Statins Suppress Ebola Virus Infectivity by Interfering with Glycoprotein Processing

ABSTRACT Ebola virus (EBOV) infection is a major public health concern due to high fatality rates and limited effective treatments. Statins, widely used cholesterol-lowering drugs, have pleiotropic mechanisms of action and were suggested as potential adjunct therapy for Ebola virus disease (EVD) during the 2013–2016 outbreak in West Africa. Here, we evaluated the antiviral effects of statin (lovastatin) on EBOV infection in vitro. Statin treatment decreased infectious EBOV production in primary human monocyte-derived macrophages and in the hepatic cell line Huh7. Statin treatment did not interfere with viral entry, but the viral particles released from treated cells showed reduced infectivity due to inhibition of viral glycoprotein processing, as evidenced by decreased ratios of the mature glycoprotein form to precursor form. Statin-induced inhibition of infectious virus production and glycoprotein processing was reversed by exogenous mevalonate, the rate-limiting product of the cholesterol biosynthesis pathway, but not by low-density lipoprotein. Finally, statin-treated cells produced EBOV particles devoid of the surface glycoproteins required for virus infectivity. Our findings demonstrate that statin treatment inhibits EBOV infection and suggest that the efficacy of statin treatment should be evaluated in appropriate animal models of EVD.

abnormalities (5,6). For example, statins have been implicated in improving survival in sepsis patients (7)(8)(9); like EVD, sepsis is characterized by inflammation, endothelial dysfunction, and coagulopathy (6). Statins are already FDA approved for reducing high cholesterol, have a favorable safety profile, and are inexpensive. Thus, they were suggested as a possible adjunct therapy for EVD patients during the 2013-2016 outbreak (10).
EBOV virions project glycoprotein (GP) spikes that are synthesized and inserted into the host cell-derived envelope during budding (33). EBOV GP is synthesized in several forms. The most abundant form of GP is a secreted protein (sGP) translated from an unedited mRNA, whereas the structural GP is a product of the edited mRNA. The monomeric EBOV GP, a type I transmembrane glycoprotein, is processed by a complex series of events (34)(35)(36). An N-glycosylated, endoplasmic reticulum (ER)-resident form of GP precursor (preGP) undergoes N,O-glycosylation maturation in the Golgi apparatus to become GP 0 (36). GP 0 is then transported further into the trans-Golgi network, where the proprotein convertase furin or a furin-like protease cleaves GP 0 at a multibasic motif that is conserved in all EBOV strains (36). Cleavage results in the mature N,Oglycosylated GP 1 and in the GP 2 subunit linked by disulfide bond (34)(35)(36)(37)(38). These subunits interact to form GP 1,2 , present on virions as trimeric spikes; GP 1 mediates receptor binding while GP 2 is critical for fusion of the EBOV envelope with the endosomal/lysosomal membrane (39,40). However, unlike other viruses, cleavage of GP 0 by furin is not required for fusion (41,42) or glycoprotein incorporation into virions (43)(44)(45)(46).
Here, we report that a statin (lovastatin) suppresses infectious EBOV production in a human hepatoma cell line (Huh7) and in primary monocyte-derived macrophages, cell types that are in vivo targets for EBOV replication. Statin treatment inhibited processing of preGP into GP 1 in EBOV-infected cells or cells transfected with plasmids encoding GP 1,2 ; the effect was reversed by adding mevalonate. EBOV particles produced in statin-treated cells were depleted of the essential glycoprotein subunit GP 1 required for virus entry, suggesting that statins reduce EBOV infectivity by inhibiting glycoprotein maturation and incorporation into virions. In addition, we have tested the effect of 5 other types of statins, fluvastatin, simvastatin, atorvastatin, rosuvastatin, and pitavastatin, on EBOV replication. Of all the statins, simvastatin and pitavastatin were the most potent in reducing EBOV infectivity. Our results suggest that statins selectively inhibit preGP maturation and should be further investigated in in vivo models for EBOV infection.

Statin treatment inhibits EBOV infection.
