Skip Navigation

This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (7)
Right arrowRequest Permissions
Google Scholar
Right arrow Articles by Gesell, J. J.
Right arrow Articles by Wyss, D. F.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Gesell, J. J.
Right arrow Articles by Wyss, D. F.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

Protein Engineering, Vol. 14, No. 8, 573-582, August 2001
© 2001 Oxford University Press

Design, high-level expression, purification and characterization of soluble fragments of the hepatitis C virus NS3 RNA helicase suitable for NMR-based drug discovery methods and mechanistic studies

Jennifer J. Gesell1, Dingjiang Liu1, Vincent S. Madison, Thomas Hesson, Yu-Sen Wang, Patricia C. Weber and Daniel F. Wyss,2

1 These two authors contributed equally to this work. Department of Structural Chemistry, Schering-Plough Research Institute, 2015 Galloping Hill Road, Kenilworth, NJ 07033, USA


   Abstract
Top
 Abstract
Introduction
 Materials and methods
 Results
 Discussion
 References
 
RNA helicases represent a family of enzymes that unwind double-stranded (ds) RNA in a nucleoside triphosphate (NTP)-dependent fashion and which are required in all aspects of cellular RNA metabolism and processing. The hepatitis C virus (HCV) non-structural 3 (NS3) protein possesses a serine protease activity in the N-terminal one-third, whereas RNA-stimulated NTPase and helicase activities reside in the C-terminal portion of the 631 amino acid residue bifunctional enzyme. The HCV NS3 RNA helicase is of key importance in the life cycle of HCV, which makes it a target for the development of therapeutics. However, neither the precise mechanism nor the substrate structure has been defined for this enzyme. For nuclear magnetic resonance (NMR)-based drug discovery methods and for mechanistic studies we engineered, prepared and characterized various truncated constructs of the 451-residue HCV NS3 RNA helicase. Our goal was to produce smaller fragments of the enzyme, which would be amenable to solution NMR techniques while retaining their native NTP and/or nucleic acid binding sites. Solution conditions were optimized to obtain high-quality heteronuclear NMR spectra of nitrogen-15 isotope-labeled constructs, which are typical of well-folded monomeric proteins. Moreover, NMR binding studies and functional data directly support the correct folding of these fragments.

Keywords: HCV NS3 RNA helicase/NMR-based screening/protein engineering/protein NMR/recombinant expression


   Introduction
Top
 Abstract
 Introduction
Materials and methods
 Results
 Discussion
 References
 
Hepatitis C virus (HCV) is recognized as the main etiological agent of parenterally transmitted non-A and non-B hepatitis (Choo et al., 1989Go; Kuo et al., 1989Go). HCV has infected an estimated 170 million people worldwide (Cohen, 1999Go), and 75–85% of people infected will develop a chronic infection. This may ultimately lead to cirrhosis in 10–20% or hepatocellular carcinoma in 1–5% of chronically infected people. More than 100 strains of the virus have been identified which are grouped into six major genotypes that tend to cluster in different regions of the world (Simmonds, 1994Go; van Doorn, 1994Go). The most effective current therapy is a combination composed of interferon-alpha and an anti-viral drug ribavirin (Ahmed and Keefee, 1999Go). The most promising antiviral targets in chronic HCV infection are the replication enzymes, RNA-binding proteins, viral entry proteins and virus maturation processes (Bartenschlager, 1997Go). Significant progress has been made in dissecting the biological functions of these targets in the viral life cycle (reviewed in Houghton, 1996Go; Neddermann et al., 1997Go; Reed and Rice, 1999Go).

HCV, a member of the Flaviviridae family, is a (+)-stranded RNA virus with a linear RNA genome of ~9.6 kb in size. Its genome encodes a single polyprotein of about 3000 amino acid residues containing 10 known proteins which occur in the following sequence NH2–C–E1–E2–p7–NS2–NS3–NS4A–NS4B–NS5A–NS5B–COOH (Takamizawa et al., 1991Go; Grakoui et al., 1993aGo; Lin et al., 1994aGo). The polyprotein undergoes subsequent proteolysis by host and viral enzymes to yield the mature viral proteins (Bartenschlager et al., 1993Go; reviewed in Reed and Rice, 1999Go). Initial cleavages catalyzed by host proteases liberate the individual envelope and capsid proteins located at the polyprotein amino terminus (Selby et al., 1994Go) and the polyprotein segment containing the non-structural (NS) proteins involved in viral replication. The NS2/NS3 junction is cleaved in an intramolecular event (in cis) by a zinc-dependent process requiring substantial segments of NS2 and NS3 flanking the NS2–NS3 scissile bond (Grakoui et al., 1993bGo; Hijikata et al., 1993Go; Santolini et al., 1995Go; Wu et al., 1998Go). NS3 is a bifunctional enzyme with a serine protease localized to the N-terminal 180 residues of the protein (Grakoui et al., 1993cGo) and an RNA helicase located in the C-terminal 451 amino acids (Jin and Peterson, 1995Go; Kim et al., 1995Go). The viral NS3 serine protease is responsible for processing the remaining polyprotein to yield an NS3 cofactor (NS4A), two proteins of unknown function (NS4B and NS5A) and an RNA-dependent RNA polymerase (NS5B). The NS3–NS4A cleavage occurs in cis, while the others occur in trans (intermolecular events). Although the NS3 protease has proteolytic activity of its own, complex formation with the viral NS4A polypeptide is essential for efficient processing of the NS3–NS4A and NS4B–NS5A sites and improves cleavage at the NS4A–NS4B and NS5A–NS5B junctions (Bartenschlager et al., 1994Go; Failla et al., 1994Go; Lin et al., 1994bGo, 1995Go; Tanji et al., 1995Go). In contrast, NS4A has recently been identified as an inhibitor of NS2–NS3 processing (Darke et al., 1999Go).

