Hepatitis C Virus (HCV) NS5B Nonnucleoside Inhibitors Specifically Block Single-Stranded Viral RNA Synthesis Catalyzed by HCV Replication Complexes In Vitro

Abstract

Replication complexes of hepatitis C virus synthesized two major species of viral RNA in vitro, double stranded and single stranded. NS5B nonnucleoside inhibitors inhibited dose dependently the synthesis of single-stranded RNA but not double-stranded RNA. Moreover, replication complexes carrying a mutation resistant to a nonnucleoside inhibitor lost their susceptibilities to the inhibitor.

The nonstructural proteins (including NS5B, a RNA-dependent RNA polymerase) of hepatitis C virus (HCV) assemble on specialized intracellular membranes into replication complexes (RCs). Several papers have shown that RCs isolated as membrane fractions from HCV replicon-containing cells are able to synthesize nascent RNA molecules by using HCV RNA templates within the complexes when reconstituted in vitro (1, 5, 7, 8, 9, 11, 13, 15). NS5B nucleoside inhibitors, in their triphosphate form, block HCV RNA synthesis catalyzed by RCs in vitro (8, 9, 11, 13). It is controversial, however, whether NS5B nonnucleoside inhibitors (NNIs) are able to do so (7, 11, 15).

As a first step of investigation, we modified the protocol reported by Lai et al. (9). We added 0.3 mM Mn²⁺ to the reaction mixtures and resolved the products on 1% native agarose gels. In brief, the reaction mixture contained 50 mM HEPES (pH 7.5); 10 mM KCl; 10 mM MgCl₂; 0.3 mM MnCl₂; 60 units of RNase inhibitor; 10 μg of actinomycin D per ml; 0.5 mM each of ATP, GTP, and UTP; 10 μCi of [α-³²P]CTP (800 Ci/mmol); and 6 μl of membrane fractions prepared from Con-1 replicon-containing cells (Huh-9-13 cells) (10) as described previously (9), in a total volume of 60 μl. Reaction mixtures were incubated at 30°C for 60 min unless otherwise indicated. RNA was isolated with TRIzol LS reagent (Invitrogen), dissolved in water, and resolved on a 1% agarose gel in 1× Tris-borate-EDTA (TBE) buffer.

The above-described modifications resulted in a consistent detection of two radiolabeled bands (Fig. 1A). These radiolabeled bands were present only in the reaction mixtures containing the membrane fractions prepared from replicon-containing cells, confirming the previous observation of their identities as HCV RNAs (9). The nature of these labeled RNA species was characterized with a pulse-chase experiment coupled with a nuclease digestion. The nascent RNA molecules were pulse-labeled for 4 min with [α-³²P]CTP and were chased with an excess amount (400-fold) of cold CTP for different durations. The samples were removed from the reaction mixture at each time point and were divided into halves, with one half loaded directly to the gel and the other loaded after digestion with mung bean nuclease, a single-stranded-specific endonuclease. As shown in Fig. 1B, the two labeled RNA species behaved differently. The small species was chased to a series of larger products (denoted “RNAs from SS”). The small species as well as its chased products was sensitive to nuclease treatment and so was mainly composed of single-stranded RNA (ssRNA). In contrast, the large species remained unchanged in position, increased somewhat in intensity during the chase, and was largely retained after nuclease digestion and so was mainly composed of double-stranded RNA.

FIG. 1.

Characterization of nascent HCV RNA synthesized by CRCs in vitro. (A) Nascent RNA synthesis with CRCs prepared from replicon-containing cells. CRCs were prepared from Huh-9-13 cells which contained a Con-1 subgenomic replicon according to the method described in reference 9. The reactions were run as described in the text. Total RNAs were extracted with TRIzol reagent, dissolved in water, and resolved on a 1% agarose gel in 1× TBE buffer. Two different preparations of CRCs (preparation 1, lane 2; preparation 2, lane 3) were used in the reactions. Two major RNA products were seen, and each was indicated as L (large) or S (small). (B) Pulse-chase labeling and nuclease sensitivity of nascent RNA synthesized by CRCs in vitro. Nascent RNA was pulsed-labeled with [α-³²P]CTP under the conditions described in the text for 4 min and was then chased with ∼400-fold cold CTP for 10, 20, 30, 45, 60, and 120 min. A portion of each reaction mixture was removed at the end of the pulse and at the end of each chase period and was immediately mixed with TRIzol reagent to stop the reaction. After purification, one-half of each sample was treated with 10 units of mung bean nuclease at 30°C for 30 min before electrophoresis. The positions of the double-stranded (DS) and the single-stranded (SS) RNA are indicated. The bracket shows the positions of RNA products derived from chasing the pulse-labeled ssRNA.

