1 Introduction

Classical swine fever (CSF) is a highly contagious disease that affects domestic and wild pigs; it is listed as one of the notifiable swine diseases by the Office International Des Epizooties (OIE) (Lin et al. 2014). Classical swine fever virus (CSFV), which belongs to the genus Pestivirus of the Flaviviridae family, is a single-stranded positive-sense RNA virus encoding four structural proteins (C, Erns, E1 and E2) and eight non-structural proteins (Npro, P7, NS2, NS3, NS4A, NS4B, NS5A and NS5B) (Lamp et al. 2011). These non-structural proteins participate in self-maturation and replication complex assembly (Sheng et al. 2012).

Plus-stranded RNA viruses replicate in the cytoplasm, and their synthesis is associated with cellular membrane structures. Specifically, the endoplasmic reticulum (ER) and Golgi apparatus as well as their resident proteins are used by many viruses to complete one or more stages in their life cycle (Ravindran et al. 2016; Wang et al. 2016). Anterograde and retrograde trafficking of cargo proteins between the ER and Golgi are mediated by the cytosolic coat protein complex II (COP II) and COP I transport systems, respectively. These systems are regulated by viruses to facilitate their multiplication (Rust et al. 2001; Richards et al. 2002; Yamayoshi et al. 2008). Viruses also regulate two important targets, i.e. Golgi-specific BFA-resistance guanine nucleotide exchange factor 1 (GBF1) and Rab2 GTPases, which are primarily involved in Golgi-to-ER retrograde trafficking (Raj 1961; Wessels et al. 2006).

Increasing evidence has shown that CSFV replication relies on cellular membrane structures. CSFV can utilize the membranes of autophagosome-like vesicles to facilitate self-replication (Pei et al. 2013). CSFV p7, a viroporin protein in the ER and plasma membrane, modifies membrane permeability (Guo et al. 2013; Largo et al. 2014). According to our previous study, CSFV NS2 and NS5A proteins localize to the ER and regulate cellular oxidative stress to promote CSFV replication (Tang et al. 2010; He et al. 2012). However, whether Golgi function and its resident proteins are involved in CSFV replication remains unknown.

In this study, chemical inhibitors, shRNA and lentiviral systems were used to regulate the function of Golgi-resident proteins. The results of these analyses indicated that GBF1 and Rab2 play important roles in CSFV replication.

2 Materials and methods

2.1 Cells and virus

Swine testicle (ST) cells were grown in high-glucose Dulbecco’s modified Eagle’s medium (DMEM; Gibco, Paisley, UK) containing 10% fetal bovine serum (FBS; HyClone, Logan, UT, USA). The CSFV Shimen strain was purchased from the Control Institute of Veterinary Bio-products and Pharmaceuticals (China) and grown in ST cells. The MOI of CSFV infection in this experiment was set at 10 TCID50.

2.2 Antibodies and chemical inhibitors

The CSFV polyclonal antibody was collected from Hog Cholera Lapinized Virus (HCLV) inoculated pigs and stored at −20°C. The anti-pig IgG-FITC antibody was purchased from Sigma Aldrich (F1638; St. Louis, MO, USA). Goat anti-GBF1 polyclonal antibody (C-16, sc-27940) was purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Rabbit anti-Rab2 polyclonal antibody (ab154729) was purchased from Abcam (Cambridge, UK). Mouse anti-GAPDH antibody was obtained from GenScript (Piscataway, NJ, USA). The secondary antibodies for Western blotting were HRP-labelled goat anti-rabbit IgG, goat anti-mouse IgG, and rabbit anti-goat IgG (Beyotime, Haimen, China). Brefeldin A (BFA) and golgicide A (GCA) were purchased from Selleck Chemicals (Houston, USA).

