLTTIVGDVGSEEGAEQAKVTDIVSSLGFSWWQGGPPHTQPVKMAPGYLNHLDLGEAVAALVEK	Paralichthys olivaceus
e	Voltage-dependent anion channel	ABH07379.1	30,310/8.65	44	264	RSEYGLTFTEKWNTDNTLGTEITVEDQIAKLTFDTTFSPNTGKMTQNNFAIGYKTGDFQLQTNVNDGAEFGGSIYQKLETAVNLAWTAGSNSTRYQLDSDATISAKVNNN SLVGVGYTQTLRPGVK	Paralichthys olivaceus

Recombinant expression of VDAC2/RACK1 and specificity of rabbit anti-VDAC2/RACK1 polyclonal antibodies.

The recombinant expression plasmids pET-32a-VDAC2 and pET-32a-RACK1 were constructed, and sequence alignment showed 100% identity with the reported VDAC2 and RACK1 nucleotide sequences, respectively. The expressions of rVDAC2 and rRACK1 were induced by isopropyl-β-d-thiogalactopyranoside, the proteins were purified by HisTrap Ni-nitrilotriacetic acid (NTA) affinity chromatography, and SDS-PAGE analysis exhibited enhancement of a protein band with a molecular weight of approximately 46 kDa (Fig. 2A and C, lanes 2) compared with the protein before induction (Fig. 2A and C, lanes 1), which was consistent with the expected molecular weight of 27.8 kDa for the protein plus a His tag of about 18 kDa. The purified rVDAC2 and rRACK1 yielded a single protein of the anticipated 46-kDa size (Fig. 2A and C, lanes 3).

FIG 2.

Expression and purification of recombinant VDAC2/RACK1 protein and specificity analysis of rabbit anti-VDAC2/RACK1 polyclonal antibodies. (A and C) Recombinant expression of VDAC2 and RACK1 in E. coli strain BL21. Lane 1, rVDAC2/rRACK1 before induction; lane 2, induced rVDAC2/rRACK1; lane 3, purified rVDAC2/rRACK1. (B) Western blot results showing that rVDAC2 and 27.8-kDa FG cell membrane proteins are recognized by anti-27.8R MAbs and anti-VDAC2 polyclonal antibodies. Lanes 1 and 2, Western blot results for rVDAC2 with anti-27.8R MAbs 2G11 and 3D9; lane 4, Western blot results for FG cell membrane proteins with a mixture of 27.8R MAbs 2G11 and 3D9; lanes 3 and 5, Western blot results for rVDAC2 and FG cell membrane proteins with anti-VDAC2 polyclonal antibodies. (D) Western blot results show that rRACK1 and 27.8-kDa FG cell membrane proteins are recognized by anti-27.8R MAbs and anti-RACK1 polyclonal antibodies. Lanes 1 and 2, Western blot results for rRACK1 with anti-27.8R MAbs 2G11 and 3D9; lane 4, Western blot results for FG cell membrane proteins with a mixture of anti-27.8R MAbs 2G11 and 3D9; lanes 3 and 5, Western blot results for rRACK1 and FG cell membrane proteins with anti-RACK1 polyclonal antibodies; lanes 6 and 8, P3U1 culture supernatant instead of anti-27.8R MAbs as a negative control; lanes 7 and 9, rabbit preimmune serum instead of anti-VDAC2/RACK1 polyclonal antibodies as a negative control. M, marker.

The specificity of the developed rabbit anti-VDAC2/RACK1 polyclonal antibodies was analyzed by Western blotting, and the results revealed that the purified rVDAC2 was recognized by anti-27.8R MAbs 2G11 and 3D9 (Fig. 2B, lanes 1 and 2) and anti-VDAC2 polyclonal antibodies (Fig. 2B, lane 3) simultaneously, and the 27.8-kDa FG cell membrane protein was recognized by a mixture of anti-27.8R MAbs 2G11 and 3D9 (Fig. 2B, lane 4) and anti-VDAC2 polyclonal antibodies (Fig. 2B, lane 5). Similarly, the purified rRACK1 was recognized by anti-27.8R MAbs 2G11 and 3D9 (Fig. 2D, lanes 1 and 2) and anti-RACK1 polyclonal antibodies (Fig. 2D, lane 3) simultaneously, and the 27.8-kDa FG cell membrane protein was recognized by the mixture of anti-27.8R MAbs 2G11 and 3D9 (Fig. 2D, lane 4) and anti-RACK1 polyclonal antibodies (Fig. 2D, lane 5), whereas no bands appeared in the negative controls, which comprised the P3-X63-Ag8U1 (P3U1) culture supernatant instead of anti-27.8R MAbs (Fig. 2B and D, lanes 6 and 8) or rabbit preimmune serum instead of the developed polyclonal antibodies (Fig. 2B and D, lanes 7 and 9), indicating that VDAC2 and RACK1 were 27.8R and that the developed rabbit polyclonal antibodies had good specificity.

Anti-VDAC2/RACK1 polyclonal antibodies inhibit LCDV infection of FG cells.

To analyze the ability of anti-VDAC2/RACK1 polyclonal antibodies to inhibit LCDV infection, the polyclonal antibodies were diluted to different concentrations and incubated with FG cells. The FACS results demonstrated that the percentages of LCDV-infected FG cells were significantly decreased with all concentrations of polyclonal antibodies, except 0.16 μg/ml anti-VDAC2 antibodies, compared with the negative control, which comprised the same concentrations of rabbit preimmune serum (Fig. 3A and B), and the lowest percentages of LCDV-infected cells were detected with 16 μg/ml anti-RACK1 and 1.6 μg/ml anti-VDAC2 polyclonal antibodies (Fig. 3A and B). The qPCR results showed that LCDV copy numbers in FG cells were significantly reduced when preincubated with 160, 16, and 1.6 μg/ml of anti-VDAC2/RACK1 polyclonal antibodies (P < 0.05), but no significant difference was observed between a concentration of 0.16 μg/ml of polyclonal antibodies and the negative control (Fig. 3C). Moreover, preincubation of FG cells with anti-VDAC2/RACK1 polyclonal antibodies obviously inhibited LCDV infection, as the green fluorescence intensity of LCDV was significantly weakened (Fig. 4A), and the development of cytopathic effect (CPE) was delayed in FG cells compared with the negative control (Fig. 4B).

