doi:10.1016/j.jmb.2006.01.032 J. Mol. Biol. (2006) 357, 1051–1057

COMMUNICATION

Novel Inhibitors of Hepatitis C Virus RNA-dependentRNA Polymerases

Gary Lee, Derek E. Piper, Zhulun Wang, John Anzola, Jay PowersNigel Walker and Yang Li*

Amgen Inc., 1120 VeteransBlvd., South San FranciscoCA 94080, USA

0022-2836/$ - see front matter q 2006 E

Abbreviations used: HCV, hepatiPOUND1, 5-(4-chlorophenylmethylsulfonylamino)-4-oxxo-2-thionothiaPOUND2, 5-(4-bromophenylmethyl(benzenesulfonylamino)-4-oxxo-2-thRdRp, RNA-dependent RNA polymthroughput screening; BVDV, bovin

E-mail address of the correspondyangl@amgen.com

Hepatitis C virus (HCV) is a major cause of chronic hepatitis, liver cirrhosis,and hepatocellular carcinoma worldwide-and is the main cause of adultliver transplants in developed nations. We have identified a class of noveland specific inhibitors of HCV NS5B RNA-dependent RNA polymerase(RdRp) activity in vitro. Characterization of two such inhibitors,COMPOUND1 (5-(4-chlorophenylmethylene)-3-(benzenesulfonylamino)-4-oxxo-2-thionothiazolidine) and COMPOUND2 (5-(4-bromophenyl-methylene)-3-(benzenesulfonylamino)-4-oxxo-2-thionothiazolidine), isreported here. With IC50 values of 0.54 mM and 0.44 mM, respectively,they are reversible and non-competitive with nucleotides. Biochemical andstructural studies have suggested that these compounds can inhibit theinitiation of the RdRp reaction. Interestingly, these inhibitors appear toform a reversible covalent bond with the NS5B cysteine 366, a residue thatis not only conserved among all HCV genotypes and a large family ofviruses but also required for full NS5B RdRp activity. This may reduce thepotential resistance of the viruses to this class of inhibitors.

q 2006 Elsevier Ltd. All rights reserved.

Keywords: HCV; RNA-dependent RNA polymerase; NS5B; inhibitor;structure

*Corresponding author

Hepatitis C virus (HCV), of the family Flavivir-idae, is a positive-strand, enveloped RNA virus andan important human pathogen.1,2 Persistent HCVinfections lead to the development of chronichepatitis, cirrhosis, and hepatocellular carcinoma,and is currently recognized as the leading cause ofchronic liver disease worldwide.2,3 The currentstandard-of-care, comprised of PEGylated inter-feron-a in combination with ribavirin,4–8 showsrelatively poor efficacy, significant side-effects, andis poorly tolerated. Therefore, intensive efforts havebeen directed at the discovery of novel drugs totreat this disease.

The HCV genome encodes a polyprotein (w3000amino acid residues) that is co and post-

lsevier Ltd. All rights reserve

tis C virus; COM-ene)-3-(benzene-zolidine; COM-ene)-3-ionothiazolidine;erase; HTS, high-

e viral diarrhea virus.ing author:

translationally processed by host and viral pro-teases into ten mature viral proteins withthe order NH2-C-E1-E2-p7-NS2-NS3-NS4A-NS4B-NS5A-NS5B-COOH.9,10 The RNA-dependent RNApolymerase (RdRp), encoded by NS5B, is thecentral enzyme in replication of the HCV genomeand represents an ideal target for small moleculedrugs.6,8,11–13 The screening for such inhibitors isnow feasible, as both the full-length and truncatedforms (NS5BD21, devoid of the C-terminal 21hydrophobic amino acid residues potentiallyimportant for anchoring NS5B to the perinuclearmembrane14) of NS5B have been purified frominsect cells and Escherichia coli,14–23 and theenzymatic activities of these various forms ofNS5B have been demonstrated on both homo-polymeric and heteropolymeric RNA templates.The structure of NS5B has been solved by X-raycrystallography.24–26 While canonical polymerasefeatures exist in the structure, NS5B displaysseveral unique and mechanistically importantfeatures. In particular, a loop region (b-loop,amino acid residues 443 to 454) may formimportant interactions with the template, rNTP,and the double-stranded RNA intermediate.24–26,31

d.