To investigate whether NNIs inhibit HCV RNA synthesis catalyzed by crude replicase complexes (CRCs) in vitro, a benzothiadiazine-based compound (compound 1) and a benzimidazole-based compound (compound 2) were chosen (Fig. 2A), each binding to a different site on NS5B (4, 14, 16, 17; R. Coulombe, P. L. Beaulieu, E. Jolicoeur, G. Kukolj, S. Laplante, and M. A. Poupart, 18 November 2004, international patent application WO 2004099241 A1). When added to Huh-9-13 cells, compound 1 was active, with a 50% effective concentration (EC₅₀) of 0.5 μM (see Fig. 4A), similar to a previously reported value (3), whereas compound 2 was not (data not shown). Several analogs of compound 2 were also reported to be inactive in replicon-containing cells, presumably due to their poor cellular permeability, resulting from the presence of a highly ionizable carboxyl group at a physiological pH (2). When the compounds were evaluated in vitro with CRCs prepared from Huh-9-13 cells, both caused a dose-dependent reduction in the yield of ssRNA, although they had no apparent effect on the yield of double-stranded RNA, with estimated IC₅₀ values of 0.1 μM and 0.3 μM, respectively (Fig. 2B). These values are similar to what has been reported in the literature for recombinant NS5B in vitro (2, 3).

FIG. 2.

Inhibition of the single-stranded RNA synthesis by benzothiadiazine and benzimidazole-based compounds. (A) Chemical structures of compound 1 and compound 2. Compound 1 and 2 were discovered by GlaxoSmithKline and Boehringer Ingelheim, respectively. No further development of the two compounds has been disclosed. (B) The synthesis of ssRNA by CRCs in vitro was blocked by compounds 1 and 2. One microliter of dimethyl sulfoxide without (untreated control) or with compound 1 or 2 was added to the reaction mixture as described in the text (the final concentration of dimethyl sulfoxide was ∼2%). The reaction mixture was incubated at 30°C for 20 min prior to the addition of the substrates, including [α-³²P]CTP. After one more hour of incubation, total RNAs were extracted with TRIzol reagent, dissolved in water, and resolved on a 1% agarose gel in 1× TBE buffer. DS, double-stranded RNA; SS, single-stranded RNA.

FIG. 4.

Effect of M414T mutation in NS5B on the inhibitory effect of benzothiadiazine. (A) Inhibitory effect of compound 1 on the wild-type (WT) replicon or the mutant replicon replication in cells. The wild type (Huh-9-13) or the M414T mutant replicon-containing cells were plated and treated with various concentrations of compound 1 (31.6, 10, 3.16, 1, 0.3, and 0.1 μM) for 72 h. The HCV replicon level following the treatment was quantified with a dot blot assay, using a ³²P-labeled negative-sense riboprobe complementary to NS5B. The dose-response curve fitting was performed by nonlinear regression analysis with Prism software (GraphPad Software, Inc.) based on the radioactivity (counts per minute) in each spot relative to that in the untreated controls (100%). Based on the analysis, the EC₅₀ values for compound 1 were 0.9 μM and >31.6 μM for the wild type and the M414T mutant replicon, respectively. (B) Inhibitory effect of compound 1 on nascent-RNA synthesis catalyzed in vitro by CRCs prepared from the M414T mutant replicon-containing cells. CRCs were prepared from the M414T mutant replicon-containing cells according to the standard protocol described in reference 9 and were examined for their sensitivities to compound 1 treatment in vitro in the presence of various concentrations of compound 1 as indicated at the top, under the conditions descried in the text. The positions of double-stranded RNA (DS) and single-stranded RNA (SS) are indicated with arrows on the right.