To test if statins affect EBOV replication, Huh7 cells were infected with the EBOV variant Mayinga (Ebola virus/H. sapiens-tc/ COD/1976/Yambuku-Mayinga) at a multiplicity of infection (MOI) of 0.05. After 1 h of virus adsorption, the cells were treated with dimethyl sulfoxide (DMSO) (vehicle control) or with 20 M or 50 M lovastatin (referred to as "statin" here unless stated otherwise), the first clinically approved statin, in medium supplemented with lipoproteindeficient serum (LPDS). LPDS eliminates the possible uptake of cholesterol from the medium (47). After 72 h postinfection (hpi), cells were fixed and viral antigen expression was evaluated by immunofluorescence assays using polyclonal anti-EBOV serum. As shown in Fig. 1A, EBOV antigen-positive staining was seen throughout infected Huh7 cells treated with DMSO only. However, EBOV-positive staining was reduced compared to controls in cells treated with statin at either concentration. To ensure that statinmediated reduction in EBOV-positive staining was not due to cytotoxicity, cell viability was assayed after 72 h of treatment. Cell viability was unaffected by either concentration of statin (Fig. 1C). These results suggest that statin reduced EBOV infection.
To determine if statin treatment can inhibit infectious EBOV production, we examined viral titers in supernatants of infected cells. High titers of infectious virus (1.5 ϫ 10 7 /ml) were detected at 72 hpi in vehicle control-treated cell culture supernatants supplemented with LPDS. Treatment with statin under the same cell culture conditions reduced EBOV titers; 20 M statin decreased the production of infectious EBOV titers by Ͼ1.1 log, and 50 M decreased EBOV titers by 1.5 log (Fig. 1B). In contrast, statin treatment under similar conditions did not affect titers of adenovirus type 5, a nonenveloped virus (see Fig. S1 in the supplemental material).
The antiviral potency of statin treatment was also evaluated in primary human monocyte-derived macrophages, since these cells represent a major target of EBOV infection. To account for donor variations, cells from 4 different donors were tested. Cell viability of macrophages treated with 10 M statin was Ͼ80% for all the donors (Fig. 1E). Untreated cells yielded infectious titers ranging between 5 ϫ 10 5 and 4 ϫ 10 6 (Fig. 1D). Statin treatment efficiently reduced EBOV titers in macrophages from each donor; 2.5 M statin reduced infectious EBOV titers by 0.5 to 1.0 log, and 10 M statin reduced EBOV titers by 1 to 2 log (Fig. 1D).
Statin inhibition of EBOV infection is reversed by exogenous mevalonate but not by LDL. Statin blocks mevalonate generation and subsequent cholesterol biosynthesis by competitively inhibiting HMG-CoA reductase (11). To investigate whether the anti-EBOV effect of statin was dependent on its ability to specifically inhibit mevalonate production, we added mevalonate during statin treatment. Since inhibition of cholesterol synthesis can be compensated for by import of low-density lipoprotein (LDL)derived cholesterol from outside the cells, we also looked at the effects of LDL supplementation during statin treatment. Huh7 cells were infected with EBOV as described above and then treated with statin with or without the indicated concentrations of LDL or mevalonate. As shown in Fig. 2A, addition of mevalonate reversed statin-mediated reduction in EBOV titers in a dose-dependent manner, while mevalonate alone had no effect on viral titers. Expression of viral glycoprotein GP 1 /preGP and nucleoprotein (NP) was also restored when mevalonate was added during statin treatment (Fig. 2B). In contrast to mevalonate, adding LDL did not reverse the effects of statin treatment on viral titers (Fig. 2C). Altogether, we showed that inhibition of EBOV infection by statin treatment was reversed by mevalonate, the immediate downstream product of the reaction catalyzed by HMG-CoA reductase. These findings are consistent with statin reducing EBOV infectivity by inhibiting HMG-CoA reductase and not via off-target effects.
