Human and Murine Antibody Recognition Is Focusedon the ATPase/Helicase, But Not the Protease Domain

of the Hepatitis C Virus Nonstructural 3 Protein

MARGARET CHEN,1,2 MATTI SALLBERG,1,2 ANDERS SONNERBORG,1,3 LING JIN,4 ASHLEY BIRKETT,5 DARRELL PETERSON,4

OLA WEILAND,3 AND DAVID R. MILICH1,6

The hepatitis C virus (HCV) nonstructural (NS) 3 proteinhas been shown to possess at least two enzymatic domains.The amino terminal third contains a serine-protease do-main, whereas the carboxy terminal two thirds is comprisedof an adenosine triphosphatase (ATPase)/helicase domain.These domains are essential for the maturation of thecarboxy-terminal portion of the HCV polyprotein and cata-lyze the cap synthesis of the RNA genome. In this report,human and murine antibody responses induced by NS3were characterized using a recombinant full-length NS3(NS3-FL) protein, or the isolated protease or ATPase/helicase domains, expressed and purified from Escherichiacoli. Sera from 40 patients with chronic HCV infection wereassayed in enzyme-linked immunoassays (EIAs) for anti-body binding to the panel of NS3 proteins. Virtually allpatient sera contained antibodies specific for NS3-FL andthe ATPase/helicase domain, whereas only 10% of serareacted with the protease domain of NS3. Human antibod-ies reactive with NS3-FL were highly restricted to theimmunoglobulin G1 (IgG1) isotype and were inhibited bysoluble ATPase/helicase, but not by the protease domain.The anti-NS3 (ATPase/helicase) reactivity decreased ondenaturation by sodium dodecyl sulfate (SDS) and b-mer-captoethanol (2ME), suggesting the recognition of nonlin-ear or conformational B-cell determinants. Similar to in-fected humans, mice immunized with NS3-FL developedhigh-titered primary antibody responses to the NS3 ATPase/helicase domain, whereas an anti-NS3 protease responsewas not observed after primary or secondary immuniza-tions. Thus, the human and murine humoral immune

responses to the HCV NS3 protein are focused on theATPase/helicase domain, are restricted to the IgG1 isotypein humans, and are conformationally dependent. Unexpect-edly, in both species, the NS3 protease domain, present inthe context of the full-length NS3, appears to possess lowintrinsic immunogenicity in terms of antibody production.(HEPATOLOGY 1998;28:219-224.)

The hepatitis C virus (HCV) is a recent member in thevirus family Flaviviridae, and has been classified as a separategenus from the pestiviruses and the flaviviruses.1,2 Theprotein sequences of the three genera share short stretches ofhomologies that, upon comparison at the molecular level, ledto the identification of the putative structural organization ofthe HCV polyprotein.3,4 This was soon followed by in vitrotranscriptional and translational analysis with the character-ization of functional properties of the predicted viral genes.Data available today indicate that the HCV polyprotein-NH2-core (C)-envelope 1 (E1)-E2-nonstructural protein 2 (NS2)-NS3-NS4a-NS4b-NS5a-NS5b-COOH is encoded by a singlepositive-stranded RNA genome of approximately 9,400 nucle-otides. Like other members in the family Flaviviridae, post-translational processing of this precursor protein is requiredto generate the HCV viral particles and the individual viralproteins needed for its replicative cycle. The processing of thecore (nucleocapsid protein), E1, and E2 (envelope proteins)involves host signal peptidases,5 while the HCV NS3 proteinis required for maturation of the NS4a, NS4b, NS5a, andNS5b proteins.6-9

The NS3 serine protease activity has been found to beassociated with the amino-terminal portion of the NS3protein. The NS2/3 boundary is cleaved by a second proteasethat includes the carboxy-terminus of NS2 and overlaps withthe serine protease.8,10 The enzymatic activity of the NS3protease is partly dependent on formation of a complex withNS4a.9,11-15 Chraracterization of the functional properties ofthe carboxy-terminal portion of NS3 of flaviviruses and HCVhave revealed adenosine phosphatase (ATPase) and helicaseactivities that are involved in viral replication.16,17

