D. J. Dargan,* C. B. Gait and J. H. Subak-Sharpe

Transcription

1 Journal of General Virology (1992), 73, Printed in Great Britain 407 The effect of cicloxolone sodium on the replication in cultured cells of adenovirus type 5, reovirus type 3, poliovirus type 1, two bunyaviruses and Semliki Forest virus D. J. Dargan,* C. B. Gait and J. H. Subak-Sharpe MRC Virology Unit, Institute of Virology, University of Glasgow, Church Street, Glasgow G 11 5JR, U.K. The effect of cicloxolone sodium (CCX) on the replication of typical representatives of different virus families [adenovirus type 5 (Ad-5), reovirus type 3 (Reo-3), Bunyamwera and Germiston viruses, poliovirus type 1 (Polio-l) and Semliki Forest virus (SFV)] in tissue culture was investigated. The Golgi apparatus inhibitor monensin (Mon) and CCX were shown to have analogous effects on some aspects of virus replication. Although the Mon-like effect of CCX played no role in the antiviral activity against Ad-5, Reo-3 or Polio-l, it could entirely account for the antiviral activity against the Bunyamwera and Germiston viruses, for which inhibition of glycoprotein processing was responsible for the antiviral activity. In the case of SFV, the Mon-like activity of CCX caused cytoplasmic assembly of fully infectious SFV within vacuoles and thus impaired virus release without altering total infectious virus yield. Fewer Ad-5 and Reo-3 progeny were produced in the presence of the drug. CCX had a dose-dependent biphasic effect on the particle:p.f.u, ratio of the Reo-3 yield. At low CCX concentration (< 50 ~tm) the virus yield contained poor quality, non-infectious virus, but at higher CCX concentration (/> ~tm) low quality virus could no longer be successfully assembled. We conclude that the antiviral effect can be manifested in three ways: (i) by a reduction in the virus particle yield produced; (ii) by a loss of quality (relative infectivity); (iii) by a virucidal effect of the drug. We have previously defined three CCX sensitivity classes. Mechanisms (i), (ii) and (iii) operate against viruses belonging to class CCXs-1 [herpes simplex virus (HSV) type 1, HSV-2 and vesicular stomatitis virus], but essentially only (i) and (ii) affect Reo-3 (CCX~-2), whereas (i) and possibly (iii) affect Ad-5 (CCX~-2). In the case of SFV (CCX~-3) none of these mechanisms operate, but relocation of assembled virus is found. Cicloxolone sodium (CCX) is a broad range antiviral compound (Gait et al., 1990) with a complex mode of action which derives from the effects of the drug on certain cell membrane functions that are required for virus replication (Dargan & Subak-Sharpe, 1986). Thus, all virus gene activities that depend on a fully functional cell membranes are affected to some extent, rather than the drug operating by specifically inhibiting the function of one essential virus gene product. Gait et al. (1990) placed viruses representing eight different virus families into three CCX sensitivity (CCX ~) classes based on the different kinetics of dose-response curves. We have studied, in some depth, the antiviral mechanism of CCX operating against the replication of the CCXs-1 viruses, herpes simplex virus types 1 and 2 (HSV-1 and -2) (Dargan & Subak-Sharpe 1985, 1986, 1991; Dargan et al., 1988) and vesicular stomatitis virus (VSV) (Dargan et al., 1992). Here we investigate the effects of CCX on the replication of viruses belonging to the CCXs-2 and CCXs-3 classes. Whereas the infectious virus yields of CCX~-I viruses decrease progressively with increasing CCX concentration, the dose-response curve of CCX~-2 viruses [adenovirus type 5 (Ad-5), reovirus type 3 (Reo- 3), Poliovirus type 1 (Polio-l), and Germiston and Bunyamwera viruses] reaches a plateau value which represents