Statin treatment does not affect EBOV entry. To assess whether reduced production of infectious virus was due to the inhibition of EBOV entry into cells, we measured the levels of cell-associated EBOV NP RNA at 3 hpi. Huh7 cells pretreated with statin were infected with EBOV at an MOI of 3 to ensure synchronous infection. Levels of viral NP RNA in the lysed cells were measured by quantitative reverse transcription PCR (qRT-PCR) and normalized to cellular glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA levels. As shown in Fig. 3A, NP RNA levels did not significantly differ among the samples, indicating that statin did not affect the levels of internalized EBOV genome. On the other hand, treatment with the positive control U18666A, a Niemann-Pick C1 protein (NPC1) inhibitor previously shown to prevent EBOV glycoproteindependent entry (25), reduced NP RNA copy numbers in a dose-dependent manner. These results are consistent with statin not affecting EBOV entry.
Statin treatment impairs EBOV infectivity. To further examine the mechanism by which statin reduces the levels of EBOV released from infected cells, the specific infectivity of virions made in the presence of statin was determined by comparing 50% tissue culture infective doses (TCID 50 ) with copy numbers of viral RNA released into culture supernatants of infected cells at 48 hpi. To maximize the initial number of infected cells, Huh7 cells were infected with EBOV at an MOI of 2.0; MOIs higher than 2.0 caused a significant drop in cell viability at 48 hpi (data not shown). Infected cells were then treated either with DMSO (vehicle control) or with 20 M or 50 M statin. After 48 h, no differences were noted in extracellular EBOV RNA copy numbers released from cells treated with statin or with vehicle control (Fig. 3B). In contrast, yield of infectious virus (measured as TCID 50 per milliliter) in supernatants from statin-treated cells was approximately~4 to 9 times lower than that in DMSO-treated samples (Fig. 3B). The ratio of TCID 50 per milliliter to viral RNA copy numbers (Fig. 3C) in statin-treated cells was reduced to~5 to 10% of controls. Taken together, these data indicate that postinfection statin treatment decreased infectivity of newly synthesized EBOV particles.
Statin treatment inhibits EBOV glycoprotein maturation. To explore the mechanism responsible for reduced particle infectivity in statin-treated cells, we examined the impact of statins on the viral proteins involved in virus assembly and budding: the matrix protein VP40 and the envelope glycoprotein GP 1,2 . We first determined whether statin treatment affected VP40 expression. Huh7 cells were transfected with plasmid expressing VP40 and then treated with statin or vehicle control; cell lysates were analyzed by Western blotting. As shown in Fig. S2, VP40 expression levels were similar in statin-treated and vehicle-treated cells. We then examined GP 1 expression in cells transfected with plasmid expressing EBOV GP 1,2 and treated with statin. Two forms of GP were detectable: a 110-kDa form sensitive to both endoglycosidase H (endo H) and peptide-N-glycosidase F (PNGase F) digestion that had previously been identified as the N-glycosylated precursor present in the endoplasmic reticulum (designated preGP) and the 140-kDa form that was sensitive only to PNGase F digestion and was identified as the mature GP 1 (Fig. 4B ), consistent with the presence of complex N-and O-glycans   (36). Most of the GP synthesized in statin-treated cells was the immature precursor glycoprotein (preGP) containing high-mannose sugar chains sensitive to endo H treatment (compare Fig. 4A with B). In contrast, the expression of mature N,O-glycosylated glycoprotein (GP 1 ), which was resistant to endo H treatment (compare Fig. 4A with B), decreased upon statin treatment. Statin did not similarly affect the glycosylation pattern of NPC1 (Fig. S3), suggesting that the observed effect on GP 1 was specific.
We next evaluated the effect of mevalonate on EBOV preGP maturation. Cells were treated with increasing concentrations of mevalonate, and GP 1 and preGP expression levels were determined. As shown in Fig. 4C, increasing mevalonate concentrations at least partially reversed statin-mediated inhibition of mature GP 1 expression. In parallel, a decrease in immature preGP expression was observed. Densitometry analysis indicated an increase in the GP 1 /(GP 1 ϩ preGP) ratio, indicating restoration of preGP maturation to GP 1 . This ratio did not change in cells treated only with mevalonate. Partial rescue of preGP processing efficiency and viral titers ( Fig. 2A) by mevalonate in statin-treated cells is consistent with statin reducing viral titers by a mechanism impeding preGP maturation.