2.3 Cell safe concentration of chemical inhibitors

An MTT assay was used to test the cytotoxicity of BFA and GCA. Various concentrations of BFA and GCA were added to ST cells, which were plated in 96-well plates at a confluence of about 60%. The cells were then maintained in DMEM with 1% FBS at 37°C in 5% CO2. After 72 h of incubation, the supernatant was removed. Each well was supplemented with 50 μL of MTT (2 μg/mL) and incubated at 37°C for 4 h. MTT was then removed, and 200 μL of DMSO was added to dissolve the formazan crystals. A microplate reader (Thermo, Waltham, MA, USA) was used to determine light absorbance for each well at 490 nm. Untreated ST cells were used as controls.

2.4 GBF1and Rab2 shRNA vector construction and cell transfection

The pGFP-V-RS shRNA plasmid (OriGene Co., Rockville, MD, USA) was used to construct Rab2 and GBF1 shRNA interference vectors. For each gene, three pairs of interference fragments were used (table 1), which were synthesized and sequenced by BGI Shenzhen (China). ST cells were transfected with the interference vector when cell confluence reached 70–90%. The transfection operation followed the protocol for the jetPRIME transfection reagent (Polyplus-transfection, Illkirch, France). At 48 h post-transfection, 5 μg/mL puromycin was added to continuously screen positive cells for about 20 days. Scrambled shRNA was constructed and transfected using the same methods.

Table 1 Short hairpin RNA (shRNA) inserts
Full size table

2.5 Real-time PCR

Cellular CSFV, Rab2 or GBF1 RNA was tested by relative quantitative PCR. CSFV-infected ST cells with or without inhibitors and cells in which genes silenced were collected. Total RNA was isolated using TRIzol (TaKaRa, Dalian, China) according to the manufacturer’s instructions. The RNA concentration was determined using a Nucleic Acid and Protein Analyzer (NanoDrop Technologies, Wilmington, DE, USA) at a wavelength of 260 nm. The PrimeScriptTM RT Reagent Kit (TaKaRa) was used for reverse transcription and cDNA synthesis. For real-time PCR, specific oligonucleotide primers (table 2) were used to determine the expression levels of different genes. The reaction was performed with SYBR Premix Ex Taq II (TaKaRa) using the Bio-Rad iQ5 system under the following conditions: pre-degeneration at 95°C for 10 min, followed by 40 cycles of degeneration at 95°C for 10 s, and annealing and extension at 60°C for 30 s. RNA expression in the samples was normalized to that of β-actin. The 2−ΔΔCT method was used to analyse the data. The ΔCT values were calculated as the CT values of the target gene minus the CT values of β-actin. GBF1 and Rab2 gene expression levels in ST cells or CSFV RNA in ST cells at the earliest time point were selected as comparative references, and so the average ΔCT values were set as the standard reference for the −ΔΔCT calculation. Then the 2−ΔΔCT values were obtained and histograms were obtained in which the height of the bar for each sample represented the average 2−ΔΔCT value.

Table 2 Primers used in this study
Full size table

2.6 Indirect immunofluorescence assay and viral titration

CSFV-infected cells at 48 hpi were observed by indirect immunofluorescence assay (IFA). Cell monolayers were fixed with 4% paraformaldehyde for 20 min, following cell penetration with 1% Triton X100 for 10 min. Subsequently, 5% skim milk was added to the cell monolayer at 4°C for 6 h. The CSFV polyclonal antibody was diluted by 100-fold with PBS and incubated with the cell monolayer at 4°C for 8 h. The FITC anti-pig antibody diluted 200-fold was used for coloration at room temperature for 1 h. The nuclei were stained with DAPI at room temperature for 5 min. Images were observed under an inverted fluorescence microscope (Nikon, Tokyo, Japan). CSFV-infected cell supernatants were collected for viral titration. The supernatants were serially diluted from 10−1.0 to 10−10.0 and added to ST cells in 96-well plates. At 72 hpi, CSFV infection was observed by IFA. The proportionate distance (PD) and the log lower dilution were calculated according to the Reed–Muench method as follows:

$$ \begin{array}{l}\mathrm{P}\mathrm{D}=\left(\%\kern0.5em \mathrm{next}\kern0.5em \mathrm{above}\kern0.5em 50\%\kern0.5em \mathrm{positive}\right)\hbox{--} 50\%/\left(\%\kern0.5em \mathrm{next}\kern0.5em \mathrm{above}\kern0.5em 50\%\kern0.5em \mathrm{positive}\right)\hbox{--} \left(\%\kern0.5em \mathrm{next}\kern0.5em \mathrm{below}\kern0.5em 50\%\kern0.5em \mathrm{positive}\right)\hfill \\ {}\mathrm{Log}\kern0.5em \mathrm{lower}\kern0.5em \mathrm{dilution}=\mathrm{dilution}\kern0.5em \mathrm{in}\kern0.5em \mathrm{which}\kern0.5em \mathrm{position}\kern0.5em \mathrm{is}\kern0.5em \mathrm{next}\kern0.5em \mathrm{above}\kern0.5em 50\kern-0.75em \%\kern0.5em \mathrm{positive}\hfill \\ {}{\mathrm{lgTCID}}_{50}=\mathrm{P}\kern-0.5em \mathrm{D}\times \mathrm{the}\kern0.5em \mathrm{difference}\kern0.5em \mathrm{of}\kern0.5em \mathrm{logarithmic}\kern0.5em \mathrm{dilution}\kern0.5em \mathrm{degrees}+\mathrm{Log}\kern0.5em \mathrm{lower}\kern0.5em \mathrm{dilution}\hfill \end{array} $$

2.7 Western blotting

Cells were collected and incubated in RIPA lysis buffer (Beyotime) containing the proteinase inhibitor PMSF (Beyotime) for 30 min on ice. A BCA Protein Assay Kit (Beyotime) was used to determine protein concentrations. Equivalent amounts of protein samples were separated by 12% SDS-PAGE and transferred to 0.45-μm PVDF membranes (Millipore, Billerica, MA, USA). Membranes were blocked in TBST containing 5% skim milk powder for 2 h at room temperature. Then, membranes were incubated with a mouse anti-GAPDH antibody or the primary antibodies at the appropriate dilutions according to the instructions overnight at 4°C. Afterwards, the HRP-conjugated secondary antibodies diluted at 1:2000 were added and incubated for 2 h at room temperature. After washing with TBST, immunoreactive bands were detected using luminol chemiluminescence reagents (Pierce) and analysed by Tanon-410 automatically gel imaging system (Shanghai Tianneng, China).

2.8 Lentivector construction and lentivirus production

The primers used for GBF1 and Rab2 gene amplification contained EcoRI and NotI restriction sites. The lentivector pCDH-CMV-MCS-EF1-GreenPuro (CD513B-1) (SBI, Mountain View, CA, USA) was used to construct the over-expression vectors pCDH-GBF1 and pCDH-Rab2. The overexpression vector was transfected into HEK-293T cells with three auxiliary plasmids (pGag/Pol, pRev, and pVSV-G). The supernatant of HEK-293T cells was replaced with advanced DMEM contain 2% FBS, 0.01 mM cholesterol, 0.01 mM lecithin, and 1× chemically defined lipid concentrate (Invitrogen, Carlsbad, CA, USA) 16 h post-transfection. The lentiviruses were collected after 48 h post-transfection. ST cells monolayers supplemented with Polybrene at a final concentration of 8 μg/mL and infected using overexpression lentiviruses (MOI=1). The culture medium was replaced with fresh medium at 24 hpi. After another 48 h of incubation, the cells overexpressing GBF1 and Rab2 were used for evaluation and subsequent tests.

2.9 Statistical analysis

Differences among groups were examined for statistical significance using Student’s t-tests. P<0.05 was considered statistically significant.

3 Results

3.1 BFA treatment inhibited CSFV proliferation in infected ST cells

The safe concentration of BFA in ST cells was tested using an MTT assay. The inhibitors were co-cultured with ST cells for 72 h. The concentration of BFA used in the antiviral experiment was 100 nM (figure 1A). After 1 h of CSFV infection, the unbound viruses were removed with the supernatant, and the cells were maintained in media with or without BFA. The cell monolayers were harvested for real-time PCR, and the supernatants were collected for viral titration after 24, 48 and 72 hpi. Cellular CSFV RNA loads decreased by about 2–6 times after BFA treatment (figure 1B). The TCID50 of CSFV exhibited a titer reduction of approximately 1.5–2.5 in the BFA treatment group compared with the untreated group (figure 1C). CSFV proliferation with BFA was evaluated by IFA. The CSFV plaques were smaller and less abundant than the untreated ST cells at all three test points (figure 1D).