1052 Novel Inhibitors of HCV RdRp

Here, we report on the discovery a class of novel,small molecule inhibitors of HCV NS5B. Ourbiochemical and structural studies have demon-strated this class of inhibitors to be selective,reversible, and non-competitive with nucleotide.Further mechanistic studies also suggest that theseinhibitors can inhibit the initiation step of thereplication process.

Purification and characterization of HCV NS5Bprotein

C-terminal His-tagged HCV NS5B (BK strain23)protein devoid of the last 21 hydrophobic aminoacid residues18,20 was expressed in bacteria andpurified to near homogeneity on a nickel-chelated(Ni-NTA) column (Figure 1(a)). The eluant from theNi-NTA column was subsequently purified on agel-filtration column, and the pooled peak fractionswere used for all experiments described here (datanot shown).

To facilitate the characterization of NS5B RdRpactivity, a radioactive assay using 96-well plates wasdeveloped with conditions similar to previouslydescribed procedures23 with the following modifi-cations. Reactions were carried out in 96-well plateswithout pre-incubation of the enzyme and template,and were spotted onto Zeta-Probe membranes(BioRad Laboratories, Hercules, CA) at the end ofa 60 min reaction to capture 32P-labeled RNAproducts. The filters were repeatedly washed with0.5 M sodium phosphate (pH 7.0) until no moreradioactivity was detected in the wash buffer.Labeled products bound to the filters were quanti-tated on a Fuji Bio-imager. Reactions contained40 nM enzyme, 50 nM RNA template, 5 mM ATP,1 mM UTP, 1 mM GTP, 0.05 mM cold CTP, 0.5 mCi of[a-32P]CTP (800 Ci/mmol; Amersham), 20 mM Tris(pH 7.5), 30 mM KCl, 2% (v/v) glycerol, and 2 mMMgCl2 in a total volume of 50 ml. In this assay format,NS5B displayed a strong preference for an RNAtemplate and exhibited very little activity on single-stranded DNA or M13 DNA template, consistent

Figure 1. (a) Purification of 6! His-tagged NS5B from bainsoluble, insoluble pellet from WCE; input, Ni-NTA loaoptimization for NS5B RdRp in vitro.

with single-stranded RNA being its natural tem-plate (data not shown).27 The single-stranded RNAtemplate used for all subsequent experimentscontained the last 567 nucleotide residues from the3 0 end of the HCV genome (including portions ofNS5B and the entire 3 0 UTR). The predominantproduct of the NS5B RdRp is a tight RNA hybridbetween the template and the newly formed strandinitiated from the 3 0-OH group of the template,similarly to earlier reports (data not shown).23,15

Evaluation of the optimal reaction temperaturesuggests that NS5B can operate under a widerange of temperatures (Figure 1(b)). From 25 8C(room temperature, RT) to 41 8C, reaction velocityincreases with temperature; the enzymatic activity islowest at RT and approaches a maximum at 41 8C.Experiments described in the subsequent sectionswere carried out at 37 8C.

Identification of benzylidene derivatives aspotent inhibitors of HCV NS5B RdRp activity

In a search for novel HCV NS5B inhibitors,benzylidene derivatives were found to be the mostpotent of the several different chemical series ofinhibitors identified by high-throughput screening(HTS) of a random chemical compound library. TheHTS was carried out with conditions described inthe legend to Figure 1 except with the followingmodification. At the end of the reaction, instead ofspotting the reaction products onto Zeta-Probemembranes, the reaction products were capturedand subsequently washed in 96-well plates withDE52 anion-exchange paper at the bottom (Poly-filtronics Inc, Rockland, MA). This format wasnecessary to accommodate automated HTS pro-cesses; although the Zeta-Probe membranes gavemore consistent and robust results and thereforewere used for all the characterizations describedhere. Two structurally related benzylidene com-pounds, COMPOUND1 and COMPOUND2, werefound to inhibit NS5B activity with IC50 values of0.54 mM and 0.44 mM, respectively (Figure 2(a) and

cteria using a Ni-NTA column. WCE, whole cell extract;d; flow thru, Ni-NTA flow through. (b) Temperature

Figure 2. Identification of benzylidene derivatives as inhibitors of HCV NS5B RdRp. (a) Chemical structures forCOMPOUND1 and COMPOUND2. (b) Titration of COMPOUND1 and COMPOUND2 in the NS5B RdRp assay.(c) Characterization of products of NS5B RdRp assay in a denaturing polyacrylamide gel in the presence and absence ofinhibitors. Lane 1, no NS5B added; lane 2, NS5B alone without inhibitors.