FIG 3.

Detection of inhibition of LCDV infection of FG cells by anti-VDAC2/RACK1 polyclonal antibodies in vitro by FACS analysis and qPCR. (A) Percentages of LCDV-infected FG cells that were preincubated with different concentrations of anti-VDAC2/RACK1 polyclonal antibodies. (B) Percentages of LCDV-infected FG cells represented in a histogram. (C) LCDV copy numbers in FG cells that were preincubated with different concentrations of anti-VDAC2/RACK1 polyclonal antibodies before LCDV infection. The same concentrations of rabbit preimmune serum were used as negative controls. Error bars represent standard deviations (SD). Differences were considered statistically significant at a P value of <0.05.

FIG 4.

Detection of inhibition of LCDV infection of FG cells by anti-VDAC2/RACK1 polyclonal antibodies in vitro by IFAs and CPE. (A) IFA results for inhibition of LCDV infection of FG cells by anti-VDAC2/RACK1 polyclonal antibodies. LCDV is stained in green, and the cell nuclei are stained in blue by DAPI. (B) Cytopathic effect in FG cells that were preincubated with different concentrations of anti-VDAC2/RACK1 polyclonal antibodies before LCDV infection. The same concentrations of rabbit preimmune serum were used as negative controls.

LCDV infection upregulates expression of VDAC2 and RACK1 in FG cells.

To investigate the mRNA expression of VDAC2 and RACK1 during LCDV infection, the relative gene expression level was calculated after LCDV infection by qPCR. The results indicated that VDAC2 expression was significantly upregulated at 2 h compared with the negative control without infection, reached the highest level at 12 h (6-fold; P < 0.05), and afterward decreased to normal levels after 72 h (Fig. 5A). RACK1 expression was upregulated at 2 h, arrived at the maximum level at 48 h (4-fold; P < 0.05), and then decreased to normal levels after 120 h (Fig. 5B). Extensive red fluorescence signals were observed by confocal microscopy (Fig. 5C), indicating that VDAC2/RACK1 was abundantly expressed in FG cells.

FIG 5.

Expression of VDAC2 and RACK1 in FG cells. (A) mRNA expression of VDAC2 in FG cells at different time points after LCDV infection. (B) mRNA expression of RACK1 in FG cells at different time points after LCDV infection. Error bars represent SD. (C) Protein expression of VDAVC2 and RACK1 in FG cells observed by confocal microscopy. VDAC2/RACK1 in FG cells is stained in red, and cell nuclei are stained in blue by DAPI. Rabbit preimmune serum instead of anti-VDAC2/RACK1 polyclonal antibodies served as the negative control.

VDAC2 and RACK1 knockdown reduces LCDV loads in FG cells.

The VDAC2-targeting short interfering RNA (VDAC2-siRNA) and RACK1-siRNA were transfected into FG cells to knock down the expression of VDAC2 and RACK1, and VDAC2 and RACK1 gene expressions were detected by qPCR. Compared with the negative control, cells transfected with a nontargeting control siRNA (NC-siRNA), the mRNA expression level of VDAC2 in VDAC2-siRNA-transfected FG cells was decreased to 0.35- and 0.14-fold at 24 h and 48 h, respectively (P < 0.05) (Fig. 6A), while the mRNA expression level of RACK1 in RACK1-siRNA-transfected FG cells was decreased to 0.51- and 0.57-fold at 24 h and 48 h, respectively (P < 0.05) (Fig. 6B). IFA results showed that the red fluorescence signals for VDAC2 and RACK1 in FG cells were weakened compared with the NC-siRNA groups, indicating that their protein expressions were also reduced (Fig. 7A and B). To elucidate the effects of silencing of VDAC2 and RACK1 on LCDV infection, VDAC2-siRNA-, RACK1-siRNA-, and NC-siRNA-transfected FG cells were further infected with LCDV at 24 h and 48 h posttransfection, and qPCR results showed that the LCDV copy numbers were significantly reduced in the VDAC2 (Fig. 6C) or RACK1 (Fig. 6D) knockdown FG cells at 48 h postinfection compared with the negative control that was transfected with NC-siRNA.

FIG 6.

Effect of VDAC2/RACK1 knockdown on LCDV infection of FG cells. (A) VDAC2 gene expression in VDAC2-siRNA-transfected FG cells. (B) RACK1 gene expression in RACK1-siRNA-transfected FG cells. (C) LCDV copy numbers in VDAC2 knockdown FG cells. (D) LCDV copy numbers in RACK1 knockdown FG cells. FG cells transfected with a nonsilencing control siRNA (NC-siRNA) served as the negative control (control). Error bars represent SD. Asterisks denote significant differences among groups (P < 0.05).

FIG 7.

VDAC2/RACK1 protein expression in VDAC2/RACK1 knockdown FG cells. (A) VDAC2 protein expression in VDAC2-siRNA-transfected FG cells. (B) RACK1 protein expression in RACK1-siRNA-transfected FG cells. VDAC2/RACK1 in FG cells is stained in red, and cell nuclei are stained in blue by DAPI. NC-siRNA instead of VDAC2-siRNA/RACK1-siRNA served as the negative control.

VDAC2 and RACK1 expression on nonpermissive EPCs confers susceptibility to LCDV infection.