A portion of the NS3 sequence expressed as a fusionprotein with superoxide dismutase (C33) appears to beserologically reactive during the early phase of HCV infec-tions, and for this reason, C33 is routinely incorporated inclinical diagnostic HCV antibody immunoassays.18 Althoughthe C33 polypeptide is derived from the NS3 region, it doesnot incorporate the entire ATPase/helicase domain and onlycontains the extreme C-terminus of the protease domain.18

Abbreviations: HCV, hepatitis C virus; NS, nonstructural; ATPase, adenosine triphos-phatase; FL, full length; E2, envelope 2; IgG, immunoglobulin G; aa, amino acid; EIA,enzyme-linked immunoassay; OD, optical density; SDS-2ME, sodium dodecyl sulfateb-mercaptoethanol.

From the 1Department of Molecular Biology, The Scripps Research Institute, La Jolla,CA; Divisions of 2Clinical Virology, and 3Infectious Diseases, Department of Immunol-ogy, Microbiology, Pathology and Infectious Diseases, Karolinska Institutet, HuddingeUniversity Hospital, Sweden; 4Department of Biochemistry and Molecular Biophysics,Virginia Commonwealth University, Richmond, VA; and 5Immune Complex Corpora-tion, San Diego, CA.

Received December 23, 1997; accepted March 17, 1998.Supported by grants from the NIH (AI 20720) and from the Swedish Medical

Research Council (B95-16X-11219-01A), and the Swedish Work Enviroment Fund.This is publication number 11328-MB from TSRI.

Address reprint requests to: David R. Milich, Ph.D., Department of Molecular BiologyCAL-2, The Scripps Research Institute, 10550 North Torrey Pines Road, La Jolla, CA92037. Fax: (619) 784-7660.

Copyright r 1998 by the American Association for the Study of Liver Diseases.0270-9139/98/2801-0028$3.00/0

219

Therefore, it has not been possible to discriminate betweenprotease- and helicase-specific antibody binding using thisreagent. Human antibodies specific for HCV proteins such asthe core, NS4, and NS5 recognize linear determinants19-21;however, recognition of linear determinants within the NS3region appears limited.22,23 This suggests that NS3 maycontain predominantly confomational determinants recog-nized by human antibodies; however, this conclusion remainscontroversial.23,24

Because HCV NS3 is a multifuctional protein essential inthe HCV life cycle, it may be a potential antiviral target for theimmune system. In this report, we characterized human andmurine antibody responses elicited to the full-length NS3(NS3-FL) protein, and to the isolated protease and ATPase/helicase domains expressed in Escherichia coli as two separatesubunits. Our results indicate that the antibody response toNS3 in both infected humans and immunized mice is almostexclusively directed to the ATPase/helicase domain. In addi-tion, the human anti-NS3 response is highly restricted to asingle immunoglobulin G (IgG) isotype, IgG1. Furthermore,denaturation of the helicase domain destroys antibody bind-ing, suggesting the presence of conformationally dependentB-cell determinants.

MATERIALS AND METHODS

Human Sera. Sera were obtained from 40 patients with chronicHCV infection of a duration of more than 2 years who were followedat the Division of Infectious Diseases, Huddinge Hospital, Stock-holm, Sweden. All patients had elevated alanine transaminase (ALT)levels and antibodies to HCV as determined by commercial assays(Anti-HCV immunoassay, second generation, Abbott Laboratory,Chicago, IL). The reactivities were confirmed by RIBA II (OrthoDiagnostics, Raritan, NJ). Serum HCV RNA was detected in allpatients by amplification of the HCV cDNA by polymerase chainreaction.25 Genotyping and quantification of the serum HCV RNAhas been performed in a previous study,26 which demonstrated that9, 4, 14, and 9 patients were infected with HCV genotype 1a, 1b, 2a,or 3, respectively, and mixed genotypes were present in 4 patients.The range of HCV-RNA levels in serum was estimated between2.5 3 104 to 16 3 107 genomes per milliliter, and the median was2.25 3 106 genomes per milliliter. Liver histology showed persistentor active hepatitis in all patients and signs of cirrhosis in severalpatients. None of these subjects had ongoing hepatitis A virus,hepatitis B virus, human immunodeficiency virus, cytomegalovirus,or Epstein-Barr virus infections, and none were treated with antiviraltherapy.