maximum antiviral activity. The infectious yield of the only identified member of the CCX~-3 class, Semliki Forest virus (SFV), is unaffected by CCX treatment (Gait et al., 1990). Experiments involving Ad- 5 were performed in HeLa cells; all other experiments employed BSC-1 cells. The CCX tolerance of BSC-1 and HeLa cells has been investigated by Gait et al. (1990), who found that the 48 h CCX concentration inhibiting cell growth by 20~o for both cell lines was > 300 IXM, and that drug-treated cell cultures could be passaged normally following removal of the drug after 48 h of SGM

2 408 Short communication 60 > 40 L) 20 0 I00 -.~ 60- e~ ~ 60,.-., > >" 60 7, Z ~ 7' CCX (as() 4 J i i i CCX (p.m) ih) } 10 2 Mon ~M) Tun (p.g/ml) i L i Mon (um) Tun ~g/ml) o treatment. When such CCX-resistant cell lines are used, antiviral activity and cytotoxicity are uncoupled (Dargan & Subak-Sharpe, 1985; Gait et al., 1990). CCX and monensin (Mon) have been shown to have analagous effects on the replication of HSV and VSV, and functions associated with the Golgi apparatus, e.g. post-translational processing of glycoproteins and their intracellular transport (Dargan & Subak-Sharpe, 1986, 1991 ; Dargan et al., 1992). Thus the Mon-like effect of CCX on the replication of different viruses can be related to the importance of Golgi apparatus functions for the replication of particular viruses. To investigate the contribution made by the Mon-like activity of CCX to its activity against different viruses, the cell-associated (CA) and cell-released (CR) infectious yields of representative viruses from several different virus families [Ad-5 (adenovirus), Reo-3 (reovirus), Polio-1 (picornavirus), Germiston and Bunyamwera (bunyaviruses), and SFV (togavirus)] were compared. Tunicamycin (Tun), an inhibitor of the initial N-linkage of sugars to precursor proteins, was included in these experiments to determine the effect of glycosylation on infectivity per se. In the case of Ad-5, Reo-3 and Polio-l, CCX did not affect the release of virus from the cell (Fig. 1 a and c). Ad-5 and Reo-3 yields were reduced about -fold, whereas those of Polio-1 were only reduced about 20-fold. Treatment with either Mon or Tun had no significant effect on CA or CR yields from Ad-5-, Reo-1- or Polio-l-infected cells (Fig. l b and d). This demonstrates that the Golgi apparatus functions that are sensitive to CCX, Mon or Tun are not implicated in the antiviral effect of CCX against these three viruses. The relative CA and CR yields of infectious Bunyamwera and Germiston viruses from CCX-treated cells were reduced in an analogous way at each concentration; therefore CCX has no effect on the release of bunyaviruses from cells (Fig. I e). The actual CA and CR infectivity was lowered 10- to 20-fold in the presence of 200 to 300 ~tm-ccx. Treatment of Bunyamwera and Germiston virus-infected cells with Mon resulted in progressive decreases in CA and CR infectivity (Fig. 1 f). The CA yield was about fourfold more sensitive to 1 gm-mon than the CR yield, but at higher concentrations the drug affected CA and CR infectivity equally. Fig. 1. The effect of CCX (a, c, e and g), Mon and Tun (b, d, fand h) on the CA (filled symbols) and CR (open symbols) infectious yields of Ad- 5, Reo-3, Polio-l, Germiston and Bunyamwera viruses, and SFV. (a, b) Ad-5 (li, I-]) and Reo-3 (O, O); (c, d) Polio-1 (am, I-q); (e,f) Germiston (0, O) and Bunyamwera (., I-q) viruses; (g, h) SFV (,, I-q). Viability of uninfected cells treated with drugs in parallel (~k). (b) and (a), (A) HeLa cells; (z~) BSC-1 cells. Dose-response experiments were performed as described by Gait et al. (1990) and the virus yields separated into CA and CR components as described by Dargan & Subak-Sharpe (1985).