EPCs were first confirmed to be nonpermissive to LCDV infection by qPCR and observation with an inverted microscope, and the results showed that the copy numbers of LCDV in EPCs were almost zero, while the copy numbers in FG cells were close to 2.5 × 10³ 48 h after LCDV infection (Fig. 8A). No CPE was observed in EPCs, and the cells looked normal during 2 weeks postinfection (Fig. 8B2) and after passage for two generations (Fig. 8B3 and B4) compared with the negative control (Fig. 8B1).

FIG 8.

VDAC2 and RACK1 expression confers LCDV susceptibility on nonpermissive EPCs. (A) LCDV copy numbers were compared in FG cells and EPCs. (B) CPE in LCDV-infected EPCs. (B1) EPCs without LCDV infection; (B2) CPE observed 2 weeks after LCDV infection; (B3 and B4) supernatant of EPCs passaged for two generations on EPCs. (C) mRNA expression of VDAC2 in pcDVDAC2-transfected EPCs detected by qPCR. (D) mRNA expression of RACK1 in pcGRACK1-transfected EPCs detected by qPCR. (E) LCDV copy numbers at different time points in pcDVDAC2-transfected EPCs. (F) LCDV copy numbers at different time points in pcGRACK1-transfected EPCs. Empty plasmid pcDNA3.1/pCIneo-GFP instead of pcDVDAC2/pcGRACK1 and EPCs treated with the transfection reagent served as negative controls. Error bars represent SD. Asterisks denote significant differences among groups (P < 0.05).

The EPCs were transfected with plasmids pcDVDAC2 and pcGRACK1, respectively, and qPCR results demonstrated that the mRNAs of pcDVDAC2 and pcGRACK1 of P. olivaceus were efficiently expressed in EPCs at high levels, whereas no expression was detected in EPCs transfected with the empty plasmid (Fig. 8C and D).

To further confirm that VDAC2/RACK1 expression on EPCs correlated with permissiveness to LCDV infection, pcDVDAC2-, pcGRACK1-, and empty plasmid-transfected EPCs were further infected with LCDV at 36 h posttransfection, and confocal microscopy results showed that positive red fluorescence of VDAC2/RACK1 and green fluorescence of LCDV were simultaneously present in the same EPCs, whereas no positive signals were observed in negative controls that were transfected with the empty plasmid (Fig. 9A and B). Moreover, qPCR results indicated that LCDV copy numbers were significantly higher in the pcDVDAC2-transfected EPCs than in negative controls that were transfected with the empty plasmid or treated with a transfection reagent 48 h and 72 h after LCDV infection, reaching a maximum value at 48 h (Fig. 8E), while in pcGRACK1-transfected EPCs, the virus copy number exhibited significantly higher levels all the times after LCDV infection than the negative controls, showing a peak at 24 h and then a second peak at 72 h (Fig. 8F). All these results indicated that VDAC2/RACK1 expression was sufficient to allow LCDV entry and infection.

FIG 9.

Costaining of VDAC2/RACK1 with LCDV in EPCs detected by confocal microscopy. (A) Costaining of VDAC2 with LCDV in pcDVDAC2-transfected EPCs. (B) Costaining of RACK1 with LCDV in pcGRACK1-transfected EPCs. Empty plasmid pcDNA3.1/pCIneo-GFP instead of pcDVDAC2/pcGRACK1 served as the negative control. VDAC2/RACK1 is stained in red, and LCDV is stained in green in pcDVDAC2/pcGRACK1-transfected EPCs. Cell nuclei were stained in blue by DAPI.

DISCUSSION

The first step for viruses to complete their life cycle is to adsorb to susceptible cells, which is determined by the structure of the virus itself and the surface composition of host cells (31, 32). For enveloped viruses such as LCDV, the virus invades host cells through interaction with cellular receptors (14, 15, 33). Research on virus cellular receptors is critical for elucidating the molecular mechanisms of virus infection, replication, and pathogenesis (34, 35). Previously, a 27.8-kDa protein in FG cells was found to be the putative receptor specific for LCDV entry and infection (14), which could interact with a 32-kDa envelope protein of LCDV functioning as a VAP (19). In the present study, we expanded the investigation to identify 27.8R, and 27.8R was identified as VDAC2 and RACK1 through two-dimensional (2D) Western blotting, 2D far-Western blotting, a 2D virus overlay protein binding assay (2D VOPBA), and mass spectrometry (MS). We found that the developed anti-VDAC2/RACK1 polyclonal antibodies could significantly decrease the percentages of LCDV-infected FG cells and LCDV copy numbers, block virus infection, and delay the development of CPE; moreover, VDAC2/RACK1 knockdown reduced LCDV copy numbers in FG cells, while ectopic expression of VDAC2/RACK1 in nonpermissive EPCs from the carp Cyprinus carpio conferred susceptibility to LCDV infection. All these results collectively revealed that VDAC2 and RACK1 are functional receptors for LCDV infection.