Figure 1
figure 1

CSFV replication was sensitive to BFA. (A) The safe concentration of BFA was tested using an MTT assay. ST cell monolayers in 96-well plates with various concentrations of BFA were cultured for 72 h at 37°C. Relative cell viability was calculated by comparing the O.D.490 values of BFA-treated cells with those of untreated control cells (set to 100%). (B) Cellular CSFV RNA with or without 100 nM BFA were examined by real-time PCR. (C) The titers of CSFV with or without BFA were examined by IFA. (D) CSFV proliferation in ST cells with or without BFA was observed by IFA at 24, 48 and 72 hpi. Results from three independent experiments are shown as means ± SD. *P<0.05 and **P<0.01 compared with the control group.

Full size image

3.2 CSFV multiplication was inhibited by GCA

GCA is a specific inhibitor of GBF1. Before the anti-CSFV test, the cytotoxicity of GCA was detected and, 400 nM was considered the maximum safe concentration for ST cells. This concentration was used in subsequent experiments (figure 2A). GCA cell treatment and the collection of cell monolayers and supernatants were performed using the same methods as those for BFA. Compared with the untreated ST cells, cellular CSFV RNA declined significantly when GCA was added during viral proliferation (figure 2B). Moreover, the viral titer and the percentage of infected ST cells were lower in GCA-treated cells than control cells (figure 2C, D).

Figure 2
figure 2

CSFV multiplication was inhibited by GCA. (A) The safe concentration of GCA was tested using the MTT assay. The operations and calculations were the same as those used for BFA. (B) Cellular CSFV RNA with or without 400 nM GCA were examined using real-time PCR. (C) The titers of CSFV with or without GCA were tested by IFA. (D) CSFV proliferation in ST cells with or without GCA was observed by IFA at 24, 48 and 72 hpi. Results from the three independent experiments are shown as means ± SD. *P<0.05 and **P<0.01 compared with the control group.

Full size image

3.3 GBF1 was involved in CSFV replication

Three shRNA vectors for the swine GBF1 gene and a scrambled shRNA vector were constructed and transfected into ST cells (figures 3A). After puromycin screening, the down-regulation of GBF1 in the transfected cells was tested. GBF1 was knocked down using three shRNA vectors. shRNA-3 was the most effective and was therefore selected for further analysis. Moreover, the transcription levels of the GBF1 were not significantly different between the untransfected cells and scrambled shRNA-transfected cells (figure 3B, C). The CSFV proliferation dynamics in ST cells, GBF1 shRNA -transfected cells, and scrambled shRNA-transfected cells were then evaluated. CSFV was incubated with cell monolayers at 37°C for 1 h, and unbound particles were removed with the supernatants. Infected cell monolayers were collected for CSFV RNA quantitative detection, and the supernatants were collected for viral titration at 24, 48 and 72 hpi. The detection of cellular CSFV RNA and viral titers showed that the CSFV proliferation dynamics in GBF1 shRNA-transfected cells were slower than those in ST cells and scrambled shRNA-transfected cells (figure 4A, B). CSFV proliferation in ST cells, GBF1 shRNA-transfected cells, and scrambled shRNA-transfected cells was observed by IFA. The percentage of CSFV-infected GBF1 knockdown cells was significantly lower than those of the infected ST cells and scrambled shRNA-transfected cells, especially at 48 and 72 hpi. GFP expression in shRNA-transfected cells was quenched after fixation with 4% paraformaldehyde, which had no effect on the observation of FITC on anti-pig IgG (figure 4C).