The results shown in Fig. 1B suggested that ssRNA was not the end product. Instead, the molecules could be converted to larger ones as “RNAs from SS.” Therefore, the disappearance of ssRNA upon addition of NNIs might be due to either the blockage of the synthesis of ssRNA or the rapid conversion of this species to the larger molecules. We conducted another pulse-chase experiment to differentiate between those two possibilities (Fig. 3). The effect of compound 1 on the yield of ssRNA was observed when the compound was present during the pulse. In contrast, no effect on nascent RNAs was detected when the compound was added after the pulse. Together, these results showed that compound 1 blocked only the initial synthesis of ssRNA and did not affect its conversion to other RNA species.

FIG. 3.

Effect of benzothiadiazine on the single-stranded RNA during pulse and chase. The pulse-chase schedule is shown in the top panel of the figure, in which a reaction mixture containing compound 1 or dimethyl sulfoxide alone is represented by a black or a white bar, respectively. The duration of each step is indicated at the right of the schedule. Lanes 1 and 2 in the middle panel show samples collected immediately after the pulse in the absence (lane 1) or the presence (lane 2) of compound 1. Lanes 3, 4, 5, and 6 show samples collected after the chase with a volume of cold CTP (∼400-fold greater than that of the labeled CTP). Lanes 3 and 4 show samples collected after a 15-min chase in the absence (lane 3) or presence (lane 4) of compound 1. Lanes 5 and 6 show samples collected after the second 15-min chase in the absence (lane 5) or presence (lane 6) of compound 1. The bracket shows the positions of RNA products derived from chasing the pulse-labeled ssRNA. DS, double-stranded RNA; SS, single-stranded RNA. The intensities of DS and SS in each lane were quantified with densitometry (Alpha Image 2200; Alpha Innotech). The intensity of SS relative to the intensity of DS in each lane was calculated as shown in the bottom panel.

To exclude the possibility that the inhibition of ssRNA synthesis was due to the effect of NNIs on a non-NS5B target in CRCs rather than on NS5B, we utilized a cell line which contained Con-1 replicons carrying a methionine-to-threonine mutation at amino acid residue 414 of NS5B. This mutation (M414T) has been reported to confer resistance to compound 1 whether it is carried by replicon molecules in vivo or by recombinant NS5B molecules in vitro (17). Hence, CRCs prepared from the mutant replicon-containing cells should not respond to the compound 1 treatment if the response observed with CRCs prepared from Huh-9-13 cells (the wild-type-replicon-containing cells) is due to the effect of compound 1 on NS5B. As expected, HCV RNA reduction in the M414T replicon-containing cells was much less than that in Huh-9-13 cells upon compound 1 treatment (the shift in EC₅₀ was >30-fold) (Fig. 4A). When the susceptibilities of CRCs prepared from the mutant replicon-containing cells were evaluated in vitro, little reduction in the yield of ssRNA was observed even at a concentration as high as 31.6 μM (Fig. 4B). These results confirmed that the inhibitory effect of compound 1 on ssRNA synthesis catalyzed by CRCs in vitro is due to its activity against NS5B.

Both benzimidazole- and benzothiadiazine-based compounds are able to inhibit the de novo initiation of RNA synthesis catalyzed in vitro by recombinant NS5B (6, 12). It was also once proposed that CRCs support the de novo initiation (1, 7). Recent reports that RNA synthesis in the system is not sensitive to several NNIs (11, 15), however, cast doubt on whether the system supports the de novo initiation or whether NNIs block the de novo initiation when it is happening in CRCs. When we performed studies under the standard conditions, i.e., by running the reactions without Mn²⁺ and resolving the products on denaturing gels, we also observed no effect of NNIs on RNA synthesis (data not shown). After the protocol optimization, however, we show for the first time a specific effect of NNIs on ssRNA synthesis. Although we are not certain that the nascent ssRNA species is derived from the de novo initiation in lack of a further characterization of the species, this finding by itself offers us an opportunity to study the mechanism of action of NNIs under an in vitro condition which is much closer to an in vivo condition than that with recombinant NS5B.