As mentioned above, VDAC2 and RACK1 have important roles in cell activity and have been reported to participate in infection by various viruses, such as nervous necrosis virus (24), Scophthalmus maximus rhabdovirus (25), infectious bursal disease virus (28–30), white spot syndrome virus (WSSV) (36, 37), mud crab reovirus (38), and influenza A virus (39). However, their roles in LCDV infection have not yet been discovered. In this study, VDAC2/RACK1 recognized by anti-27.8R MAbs also reacted with the 32-kDa VAP and LCDV particles, while the anti-VDAC2/RACK1 polyclonal antibodies and anti-27.8R MAbs could simultaneously recognize the rVDAC2/rRACK1 and 27.8-kDa proteins of FG cells, indicating that VDAC2 and RACK1 were 27.8R. After FG cells were preincubated with different concentrations of anti-VDAC2/RACK1 polyclonal antibodies before LCDV infection, the percentages of LCDV-infected cells and virus copy numbers were decreased, and CPE development and LCDV infection were inhibited, consistent with our previous results for anti-27.8R MAbs (16), revealing that these two proteins were both involved in LCDV entry and infection. However, LCDV infection could not be blocked completely, even with a high concentration of antibodies, which might suggest the existence of other LCDV cellular receptors, e.g., the 37.6-kDa cellular receptor (15), so further research about LCDV cellular receptors is needed. Moreover, whether simultaneous preincubations of anti-VDAC2 and anti-RACK1 polyclonal antibodies could exert a greater inhibiting effect also needs to be clarified. On the other hand, VDAC2/RACK1 was abundantly expressed on FG cells, and upregulated expression reached a higher peak value at 12 h for VDAC2 and a relatively lower peak at 48 h for RACK1, which might imply their functional difference in LCDV infection, but more studies on the difference in their functions are also needed. In our previous research, transcript analysis of flounder gills after LCDV infection showed the differential expression of apoptosis-related genes and pathways (40), so VDAC2 and RACK1 might work as receptors in the process of apoptosis in LCDV infection considering our present results and their biological functions (25, 41, 42). LCDV-infected cells were found to be inhibited in apoptotic death and cell division before enlargement in the early stage of lymphocystis cell formation (43); therefore, as cellular receptors for LCDV infection, how VDAC2 and RACK1 mediate cell apoptosis when initiating virus infection is unknown. LCDV isolated in China is reported to contain a putative homolog of cellular G protein-coupled receptors, which may inhibit apoptosis, as the viral G protein-coupled receptors can help the virus escape from host immune surveillance and contribute to viral pathogenesis (44), but the specific mechanism of VDAC2/RACK1-mediated LCDV infection and infected cell apoptosis inhibition requires further research.

Although VDAC2 and RACK1 have been reported to participate in various virus infections, there has been no report that the two proteins serve as cellular receptors. Prohibitin was identified as a receptor protein of dengue virus, and siRNA-mediated knockdown of prohibitin expression significantly reduced infection levels and subsequent virus production in permissive cells (45). In the present study, VDAC2 and RACK1 expressions were downregulated in FG cells after transfection with specific siRNAs, and LCDV copy numbers were significantly decreased in the knockdown FG cells, indicating that LCDV infection of FG cells is correlated with VDAC2 and RACK1 expressions. However, since we knocked down solely VDAC2 and RACK1 in this study, whether simultaneous knockdown of the two proteins would exert a greater impact on LCDV infection is not known, so more research is worth doing in the future. Moreover, VDAC2 is one of the isoforms of VDACs, along with VDAC1 and VDAC3. In a study of grass carp reovirus, VDAC2 was suggested to functionally complement VDAC1 (38). In the present study, we also found that VDAC1 (spot b in MS results) could react with 27.8R MAbs, but it did not react with LCDV or the 32-kDa VAP that specifically interacted with 27.8R, which suggested that VDAC1 was not involved in LCDV infection, although VDAC1 might have an antigen epitope similar to that of VDAC2, resulting in the reaction with 27.8R MAbs. RACKs were shown to anchor individual protein kinase C isozymes (46). In flounder, RACK (GenBank accession number AAT35603.1) and RACK1 (GenBank accession number XP_019964456.1) have been reported. We checked the sequences of the two proteins and found that the sequences were basically the same; only three peptides were different, so we speculate that the two reported proteins might be the same one, but more work is needed to confirm whether there are other paralogous proteins of RACK1 that might have the ability to bind LCDV. Besides gene interference, receptor reconstruction on nonpermissive cells is another acceptable method to clarify whether a protein is a virus cellular receptor (34); for example, JAM-1 expression in nonpermissive cells led to infection by feline calicivirus and reovirus (35, 47). To further elucidate the roles of VDAC2 and RACK1 in LCDV infection, reconstruction of flounder VDAC2/RACK1 was conducted on LCDV-nonpermissive EPCs, and the number of LCDV copies was significantly increased at 48 h and 72 h in VDAC2-transfected EPCs, and from 24 h to 96 h in RACK1-transfected EPCs, to levels higher than those of the negative controls. Meanwhile, VDAC2/RACK1 and LCDV were doubly stained in the same EPCs, suggesting that VDAC2 and RACK1 expression was sufficient to allow LCDV entry and infection, and LCDV could infect host cells via one of the two receptors; of course, this also raised the questions of whether coexpression of VDAC2 and RACK1 in EPCs could enhance LCDV infectivity and their respective roles in virus infection. Moreover, there were obvious differences between VDAC2 and RACK1 expressions in LCDV-infected FG cells, siRNA-transfected FG cells, and transfected EPCs; accordingly, LCDV copy numbers and their peak times were also different between VDAC2 and RACK1 knockdown FG cells as well as VDAC2- and RACK1-transfected EPCs. These results implied that the roles of VDAC2 and RACK1 in LCDV infection might have some differences, but more studies are needed to clarify this. Unexpectedly, the LCDV copy numbers were also detected in the negative controls comprised of empty plasmid-transfected or Lipofectamine 3000 reagent-treated EPCs; even though VDAC2/RACK1- and LCDV-positive fluorescence was not observed, this might be because the liposomes in the reagent could penetrate the cell membrane and lead to nonspecific binding of LCDV, since the copy numbers decreased over time. Furthermore, the interaction of the two proteins with the LCDV 32-kDa VAP would be helpful in elucidating the molecular mechanism underlying LCDV infection. In addition to facilitating attachment to the cells, virus receptors may transmit signals required for viral entry, mediate internalization of the virus, or trigger the release of the viral genome into the cell (48–50), so further research on the function of VDAC2 and RACK1 in LCDV internalization, uncoating, or genome release needs to be performed.

