Inhibition of human cytomegalovirus replication by benzimidazole nucleosides involves three distinct mechanisms

Abstract

The benzimidazole nucleosides 2-bromo-5,6-dichloro-1-(beta-d-ribofuranosyl)benzimidazole (BDCRB) and 2-isopropylamino-5,6-dichloro-1-(beta-l-ribofuranosyl)benzimidazole (1263W94, or maribavir) are potent and selective inhibitors of human cytomegalovirus (HCMV) replication. These inhibitors act by two different mechanisms: BDCRB blocks the processing and maturation of viral DNA, whereas maribavir prevents viral DNA synthesis and capsid nuclear egress. In order to determine by which of these two mechanisms other benzimidazole nucleosides acted, we performed time-of-addition studies and other experiments with selected new analogs. We found that the erythrofuranosyl analog and the alpha-lyxofuranosyl analog acted late in the viral replication cycle, similar to BDCRB. In marked contrast, the alpha-5'-deoxylyxofuranosyl analog of 2,5,6-trichloro-1-(beta-d-ribofuranosyl)benzimidazole (compound UMJD1311) acted early in the replication cycle, too early to be consistent with either mechanism. Similar to other reports on early acting inhibitors of herpesviruses, compound 1311 was multiplicity of infection dependent, an observation that could not be reproduced with UV-inactivated virus. HCMV isolates resistant to BDCRB and maribavir were sensitive to compound 1311, as were viruses resistant to ganciclovir, cidofovir, and foscarnet. The preincubation of host cells with compound 1311 and removal prior to the addition of HCMV did not produce an antiviral cellular response. We conclude that this newly discovered early mode of action occurs at a stage of viral replication after entry to cells but prior to viral DNA synthesis, thereby strongly suggesting that the trisubstituted benzimidazole nucleoside series possesses three distinct biochemical modes of action for inhibition of HCMV replication.

FIG. 1.

Structures of compounds used in this study. Compounds were synthesized, purchased, or obtained as gifts as described in Materials and Methods.

FIG. 2.

HCMV time-of-addition study under conditions that emphasize the differences between viral DNA synthesis and DNA processing. Benzimidazole nucleosides 853, 1311, maribavir, and BDCRB and a GCV control were added to cells infected with HCMV at time zero and at the indicated time points. To emphasize the differences between late-acting inhibitors, virus-infected cells were incubated with a lower amount of FBS at 34°C instead of 37°C, which is ordinarily used as in the experiment shown in Fig. 3. The inhibition of viral replication was measured by plaque reduction. Data are presented as the means ± SDs of duplicate experiments. The concentrations of compounds used in the experiments were below cytotoxic levels and were as follows: 30 μM BDCRB, 10 μM 1311, 20 μM 853, 20 μM maribavir, and 30 μM GCV.

FIG. 3.

HCMV time-of-addition study for benzimidazole nucleosides 1311 and BDCRB with GCV and pyrrolo[2,3-d]pyrimidine 1028 control compounds. Drugs were added to cells infected with HCMV at time zero and at the indicated time points. Inhibition of viral replication was measured by plaque reduction assays under standard conditions. Data are presented as the means ± SDs of triplicate experiments. The concentration of compound 1311 was 15 μM; the concentration of each control compound was 75 μM.

FIG. 4.

Yield reduction potencies for compound 1311 and GCV when compounds were added at different times. Compounds were added to cells at eight different concentrations (from 100 to 0.05 μM) at the indicated times, and virus was added at time zero. Dose-response experiments were performed at each time point and were used to determine IC₉₀s. Data are presented as the means ± SDs of triplicate experiments. When added at 48 h post infection, the IC₉₀ of compound 1311 was >100 μM (*).

FIG. 5.

Interaction against HCMV between compounds 1311 and GCV. The extent of viral inhibition from combinations of compound 1311 and GCV was assessed by ELISA. Data are presented as a three-dimensional surface area of inhibition above the level expected based on a model of additive interaction (29). The results shown are those computed to be statistically significant at a 95% confidence interval.

FIG. 6.

The effect of preincubation and removal of compounds prior to infection compared to adding drug at 1 h postinfection. Shaded bars quantify the percent inhibition compared to the control without drug for a 24-h preincubation with compound, which was then washed out prior to infection with virus. Open bars quantify the percent inhibition compared to the control without drug for the plaque reduction assay under standard conditions; the same concentrations of drugs were added at 1 h postinfection.