Novel Inhibitors of HCV RdRp 1053

(b)). These two compounds also showed a highdegree of selectivity against HCV NS5B comparedto bovine viral diarrhea virus (BVDV) NS5B andalfalfa mosaic virus (AMV) reverse transcriptase,both BVDV NS5B and AMV RT had IC50 valuesgreater than 30 mM (data not shown).

The products of the reactions containing thesetwo inhibitors were separated and characterized ona denaturing gel with the following reactionconditions: the amount of enzyme was reduced to20 nM, RNA template to 10 nM, and [a-32P]CTP to0.1 mCi. Assays were incubated for 30 min, andthen stopped with EDTA and then separated in adenaturing 5%(w/v) polyacrylamide gel (acryl-amide/bis-acrylamide 19:1, w/w) containing 7 Murea in 1!89 mM Tris–borate (pH8.1), 2 mM EDTA(TBE). Full-length product formation decreasedwith increasing concentrations of inhibitors(Figure 2(c)). Interestingly, reactions containinginhibitors did not appear to produce additionalshorter products due to abortive initiation orpremature termination.

Mechanism of inhibition by COMPOUND1and COMPOUND2

To better understand how these compoundsinhibit the polymerase activity, COMPOUND2

was co-crystallized with NS5B. The NS5B/inhibitorcomplex was solved to a resolution of 2.15 A, thedetails of which will be described elsewhere.32 TheCOMPOUND2 binding pocket is found betweenthe thumb and fingers domains, and includesresidues from the b-loop and the C-terminal regionthat precedes the hydrophobic tail.25 The majorityof the interactions between COMPOUND2 and theNS5B protein are van der Waals interactions, withone direct hydrogen bond and two water-mediatedhydrogen bonds also observed between the com-pound and the protein (Figure 3(a) and (b)). Mostinterestingly, the benzylidene alkene group ofCOMPOUND2 forms a covalent bond withCys366 of NS5B in the crystal structure(Figure 3(b)). Cys366 is located in motif E, whichis found only in RNA-dependent polymerases andis located at the junction between the palm andthumb sub-domains.26,28 The covalent bond formedwith the Cys366 residue is also seen in the solvedcrystal structures of all other compounds from thebenzylidene series bound to the NS5B protein.32 Itis, therefore, reasonable to assume that COM-POUND1 will also form this covalent bond withCys366, and this might be a general feature of thisbenzylidene class of inhibitors.

The crystal structure described here suggeststwo possible effects of compound binding on

Figure 3. Mechanism of action studies of COMPOUND1 and COMPOUND2. (a) Structure of NS5B bound toCOMPOUND2. NS5B is shown in the standard orientation, with the fingers domain (violet) at the top left, the thumbdomain (light brown) at the top right and the palm domain (cyan) at the bottom. COMPOUND2 is shown in green sticksbelow the b-loop (yellow). (b) The NS5B/COMPOUND2 complex. COMPOUND2 is shown in green, residues of NS5Bthat interact with the compound are shown in black, and the covalent bond between Cys366 and COMPOUND2 isshown in red. Hydrogen bonds are indicated by broken lines. (c) Comparison of wild-type (wt) and C366G mutant NS5B

1054 Novel Inhibitors of HCV RdRp

Novel Inhibitors of HCV RdRp 1055

polymerase activity. First, the interaction betweenthe compound and the b-loop likely stabilizes theposition of the b-loop to prevent the binding of thetemplate to the polymerase. As will be discussedbelow, data presented in Figure 3(f) support thishypothesis. Second, modeling an RNA template/primer into the NS5B/inhibitor complex suggeststhat COMPOUND2 binding would not overlapwith template binding, but would block theinitiation of a newly synthesized strand.32 Theseobservations suggest that COMPOUND2, and byextrapolation the benzylidene series, may inhibitthe initiation step of the RdRp reaction. Consistentwith this speculation, no abortive or prematurelyterminated products were found in the reactionscontaining these inhibitors (Figure 2).