Animals. C57BL/10 (B10) xB10.S F1 mice were obtained from thebreeding colony at the Scripps Research Institute. All animalsreceived humane care according to the criteria outlined in the Guidefor the Care and Use of Laboratory Animals. Groups of three to fivemice were immunized intraperitoneally with 20 µg of the NS3-FL orATPase/helicase domain, and 100 µg of the serine protease domainemulsified in Freund’s complete adjuvant. Booster injections weregiven at day 14 in incomplete adjuvant. Serum samples werecollected 2 weeks after each immunization.

Recombinant HCV NS3 Antigens and Synthetic Peptides. Full-lengthNS3 (amino acids [aa] 1006-1612), the serine protease domain (aa1026-1207), and the ATPase/helicase domain (aa 1207-1612) of NS3were expressed in transformed E. coli as described27 using standardprotocols. Solubilized proteins were purified and concentrationswere determined by spectrophotometry. The serine protease orhelicase enzymatic activities were verified27 (L. Jin and D. L.Peterson, Unpublished observation, June 1996). Before use asantigens, the preparations were extensively dialyzed against phos-phate-buffered saline.

The production of a total of 39 synthetic peptides, each 16 aa long

with a 6-aa overlap, spanning the complete HCV NS3 sequence (aaresidues 1007-1606) has been described previously.28

HCV NS3 Specific Solid-Phase Enzyme-Linked Immunoassays for Humanor Mouse Sera. Recombinant proteins and synthetic peptides wereused as solid-phase antigens passively adsorbed to microtiter platesin 50 mmol/L sodium carbonate buffer (pH 9.6) at 0.4 µg/mL and10 µg/mL, respectively. The enzyme-linked immunoassay (EIA)protocols were as described.29 In brief, dilutions of human sera of1:100, or serial dilutions of mouse sera starting at 1:40, were addedto the coated microtiter plates. Bound human antibodies weredetected by peroxide-labeled goat anti-human IgG antibody (Sigma,St. Louis, MO). Bound mouse antibodies were detected by peroxidase-conjugated anti-mouse IgG 1 IgM (Boehringer Mannheim, India-napolis, IN). After adding substrate, the optical density (OD) wasdetermined at 480 nm in a spectrophotometer. The assay cut-off forthe human EIA was calculated by adding the mean OD480 value ofeight anti-HCV–negative human sera plus 3 SD. The cut-off valuefor the murine EIA was three times the OD480 values of preimmu-nized mouse sera.

EIA for IgG Isotyping the NS3 Reactive Sera. The EIA protocol forhuman IgG isotyping was performed as described above, except thatthe detection of bound human IgG subclasses was performed byusing mouse anti-human IgG1 (1:3,000), IgG2 (1:500), IgG3(1:3,000), or IgG4 (1:500) specific monoclonal antibodies obtainedfrom Zymed Laboratories (San Francisco, CA). The reaction wasfollowed by incubation with peroxidase-conjugated goat anti-mouseIgG 1 IgM diluted 1:2,000. The mouse IgG isotyping was performedby using goat anti-mouse IgG1, IgG2a, IgG2b, or IgG3 (1:1,000) assecond antibodies, followed by incubation with peroxidase-labeledswine anti-goat IgG antibodies (1:20,000). The OD480 values weredetermined as above.

Inhibition of Anti-NS3 Reactivity by Soluble Antigens. Patients’ serawere diluted to achieve an optical absorbance of approximately 0.5at 480 nm in the respective EIAs. This was followed by preincuba-tion with varying concentrations of either soluble protease orATPase/helicase for 1 hour at room temperature. The mixtures werethen transferred to wells of microtiter plates that were coated withfull-length NS3 protein at 0.4 µg/mL. Detection of the remaininganti-NS3 IgG reactivity was perfomed according to the protocoldescribed above.

Denaturation of Recombinant Antigens. The recombinant serine pro-tease and ATPase/helicase proteins were treated with varying concen-trations of SDS–b-mercaptoethanol (2ME) at 14°C for 2 hours.They were then immediately diluted in 50 mmol/L sodium carbon-ate buffer (pH 9.60) and passively adsorbed overnight to the wells ofmicrotiter plates. EIAs were performed as described above.