3 Short communication Tun (2 ~tg/ml) added to infected cell cultures gave results similar to those obtained with 300 ktm-ccx and 10 ~tmmon (Fig. 1 f ). The data are consistent with the Monlike activity of CCX accounting for its antiviral effect against the bunyaviruses, which suggests that inhibition of glycoprotein processing is probably responsible. Treatment of SFV (CCXs-3)-infected cells with CCX resulted in an approximately 10-fold increase in CA infectivity (between and 200 ~tm-ccx) accompanied by a corresponding decrease in CR infectivity. Thus, the total yield of infectious virus remained unchanged, whereas the distribution between CA and CR was inverted, suggesting that release of SFV from the cells was impaired (Fig. l g). The dose-response curves obtained with Mon exhibited a similar inversion (Fig. 1h), suggesting that the Mon-like activity of CCX is responsible for the impairment of SFV release. Treatment of SFV-infected cells with Tun had a very different effect, resulting in a 0-fold reduction in both CA and CR infectivity. The data suggest that core sugar processing of SFV glycoproteins is essential for infectivity and that, quite separately, a CCX/Mon-sensitive Golgi apparatus function(s) is required for virus release. These two aspects are therefore uncoupled. The intracellular accumulation of infectious SFV in BSC-I cells treated with 300p.M-CCX or 10ktM-Mon have been studied by electron microscopy (EM). Infected cells (m.o.i. 50 p.f.u./cell) were harvested 18 h post-infection (p.i.) and processed for EM as described by Dargan et al. (1988). In the absence of drug, SFV was observed to assemble at and bud from the plasma membrane of infected cells (Fig. 2a). In contrast, CCX (Fig. 2b) or Mon (Fig. 2c) treatment resulted in SFV particle assembly not only at the plasma membrane but also at membranes bounding cytoplasmic vacuoles in which mature SFV particles accumulated. Association of SFV particles with cytoplasmic vacuoles was never observed in drug-free infected cell cultures. Thus, the impairment of SFV release from cells treated with CCX or Mon (Fig. 1g and h) is due to assembly of infectious SFV at intracellular sites, from which it is either not transported or transported inefficientlyto the cell surface for release into the extracellular medium. The infectious Ad-5 and Reo-3 yields from cells treated with three different plateau value concentrations of CCX (maximum antiviral effect) were investigated by second generation dose-response experiments for the presence of drug-resistant variants. However, no evidence for progeny virus being genetically resistant to CCX could be obtained (data not shown). Thus the plateau appears to reflect either partial impairment of an essential process, or the blocking of a virus- or cellspecified function which modifies, but is not absolutely essential, for virus infectivity. 409 Fig. 2. Assembly in and budding of SFV from BSC-1 cells in the absence or presence of either CCX or Mon. (a) In the absence of any drug SFV assembly occurs normally at the plasma membrane of infected cells. Bar marker represents nm. (b) In the presence of 300 ~tm-ccx SFV particles are associated with cytoplasmic vacuoles. Virus particles are arranged around the inner surface of the membrane and are also present within the lumen of the vacuole. Bar marker represents 150 nm. (c) In the presence of 10 p.m-monsfv particles are associated with cytoplasmic vacuoles. Virus particles are arranged around the inner and outer surfaces of cytoplasmic vacuoles, and are also present within the lumen of some vacuoles. Bar marker represents 150 nm. The effects of CCX on the yields of Ad-5 and Reo-3 particles, infectivity (p.f.u.) and the particle :p.f.u. ratio are shown in Table 1. The data for Ad-5 show that 150