A cysteine-to-glycine mutation (C366G) wasconstructed for NS5BD21 to study the importanceof Cys366 on inhibitor binding. This C366Gmutation had a profound effect on NS5B RdRpactivity, as only 20% of the wild-type activity couldbe detected at the highest level of enzyme tested(Figure 3(c)). The importance of this cysteineresidue for RdRp can be inferred from its conser-vation among not only Flaviviridae (Figure 3(d)) butalso a large family of RNA viruses.28 It is notsurprising then to see that the equivalent cysteineresidue (Cys497) in BVDV, when mutated toalanine, also reduced the RdRp activity to 10% ofwild-type levels.29 The C366G mutation had aprofound effect on the ability of COMPOUND1and COMPOUND2 to inhibit NS5B RdRp activity,with IC50 values increased to 30.3 mM and 31.8 mM,respectively (Figure 3(e)). Consistent with thisobservation, the removal of the double bond fromthe benzylidene alkene group also similarlyincreased the IC50 values of those inhibitors.32 TheIC50 value of an inhibitor from diketoacid deriva-tives, a different structural class from the benzyl-idene series, was unaffected by this mutation(unpublished observations),33 suggesting that theloss of inhibition by COMPOUND1 and COM-POUND2 could not simply be attributed tomisfolding of the mutant enzyme. These resultssuggest that Cys366 is indeed very important forinhibitor action. Other specific interactions betweeninhibitors and enzyme also contribute to binding,since COMPOUND1 and COMPOUND2 did notlose their ability to inhibit NS5B C366G enzymecompletely (Figure 3(e)).

The binding properties of the inhibitors wereexamined further. NS5B was pre-incubated witheither inhibitor or template first before the additionof the remaining components. Pre-incubation of

RdRp activities. (d) Alignment of sequences surrounding Cysvirus (BVDV), dengue fever virus type 4 (DEN4), and yellow(*), and motif E designation are indicated at the top.26,28 (e) TiRdRp assay against NS5B C366G mutant. (f) Effects of pre-incRNA template; I, inhibitor COMPOUND1; Sub, nucleotide suThe IC50 values of COMPOUND1 from the three different treaanalysis at different COMPOUND1 concentrations.

COMPOUND1 with NS5B before the addition oftemplate resulted in a modest shift in IC50 valuefrom 0.54 mM to 0.29 mM. If the pre-incubatedinhibitor/NS5B complex was dialyzed for 1 hprior to the start of the reaction, NS5B RdRp activitycould be fully recovered (unpublished obser-vations). These results suggest that if a covalentbond is formed in solution between COMPOUND1and Cys366 as seen in the crystal structures betweencompounds and NS5B (Figure 3(b)), it is reversibleunder the in vitro RdRp reaction conditions(Figure 3(f)). Incubation of NS5B and templatebefore the addition of COMPOUND1 resulted in anupward IC50 shift to 3.1 mM (Figure 3(f)). Thisobservation is consistent with crystal structure data,indicating that these inhibitors should bind morefavorably to NS5B in the absence of template, andlends additional support for the hypothesis thatthese compounds can block initiation of thepolymerase reaction as discussed before. Thereaction kinetics were determined and Lineweaver–Burk plots of the 1/initial velocity (1/vo) versus 1/[substrate] (1/[S]) suggest that the mode ofinhibition by COMPOUND1 is non-competitiverespective to nucleotide substrates, as all of thelines converged reasonably well at the intercepts onthe 1/[S] axis (Figure 3(g)).

Here, we have described the discovery and initialcharacterization of benzylidene derivatives as noveland potent HCV NS5B RdRp inhibitors. Biochemi-cal and structural studies with COMPOUND1 andCOMPOUND2, representative examples from theseries, suggest that the compounds bind to thepolymerase in a reversible fashion, are non-competitive with nucleotide substrates and caninhibit the initiation of RNA synthesis. Our resultsalso demonstrate the importance of the Cys366residue for benzylidene-class inhibitors and a novelsite on the NS5B enzyme that could be targeted forfuture inhibitor development. The conservation ofCys366 among not only different HCV genotypesbut also a large family of viruses suggests thatbenzylidene inhibitors may have the potential toinhibit other HCV genotypes as well. This alsoraises an interesting possibility that benzylideneinhibitors could be developed for other human andanimal pathogens. Indeed, certain benzylidenederivatives has been shown to inhibit the RdRp ofmembers of pestiviruses and flaviviruses.27,30 Inaddition, the importance of Cys366 to RdRppolymerase activity may reduce the potentialresistance of the viruses to this class of inhibitors.Additional experiments are required to explorethese possibilities further. The effects of this class of

366 of HCV to similar regions from bovine viral diarrheafever virus (YFV). Identical residues, conserved residues

tration of COMPOUND1 and COMPOUND2 in the NS5Bubation on COMPOUND1 activities. E, NS5B enzyme; T,bstrates. The time of incubation for each step is indicated.tments are shown at the bottom. (g) Lineweaver–Burk plot

1056 Novel Inhibitors of HCV RdRp

inhibitors on HCV viral replication and thepotential development of novel therapeutics fromthis class of compounds for HCV infection will bethe subject of future studies.