RESULTS

Specificity of NS3-Specific Antibodies in Patients With ChronicHCV Infection. Sera from the 40 patients with chronic HCVinfection were analyzed in solid-phase immunoassays for thepresence of antibody to the three recombinant NS3 proteins:NS3-FL, the ATPase/helicase domain, and the protease do-main. Data obtained from this antibody screening showedthat 36 of 40 (90%) of the sera contained antibodies thatbound the NS3-FL, and 39 of 40 (98%) of the sera bound theATPase/helicase domain, whereas antiprotease reactivitiy wasobserved in only 10% of the sera (Fig. 1). Furthermore, asshown in Table 1, antibodies directed to all three NS3antigens were primarily restricted to the IgG1 isotype.Low-level IgG3 antibodies specific for the ATPase/helicase orNS3-FL were observed in 4 and 1, respectively, of 40 sera,whereas no IgG3 antiprotease antibodies were observed.None of the other IgG isotypes were detected.

To further characterize the antigenicity of the two func-tional domains of full-length NS3, the protease and the

220 CHEN ET AL. HEPATOLOGY July 1998

ATPase/helicase domains were compared as soluble inhibitorsof human anti-NS3 sera. Patient sera were preincubated withthe soluble protease domain or the soluble ATPase/helicasedomain, followed by determination of the anti–NS3-FLreactivity in the solid-phase EIA. Consistent with the resultsusing solid-phase antigens, only the soluble ATPase/helicasedomain was able to inhibit antibody binding to the NS3-FL(Fig. 2). As shown in Fig. 2A, at a concentration of 0.5 µg/mLof soluble ATPase/helicase, the anti–NS3-FL reactivity wasblocked by .50% as compared with the well withoutinhibitor. By increasing the ATPase/helicase concentration to2.5 µg/mL, the majority of the anti–NS3-FL reactivity waslost. In contrast, antibody binding to NS3-FL was not affectedby the addition of the soluble protease domain even at thehighest concentrations (Fig. 2B). These data confirm thatmost antibodies elicited by NS3 during natural chronic HCVinfection are directed to the ATPase/helicase and not to theprotease domain. Analysis of acute-phase sera of 12 HCV-infected patients, 8 of whom became chronically infected and4 who resolved, revealed that the only anti-NS3 antibodydetected was specific for the ATPase/helicase and not theprotease domain (manuscript submitted).

Antigenicity of NS3 is Conformation-Dependent. As reportedpreviously, we also observed that antibody recognition ofNS3-derived peptidic epitopes in the sera of chronic HCVpatients is minimal to absent as compared with recognition ofHCV core or NS4 peptides.22 For example, using 39 syntheticpeptides each 16 aa long with a 6-aa overlap spanning theentire NS3 sequence, we were unable to define a singleantibody epitope (data not shown). We therefore analyzedthe antigenicity of the ATPase/helicase domain after denatur-ation with SDS-2ME. The antigenicity of the treated ATPase/helicase was measured by solid-phase EIA in relation to thereactivity of a nondenatured preparation. The results showedthat treatment with as little as 0.025% of SDS-2ME was able

to affect human antibody binding to the ATPase/helicasedomain (Fig. 3), whereas the relatively poor antigenicity ofthe protease domain was not affected by denaturation (datanot shown). Note the apparent increase in ATPase/helicaseantigenicity at 0.4% SDS-2ME. This most likely reflects theincreased solubility of the ATPase/helicase recombinant pro-tein under these conditions.

