Crimean-Congo hemorrhagic fever virus entry and replication is clathrin-, pH- and cholesterol-dependent

Crimean-Congo hemorrhagic fever virus (CCHFV) (genus Nairovirus, family Bunyaviridae) is a human pathogen that causes severe disease with high mortality (3–30 %) (Ergönül, 2006). The infection, acquired by tick bite, through contact with infected tissue/blood or nosocomially, is manifested by symptoms such as fever, prostration and severe haemorrhage (Ergönül, 2006; Hoogstraal, 1979). At present, CCHFV is the most widespread tick-borne pathogen and its occurrence correlates well with the widespread geographical distribution of its tick vector (genus Hyalomma) (Ergönül, 2006). Although CCHFV exists endemically in limited parts of the world, there is a concern that future climate changes will bring the tick vectors across their current borders, increasing the pool of susceptible humans and animals. Due to limited research facilities and lack of animal models, the progress of detailed investigation on CCHFV is slow, yet necessary.

All viruses are obligate intracellular organisms that must find their way into target cells for successful replication. To enter cells, viruses bind to a variety of host cell receptors, which determine the entry mechanism, host range and pathogenesis of each virus (Marsh & Helenius, 2006). Following binding, viruses gain access to the intracellular compartment by fusing with the plasma membrane or by internalization in endosomes (Sieczkarski & Whittaker, 2002). Several viruses use the benefits of endocytosis since it provides a protected environment and permits intracellular movement (Marsh & Helenius, 2006). The most well-studied and commonly used cellular endocytosis pathway, also used by many viruses, is clathrin-dependent endocytosis (Marsh & Helenius, 2006).

The receptor for CCHFV is presently unknown; however, hantaviruses (genus Hantavirus, family Bunyaviridae) bind to host cell integrins before they are internalized by clathrin-dependent endocytosis (Gavrilovskaya et al., 1998; Jin et al., 2002). Because no such investigations have been conducted on CCHFV or other nairoviruses, we investigated the role of clathrin, caveolin-1 and cholesterol in the CCHFV entry and replication cycle.

Vero E6 cells were pretreated with sucrose or chlorpromazine (CPZ) for 30–60 min at 37 °C, before infection with an m.o.i. of 1 of CCHFV, severe acute respiratory syndrome coronavirus (SARS-CoV) or simian virus 40 (SV40) for 6 h, in the presence of these drugs (Fig. 1a⇓). CPZ and sucrose inhibit the clathrin-dependent pathway by reducing the number of coated-pit-associated receptors at the cell surface (Heuser & Anderson, 1989; Wang et al., 1993). After 6 h, cells were washed and infection was maintained in fresh medium until 24 h post-infection (p.i.) before cells were harvested by lysis. Lysates were resolved in a tris/glycine polyacrylamide gel, transferred onto a PVDF membrane and probed with primary antibodies [CCHFV nucleocapsid protein (NP) (Andersson et al., 2004), SARS-CoV NP (Invitrogen), SV40-T-Ag (BD-Biosciences), calnexin (Simon et al., 2006), clathrin or caveolin-1 antibodies (Santa-Cruz)] and horseradish peroxidase-conjugated secondary antibodies, before proteins were detected with the enhanced chemiluminescent plus Western blotting detection kit. Results showed a decrease in CCHFV NP levels in response to sucrose and CPZ (Fig. 1a⇓). The reduction in NP was not due to the toxic effect of sucrose and CPZ, since cellular viability and protein synthesis were not affected (measured using an MTT assay and ³⁵S-labelled proteins; data not shown). Consistent with clathrin-dependent entry, SARS-CoV NP levels were also reduced while SV40-T-Ag, entering the cell in a non-clathrin-dependent manner, was unaffected (Fig. 1a⇓) (Damm et al., 2005; Inoue et al., 2007).

CCHFV clathrin-mediated entry was further analysed using RNA interference and clathrin knockdown. First, Hela, Vero E6 and Vero cells were transfected with control or clathrin small interfering RNA (siRNA) (Invitrogen). The next day, cells were infected and maintained for 24 h before cultures were subjected to Western blot and RNA quantification (Fig. 1b, c⇑). RNA was purified, reverse transcribed and subjected to quantification by real-time PCR (qPCR) using gene-specific primers for CCHFV S-segment and clathrin. RNA in control cells was set to 1 and relative RNA levels were calculated from the cycle threshold (C_t) value of each sample according to the ΔΔC_t method (Applied Biosystems). Mock-infected samples were included in all qPCR assays and were always negative for CCHFV. As determined by Western blot and qPCR, clathrin was successfully knocked down in all three cell lines. As shown in Fig. 1⇑, we could clearly demonstrate a significant reduction of CCHFV RNA and NP levels in cells transfected with clathrin siRNA (Fig. 1b, c⇑). It should be mentioned that in Vero E6 cells, control siRNA reduced the NP levels to a similar degree as clathrin siRNA (Fig. 1b⇑). At this stage we have no explanation for this observation; however, while NP was reduced (Fig. 1b⇑), S-segment RNA was not affected by control siRNA (Fig. 1c, d⇑), suggesting that siRNA-mediated inhibition might be post-transcriptional, occurring after CCHFV entry. It is possible that protein translation could be inhibited by the non-specific activation of protein kinase R by control siRNA in Vero-E6 cells.