Figure 3
figure 3

Construction and selection of cells with down-regulated GBF1. (A) ST cells transfected with scrambled shRNA or GBF1 shRNA were selected by adding 5 μg/mL puromycin for about 20 days (only GBF1 shRNA-3 transfected cells are shown). (B) GBF1 RNA transcription in ST cells, scrambled shRNA, and GBF1 shRNA-transfected cells were examined by real-time PCR. (C) GBF1 protein in ST cells, scrambled shRNA, and GBF1 shRNA-transfected cells were examined by Western blotting. Results from the three independent experiments are shown as means ± SD. *P<0.05 and **P<0.01 compared with the control group.

Full size image
Figure 4
figure 4

CSFV proliferation dynamic in GBF1 knockdown cells. (A) CSFV proliferation dynamics in ST cells, scrambled shRNA-transfected cells, and GBF1 shRNA-transfected cells were examined by real-time PCR. (B) CSFV titers in ST cells, scrambled shRNA-transfected cells, and GBF1 shRNA-transfected cells were examined by IFA. (C) CSFV proliferation in ST cells, scrambled shRNA-transfected cells, and GBF1 shRNA-transfected cells observed by IFA at 24, 48 and 72 hpi. Results from the three independent experiments are shown as means ± SD. *P<0.05 and **P<0.01 compared with the control group.

Full size image

3.4 CSFV proliferation required the assistance of Rab2

shRNA vectors and transfected cells specific to the swine Rab2 gene were constructed and selected using the same methods used for the GBF1 gene. Rab2 shRNA-1 demonstrated the optimal down-regulation of expression (figure 5B, C), and the corresponding transfected cells were used for the anti-CSFV test. Cellular CSFV RNA tested by real-time PCR at three time points showed that the down-regulation of the Rab2 gene significantly slowed CSFV RNA replication. Cellular CSFV RNA in Rab2 knockdown cells declined by about 10 times relative to the levels in ST cells and scrambled shRNA-transfected cells (figure 6A). In addition, viral titration showed that the production of CSFV virions declined by about 1–2.5 titers by Rab2 knock-down (figure 6B). Furthermore, CSFV-infected Rab2 interference cells were less abundant than cells of the control groups at the three testing points (figure 6C).

Figure 5
figure 5

Construction and selection of cells with downregulated Rab2. (A) Rab2 shRNA was stably transfected to ST cells (only Rab2 shRNA-1-transfected cell are shown). (B) Relative expression of Rab2 in ST cells and scrambled shRNA- or Rab2 shRNA-transfected cells were examined using real-time PCR. (C) Rab2 protein levels in ST cells, scrambled shRNA-transfected cells, and GBF1 shRNA-transfected cells were analysed by Western blotting. Results from the three independent experiments are shown as means ± SD. *P<0.05 and **P<0.01 compared with the control group.

Full size image
Figure 6
figure 6

CSFV proliferation dynamic in Rab2 knockdown cells. (A) CSFV proliferation dynamics in ST cells, scrambled shRNA-transfected cells, and Rab2 shRNA-transfected cells were examined by real-time PCR. (B) CSFV titers in ST cells, scrambled shRNA-transfected cells, and Rab2 shRNA-transfected cells were examined by IFA. (C) CSFV proliferation in ST cells, scrambled shRNA-, and Rab2 shRNA-transfected cells observed by IFA at 24, 48 and 72 hpi. Results from the three independent experiments are shown as means ± SD. *P<0.05 and **P<0.01 compared with the control group.

Full size image

3.5 GBF1 and Rab2 overexpression facilitated CSFV proliferation

GBF1 and Rab2 were overexpressed using the lentiviral vector system. GBF1 and Rab2 expression increased at both the mRNA (figure 7B, C) and protein levels (figure 7D, E). CSFV proliferation was higher after the upregulation of GBF1 and Rab2 compared with ST cells and the mock lentivirus group (figure 8).

Figure 7
figure 7

Overexpression of GBF1 and Rab2 using a lentivirus (LV) system. (A) Fluorescent photographs of ST cells infected with lentiviruses. LV Mock: lentivirus without exogenous gene, LV GNF1: lentivirus carrying the GBF1 gene, LV Rab2: lentivirus carrying the Rab2 gene. (B–C) Overexpression of GBF1 or Rab2 was evaluated by real-time PCR. (D–E) Overexpression of GBF1 or Rab2 was evaluated by western blotting. Results from the three independent experiments are shown as means ± SD. *P<0.05 and **P<0.01 compared with the control group.