The Murine Antibody Response to the NS3 Proteins. In vitrostudies have shown that the HCV NS3 protease domain andNS4a form a complex that enhances the processing down-stream of the precursor protein. This could limit B-cellrecognition of the protease domain because of masking orsteric hindrance. To test the immunogenicity of the proteasedomain in the absence of NS4a, (B10 x B10.S) F1 mice wereimmunized with either NS3-FL or the protease domain alone.Groups of mice were primed and boosted with 100 µg of theprotease, and 20 µg of NS3-FL or the ATPase/helicasedomain. The primary antibody responses are shown in Fig.4A. Mice immunized with 20 µg of the isolated ATPase/helicase domain produced antibodies reactive with the ATPase/helicase and with the NS3-FL. Immunization with 20 µg ofthe isolated protease domain did not elicit a primary antibodyresponse (data not shown); therefore, mice were injectedwith a 100 µg dose of the NS3 protease. Only low-levelantiprotease antibody (1:160) was produced, and could onlybe detected on solid-phase NS3-FL (Fig. 4A). Most impor-tantly, immunization of mice with 20 µg of NS3-FL elicitedantibodies reactive with NS3-FL (the immunogen) and withthe ATPase/helicase domain, but not with the proteasedomain. All antibody responses were boosted by secondaryimmunizations, and even protease-specific antibodies couldbe detected (Fig. 4B). However, note that even after the boost,the NS3-FL–immunized mice did not produce protease-specific antibodies, whereas the ATPase/helicase-specific anti-body titer was equivalent to the NS3-FL–specific titer (Fig. 4B).

Because the human anti–NS3-FL antibody response washighly IgG isotype–restricted, the IgG isotype distribution ofmurine anti-NS3 antibodies elicited by immunization withNS3-FL was examined. Antibodies of the IgG1, IgG2a, andIgG2b isotypes specific for NS3-FL were detected after theprimary immunization. After the secondary immunization,the IgG3 isotype appeared and the titers of the other isotypesincreased (Fig. 5). Therefore, induction of an IgG1-restrictedantibody response is not an intrinsic characteristic of theNS3-FL protein.

TABLE 1. Distribution of Serum IgG Isotypes Specific for theATPase/Helicase, Protease, and NS3-FL Proteins in 40 Patients With

Chronic HCV Infection

Antigens

No. of Reactive Sera

IgG1 IgG2 IgG3 IgG4

NS3-FL 35 0 1 0NS3-protease 4 0 0 0NS3-ATPase/helicase 36 0 4 0

FIG. 1. Serum antibody reactiv-ity to the recombinant ATPase/heli-case, protease, and the NS3-FL pro-teins. Sera (dilution 1:100) from 40patients with chronic HCV infectionwere tested. Antibody binding wasanalyzed in EIA using ATPase, prote-ase, or the NS3-FL as ligands. TheOD of the EIA reaction was mea-sured at 490 nm.

HEPATOLOGY Vol. 28, No. 1, 1998 CHEN ET AL. 221

DISCUSSION

The HCV NS3 protein is associated with multiple func-tional properties such as serine protease activity and ATPase/helicase activity. At the molecular level, the NS3 region ishighly conserved and induces antibodies in HCV infection.30

A previous study examined the immunogenicity of theATPase/helicase domain in a mouse model,29 and found thatthis region of NS3 was immunogenic at the B-cell and T-celllevels. Although the presence of NS3-specific antibodies hasnot been shown to correlate with clearance, active T-cellproliferation specific for the NS3 region31 and to otherstructural and nonstructural HCV proteins during the acutephase has been suggested to be important.32 The HCV NS3protein corresponds to p125 of the pestivirus BVDV.3 The

serine-protease domain (p80) of BVDV p125 has been shownto induce strong humoral responses, which correlate withvirus neutralization in both vaccinated and infected ani-mals.33,34 Recently, it was proposed that antibodies to HCVNS3 are mostly conformation-dependent, because only low-level reactivities have been detected using NS3 peptides assolid-phase ligands.22,23 This is one reason why the finespecificity of B-cell recognition of the domains of NS3 are notas well characterized as compared with other HCV proteinscontaining multiple linear B-cell determinants. However, theconformational dependence of B-cell epitopes of the ATPase/helicase domain remains controversial. At least two groupshave used recombinant, truncated fragments of NS3 and wereable to detect reactivity on Western blot under highlydenaturating conditions.23,24 Despite possible low-level refold-ing under these conditions, these experiments would argueagainst a strict conformational dependence of all ATPase/helicase-related B-cell determinants.