In the present study, the two anti-27.8R MAbs (2G11 and 3D9) could recognize the identified RACK1 and VDAC2, indicating that 27.8R was a mixed protein. Previously, 27.8R was identified to have an association with β-actin of Mus musculus by MS using polyclonal anti-27.8-kDa protein antiserum, and the β-actin antibody reacted strongly with a 42.7-kDa protein and weakly with the 27.8-kDa protein band of FG cells (14), whereas in this study, 2D Western blotting of FG cell membrane proteins with anti-27.8R MAbs also identified an ∼43-kDa protein, which was identified as β-actin by MS (data not shown), which might suggest that 27.8R shared epitopes with the β-actin protein. However, 2D far-Western blotting and a 2D VOPBA indicated that the 32-kDa VAP and LCDV reacted with just the two protein dots that were confirmed to be RACK1 and VDAC2; therefore, the β-actin protein should not be the binding protein of LCDV as a cellular receptor.

In conclusion, we confirmed that VDAC2 and RACK1 were 27.8R, and anti-VDAC2/RACK1 polyclonal antibodies could recognize the 27.8-kDa protein in FG cells and rVDAC2/rRACK1 and inhibit LCDV infection of FG cells. The mRNA expressions of the two proteins were upregulated after LCDV infection, and knockdown of VDAC2 and RACK1 expression in FG cells decreased the LCDV copy numbers, while reconstruction of VDAC2 and RACK1 expression on LCDV-nonpermissive EPCs conferred susceptibility to LCDV infection. These results revealed that VDAC2 and RACK1 expression was sufficient to allow LCDV entry and that they were functional receptors that initiated LCDV infection, which leads to a better understanding of the molecular mechanism underlying LCDV infection and virus-host interactions.

MATERIALS AND METHODS

Ethics statement.

This study was conducted strictly in line with the procedures in the guide for the use of experimental animals of Ocean University of China. Animal experiments were approved by the Institutional Animal Care and Use Committee of Ocean University of China (permit number 20151201). All efforts were made to minimize suffering.

Cells, virus, proteins, and antibodies.

FG cells were grown in minimal essential medium (MEM; Gibco, Germany) supplemented with 10% fetal bovine serum (FBS; Gibco), 100 IU/ml penicillin, and 100 μg/ml streptomycin (Gibco) and cultivated at 22°C with 2% CO₂. The dosage of FBS was reduced to 2% in maintenance medium after LCDV infection.

The LCDV particles were prepared through sucrose discontinuous density gradients (51). FG cell membrane proteins were prepared using gradient centrifugation as described in our previous study (14). Escherichia coli strain BL21(DE3) expressing the 32-kDa VAP of LCDV was grown in LB broth, and recombinant VAP (rVAP) protein was purified by using a Ni-NTA column (19). White spot syndrome virus (WSSV) particles and recombinant pET28a-VP26 of WSSV were produced previously by our laboratory (52, 53).

Mouse anti-27.8R MAbs (3D9 and 2G11) (16), anti-LCDV MAbs (1A8 and 1B2) (51), and anti-LCDV 32-kDa VAP MAbs (1C6, 1C8, and 3H10) (20) were previously produced by our laboratory.

Two-dimensional electrophoresis of FG cell membrane proteins.

The concentrations of FG cell membrane proteins were determined by a bicinchoninic acid (BCA) protein assay kit (Beyotime) according to the manufacturer’s instructions and adjusted to 1.2 mg/ml. A total of 100 μl FG cell membrane proteins was used for two-dimensional electrophoresis. Protein desalination was carried out using 2D cleanup kits (GE Healthcare, USA) according to the manufacturer’s directions. First-dimension isoelectric focusing (IEF) was performed with the Ettan IPGphor II system (Amersham Biosciences), and 300 μl rehydration buffer {8 M urea, 4% 3-[(3-cholamidopropyl)-dimethylammonio]-1-propanesulfonate (CHAPS), 30 mM dithiothreitol, 0.001% bromophenol blue} containing 120 μg membrane protein was loaded onto immobilized pH gradient strips (Immobiline DryStrip, nonlinear, pH 3 to 10, 13 cm; GE Healthcare). IEF was performed at 50 V for 12 h, and the voltage was then increased from 100 V to 8,000 V for 7 h (100 V for 30 min, 200 V for 30 min, 500 V for 30 min, 1,000 V for 30 min, and 8,000 V for 5 h), followed by 500 V for 2 h; all the procedures were conducted at 20°C. In the second dimension, the focused strips were first equilibrated in equilibration solution (6 M urea, 30% glycerol, 2% SDS, 50 mM Tris-HCl, 0.002% bromophenol blue [pH 8.8]) containing 1% (wt/vol) dithiothreitol for 14 min and further equilibrated in equilibration solution containing 2.5% (wt/vol) iodoacetamide for 14 min. The strips were subjected to 12.5% SDS-PAGE and sealed with 0.5% melted agarose. Electrophoresis was ended when bromophenol blue dye migrated to the bottom of the gels. Finally, the gels were subjected to silver staining and scanned using a scanning densitometer (GE Healthcare).

2D Western blotting, 2D far-Western blotting, and 2D VOPBA.

FG cell membrane proteins were first subjected to two-dimensional electrophoresis as described above, and the gels were then equilibrated in transfer buffer (0.025 M Tris base, 0.19 M Gly, 20% methanol) for 30 min and transferred to a PVDF membrane using a semidry transfer unit (GE Healthcare) at 132 mA for 1.5 h. For 2D Western blotting, the membrane was blocked with 4% bovine serum albumin (BSA) at 4°C overnight, washed three times with phosphate-buffered saline (PBS) containing 0.05% Tween 20 (PBST), incubated with anti-27.8R MAbs (1:1,000; 1:1 mixture of 2G11 and 3D9) at 37°C for 1.5 h, washed with PBST, and then incubated with alkaline phosphatase (AP)-conjugated goat anti-mouse Ig (1:1,000; Sigma, USA) at 37°C for 1 h. After washing as described above, the membrane was placed into a substrate solution (100 mM NaCl, 100 mM Tris, and 5 mM MgCl₂ [pH 9.5]) containing nitro blue tetrazolium and 5-bromo-4-chloro-3-indolylphosphate (NBT-BCIP) substrates (Sigma) for staining. A myeloma cell (P3-X63-Ag8U1 [P3U1]) culture supernatant instead of MAbs against 27.8R was used as a negative control.