Full size image
Figure 8
figure 8

CSFV proliferation was accelerated by the overexpression of GBF1 and Rab2. (A) CSFV proliferation dynamics in ST cells and LV Mock, LV GBF1, and LV Rab2 infected cells were evaluated by real-time PCR at 24 hpi. (B) CSFV titers in ST cells and LV Mock, LV GBF1, and LV Rab2 infected cells were examined by IFA at 24 hpi. (C) CSFV proliferation in ST cells and LV Mock, LV GBF1, and LV Rab2 infected cells were observed by IFA at 24 hpi. Results from the three independent experiments are shown as means ± SD. *P<0.05 and **P<0.01 compared with the control group.

Full size image

3.6 GBF1 and Rab2 regulated CSFV RNA replication

Cellular CSFV RNA in the first 12 h was measured by real-time PCR to evaluate whether GBF1 and Rab2 regulate CSFV RNA replication. CSFV RNA replication was suppressed when GBF1 and Rab2 were disturbed by inhibitors or shRNA (figure 9A, B). Conversely, overexpression of GBF1 and Rab2 could accelerate CSFV RNA replication (figure 9C).

Figure 9
figure 9

GBF1 and Rab2 regulated CSFV RNA replication. (A) CSFV replication with or without BFA and GCA was tested by real-time PCR at 3, 6, 9 and 12 hpi. (B) CSFV replication in ST cells, scrambled shRNA-transfected cells, GBF1 shRNA-transfected cells, and Rab2 shRNA-transfected cells were examined by real-time PCR at 3, 6, 9 and 12 hpi. (C) CSFV replication in ST cells and LV Mock, LV GBF1 and LV Rab2 infected cells were evaluated by real-time PCR at 3, 6, 9 and 12 hpi. Results from three independent experiments are shown as means ± SD. *P<0.05 and **P<0.01 compared with the control group.

Full size image

4 Discussion

The functions of CSFV non-structural proteins have been studied extensively in recent years (Lamp et al. 2011; Sheng et al. 2012); nevertheless, the CSFV replication mechanism and the subcellular location of the replication complex remain unclear. Studies on the replication membranes of positive-stranded viruses demonstrated that they can induce the alteration of intracellular membranes to facilitate the replication cycle (Miller and Krijnse-Locker 2008). CSFV NS2 and NS5A are located in the ER (Tang et al. 2010; He et al. 2012). NS5B, which is an RNA-dependent RNA polymerase of CSFV, is located in the cytoplasm (Zhao et al. 2006). The locations of these CSFV replication-associated proteins suggest that the complex is located in the cytoplasm, and the ER may support CSFV replication. However, it is not clear whether other intracytoplasmic membrane structures and their resident proteins are involved in CSFV replication.

In this study, we found that CSFV replication is sensitive to BFA. Both CSFV RNA replication and virus particle generation were inhibited by BFA treatment. BFA inhibited protein transport between the ER and the Golgi by suppressing the formation of COP I-mediated transport vesicles. The proliferation of viruses, such as hepatitis C virus (HCV) (Goueslain et al. 2010) and dengue virus (Carpp et al. 2014), is sensitive to BFA treatment owing to the inhibitory effect of BFA on GBF1 and the consequent inhibition of COP I vesicle formation. However, three BFA-sensitive large Arf-GEFs (GBF1, BIG1 and BIG2) are present in mammalian cells (Li et al. 2014). Therefore, GCA and RNA interference were used for further tests. GCA is a highly specific inhibitor of GBF1 and had the same inhibitory effect on CSFV proliferation as BFA. Furthermore, CSFV replication in GBF1 shRNA-transfected cells was significantly inhibited compared with that in GBF1-undisturbed cells. Conversely, CSFV multiplication was increased in cells overexpressing GBF1. Therefore, GBF1 participated in CSFV multiplication.