Statin treatment inhibits preGP glycan maturation. EBOV GP 1,2 is cleaved into subunits GP 1 and GP 2 by the proprotein convertase furin at the RRTRR 501 site. To investigate whether statin inhibits maturation of preGP glycan, we investigated the effect of statin on mutant EBOV GP 1,2 resistant to furin cleavage. To generate this mutant, we replaced the RRTRR cleavage site with AGTAA, as described previously (44,45). Huh7 cells were transfected with plasmids encoding wild-type or furin-resistant mutant EBOV GP 1,2 and treated with either statin or furin-like protease inhibitor (proprotein convertase inhibitor). GP 1,2 levels were determined in cell lysates by Western blotting. Unlike wild-type EBOV GP 1,2 , which was processed into GP 1 , mutant GP 1,2 was observed as a higher-molecular-weight form of preGP consistent with the Statin Inhibition of Ebola Virus ® molecular weight of GP 0 (36) (Fig. 4D). Treating mutant EBOV GP 1,2 -transfected cells with statin resulted in a dose-dependent decrease in GP 0 level compared to vehicle control, similar to samples with wild-type GP 1,2 . No changes in preGP levels were detected in cells expressing either wild-type or furin-cleavage-resistant EBOV GP 1,2 . In addition, the GP 0 /(GP 0 ϩ preGP) ratio in cells expressing the mutant EBOV GP 1,2 was similar to that in cells expressing wild-type GP 1,2 , indicating that statin treatment affected GP 1 glycan maturation independently of GP 0 cleavage by furin.
Statin treatment results in GP 1 -deficient virions. To confirm the effect of statin treatment on EBOV GP maturation, we infected Huh7 cells with EBOV, treated the cells with DMSO or statin, and partially purified the viral particles released from infected cells. GP 1,2 and VP40 expression in cell lysates and corresponding viral particles from statin-treated cells was analyzed by Western blotting. As shown in Fig. 5, blotting cell lysate samples from EBOV-infected, vehicle-treated cells showed bands of preGP (~110 kDa) and GP 1 (~140 kDa), as expected. Treatment with statin resulted in a selective decrease in GP 1 levels compared to preGP and VP40. These results are consistent with statin inhibiting preGP maturation of GP 1 , as was observed in GP 1,2transfected cells.
Analysis of viral particles from supernatants showed GP 1 migrating at~140 kDa in the pelleted virions, and no uncleaved preGP or GP 0 was detected (Fig. 5). In cells treated with statin, GP 1 levels decreased in a dose-dependent manner while VP40 levels were unchanged. These results are consistent with the concept that statin treatment resulted in production of VP40-containing EBOV particles deficient in GP 1 . Taken together, these results indicate that the lowered expression of GP 1 in statin-treated, EBOV-infected cells results in reduced incorporation of GP 1 into EBOV particles, leading to lower infectivity.
Multiple statins show antiviral activity against EBOV. To investigate the efficacy of other commonly prescribed statins, we compared the antiviral activities of lovastatin, fluvastatin, simvastatin, atorvastatin, rosuvastatin, and pitavastatin in Huh7 cells. As shown in Fig. 6A, at all doses tested, simvastatin and pitavastatin reduced EBOV infectious particle production most potently: 50 M simvastatin or pitavastatin reduced viral titers by 2.5 log. Rosuvastatin inhibited EBOV production least effectively, reducing viral titers by~1 log, while lovastatin, atorvastatin, and fluvastatin were moderately effective, reducing titers by 1.7, 1.5, and 1.4 log, respectively. The viability of untreated cells was similar to that of cells treated with each statin (Fig. 6B).