For 2D far-Western blotting, the FG cell membrane proteins were subjected to two-dimensional electrophoresis and transferred to a PVDF membrane as described above. The membranes were blocked with 4% BSA at 4°C overnight, washed three times with PBST, and incubated with the purified rVAP at a concentration of 1 mg/ml at 4°C for 20 h. Following washing as described above, the membrane was incubated using anti-32 kDa VAP MAbs (1C8, 3B5, and 3H10 in a 1:1:1 mixture) as the primary antibodies and AP-conjugated goat anti-mouse Ig (1:1,000) as the secondary antibody and then stained with a substrate solution containing NBT-BCIP. Recombinant pET28a-VP26 of WSSV in place of the rVAP and the P3U1 culture supernatant instead of anti-32-kDa VAP MAbs were used as negative controls.

For the 2D VOPBA, the FG cell membrane proteins were subjected to two-dimensional electrophoresis and transferred to a PVDF membrane as described above. The VOPBA was performed as follows. The membrane was blocked with 4% BSA at 4°C overnight, washed with PBST, and then incubated with purified LCDV particles at a concentration of 1 mg/ml at 4°C for 20 h. After washing as described above, the membrane was incubated using anti-LCDV MAbs (1A8 and 1B2 in a 1:1mixture) as primary antibodies and AP-conjugated goat anti-mouse Ig (1:1,000) as the secondary antibody and then stained as described above. The same concentration of WSSV instead of LCDV and the P3U1 culture supernatant instead of anti-LCDV MAbs served as negative controls.

Mass spectrometry analysis.

According to the positive spots in PVDF membranes, the relevant spots between 25 kDa and 35 kDa that specifically reacted with the MAbs against 27.8R were excised from gels for protein identification using MS by Shanghai Applied Protein Technology (China). The sequences of possible receptor proteins that simultaneously reacted with the anti-27.8R MAbs, 32-kDa VAP, and LCDV particles were obtained according to the results of amino acid sequence alignment.

Recombinant expression of VDAC2 and RACK1.

Based on the result of MS analysis and sequence alignment, the recombinant plasmids pET-32a-VDAC2 (24) and pET-32a-RACK1 were constructed. Briefly, the VDAC2 and RACK1 genes were amplified from the cDNA of FG cells using the primers shown in Table 2, the amplified VDAC2 and pET-32a plasmids were doubly digested by KpnI and HindIII restriction enzymes (TaKaRa, Japan), the amplified RACK1 and pET-32a plasmids were doubly digested by KpnI and XhoI restriction enzymes (TaKaRa, Japan), and the doubly digested gene and plasmid were then linked together using T4 DNA ligase (TaKaRa, Japan). The recombinant plasmid was transformed into competent E. coli BL21(DE3) cells, cultured in LB broth, and then detected by colony PCR. DNA sequencing of positive clones was performed by Shanghai Sangon Biotech (China). E. coli BL21(DE3) cells containing recombinant plasmid pET-32a-VDAC2/pET-32a-RACK1 were cultured, and the expression of rVDAC2/rRACK1 was induced by adding 100 mM isopropyl-β-d-thiogalactopyranoside (1:100) (Solarbio, China). rVDAC2 and rRACK1 were purified using HisTrap Ni-NTA affinity chromatography (GE Healthcare, USA) with a protein purification apparatus according to the manufacturer’s instructions.

TABLE 2.

Primers and siRNAs used in this study

Primer	Primer sequence (5′–3′)	Usage
yVDAC2-F	AATTTGGTACCGATACAATGGCCG	pET-32a-VDAC2 plasmid construction
yVDAC2-R	CTCGAAAGCTTAGGCTTCCAGTTCC	pET-32a-VDAC2 plasmid construction
yRACK1-F	GGGGTACCATGACCGAGCAGATGACAGTGAG	pET-32a-RACK1 plasmid construction
yRACK1-R	CCCTCGAGGACTTGCCACACTCTGATCAGGTTG	pET-32a-RACK1 plasmid construction
VDAC2-F	CCAAGCTTACCATGGATACAATGGCCG	pcDNA3.1-VDAC2 plasmid construction
VDAC2-R	GGGGTACCTTAAGGCTTCCAGTTCC	pcDNA3.1-VDAC2 plasmid construction
RACK1-F	CCCTCGAGATGACCGAGCAGATGACAGTGAG	pCIneo-GFP-RACK1 plasmid construction
RACK1-R	CGGAATTCTGACTTGCCACACTCTGATCAGGTTG	pCIneo-GFP-RACK1 plasmid construction
qVDAC2-F	ACAGCACACGCTTCGGTATT	qPCR (mRNA expression of VDAC2)
qVDAC2-R	CCTCCAGCGTTGATGTTCTT	qPCR (mRNA expression of VDAC2)
qRACK1-F	GGACCTGAATGAGGGAAAGCAC	qPCR (mRNA expression of RACK1)
qRACK1-R	AGCAGACCATGCCAGGGAAGTA	qPCR (mRNA expression of RACK1)
beta-actin-F	CACTGTGCCCATCTACGAG	qPCR
beta-actin-R	CCATCTCCTGCTCGAAGTC	qPCR
LCDV-038F	TCTTGTTCAGCATTTACTTCTCGGC	qPCR (LCDV copy no.)
LCDV-038R	TCTTCTCCTTTAGATGATTTCCC	qPCR (LCDV copy no.)
VDAC2-siRNA	GCAGGUAGCCUGGAAACUATT	RNA interference assay
RACK1-siRNA	CCGAAUCCAUCCUCGUCAATT	RNA interference assay
NC-siRNA	UUCUCCGAACGUGUCACGUTT	RNA interference assay

Polyclonal antibody preparation and specificity analysis.