To further demonstrate the effects of clathrin knockdown on CCHFV entry and early infection, transfected cells were infected for 6 h (Fig. 1d⇑). Because CCHFV NP and progeny virus could not be detected, and since CCHFV RNA increased 10-fold in control cells within this time period (data not shown), infection was determined by relative RNA levels. Similar to 24 h infections, CCHFV RNA was significantly reduced in 6 h cultures (Fig. 1d⇑). The reduction in CCHFV RNA levels 6 h p.i. was less than at 24 h p.i. in cells transfected with clathrin siRNA, which is most likely due to additional rounds of the CCHFV replication cycle at 24 h p.i.

In addition to clathrin-dependent endocytosis, there are several clathrin-independent entry pathways. One such pathway is caveolae endocytosis, a subgroup of lipid-raft-dependent endocytosis that is stabilized by the oligomerization of caveolin-1, the major constituent of caveolae (Sandvig et al., 2008). Depletion of caveolin-1 abolishes the formation of caveolae and can be used to specifically address caveolae endocytosis.

To investigate the role of caveolae in CCHFV entry, Hela, Vero E6 and Vero cells were transfected with caveolin-1 siRNA (Invitrogen) and infected with CCHFV, as described above for clathrin. Despite the successful reduction of caveolin-1 in HeLa and Vero cells, no decrease was observed in CCHVF NP levels, suggesting that caveolae endocytosis is not an entry route employed by CCHFV (Fig. 1e⇑). Moreover, we could not see any differences in progeny virus titres compared to the control when supernatants were collected from infected cells transfected with two different caveolin-1 siRNAs (data not shown). Also, we could not detect caveolin-1 in Vero E6 cells by Western blot analysis.

Caveolae endocytosis is currently questioned as a portal for virus entry. It appears that caveolae are rather immobile structures in the plasma membrane, stabilized by caveolin-1 that prevents, rather than promotes, caveolae from pinching off (Le et al., 2002; Thomsen et al., 2002). In line with this, cholera toxin and SV40, which has been shown to use caveolae endocytosis, also enter target cells in a caveolae-independent manner (Damm et al., 2005; Torgersen et al., 2001). It is therefore proposed that, in cells without detectable caveolae, there might be an internalization pathway similar to caveolae endocytosis, except that it is constitutive and independent of caveolin-1 (Damm et al., 2005). Whether this pathway could be employed by CCHFV as well remains to be investigated.

Observing the reducing effects of clathrin inhibitors and clathrin siRNA prompted us to further investigate clathrin-dependent endocytosis. Following scission from the plasma membrane, vesicles fuse with endosomes where acid-triggered processes induce the necessary conformational changes that lead to viral escape (Marsh & Helenius, 2006). To investigate a potential pH-dependent step in CCHFV infection, Vero E6 cells were pretreated with increasing concentrations of bafilomycin A1 for 30–60 min and infected with an m.o.i of 1 of CCHFV in the presence of bafilomycin A1 before cells and supernatants were harvested 24 h p.i. Results showed that there was a dose-dependent reduction in progeny virus titres and NP levels (data not shown). Alternatively, cells were pretreated with 50 nM bafilomycin A1 or 25 mM NH₄Cl and infected with CCHFV, SARS-CoV or SV40 in the presence of these drugs for 6 h (Fig. 2a⇓). Cultures were harvested 24 h p.i. for Western blot analysis. Results showed that interference with endosome acidification strongly reduced CCHFV NP levels (Fig. 2a⇓). The reduction was not due to the adverse effects of the inhibitors bafilomycin A1 and NH₄Cl, since cell viability and overall protein synthesis were not affected. The specificity of the inhibitors was confirmed using SARS-CoV and SV40, viruses that are known to use pH-dependent and pH-independent entry pathways, respectively (Fig. 2a⇓) (Inoue et al., 2007; Pelkmans et al., 2001).