Acknowledgements

We thank Haoda Xu, Xiaohong Liu, and MeganVan Overbeek for technical support, Juan Jaen andGreg Peterson for critical reading of the manuscriptand helpful discussions.

References

1. Major, M. E. & Feinstone, S. M. (1997). The molecularvirology of hepatitis C. Hepatology, 25, 1527–2538.

2. Leyssen, P., De Clercq, E. & Neyts, J. (2000).Perspectives for the treatment of infections withFlaviviridae. Clin. Microbiol. Rev. 13, 67–82.

3. Alter, M. J., Margolis, H. S., Krawczynski, K.,Judson, F. N., Mares, A., Alexander, W. J. et al.(1992). The natural history of community-acquiredhepatitis C in the United States. N. Engl. J. Med. 327,1899–1905.

4. Myles, D. C. (2001). Recent advances in the discoveryof small molecule therapies for HCV. Curr. Opin. DrugDiscov. Dev. 4, 411–416.

5. Wang, Q. M. & Heinz, B. A. (2001). Recent advances inprevention and treatment of hepatitis C virusinfections. Prog. Drug Res. Spec. 00, 79–110.

6. De Francesco, R. & Rice, C. M. (2003). New therapieson the horizon for hepatitis C: are we close? Clin. LiverDis. 7, 211–242.

7. Dillon, J. F. (2004). Hepatitis C: what is the besttreatment? J. Viral Hepat. 11, 23–27.

8. Ni, Z. J. & Wagman, A. S. (2004). Progress anddevelopment of small molecule HCV antivirals. Curr.Opin. Drug Discov. Dev. 7, 446–459.

9. Moradpour, D., Wolk, B., Cerny, A., Heim, M. H. &Blum, H. E. (2001). Hepatitis C: a concise review.Minerva. Med. 92, 329–339.

10. Shi, S. T. & Lai, M. M. (2001). Hepatitis C viral RNA:challenges and promises. Cell. Mol. Life Sci. 58,1276–1295.

11. Hagedorn, C. H., van Beers, E. H. & De Staercke, C.(2000). Hepatitis C virus RNA-dependent RNApolymerase (NS5B polymerase). Curr. Top. Microbiol.Immunol. 242, 225–260.

12. Hugle, T. & Cerny, A. (2003). Current therapy andnew molecular approaches to antiviral treatment andprevention of hepatitis C. Rev. Med. Virol. 13, 361–371.

13. Wu, J. Z. & Hong, Z. (2003). Targeting NS5B RNA-dependent RNA polymerase for anti-HCV chemother-apy. Curr. Drug Targets Infect. Disord. 3, 207–219.

14. Yamashita, T., Kaneko, S., Shirota, Y., Qin, W.,Nomura, T., Kobayashi, K. & Murakami, S. (1998).RNA-dependent RNA polymerase activity of thesoluble recombinant hepatitis C virus NS5B proteintruncated at the C-terminal region. J. Biol. Chem. 273,15479–15486.

15. Behrens, S. E., Tomeri, L. & DeFrancesco, R. (1996).Identification and properties of the RNA-dependentRNA polymerase of hepatitis C virus. EMBO J. 15,12–22.

16. De Francesco, R., Behrens, S. E., Tomei, L., Altamura,S. & Jiricny, J. (1996). RNA-dependent RNA poly-merase of hepatitis C virus. Methods Enzymol. 275,58–67.

17. Lohmann, V., Korner, F., Herian, U. & Bartenschlager,R. (1997). Biochemical properties of hepatitis C virusNS5B RNA-dependent RNA polymerase and identi-fication of amino acid sequence motifs essential forenzymatic activity. J. Virol. 71, 8416–8428.