The anti-VDAC2/RACK1 polyclonal antibodies were obtained by immunizing New Zealand White rabbits weighing about 2.5 kg. For the first immunization, the rabbits were injected at six sites in the backside with 1.2 mg recombinant proteins emulsified with complete Freund’s adjuvant (Sigma, USA) (1:1). Two weeks later, a booster immunization was performed by injecting 1.2 mg recombinant proteins emulsified with incomplete Freund’s adjuvant (Sigma, USA) (1:1) at six sites in the rabbit backside. The rabbits were thereafter given two booster immunizations with 1.2 mg recombinant proteins through ear vein injection at 1-week intervals. One week after the last immunization, blood was obtained from the rabbit heart and placed at room temperature for 1 h and then at 4°C overnight. The samples were centrifuged at 10,000 × g for 30 min, and the supernatant was collected, purified by a caprylic acid-ammonium sulfate precipitation method (54), completely dialyzed in PBS at 4°C, and stored at −80°C until use.

To analyze the specificity of the developed anti-VDAC2 polyclonal antibodies, 15 μl purified rVDAC2 and 15 μl FG cell membrane proteins were mixed with an equal volume of protein loading buffer separately and denatured for 5 min at 100°C, and the mixtures were subjected to SDS-PAGE, followed by transfer to a PVDF membrane and incubation with 4% BSA at 4°C overnight. After washing with PBST, the membrane was separately incubated using anti-27.8R MAbs 2G11 and 3D9 and anti-VDAC2 polyclonal antibodies (1:500) as primary antibodies and AP-conjugated goat anti-mouse Ig (1:1,000) or goat anti-rabbit Ig (1:1,000) (Sigma, USA) as the secondary antibody and then stained using a substrate solution containing NBT-BCIP. The P3U1 culture supernatant instead of anti-27.8R MAbs and rabbit preimmune serum instead of anti-VDAC2 polyclonal antibodies were used as negative controls. Similarly, the specificity of the developed rabbit anti-RACK1 polyclonal antibodies was analyzed.

Ability of anti-VDAC2/RACK1 polyclonal antibodies to inhibit LCDV infection in vitro.

FG cells were grown in 75-cm² cell culture flasks at 22°C, and monolayer cultures were washed twice with MEM. Anti-VDAC2/RACK1 polyclonal antibodies were diluted with MEM into serial concentrations (160 μg/ml, 16 μg/ml, 1.6 μg/ml, and 0.16 μg/ml) and incubated with FG cells at 22°C for 2 h. After the supernatant was removed, the cells were further infected with LCDV at 4 50% tissue culture infective doses (TCID₅₀)/ml at 22°C for 1 h. Following this, the LCDV particles were removed, maintenance medium was added, the cells were cultured at 22°C with 2% CO₂, and the cytopathic effect (CPE) in FG cells was monitored by using an inverted microscope. Rabbit preimmune serum with the same concentration gradients instead of anti-VDAC2/RACK1 polyclonal antibodies served as the negative control. After 48 h, the FG cells were digested using trypsin, washed with PBS, and thereafter fixed with 4% paraformaldehyde for 15 min at room temperature. Subsequently, the cells were incubated with MAbs against the 32-kDa VAP of LCDV for 1.5 h at 37°C, washed three times with PBS, and then incubated with goat anti-mouse Ig-fluorescein isothiocyanate (FITC) (1:256) (Sigma) for 45 min at 37°C. After three washes with PBS, the percentages of LCDV-infected FG cells were analyzed by FACS analysis using a flow cytometer (Beckman Coulter, USA). The DNA of FG cells in each treatment group was extracted using a TIANamp marine animal DNA kit (Qiagen, Germany), and LCDV copy numbers were detected by qPCR using specific primers LCDV038F and LCDV038R (Table 2) and calculated according to a standard curve as described in a previous study (19). Furthermore, FG cells grown on circular coverslips in 24-well plates were incubated with 1.6 μg/ml anti-VDAC2 polyclonal antibodies or 16 μg/ml anti-RACK1 polyclonal antibodies and then infected with LCDV as described above. LCDV infection of FG cells was detected by an IFA, using anti-32-kDa VAP MAbs as primary antibodies and goat anti-mouse Ig-FITC (1:256) as a secondary antibody. Cell nuclei were visualized with 4′,6-diamidino-2-phenylindole (DAPI; Roche). Rabbit preimmune serum instead of primary antibodies with the same concentrations served as the negative control. The experiments were repeated in triplicate.

Expression of VDAC2 and RACK1 in FG cells after LCDV infection.

FG cells grown in 25-cm² cell culture flasks were infected with LCDV particles at 4 TCID₅₀/ml at 22°C for 1 h, the virus was then removed, and maintenance medium was added. Triplicate flasks were sampled at 0, 2, 6, 12, 24, 36, 48, 72, 96, 120, and 168 h postinfection. After the culture medium was removed, the cells of each flask were collected, and total RNA of FG cells was extracted using RNAiso (TaKaRa, Japan) and reverse transcribed into cDNA using a PrimeScript RT-PCR kit (TaKaRa). Primers in Table 2 were used to detect VDAC2 and RACK1 mRNA expression by real-time PCR. The reaction mixture for qPCR contained 10 μl SYBR green I (Roche, Switzerland), 50 ng sample cDNA, and 2 μl detection primers, with sterile distilled water added to a final volume of 20 μl. The reaction procedures were as follows: 10 min at 95°C, followed by 45 cycles of 10 s at 95°C, 10 s at 55°C, and 20 s at 72°C. The relative expression levels of VDAC2 and RACK1 genes were calculated according to $2^{-\Delta \Delta C_T}$ methods, with the expression of β-actin as an internal control.