RNA helicases are required in all aspects of cellular RNA metabolism and processing (De la Cruz et al., 1999Go). They are also essential for the replication of many viruses which makes them targets for the development of therapeutics (Kadaré and Haenni, 1997Go). RNA helicases represent a family of enzymes that unwind dsRNA in a nucleoside triphosphate (NTP)-dependent fashion and, in most cases, this activity is stimulated in vitro by the addition of RNA or DNA. However, neither the precise mechanism nor the substrate structure has been defined for these enzymes. Recently, the basic mechanism for RNA duplex unwinding by the DExH RNA helicase NPH-II has been described (Jankowsky et al., 2000Go).

Helicases, including DNA and RNA helicases, are grouped into two major superfamilies (SFI and SFII) on the basis of the occurrence of seven conserved motifs, a smaller superfamily (SFIII) and two smaller families (Gorbalenya and Koonin, 1993Go). RNA helicases are mostly of the SFII superfamily and can be further classified into families on the basis of particular consensus sequences in the conserved motifs (De la Cruz et al., 1999Go). The HCV NS3 RNA helicase is classified as a DExH protein of the SFII superfamily. The functions of some of these motifs have been elucidated by studies of the effects of mutations on NTP and RNA binding, NTP hydrolysis and unwinding activity (Pause and Sonenberg, 1993Go).

Three-dimensional structures have been determined for a number of HCV enzymes that are essential in the viral life cycle; including the NS3 protease, both by itself (Love et al., 1996Go; Barbato et al., 1999Go) and in complex with NS4A-derived peptides (Kim et al., 1996Go; Yan et al., 1998Go), the RNA helicase portion of NS3 alone (Yao et al., 1997Go; Cho et al., 1998Go) and in complex with a oligonucleotide (Kim et al., 1998Go) and the RNA-dependent RNA polymerase (NS5B) (Ago et al., 1999Go; Bressanelli et al., 1999Go; Lesburg et al., 1999Go). The recently solved crystal structure of an engineered single chain NS4A–NS3 (Yao et al., 1999Go) revealed that the overall folds of the NS3 protease and NS3 helicase portions are maintained in full-length HCV NS3 and their catalytic centers are segregated in the bifunctional enzyme. The NTP and the single-stranded (ss) RNA-binding sites, as well as the deep surface grooves that define the helicase active-site region, are oriented away from the protease–helicase interdomain interface and are exposed to solvent. Therefore, the HCV NS3 helicase domain can be targeted in the absence of its N-terminal serine protease domain for structure-based drug design.

Development of specific inhibitors of HCV NS3 helicase might lead to the identification of clinically relevant antiviral agents in the treatment of hepatitis C virus infection. NMR has recently been recognized as a powerful tool in the identification and optimization of non-peptidic drug-like lead compounds in drug discovery programs (Shuker et al., 1996Go; Hajduk et al., 1997aGo). Examples have been reported where high-throughput screening of large numbers of compounds failed to identify suitable leads, but NMR methods were successfully employed to develop lead compounds (Hajduk et al., 1997aGo–cGo). Moreover, NMR-based screening has been useful in identifying novel scaffolds which subsequently were optimized by the parallel synthesis of a large number of analogues (Hajduk et al., 1999aGo) or incorporated into existing leads to replace moieties with undesirable properties (Hajduk et al., 1999bGo). This NMR-based technology is, however, currently limited to smaller proteins with molecular weights up to ~30–40 kDa and requires large amounts of 15N-labeled protein (Hajduk et al., 1997aGo). Hence it cannot be applied to the 631 amino acid residue HCV NS3 bifunctional enzyme and is not practical for the 451 amino acid residue HCV NS3 helicase domain. Nevertheless, we are interested in using NMR techniques as an aid in the discovery of potent inhibitors for the HCV helicase. Therefore, we designed several truncated versions of the HCV NS3 helicase based on available structure and functional data for making samples appropriate for both NMR structural investigations and NMR-based drug discovery techniques. In addition, such fragments are suitable to probe NTP and oligonucleotide binding sites of the HCV NS3 helicase by NMR and crystallography, which together with mechanistic studies will provide insights into the enzyme's mode of unwinding.