18. Al, R. H., Xie, Y., Wang, Y. & Hagedorn, C. H. (1998).Expression of recombinant hepatitis C virus non-structural protein 5B in Escherichia coli. Virus Res. 53,141–149.

19. Lohmann, V., Roos, A., Korner, F., Koch, J. O. &Bartenschlager, R. (1998). Biochemical and kineticanalyses of NS5B RNA-dependent RNA polymeraseof the hepatitis C virus. Virology, 249, 108–118.

20. Ferrari, E., Wright-Minogue, J., Fang, J. W. S.,Baroudy, B. M., Lau, J. Y. N. & Hong, Z. (1999).Characterization of soluble hepatitis C virus RNA-dependent RNA polymerase expressed in Escherichiacoli. J. Virol. 73, 1649–1654.

21. Oh, J. W., Ito, T. & Lai, M. M. (1999). A recombinanthepatitis C virus RNA-dependent RNA polymerasecapable of copying the full-length viral RNA. J. Virol.73, 7694–7702.

22. Zhong, W., Uss, A. S., Ferrari, E., Lau, J. Y. & Hong, Z.(2000). De novo initiation of RNA synthesisby hepatitis C virus nonstructural protein 5Bpolymerase. J. Virol. 74, 2017–2222.

23. Carroll, S. S., Sardana, V., Yang, Z., Jacobs, A. R.,Mizenko, C., Hall, D. et al. (2000). Only a smallfraction of purified hepatitis C RNA-dependent RNApolymerase is catalytically competent: implicationsfor viral replication and in vitro assays. Biochemistry,39, 8243–8249.

24. Lesburg, C. A., Cable, M. B., Ferrari, E., Hong, Z.,Mannarino, A. F. & Weber, P. C. (1999). Crystalstructure of the RNA-dependent RNA polymerasefrom hepatitis C virus reveals a fully encircled activesite [In Process Citation]. Nature Struct. Biol. 6,937–943.

25. Ago, H., Adachi, T., Yoshida, A., Yamamoto, M.,Habuka, N., Yatsunami, K. & Miyano, M. (1999).Crystal structure of the RNA-dependent RNA poly-merase of hepatitis C virus. Struct. Fold. Des. 7,1417–1426.

26. Bressanelli, S., Tomei, L., Roussel, A., Incitti, I.,Vitale, R. L., Mathieu, M. et al. (1999). Crystal structureof the RNA-dependent RNA polymerase of hepatitisC virus. Proc. Natl Acad. Sci. USA, 96, 13034–13039.

27. Ranjith-Kumar, C. T., Gajewski, J., Gutshall, L.,Maley, D., Sarisky, R. T. & Kao, C. C. (2001). Terminalnucleotidyl transferase activity of recombinant Flavi-viridae RNA-dependent RNA polymerases: impli-cation for viral RNA synthesis. J. Virol. 75, 8615–8623.

28. Koonin, E. V. (1991). The phylogeny of RNA-dependent RNA polymerases of positive-strandRNA viruses. J. Gen. Virol. 72, 2197–2206.

29. Lai, V. C., Kao, C. C., Ferrari, E., Park, J., Uss, A. S.,Wright-Minogue, J. et al. (1999). Mutational analysis ofbovine viral diarrhea virus RNA-dependent RNApolymerase. J. Virol. 73, 10129–10136.

30. Bailey, T. R. & Young, D. C. (2000). Compounds,Compositions and Methods for Treating or PreventingViral Infections and Associated Diseases WO-00010573,ViroPharma, Exton, PA.

31. O’Farrell, D., Trowbridge, R., Rowlands, D. & Jager, J.(2003). Substrate complexes of hepatitis C virus

Novel Inhibitors of HCV RdRp 1057

RNA polymerase (HC-J4): structural evidence fornucleotide import and de novo initiation. J. Mol. Biol.326, 1025–1035.

32. Powers, J. P., Anzola, J., Chen, J., Jaen, J. C., Li, Y., Lee,G., et al. (2006). SAR and mechanism of novel HCV

NS5B RNA polymerase inhibitors. J. Med. Chem. In thepress.

33. Altamura, S., Tomei, L., Koch, U., Neuner, P. J. S. &Summa, V. (2000). Diketoacid-derivatives as Inhibitors ofPolymerases WO-0006529, ViroPharm, Exton, PA.

Edited by J. Karn

(Received 3 November 2005; received in revised form 5 January 2006; accepted 6 January 2006)Available online 30 January 2006