DISCUSSION
Statins, well-known cholesterol-lowering drugs, have been proposed as therapeutic agents against certain viruses (16-18, 21, 22, 48-59). Statins are known for their anti-inflammatory and immunomodulatory effects as well as for preserving endothelial integrity; inflammation, immune system dysregulation, and endothelial dysfunction are major contributors of EVD pathogenesis (3,(60)(61)(62). Statin use was suggested as an adjunct therapy for EVD during the 2013-2016 outbreak (63). A clinical trial evaluating atorvastatin for use in EVD was registered with clinicaltrials.gov (NCT02380625), but the trial was never initiated, presumably because the outbreak was waning prior to the scheduled study start date. Here, we provide evidence of the antiviral effects of statin treatment in a human liver cell line and in primary human macrophages, both major target cells of EBOV. The antiviral activity of statin in Huh7 cells was due to loss in particle infectivity rather than inhibition of viral entry (Fig. 3). Statin reduced the levels of GP 1 , the envelope glycoprotein responsible for receptor binding and entry into cells, in GP 1,2 -transfected (Fig. 4) and EBOV-infected (Fig. 5) cells. Finally, we found that virus particles produced in statin-treated cells had lower levels of GP 1 relative to VP40 matrix protein than did control cells. Thus, statin's antiviral activity was due to its interference in GP 1 maturation, leading to production of EBOV particles with impaired infectivity.
While GP 1 levels were reduced in infected cells and in released EBOV particles, VP40 levels were not (see Fig. S2 in the supplemental material). Similarly, qRT-PCR analysis showed no changes in the extracellular levels of EBOV RNA after statin treatment, while TCID 50 values were reduced upon statin treatment (Fig. 3). Interaction of VP40 with minigenome RNA has been reported to be sufficient for packaging RNA into virus-like  (64). Our observation that the levels of extracellular VP40 and genomic RNA did not change even in the presence of little GP 1 is consistent with this report (64) and supports the idea that an interaction between the ribonucleoprotein components and VP40 is a critical step for the budding and release of viral particles (65).
Despite the abundant preGP present in cells transfected with furin-cleavageresistant EBOV GP 1,2 , statin treatment resulted in decreased GP 0 levels similar to those seen in cells transfected with wild-type EBOV GP 1,2 (Fig. 4D). This indicates that statin affects a step prior to GP 0 cleavage, possibly by blocking transport of preGP out of the ER (36,38). The observed decrease in the GP 1 steady-state levels could be due to degradation of GP 1 (via ER-associated or proteasomal degradation) resulting from its prolonged ER residency because of improper or insufficient maturation of preGP. Since cholesterol levels are lowest in the ER in the secretory pathway (66), ER-resident events involving transmembrane proteins might be particularly sensitive to very small deviations in cholesterol levels from a critical threshold. While most of statins' effect is associated with lowering cellular cholesterol levels, statins also blunt the nonsterol branch of the mevalonate pathway, decreasing formation of isoprenoids and altering protein prenylation, an often critical event in posttranslational modulation of proteins (67). Inhibitors of isoprenoid intermediates, such as geranylgeranyltransferase inhibitor (GGTI), which inhibits prenylation of Rho proteins, or farnesyltransferase inhibitor (FTI), which inhibits the prenylation of the Ras proteins geranyltransferase and farnesyltransferase, have been effective against certain viruses (19,51). Whether statin's effects on GP 1 processing are mediated through the isoprenoid pathway is currently unclear and needs further investigation.
All statins tested in our study reduced EBOV titers, although with variable efficacy. Simvastatin and pitavastatin inhibited EBOV production most potently (0.5-to 2.5-log reduction at 5 to 50 M concentrations). Differences in the antiviral effects of individual statins may be due to many factors, such as chemical structures of each compound affecting pharmacokinetics and pharmacodynamics (68). One limitation of our study is that higher concentrations of statins were required to inhibit EBOV replication in vitro than are achievable in humans using current dosing regimens, as plasma levels of statins are usually low (maximum concentration of drug in serum [C max ], 0.019 to 0.031 M for simvastatin and 0.005 M for lovastatin, based on 40-mg oral dose [69]). Although statin concentrations are likely to be much higher in the liver (69), a major site of EBOV replication, comprehensive in vivo studies in appropriate animal models are required. Unfortunately, statins do not reliably decrease circulating cholesterol concentrations in rodents (70)(71)(72), and thus, such studies would require nonhuman primate models that recapitulate human EVD signs (73).