GBF1 participates in the retrograde Golgi-to-ER transportation by activating the ADP-ribosylation factor (ARF) to regulate cytosolic COP I recruitment (Zhao et al. 2006; Lowery et al. 2013). GBF1 is involved in the multiplication of many viruses, such as severe acute respiratory syndrome coronavirus (SARS-CoV) (de Wilde et al. 2015), dengue virus (Carpp et al. 2014) and influenza virus (Watanabe et al. 2014). One of the mechanisms by which GBF1 regulates viral multiplication involves the GBF1-ARF1-COP I pathway. HCV replication is suppressed when the functions of GBF1 and ARF1 are disturbed (Goueslain et al. 2010; Matto et al. 2011). Dengue virus uses the GBF1-ARF1/ARF4-COP I pathway to form capsids around lipid droplets (Iglesias et al. 2015). GBF1-mediated ARF1 activation is crucial for mouse hepatitis coronavirus (MHV) RNA replication (Verheije et al. 2008), Ebola virus virion formation (Yamayoshi et al. 2010), and diverse families of negative-stranded RNA viruses (Panda et al. 2011). Other metabolic pathways are also regulated by GBF1 and are involved in virus proliferation. GBF1 simultaneously regulates ARF4 and ARF5 to modulate HCV replication (Farhat et al. 2016). GBF1 has other functions in HCV replication in BFA-resistant hepatoma-derived cells (Farhat et al. 2012). Here, we discovered that both CSFV RNA replication and virus particle formation were inhibited by the disturbance of GBF1 function. However, the CSFV proliferation dynamics in ARF1 shRNA-transfected cells were not different from those in control cells (data not shown). Therefore, GBF1 may regulate some metabolic pathways without ARF1 to modulate CSFV proliferation. The detailed function of GBF1 in CSFV proliferation requires further investigation.

Rab2 is a component of pre-Golgi intermediates that mainly regulates retrograde trafficking from the ERGIC (ER-Golgi intermediate compartment) to the ER (Ortiz Sandoval and Simmen 2012). Moreover, Rab2 is essential for Golgi organization (Haas et al. 2007) and can regulate COP I-mediated retrograde transport to intervene with the release of HCV particles (Raj 1961). In the present study, CSFV replication was suppressed when Rab2 was down-regulated. The overexpression of Rab2 can accelerate CSFV proliferation. The mechanisms by which Rab2 influences CSFV proliferation must be further investigated.

As the cargo sorting station within the cell, the Golgi apparatus participates in protein and vesicular transport (Mironov et al. 2013; Guo et al. 2014). Additionally, the secretory pathways involved in the Golgi apparatus are always utilized by viruses for egress on infected cells. Transport from early endosomes to the trans-Golgi network is important for vaccinia virus membrane wrapping and egress (Sivan et al. 2016). Human herpesvirus-6 (HHV-6) can egress via the exosomal release pathway by budding at trans-Golgi network vacuoles and inducing multivesicular body formation (Mori et al. 2008). Furthermore, the hepatitis E virus and HCV also utilize the exosome release pathway for egress (Tamai et al. 2011; Nagashima et al. 2014). However, viral egress is a complicated process that involves many pathways and related proteins. Reports have shown that HCV can be released from Huh7 cells via a trans-Golgi network-to-endosome pathway (Jamel et al. 2016). Additionally, HCV particles are found in endosomal compartments and released via a non-canonical secretory route that is unrelated to the Golgi-associated secretory route and may be an undescribed cellular secretion route (Bayer et al. 2016). Viral egress has a close connection with the virus particle formation process. CSFV is a single-stranded positive-sense RNA virus that has no cytopathic effect in infected cells. The CSFV assembly process is still unclear and mysterious because its core protein is dispensable for CSFV assembly (Riedel et al. 2012). Additional experiments are urgently needed to determine the mechanisms of CSFV assembly and egress.

In summary, we demonstrated that GBF1 and Rab2 play important roles in CSFV proliferation. Golgi and pre-Golgi intermediates are also important in CSFV replication. Nevertheless, the effect of COP I-mediated retrograde trafficking on CSFV replication must be further investigated.