In this report, we have characterized the human andmurine humoral immune reponse to E. coli–derived, recombi-nant full-length HCV NS3 protein and the amino- andcarboxy-terminal domains (i.e., the protease and ATPase/helicase domains). The results indicate that the ATPase/

FIG. 2. Inhibition of human anti–NS3-FL antibodies by the solublerecombinant ATPase/helicase or pro-tease domains. Sera (dilutions yield-ing OD values of approximately 0.5)from patients with chronic HCV in-fection designated K, D, or L werepreincubated with serially dilutedsoluble rHCV ATPase/helicase orprotease, and were then tested forantibody reactivity to the NS3-FL inan EIA. (h), K; (s), D; (n), L.

FIG. 3. Effect of denaturation of the ATPase/helicase domain on antibodybinding. Serum dilutions (1:100) of patients with chronic HCV infectiondesignated W, S, V, N, or Y were tested in an EIA for antibody binding to theATPase/helicase treated with varying concentration of SDS-2ME (0 to 0.4%).(h), W; (j), S; (s), V; (d), N; (e), Y.

FIG. 4. Primary (A) and secondary (B) murine IgG responses to theATPase/helicase, protease, or NS3-FL proteins. Groups of 3 (B10 x B10.S)F1

mice were immunized with the recombinant ATPase/helicase (20 µg)protease (100 µg) or NS3-FL (20 µg) proteins (immunogens). Sera collected14 days after the first (day 0) and second immunization (day 14) werepooled, serially diluted, and analyzed in EIA for antibody binding to thethree recombinant proteins (antigens). Each bar represents pooled mousesera from three mice. End-point titers of IgG are given.

222 CHEN ET AL. HEPATOLOGY July 1998

helicase is the major antigenic region recognized by B cellsand appears to contain mainly discontinuous B-cell determi-nants. In contrast, only a poor humoral response to theprotease domain is induced in both infected humans andimmunized mice. With respect to humans, one could arguethat the formation of complexes between NS4a and theprotease domain of NS3 may mask B-cell determinants withinthis domain, which may be the reason for the absence ofhuman protease domain–specific antibodies. To begin to testthis hypothesis, we immunized mice with both NS3-FL andthe isolated protease domain. Because the mice immunizedwith NS3-FL produced only ATPase/helicase-specific antibod-ies and the isolated protease domain itself, used at a fivefoldhigher dose, was a poor B-cell immunogen, it is not at allcertain that NS4a interferes with B-cell recognition of theprotease domain during a natural HCV infection. Rather, theprotease domain may simply represent a very weak B-cellimmunogen. The clinical relevance of this lack of an antibodyresponse to the NS3 protease domain in acute- and chronic-phase sera is uncertain.

The present observation of limited anti-NS3 IgG isotypediversity is consistent with another recent study in whichIgG1 was found to be the predominant IgG isotype elicited bya number of other HCV antigens including E2, the E2 HVR-1region, NS3, NS4a, and NS5 in humans chronically infectedwith HCV (manuscript submitted). However, the primaryresponses to NS3-FL and the ATPase/helicase domain inimmunized mice consisted of a broad array of IgG isotypes.Therefore, the lack of IgG isotype diversity in the humanhumoral response to the NS3 region is not intrinsic to theprotein, but may be caused by the level of expression or bythe stability of this HCV protein in natural infection.

In summary, our data suggest that, regardless of the host,B-cell recognition of the NS3 protein is directed to theATPase/helicase domain. Also, human antibodies to theATPase/helicase domain are restricted to the IgG1 isotype inchronic patients and appear to be predominantly directedagainst conformationally dependent determinants. The rea-son for the limited immunogenicity of the protease domain ofNS3 deserves further investigation.

REFERENCES

1. Choo Q-L, Kuo G. Weiner AJ, Overby LR, Bradley DW, Houghton M.Isolation of a cDNA clone derived from a blood borne non-A, non-B viralhepatitis genome. Science 1989;244:359-362.

2. Kuo G, Choo QL, Alter HJ, Gitnik GL, Redeker AG, Purcell RH,Miyamura T, et al. An assay for circulating antibodies to a major etiologicvirus of human non-A, non-B hepatitis. Science 1989;244:362-364.

3. Choo QL, Richman KH, Han JH, Berger K, Lee C, Dong C, Gallegos C, etal. Genetic organization and diversity of the hepatitis C virus. Proc NatlAcad Sci U S A 1991;88:2451-2455.