4 410 Short communication Table 1. The effect of increasing concentrations of CCX on the yield of Reo-3 and Ad-5 particles, infectious virus and the particle :p.f.u. ratio* Particles P.f.u. Virus CCX ~M) ( 10-9) (x 10-6) Particle:p.f.u. Ad Reo Expt Reo Expt * The yields of Reo-3 and Ad-5 were determined at 24 and 48 h respectively. ~M- or 300 pm-ccx results in a 60- to -fold reduction in both infectivity and particle numbers with the particle:p.f.u, ratio remaining unchanged, suggesting that CCX has no effect on the quality (relative infectivity) of the Ad-5 progeny produced in HeLa cells. In the case of Reo-3, 300 ktm-ccx treatment of infected BSC-1 ceils resulted in a 40- to 50-fold reduction in virus particle numbers and a simlar 30- to 60-fold reduction in p.f.u. However, at low concentrations (10 txm- and 50 ~tm- CCX), the particle:p.f.u, ratio consistently increased, suggesting the assembly of a relatively greater proportion of non-infectious (poor quality) Reo-3 particles. We propose that the biphasic effect of CCX on the Reo-3 particle :p.f.u. ratio reflects a biphasic effect of the drug on a cell membrane-associated function which is essential for progeny virus infectivity. In this context the demonstrated biphasic effect of CCX on the protein permeability of the plasma membranes of BHK and Flow 2002 cells should be noted (Dargan & Subak- Sharpe, 1986). To investigate whether CCX has a virucidal effect on Reo-3 and Ad-5, suspensions of virus with or without 300 gtm-ccx were incubated at 4 C or 37 C. In the absence of CCX the 24 h stability of Reo-3 was about 90~ of the 0 h control value at both temperatures; the presence of 300 gtm-ccx in the suspension medium only marginally altered the infectivity. Thus, CCX has negligible virucidal activity against Reo-3. In two independent experiments, the 48 h stability of Ad-5 in the presence of 300 gm-cc X was 48.5~o and 64.2~ of the drug-free control at4 C, and 89~ and 46.8~o at 37 C. El 94g p620 E1/E2o Fig. 3. SDS-PAGE profile on a 10~ gel of the polypeptides induced in m.l (lanes 1 to 6) or SFV-infected BSC-1 cells (lanes 7 to 12). Cells were either drug-free (lanes 1 and 7), or treated with, 200 or 300 ktm-ccx (lanes 2, 3 and 4, and 8, 9 and 10, respectively); 10 IxM-Mon (lanes 5 and 11); or 2 ~tg/ml Tun (lanes 6 and 12). Cells were labelled with [3sS]- methionine from 1 to 24 h p.i. Cellular stress proteins (D) and SFVspecified proteins (mr) are identifed, p62 (polyprotein E2 + E3; 62K); El/E2 (El, 50K; E2, 52K); C (capsid protein, 33K); p620 (nonglycosylated precursor); E10/E2o (non-glycosylated precursors). This is suggestive of, but not conclusive evidence for virucidal activity of CCX against Ad-5. Polypeptides labelling with [35S]methionine in Ad-5- and Reo-3-infected cells treated with increasing concentrations of CCX were investigated by SDS-PAGE (data not shown). The intensity of most Ad-5 and Reo-3 polypeptide bands decreased with increasing CCX concentration, consistent with a general reduction in the transcription and/or translation of virtually all virus polypeptides. However, the Ad-5 13K polypeptide was exceptional in that its intensity actually increased with CCX concentration. Changes in the electrophoretic mobility of some Ad-5 (72K and 65K) and Reo-3 ~ and tr region) polypeptides indicated an effect of CCX on polypeptide processing. We conclude that reduced synthesis of Ad-5 and Reo-3 proteins coupled with specific effects of CCX on the processing of some polypeptides limits the numbers of Ad-5 and Reo-3 progeny particles assembled. The effect of increasing CCX concentration, compared with that of Mon and Tun, on polypeptides labelling with [3SS]methionine in mock-infected (m.i.) and SFV-infected BSC-1 cells is shown in Fig. 3. From 3 h before infection with SFV (50 p.f.u./cell) and throughout, the cells were treated with 5 gtg/ml actinomycin D. Infection and treatment with CCX was as described for dose-response experiments (Galt et al., 1990), except that 10 ~tg/ml [35S]methionine was present

5 Short communication 411 between 1 and 24 h p.i. Extraction of proteins and SDS- PAGE were as described in Marsden et al. (1976) except that 10~ acrylamide gels were used. Little or no change was detected in the pattern of polypeptides extracted from m.i. cells except when they were treated with Tun, which induced the synthesis of cellular stress proteins of 120K, 94K, 78K and 40K (lane 6). The only SFV proteins detected in infected cell extracts (lane 7) were the p62, El/E2 and C structural proteins (Melancon & Garoff, 1987; Simons & Garoff, 19). Treatment with increasing concentrations of CCX (lanes 8, 9 and 10) and 10 ktm-mon (lane 11) resulted in an increase in the intensity of the p62 band and a decrease in that of the C protein band. In extracts of cells treated with Tun (lane 12), all the virus bands increased in intensity, with the p62 and E 1/E2 glycoprotein bands (Hakimi & Atkinson, 1982; Simons & Garoff, 19) showing markedly increased electrophoretic mobility. The mobility of the p62 and El/E2 glycoprotein bands appeared to be unaffected by CCX or Mon treatment, leading us to conclude that neither drug had an appreciable effect on SFV glycoprotein processing in BSC-1 cells. Our results (Fig. 3) suggest that transport is the CCX/Mon-sensitive Golgi apparatus function affected by the drug and which results in intracellular assembly (Fig. 2) of SFV. Griffiths et al. (1983) have also reported intracellular assembly of SFV in BHK cells treated with Mon, and suggest that the virus buds into vacuoles derived from Golgi apparatus membranes. The presence of CCX or Mon in the culture medium of infected cells impairs the transport of both VSV (Dargan et al., 1992) and SFV (this paper) glycoproteins through the Golgi apparatus, but the outcomes are quite different. With VSV the result is greatly reduced assembly of particles at the plasma membrane, whereas SFV assembly is relocated from the plasma membrane to internal vesicles. The differential CCX sensitivity of VSV and SFV indicates that the mechanism of antiviral action operates before virus assembly. Finally we can discern three different antiviral effects of CCX: (i) lower virus particle yield; (ii) reduced quality of the virus progeny produced; (iii) a direct virucidal effect. All three operate with CCX~-I class viruses (HSV-1, HSV-2 and VSV), which are the viruses most sensitive to the drug. Only two operate against the CCXs-2 class viruses analysed; in the case of Reo-3 these are (i) and (ii), whereas for Ad-5 they are (i) and (iii). In the case of the CCXs-3 class virus (SFV), none of these effects operate, and only relocation of assembled virus between plasma membrane cytoplasmic vesicles is found. We thank Miss L. Baxendale and Dr P. Thornton of the Biorex Laboratories Ltd who provided a Ph.D. studentship for Mrs C. Gait and supplied us with CCX, and Mr J. Aitken who performed the electron microscope work. References DARGAN, D. J. & SUBAK-SHARPE, J. H. (1985). The effect of triterpenoid compounds on uninfected and herpes simplex virusinfected cells in culture. I. Effect on cell growth, virus particles and virus replication. Journal of General Virology 66, DARGAN, D. J. & SUBAK-SHARPE, J. H. (1986). The effect of triterpenoid compounds on uninfected and herpes simplex virusinfected cells in culture. II. DNA and protein synthesis, polypeptide processing and transport. Journal of General Virology 67, DARGAN, D. J. & SUBAK-SHARPE, J. H. (199l). The difference in sensitivity to cicloxolone sodium between herpes simplex virus types 1 and 2 maps to the locations of genes UL22 (gh) and UL44 (gc). Journal of General Virology 72, DARGAN, D. J., AITKEN, J. D. & SUBAK-SHARPE, J. H. (1988). The effect of triterpenoid compounds on uninfected and herpes simplex virus-infected cells in culture. III. Ultrastructural study of virion maturation. Journal of General Virology 69, DARGAN, D. J., GALT, C. B. & SUBAK-SHARPE, J. H. (1992). The effect of cicloxolone sodium on the replication of vesicular stomatitis virus in BSC-1 cells. Journal of General Virology 73, GALT, C., OAR.DAN, D. J. ~: SUBAK-SHARPE, J. H. (1990). In vitro studies of the antiviral range of cicloxolone sodium and identification of cell lines tolerant to the drug. Antiviral Chemistry and Chemotherapy 1, GRIFFIYHS, G., QUINN, P. & WARREN, G. (1983). Dissection of the Golgi complex. 1. Monensin inhibits the transport of viral membrane proteins from medial to trans Golgi cisternae in baby hamster kidney cells infected with Semliki Forest virus. Journal of Cell Biology 96, HAKIMI, J. & ATKINSON, P. H. (1982). Glycosylation of intracellular Sindbis virus glycoproteins. Biochemistry 21, MARSDEN, H. S., CROMBIE, I. K. & SUBAK-SHARPE, J. H. (1976). Control of protein synthesis in herpesvirus-infected cells: analysis of the polypeptides induced by wild type and sixteen temperaturesensitive mutants of HSV strain 17. Journal of General Virology 31, MELANCON, P. & GAROFF, H. (1987). Processing of the Semliki Forest virus structural polyprotein: role of the capsid protease. Journal of Virology 61, SIMONS, K. & GAROFF, H. (19). The budding mechanisms of enveloped animal viruses. Journal of General Virology 50, (Received 12 June 1991; Accepted 14 October 1991)