To detect the protein expression of VDAC2/RACK1 in FG cells, FG cells grown on circular coverslips were fixed with 4% paraformaldehyde and incubated with rabbit anti-VDAC2/RACK1 polyclonal antibodies (1:500) as primary antibodies and Cy3-labeled goat anti-rabbit Ig (Sigma) (1:400) as a secondary antibody, cell nuclei were visualized with DAPI, and the expression of VDAC2/RACK1 was detected by confocal microscopy. Rabbit preimmune serum was used as a negative control.

RNAi knockdown of VDAC2/RACK1 in FG cells.

Short interfering RNAs (siRNAs) designed by Shanghai GenePharma Company (China) were used to knock down VDAC2/RACK1 in FG cells. The siRNA sequences of VDAC2/RACK1 (VDAC2-siRNA/RACK1-siRNA) and a nonsilencing control siRNA (NC-siRNA) are shown in Table 2. FG cells were grown in six-well plates, and the siRNAs were transfected into FG cells using Lipofectamine 3000 reagent (Invitrogen, USA) according to the manufacturer’s instructions. Briefly, 5 μl (20 μM) of siRNAs and 3.25 μl Lipofectamine 3000 reagent were diluted with 125 μl Opti-MEM (Gibco, Germany) separately, mixed, and incubated for 20 min at room temperature, and 2 ml Opti-MEM was then added to the mixtures. After incubation for 6 h, the mixtures were removed, and maintenance medium without penicillin and streptomycin was added to the FG cells. At 12 h, 24 h, and 48 h posttransfection, the FG cells were collected, total RNA of FG cells was extracted, and VDAC2/RACK1 mRNA expression was detected by real-time PCR as described above. In addition, the VDAC2/RACK1-knocked-down FG cells at 36 h were dropped onto coverslips and allowed to settle for 2 h, and the protein expression of VDAC2/RACK1 was detected by an IFA as described above.

To confirm whether VDAC2/RACK1 knockdown can reduce LCDV infection, 24 h and 48 h after transfection with VDAC2-siRNA, RACK1-siRNA, and NC-siRNA, FG cells were infected with LCDV particles at 4 TCID₅₀/ml at 22°C for 1 h. Forty-eight hours after LCDV infection, the cells were collected, DNA was extracted using a TIANamp marine animal DNA kit (Qiagen, Germany), and LCDV copy numbers were detected by qPCR and calculated by using a standard curve according to the threshold cycle (C_T) values (19). The experiments were carried out in triplicate.

LCDV infection of pcDVDAC2/pcGRACK1-transfected EPCs.

EPCs are nonpermissive to LCDV infection according to a previous study by Alonso et al. (55). To further confirm this, EPCs and FG cells cultured in 25-cm² culture flasks were inoculated with LCDV and cultured for 48 h, the cells were collected, and DNA was extracted to detect LCDV copy numbers by qPCR as described above. The inoculated EPCs were passaged in a blind manner for two generations, and the CPE in EPCs was observed for 2 weeks under an inverted microscope.

To determine whether the expression of VDAC2 or RACK1 on nonpermissive EPCs could confer susceptibility to LCDV infection, the plasmids pcDVDAC2 and pcGRACK1 were constructed with primers shown in Table 2. Briefly, the amplified VDAC2 and pcDNA3.1 plasmids were doubly digested by HindIII and KpnI restriction enzymes (TaKaRa, Japan), the amplified RACK1 and pCIneo-GFP plasmids were doubly digested by XhoI and EcoRI restriction enzymes (TaKaRa), and the doubly digested genes and plasmids were then linked together using T4 DNA ligase (TaKaRa). The recombinant plasmid was transformed into competent E. coli DH5α cells (TaKaRa), followed by detection by colony PCR and sequencing. The recombinant plasmid was extracted using EndoFree miniplasmid kit II (Tiangen, Beijing), and the concentration was adjusted to 500 ng/μl.

The EPCs were seeded onto six-well plates, cultured for 24 h before transfection, and then transfected with the pcDVDAC2/pcGRACK1 recombinant plasmid using Lipofectamine 3000 reagent. Briefly, 5 μl (2.5 μg) of the pcDVDAC2/pcGRACK1 recombinant plasmid and 5 μl P3000 reagent were added to 125 μl Opti-MEM, 3.25 μl Lipofectamine 3000 reagent was added to 125 μl Opti-MEM separately, the two mixtures were mixed and incubated for 20 min at room temperature, and 2 ml Opti-MEM was then added to the mixtures, which were used to incubate EPCs for 6 h. EPCs transfected with the empty plasmid or treated with Lipofectamine 3000 reagent served as negative controls. Thirty-six hours after pcDVDAC2/pcGRACK1 transfection, the treated EPCs were infected with LCDV at 4 TCID₅₀/ml at 22°C for 2 h, the virus liquid was then removed, and maintenance medium was introduced. At 24 h, 48 h, 72 h, and 96 h postinfection, the mRNA expressions of VDAC2 and RACK1 were detected by qPCR, and the DNA of EPCs was extracted using a TIANamp marine animal DNA kit for detection of LCDV copy numbers by qPCR (19). Furthermore, costaining of RACK1/VDAC2 with LCDV was detected by confocal microscopy 72 h after LCDV infection. Briefly, EPCs transfected with pcGRACK1/pcDVDAC2 were dropped onto the slides, allowed to settle for 2 h, and then fixed with 4% paraformaldehyde for 15 min at room temperature. The cells were thereafter incubated with rabbit anti-RACK1/VDAC2 polyclonal antibodies (1:500) paired with mouse anti-LCDV 32-kDa VAP MAbs for 1.5 h at 37°C. After washing with PBST three times, the cells were incubated with Cy3-labeled goat anti-rabbit Ig (1:400) paired with FITC-conjugated goat anti-mouse Ig (1:256) for 1 h at 37°C. Following three washes with PBST, the cell nuclei were visualized by using DAPI for 20 min at 37°C, and costaining of RACK1/VDAC2 and LCDV was observed under a confocal microscope. The experiments were repeated in triplicate.