Considering the early role of endosomes in viral infections, we were interested in whether bafilomycin A1 or NH₄Cl could interfere with CCHFV endosomal escape. To investigate binding, Vero E6 cells were pretreated with bafilomycin A1 and NH₄Cl as described above and incubated with CCHFV for 1 h at 4 °C in the presence of the drugs (Fig. 2b⇑, 0 h). After 1 h, cells were harvested and subjected to qPCR analysis, as described above. In agreement with the notion that pH-dependent steps occur after vesicle internalization, we observed no impairment in CCHFV binding in response to either bafilomycin A1 or NH₄Cl compared to the control. Next, we investigated CCHFV infection in response to the inhibitors during the first 6 h of infection. Vero E6 cells were pretreated as described above and infected for 6 h in the presence of bafilomycin A1 and NH₄Cl (Fig. 2b⇑, −1 h). Alternatively, cells were first infected and treated 1.5 h p.i., and were analysed by qPCR at 6 h p.i. (Fig. 2b⇑, +1.5 h). We found that bafilomycin A1 and NH₄Cl significantly reduced CCHFV RNA levels when they were present during the very early events in the replication cycle (up to 1.5 h p.i.), but not following their addition at 1.5 h p.i. (Fig. 2b⇑). This suggests that CCHFV has a pH-dependent step within the 1.5 h following virus binding. Although we have not investigated the co-localization of CCHFV virus particles with early endosomes, our results imply that CCHFV uses pH-dependent endocytosis. There is evidence that other bunyaviruses also require low pH for productive infection (Gonzalez-Scarano et al., 1984; Hacker & Hardy, 1997; Pekosz & Gonzalez-Scarano, 1996; Whitfield et al., 2005).

As demonstrated previously, caveolae endocytosis, a subgroup within lipid-raft-dependent endocytosis pathways, is unlikely to constitute the major entry pathway for CCHFV (Fig. 1e⇑). In general, lipid rafts are membrane microdomains that are rich in cholesterol, sphingolipids and various signalling proteins and receptors (Simons & Toomre, 2000). In terms of viral entry, lipid rafts are commonly used as entry portals (Campbell et al., 2001; Chung et al., 2005; Damm et al., 2005; Lee et al., 2008; Lu et al., 2008; Medigeshi et al., 2008; Rojek et al., 2008; Wang et al., 2008). Since no bunyaviruses, including CCHFV, have been investigated for their requirement for lipid rafts during entry or replication, we examined the role of cholesterol in CCHFV infection by using the cholesterol-depleting drug methyl-β-cyclodextrin (MβCD). Vero E6 cells were pretreated with 1, 5 or 10 mM MβCD for 30–45 min at 37 °C before infection with an m.o.i of 1 of CCHFV in the absence of MβCD. Alternatively, cells were pretreated with 5 mM MβCD and replenished with 100 or 200 μM exogenous cholesterol for 30–45 min at 37 °C before infection. After 24 h, cells were harvested or supernatants were collected for Western blot or progeny virus determination, respectively (Fig. 3a, b⇓). We found that 5 mM MβCD reduced the cell cholesterol content to 60 % of control cells within 30 min and it remained at similar levels for the next 1.5 h (data not shown).

Results showed that MβCD treatment reduced CCHFV NP and progeny virus levels in a dose-dependent but reversible manner, since both protein and progeny virus levels were reversed when exogenous cholesterol was added (Fig. 3a, b⇑). These results were not due to the toxic effect of MβCD or cholesterol, as determined by cell viability assay and protein labelling. However, because cholesterol is also important in virus assembly and budding, our results could reflect some late and cholesterol-dependent events in the life cycle of CCHFV. In agreement with cholesterol-dependent entry, SARS-CoV and SV40 protein levels were also reduced when cells were pretreated with 5 mM MβCD but not when cells were depleted and replenished with cholesterol before infection (data not shown).

To investigate CCHFV early RNA expression and thereby determine entry efficiency in response to MβCD treatment, Vero E6 cells were depleted of cholesterol and subjected to binding assay as described above. In parallel, cells were either depleted of cholesterol or depleted and replenished with cholesterol before infection for 6 h in the presence of MβCD or cholesterol (Fig. 3c⇑, −0.5 h and −0.5 h*). Alternatively, cells were infected and incubated with MβCD between 1 and 6 h p.i. (Fig. 3c⇑, +1 h). RNA was purified and relative levels were determined as described above. We found that while MβCD treatment did not impair CCHFV binding, CCHFV RNA levels were significantly lower in cells depleted of cholesterol but not in cells replenished with cholesterol, suggesting that cholesterol is important early in CCHFV infection (Fig. 3c⇑). Moreover, it seemed that cholesterol was not important for CCHFV internalization either, since cholesterol depletion 1 h p.i. resulted in RNA levels similar to pretreatment (Fig. 3c⇑, −0.5 h versus +1 h), implying that cholesterol is required in events following CCHFV binding and internalization. One possible explanation could be viral entrapment in endosomes. For adenoviruses, it was shown that cholesterol depletion trapped virus particles within endosomes, resulting in poor infection (Imelli et al., 2004). Alternatively, cholesterol depletion may interfere with clathrin-dependent endocytosis (Rodal et al., 1999; Subtil et al., 1999).

Taken together, the results presented here demonstrate that CCHFV entry is mediated by clathrin- but not caveolin-1-dependent pathways. Clathrin-dependent entry was further confirmed using chemical inhibitors of clathrin and endosome acidification. We also hypothesize that cholesterol is involved in the early events of the CCHFV replication cycle. Although this study provides new data on CCHFV entry, further investigations are needed to characterize CCHFV entry in detail, with special focus on antiviral development.