4. Miller RH, Purcell RH. Hepatitis C virus shares amino acid sequencesimilarity with pestiviruses and flaviviruses as well as members of twoplant virus supergroups. Proc Natl Acad Sci U S A 1990;87:2057-2061.

5. Hijikata M, Kato N, Ootsuyama Y, Nakagawa M, Shimotohno K. Genemapping of the putative structural region of the hepatitis C virus genomeby in vitro processing analysis. Proc Natl Acad Sci U S A 1991;88:5547-5551.

6. Grakoui A, McCourt DW, Wychowski C, Feinstone SM, Rice CM.Characterization of the hepatitis C virus–encoded serine proteinase:determination of proteinase-dependent polyprotein cleavage sites. JVirol 1993;67:2832-2843.

7. Grakoui A, Wychowski C, Lin C, Feinstone S, Rice CM. Expression andidentification of hepatitis C virus polyprotein cleavage products. J Virol1993;67:1385-1395.

8. Hijikata M, Mizushima H, Tanji Y, Mori S, Komoda Y, Hirowatari Y,Akagi T, et al. Proteolytic processing and membrane association ofputative nonstructural proteins of hepatitis C virus. Proc Natl Acad SciU S A 1993;90:10773-10777.

9. Bartenschlager R, Ahlborn-Laake L, Mous J, Jacobsen H. Kinetic andstructural ananlyses of hepatitis C virus polyprotein processing. J Virol1994;68:5045-5055.

10. Grakoui A, McCourt D, Wychowski C, Feinstone S, Rice C. A secondhepatitis C virus–encoded protease. Proc Natl Acad Sci U S A 1993;90:10583-10587.

11. Failla C, Tomei L, Defrancesco R. Both NS3 and NS4A are required forproteolytic processing of hepatitis C virus nonstrucral proteins. J Virol1994;68:3753-3760.

12. Lin C, Pragai B, Grakoui A, Xu J, Rice CM. The hepatitis C virus NS3serine proteinase: trans processing requirements and cleavage kinetics. JVirol 1994;68:8147-8157.

13. Lin C, Rice CM. The hepatitis C virus NS3 serine proteinase and NS4Acofactor: establishment of a cell-free trans-processing assay. Proc NatlAcad Sci U S A 1995;92:7622-7626.

14. Lin C, Thompson J, Rice CM. A central region in the hepatitis C virusNS4A protein allows formation of an active NS3-NS4A serine proteinasecomplex in vivo and in vitro. J Virol 1995;69:4373-4380.

15. Sato S, Tanji Y, Hijikata M, Kimura K, Shimotohno K. The N-terminalregion of hepatitis C virus nonstructural protein (NS3) is essential forstable complex formation with NS4A. J Virol 1995;69:4255-4260.

16. Suzich JA, Tamura JK, Palmer-Hill F, Warrener P, Grakoui A, Rice CM,Feinstone SM, et al. Hepatitis C virus NS3 protein polynucleotide-stimulated nucleoside triphosphate and comparison with the relatedpestivirus and flavivirus enzymes. J Virol 1993;67:6152-6158.

17. Wengler G, Wengler G. The NS3 nonstructural protein of flaviviruscontains and RNA triphosphatase activity. Virology 1993;197:265-273.

18. Chien DY, Choo QL, Tabrizi A, Kuo C, Mcfarland J, Berger K, Lee C, etal. Diagnosis of hepatitis-C virus (HCV) infection using an immunodomi-nant chimeric polyprotein to capture circulating antibodies—reevalua-tion of the role of HCV in liver disease. Proc Natl Acad Sci U S A1992;89:10011-10015.

19. Simmonds P, Rose KA, Graham S, Chan SW, McOmish F, Dow BC, FollietEAC, et al. Mapping of serotype-specific immunodominant epitopes inthe NS4 region of hepatitis C virus (HCV): use of type-specific peptidesto serologically differentiate infection with HCV types 1, 2, and 3. J ClinMicrobiol 1993;31:1493-1503.

20. Sallberg M, Ruden U, Wahren B, Magnius LO. Immunodominant regionswithin the hepatitis-C cirus core and putative matrix proteins. J ClinMicrobiol 1992;30:1989-1994.

