Overdependence on the host—an Achilles' heel of HCV?



Hepatitis C virus (HCV) RNA replication depends on viral protein association with intracellular membranes, but the influence of membrane composition on viral replication is unclear. We report that HCV RNA replication and assembly of the viral replication complex require geranylgeranylation of one or more host proteins. In cultured hepatoma cells, HCV RNA replication was disrupted by treatment with lovastatin, an inhibitor of 3-hydroxy-3-methyglutaryl CoA reductase, or with an inhibitor of protein geranylgeranyl transferase I, each of which induced the dissolution of the HCV replication complex. Viral replication was not affected by treatment of cells with an inhibitor of farnesyl transferase. When added to lovastatin-treated cells, geranylgeraniol, but not farnesol, restored replication complex assembly and viral replication. Inasmuch as the HCV genome does not encode a canonical geranylgeranylated protein, the data suggest the involvement of a geranylgeranylated host protein in HCV replication. Inhibition of its geranylgeranylation affords a therapeutic strategy for treatment of HCV infection.

Disruption of hepatitis C virus RNA replication through inhibition of host protein geranylgeranylation Ye J, Wang C, Sumpter R Jr., Brown MS, Goldstein JL, Gale M Jr. Proc Natl Acad Sci USA 2003; 100: 1586515870.

Jeffrey S. Glenn M.D., Ph.D.*, * Division of Gastroenterology and Hepatology, Stanford University School of Medicine, Palo Alto, CA.


The aforementionedwork by Ye et al. is in an exciting area at the intersection of two fields: virology and prenylation. The virus is hepatitis C (HCV), which flourishes in over 150 million human hosts and is well known to readers of this journal. In spite of hard-fought advances, current therapies remain suboptimal for many patients. What about prenylation?

Prenylation is a site-specific, posttranslational modification of proteins whereby a prenyl lipid is covalently attached to a cysteine residue at or near the carboxyl end of the protein.1 The prenyl lipids consist of farnesyl or geranylgeranyl. Both are derived in a series of reactions starting from mevalonate. The fully synthesized prenyl lipids are then attached to protein substrates in reactions catalyzed by farnesyltransferase (FTase) or geranylgeranyltransferase (GGTase). The substrates recognized by these enzymes have characteristic sequence motifs. FTase and GGTase-I modify proteins having a “CXXX-box” motif, where C is a cysteine and X is one of the last three amino acids at the protein's carboxyl terminus. GGTase-II has a more complex substrate recognition and adds geranylgeranyl to cysteines contained in proteins ending with motifs such as CXC or CC. Lamin B and Ras are examples of farnesylated proteins. Geranylgeranylated proteins include the γ-subunit of G-proteins and the Rab proteins. One important consequence of prenylation is to make the modified protein more hydrophobic, which can serve to promote its interaction with membranes. The prenyl moiety can also function as a ligand, mediating interaction with a specific protein receptor.2 Host cells have evolved a variety of cell signaling and vesicular membrane trafficking pathways that exploit prenylation. The latter has not escaped the parasitic penchants of viruses as well. Certainly the aforementioned types of host cell functions are ideal candidates to be hijacked to establish or maintain the HCV membrane-associated replication complex. Indeed, Ye et al. now describe yet another example of how prenylation appears to be too seductive of a posttranslational modification for viruses to ignore.

The cells used for their experiments are human liver tumor-derived Huh-7 cells that have been modified to contain self-replicating HCV replicons. The latter consist of bicistronic RNA molecules, which encode the HCV nonstructural proteins thought to constitute the viral replication machinery, as well as the conserved 5′ and 3′ nontranslated regions at the ends of the HCV genomic RNA, which are presumed recognition sequences for the viral replication complex. Although they are not a complete substitute for a culture system permitting infection with natural HCV virions, replicon cells represent the state-of-the-art for studying HCV replication. Their development3, 4 has revolutionized the field, has been the subject of an excellent recent review in this series,5 and has provided an ideal system in which to begin to explore the role of prenylation in the HCV life cycle.

For this study, the authors began with a pharmacological experiment. The first in a long series of reactions leading to protein prenylation is the reduction of hydroxy-3-methyl-glutaryl-CoA (HMG-CoA) by HMG-CoA reductase to yield mevalonate. Many inhibitors of this reaction are available because mevalonate production is also the committed step in cholesterol biosynthesis. One of the earliest developed specific inhibitors of HMG-CoA reductase is lovastatin. Treatment of replicon-harboring cells with 50 μmol/L lovastatin resulted in decreased HCV RNA levels. This decrease was accompanied by decreases in accumulated levels of HCV proteins relative to control actin measurements. These decreases were observed with 3 different replicon clones, suggesting they were not dependent on the presence of a specific HCV construct.

The effect of lovastatin administration on HCV RNA levels could be mitigated by cotreatment with either mevalonate or geranylgeraniol, but not farnesol or oleate. This result suggested that HCV replication is dependent on a molecule derived from, or containing either of, the first two compounds.