In summary, we provide evidence that statin treatment decreases production of infectious EBOV virions in a human liver cell line (Huh7) and primary human macrophages, both of which are primary target cells for EBOV infection. Statin reduced production of infectious EBOV particles in Huh7 cells by interfering with GP processing and reducing the amount of GP 1 incorporated into virus particles. The results of this study clearly show that statin inhibits EBOV infection. Our results, combined with statins' known role in suppressing inflammation (74) and preserving endothelial integrity (75), pathways that are impaired in EVD, argue for a potential benefit of using statins as adjunctive therapy in patients with EVD. Ideally, the use of an antiviral that exhibits additional effects in combination with a statin has the potential both to block virus replication and to decrease the deleterious effects of inflammation on the host. Clearly, the next step for evaluating statins for use in EVD would require testing in a nonhuman primate model of disease to ensure both safety and potential efficacy. Louis, MO, USA; not heat inactivated) was used instead of medium with FCS to eliminate the possible uptake of cholesterol from the medium, as previously reported (47). Human monocyte-derived macrophages were isolated as described previously (76). Vero-E6 cells were obtained from the CDC core facility and maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% (vol/vol) FCS and 1% penicillin-streptomycin (Life Technologies). Wild-type EBOV variant Mayinga (Ebola virus/H. sapiens-tc/COD/1976/Yambuku-Mayinga) was from the CDC Viral Special Pathogens Branch reference collection. Lovastatin was from Calbiochem (Billerica, MA). Human LDL was purchased from Sigma-Aldrich (St. Louis, MO, USA). Mevalonate was synthesized according to the method of Goldstein et al. (77).

Biosafety
The plasmid encoding EBOV GP 1,2 was designed to express a human codon-optimized synthetic gene (GPco) corresponding to that of the EBOV Mayinga isolate. Briefly, the GPco (without RNA editing site) gene was purchased from GenScript (Piscataway, NJ) and cloned into the polymerase II (PolII) expression vector pCAGGS (78). EBOV GP resistant to furin cleavage (RRTRR 501 cleavage site replaced with AGTAA) was purchased from GenScript as described previously (44,45). The plasmid encoding Flag-tagged VP40 has been described previously (79).
The following antibodies were used in this study: rabbit polyclonal antibody against EBOV NP (IBT Bioservices, Rockville, MD), rabbit polyclonal antibody against EBOV GP (IBT Bioservices), and rabbit polyclonal antibody against EBOV for immunofluorescence (in-house reagent 70331; Viral Special Pathogens Branch, CDC, Atlanta, GA). The anti-Flag antibody and mouse monoclonal anti-actin antibody were from Sigma (Sigma-Aldrich, St. Louis, USA). NPC1 antibody was from Abcam (Cambridge, MA).

Transfection and infection.
To determine the effects of statin on GP processing or VP40 expression, 2.0 ϫ 10 5 Huh7 cells plated in 12-well plates were transiently transfected with plasmids expressing either EBOV GP or Flag-tagged VP40; transfections were done using LT-1 reagent according to the manufacturer's instructions (Mirus, Madison, WI). After 24 h, cells were treated with statin in LPDS-containing DMEM; cells were harvested 48 h posttransfection. For EBOV infection, Huh7 cells were plated at 2 ϫ 10 5 cells per well in 12-well plates. The next day, cells were infected with EBOV at the indicated MOI for 1 h. For control experiments, Huh7 cells were infected with adenovirus type 5 (ATCC, Manassas, VA) at an MOI of 0.05 for 1 h. Virus inoculum was removed, and cells were washed with serum-free medium. Fresh medium containing 10% LPDS and with or without statin was then added. Culture supernatants and cell lysates were harvested and analyzed as indicated.
TCID 50 and cell viability determination. Supernatants from EBOV-infected Huh7 cells and monocyte-derived macrophages were harvested 72 hpi, and virus titrations were performed in Vero-E6 cells. Three days postinfection, the cells were fixed, permeabilized, and stained to visualize viral proteins. For adenovirus type 5 titer, Vero-E6 cells were treated with 8 serial 10-fold dilutions of supernatants of infected Huh7 cells. After 10 days, the wells with cytopathic effects were counted for each dilution after crystal violet staining. Endpoint viral titers were determined, and TCID 50 was calculated as described previously (80). Results represent mean titers, with error bars indicating standard deviations calculated from 3 independent experiments. For human monocyte-derived macrophages, results represent mean titers with standard deviations from 3 replicate wells, representative of 2 independent experiments.