21. Zhang ZX, Yun ZB, Chen M, Sonnerborg A, Sallberg M. Evaluation of amultiple peptide assay for typing of antibodies to the hepatitis C virus:relation to genomic typing by the polymerase chain reaction. J Med Virol1994;45:50-55.

22. Khudyakov Y, Khudyakova NS, Jue DL, Lambert SB, Fang S, Fields HA.Linear B-cell epitopes of the NS3-NS4-NS5 proteins of the hepatitis Cvirus as modeled with synthetic peptides. Virology 1995;206:666-672.

FIG. 5. Murine IgG isotypes specific for the NS3-FL protein after primaryor secondary immunization with NS3-FL. Mice were immunized withNS3-FL and boosted at day 14. Fourteen days after each immunization serawere collected and analyzed for IgG isotype production to the NS3-FLprotein. Each bar represents the mean end-point titer of sera pooled fromthree mice.

HEPATOLOGY Vol. 28, No. 1, 1998 CHEN ET AL. 223

23. Mondelli M, Cerino A, Boender P, Oudshoorn P, Middeldorp J, FipaldiniC, La Monica N, et al. Significance of the immune response to a majorconformational B-cell epitope on the hepatitis C virus NS3 regiondefined by a human monoclonal antibody. J Virol 1994;68:4829-4836.

24. Hwang LH, Yang PM, Lai MY, Chiang BL, Kao JH, Wang JT, Lee SY, et al.Identification of humoral antigenic determinants in the hepatitis C virusNS3 protein. J Infect Dis 1996;174:173-176.

25. Yun Z, Lindh G, Weiland O, Johansson B, Sonnerborg A. Hepatitis Cvirus (HCV) RNA related to HCV antibody pattern in serum and liverhistology in Swedish blood donors. J Med Virol 1993;1:37-42.

26. Weiland O, Chen M, Lind G, Mattson L, Schvarcz R, Sonnerborg A, WahlM, et al. Efficacy of human leucocyte alpha-interferon treatment forchonic hepatitis C virus infection. Scand J Infect Dis 1995;27:319-324.

27. Jin L, Peterson DL. Expression, isolation, and characterization of thehepatitis C virus ATPase/RNA helicase. Arch Biochem Biophys 1995;323:47-53.

28. Zhang ZX, Sonnerborg A, Sallberg M. A cell-binding Arg-Gly-Aspsequence is present in close proximity to the major linear antigenicregion of HCV. Biochem Biophys Res Commun 1994;202:1352-1356.

29. Sallberg M, Zhang Z-X, Chen M, Jin L, Birkett A, Peterson DL, MilichDR. Immunogenicity and antigenicity of the ATPase/helicase domain of

the hepatitis C virus non-structural 3 protein. J Gen Virol 1996;77:2721-2728.

30. Houghton M, Weiner A, Han J, Kuo G, Choo QL. Molecular biology ofthe hepatitis C viruses: implications for diagnosis, development andcontrol of viral disease. HEPATOLOGY 1991;14:381-388.

31. Diepolder H, Zachoval R, Hoffmann R, Wierenga E, Santantonio T, JungM, Eichenlaub D, et al. Possible mechanism involving T lymphocyteresponse to non-structural protein 3 in viral clearance in acute hepatitisC virus infection. Lancet 1995;346:1006-1007.

32. Missale G, Bertoni R, Lamonaca V, Valli A, Massari M, Mori C, Rumi MG,et al. Different clinical behaviors of acute hapatitis C virus infection areassociated with different vigor of anti-viral cell-mediated immuneresponse. J Clin Invest 1996;98:706-714.

33. Lecomte C, Vandenbergh D, Vanderheijden N, De Moerlooze L, Pin JJ,Chappuis G, Desmettre PH, et al. Identification and production ofpestivirus protein for diagnostic and vaccination purposes. Arch VirolSuppl 1991;3:149-156.

34. Vanderheijden N, De Moerlooze L, Vandenbergh D, Chappuis G, RenardA, Lecomte C. Expression of the bovine viral diarrhoea virus p80protein: its use as ELISA antigen for cattle serum antibody detection. JGen Virol 1993;74:1427-1431.

224 CHEN ET AL. HEPATOLOGY July 1998