In addition to allowing for measurements of steady-state viral RNA levels, cells harboring HCV replicons permit evaluation of other parameters associated with presumed authentic HCV replication. In particular, the viral nonstructural proteins display a characteristic morphology consisting of a specklelike pattern in the cell cytoplasm. Bromouridine labeling of nascent RNA-directed RNA synthesis also colocalizes to these speckles, supporting the notion that they represent the sites of HCV replication complexes.6 Treatment of cells for 48 hours with 50 μmol/L lovastatin resulted in disruption of the speckle pattern. Instead, an intracellular redistribution of the HCV proteins into a more diffuse staining was observed by specific immunofluorescence. Again, this effect was reversible by the addition of geranylgeraniol but not farnesol. One logical speculation was that inhibition of the prenylation of at least one protein that is normally modified by geranylgeranyl is required for maintenance of the HCV replication complex and, by extension, viral RNA levels.

To more directly test this hypothesis, cells were treated with a GGTase-I inhibitor (GGTI) and the effect on HCV RNA levels and speckle formation assessed. As a control, an inhibitor of FTase was added to parallel cultures of cells. Although the specificity of the inhibitors was suboptimal (with both drugs inhibiting farnesylation and geranylgeranylation of transfected marker proteins), the GGTI displayed a greater degree of geranylgeranylation inhibition,which correlated with a greater degree of inhibition of HCV RNA levels and presumably complex assembly. The authors concluded that geranylgeranylation of host protein(s) mediates HCV replication. Overall, this is an exciting study from a group of outstanding scientists who collectively have made major contributions over the years to the study of both HCV and prenylation. A few limitations, however, should be noted.

Lovastatin is a specific inhibitor of HMG-CoA-reductase, which catalyzes an early step in the cholesterol and prenyl lipid biosynthetic pathways. Inhibition of HMG-CoA-reductase is not, however, a very selective way to inhibit protein prenylation, because several of the intermediates from mevalonate to farnesyl (or geranylgeranyl) also feed other important metabolic pathways such as dolichol and ubiquinone (coenzyme Q) synthesis. As such, the concentration of lovastatin needed to inhibit downstream reactions such as protein prenylation is associated with significant cytotoxicity. Indeed, as the authors point out, the concentration of lovastatin used in their in vitro experiments would be too high for use in vivo. Even in the in vitro studies, protein synthesis was inhibited by 50%. Under these conditions, nonspecific or indirect effects and mechanisms besides simply inhibiting the prenylation of a particular protein may be operative. Similar caveats need to be considered with inhibitors of protein prenyltransferases, although the latter are expected to have less pleiotropic effects than lovastatin. Pure FTase inhibitors (FTIs) may be better tolerated than GGTIs because more host cell proteins are geranylgeranylated than farnesylated. Nevertheless, simultaneous assessment for nonspecific effects on cell growth, metabolism, or protein synthesis would help strengthen the case for a selective anti-HCV effect of the GGTI used. In addition, the total decrease in HCV RNA levels appeared to be on the order of 1-2 logs, which by itself is on the lower end of the spectrum for general antivirals. With optimization, more potent compounds, or in the setting of combination therapy, however, this efficacy could be significantly improved. Finally, no direct experiments were presented to examine for prenyl lipid modification of viral proteins or how such modifications may have been affected by some of their drug treatments.

The work by Ye et al. does add to a growing list of viruses that have been reported to depend on prenylation. For example, somewhat similar to the current study, antiviral activity of lovastatin against respiratory syncytial virus has been reported.7 Disrupting geranylgeranylation of the host cell protein RhoA was proposed as the mechanism. A dependence of herpes simplex virus on farnesylation of host cell activated Ras protein has been described.8 In addition to interactions with host cell prenylated proteins, direct binding of viral proteins with a host cell enzyme involved in prenyl lipid synthesis has also been observed.9 Finally, as alluded to above, the prenylated protein exploited by the virus can be encoded in the viral genome as well.

Indeed, an impressive array of viruses encode proteins with the requisite substrates for prenylation.10 A prototypical example is hepatitis delta virus, where prenylation of its coatlike protein delta antigen is essential for virus assembly.11 The availability of specific prenylation inhibitors has enabled direct testing of the hypothesis that pharmacological inhibition of prenylation can successfully inhibit prenylation of a viral protein with dramatic disruption of virus assembly in both in vitro12 and in vivo13 models.

The use of prenylation inhibitors to inhibit the prenylation of a viral protein, or a host protein needed by the virus, are not mutually exclusive. In either case, this represents an example of a new approach to antiviral therapy. In contrast to classical antivirals, which target a virus-specific enzyme, prenylation inhibitors target a host cell enzyme and thereby seek to deprive the virus of access to a host cell function. One interesting consequence may be to make the development of resistance a more difficult task, as the targeted locus is not under genetic control of the virus.

What are the specific proteins whose impaired prenylation inhibits HCV? The answer may soon be forthcoming and should be interesting. Whatever the target, prenylation appears to serve a critical purpose in the HCV life cycle. Yet, like most pathogens, HCV does not encode its own prenylating enzymes. Perhaps the overreliance on its host cell can be exploited to accelerate its demise.