Cell viability was determined on statin-treated and mock-infected cells, using CellTiter-Glo (Promega) according to the manufacturer's instructions.
qRT-PCR. Huh7 cells were infected with EBOV for the indicated times, and then RNA was isolated from cells or from supernatants of infected cells using the MagMAX-96 total RNA isolation kit (Thermo Fisher Scientific). To determine viral RNA copy numbers, RNA was extracted from supernatants of infected cells. Absolute quantification of viral RNA copy numbers was done by measuring EBOV NP copy numbers using a standard curve with known viral titers serially diluted 5-fold. qRT-PCR was performed with the EBOV NP assay (81). To measure cell-associated EBOV NP gene levels, Huh7 cells pretreated with statin for 48 h or with U18666A (positive control) for 1 h were infected with EBOV at an MOI of 3. Cells were harvested after 3 h, and RNA was isolated from the cells. NP RNA levels were measured by qRT-PCR as described above and normalized to GAPDH mRNA. Results represent mean percent normalized NP RNA levels, with error bars indicating standard deviations calculated from 3 independent experiments.
Western blotting. At indicated times after transfection or infection, cell lysates were harvested by adding lysis buffer containing 50 mM NaCl, 5 mM EDTA, 1% NP-40, 1.0% SDS, and 0.5% sodium deoxycholate supplemented with a protease inhibitor cocktail (Sigma-Aldrich). Lysates from infected cells were gamma irradiated at 2 ϫ 10 6 rads using a high-energy 60 Co source to ensure complete virus inactivation, allowing work at biosafety level 2. Proteins were electrophoretically separated on either 3 to 8% Tris-acetate or 4 to 12% NuPAGE gels (Invitrogen) and transferred to nitrocellulose membranes. The membranes were blocked for 1 h with buffer containing Tris-buffered saline, 0.1% Tween 20, and 5% nonfat dry milk and then probed overnight at 4°C with primary antibodies. Membranes were developed using horseradish peroxidase-conjugated secondary antibodies and enhanced chemiluminescence. After detection of primary antibodies, the membranes were stripped and reprobed with antiactin antibody as a loading control. Results shown are representative of 3 independent experiments.
Immunofluorescence. At 72 hpi, the cells were washed twice with phosphate-buffered saline (PBS) and fixed with 10% formalin at room temperature for 20 min. After formalin fixation, the cells were washed three times with PBS and permeabilized with 0.1% Triton X-100 for intracellular staining. The primary antibodies were added at a 1:1,000 dilution in 1% bovine serum albumin in PBS for 1 h. The cells were then washed 3 times with PBS and incubated for 30 min with the secondary antibodies diluted 1:1,000 in 1% bovine serum albumin in PBS. Multiple final washes were done, and the images were taken using a Nikon Eclipse Ti-S. Virion purification. EBOV virions were partially purified similarly to the procedure reported for Lassa virus (82). Briefly, supernatants from DMSO-or statin-treated and EBOV-infected cells were clarified by centrifugation at 1,500 ϫ g for 30 min. Clarified supernatants were subjected to ultracentrifugation (100,000 ϫ g for 90 min at 4°C) through a 20% sucrose cushion to collect EBOV virions. Virions were suspended in 2ϫ Western lysis buffer, gamma irradiated at 5 ϫ 10 6 rads using a high-energy 60 Co source, and analyzed by Western blotting to detect GP and VP40.
Endo H and PNGase F treatment. In order to investigate the modifications of EBOV glycoproteins, cell lysates were digested with endo H or PNGase F (New England Biolabs, Ipswich, MA) according to the manufacturer's instructions. The digested proteins were resolved by SDS-PAGE under reducing conditions and were analyzed by Western blotting.
Statistical analysis. Error bars in graphs represent standard deviations of the means from comparing Student's t tests for paired samples. Differences were considered significant for P values of Ͻ0.005.