Here we report the engineering, preparation and characterization of several truncated constructs of the HCV NS3 RNA helicase which retain native protein structure and are suitable for detailed NMR studies.


   Materials and methods
Top
 Abstract
 Introduction
 Materials and methods
Results
 Discussion
 References
 
Materials

Native gene sequences were originally amplified from plasmid pJC84 (Grakoui et al., 1993aGo) which encodes the entire NS3 region of the 1a strain of HCV. DNA primers were obtained from Sigma Genosys (The Woodlands, TX). A Quickchange mutagenesis kit used for polymerase chain reaction (PCR) reactions and DNA polymerase were purchased from Stratagene (La Jolla, CA). Restriction enzymes and T4 DNA ligase were obtained from New England Biolabs (Beverly, MA). A Wizard Plus SV Minipreps DNA purification system was purchased from Promega (Madison, WI). Geneclean and MERmaid were obtained from Q·Biogene (Carlsbad, CA). The genes of interest were ligated into pET28b(+) from Novagen (Madison, WI) and transformed into DH5{alpha} cells from Gibco BRL (Rockville, MD). Expression host Escherichia coli strain BL21(DE3) was purchased from Novagen. BPER (Bacterial Protein Extraction Reagent) was obtained from Pierce Chemical (Rockford, IL). Ni2+ resin was acquired from Qiagen (Valencia, CA). Superdex 200 size-exclusion columns were purchased from Amersham Pharmacia Biotech (Piscataway, NJ). Thrombin was acquired from Human Research Laboratories (South Bend, IN). Protease inhibitor cocktail sets I, II and III were obtained from Calbiochem (La Jolla, CA). Gels were acquired from Novex (San Diego, CA). Isotopically enriched 15NH4Cl and 2H2O were purchased from Cambridge Isotope Laboratories (Andover, MA). FELIX980 was obtained from Molecular Simulations (San Diego, CA). Sybyl was obtained from Tripos (St. Louis, MO). The statistical package SAS was acquired from Institute (Cary, NC).

Design of a truncated ß-hairpin in subdomain 2

The Loop Search module of Sybyl was used to search known protein structures in the PDB (Bernstein et al., 1977Go) for hexapeptide ß - turns whose terminal residues match the positions of Thr430 and Pro452 in subdomain 2. From the initial set of 200 turns, 13 were selected that sequentially matched the C{alpha} positions of Thr430, Cys431, Val432, Thr450, Leu451 and Pro452 (Table IGo). The i to i + 3 positions of ß-turns correspond to the central four residues in Table IGo. For these positions in the 13 turns, the most frequent residues are four Arg at i, six Asp/Asn at i + 1, 10 Gly at i + 2 and five Lys at i + 3. To replace the extended hairpin (Thr430–Cys431 ... Pro452), a Thr430–Ser–Asp–Gly–Lys–Pro452 ß-turn was selected. Within this insertion, Ser is homologous to Cys431 while Asp–Gly–Lys matches the consensus turn sequence while adding two charges that may help position the turn on the surface and increase protein solubility.


View this table:
[in this window]
[in a new window]
  		Table I. . ß-Turns from the PDB
 
Plasmid construction

The DNA sequences encoding for both HCV NS3 helicase subdomain 1 (d1-HCVh) and full-length HCV NS3 helicase (d1,2,3-HCVh) [i.e., residues 181–324 and 181–631, respectively, of HCV NS3 from HCV-1a (Table IIGo)] were PCR amplified from pJC84 using primers which incorporated a NdeI site at the 5' end of the genes (PFF: 5'-GGGAATTCGCATATGTCCCC GGTGTTCACGGAC-3') and a HindIII site at the 3' end [(PD1R: 5'-GGGCCAAGCTTAGGTAGCAGTGGCGAGCAC-3') for d1-HCVh and (PFR: 5'-GGGACAAGCTTACGTGAC GACCTCCAGGTC-3') for d1,2,3-HCVh]. The resulting PCR products were digested with appropriate restriction enzymes, isolated with Geneclean and ligated into NdeI/HindIII-digested pET28b(+) vectors. The ligation reactions were used to transform competent E.coli DH5{alpha} which were selected on Lauria–Bertoli (LB) agar plates with kanamycin (30 µg/ml). Recombinant clones were identified and confirmed by PCR gene amplification and DNA sequencing. The final vector constructs, pNS3d1 and pNS3d1,2,3, encoded fusion proteins of HCV NS3 residues 181–324 and 181–631, respectively, C-terminal to a polyhistidine-tag and thrombin cleavage site. Following thrombin cleavage both protein fragments had additional N-terminal GSHM residues.


View this table:
[in this window]
[in a new window]
  		Table II. . Amino acid sequence of the HCV NS3 RNA helicase domain (strain 1a)

 
The DNA sequences encoding for HCV NS3 helicase subdomain 2 (d2{Delta}-HCVh) and subdomains 1 and 2 (d1,2{Delta}-HCVh) (i.e. residues 327–481 and residues 181–481, respectively, of HCV NS3 from HCV-1a (Table IIGo) with residues 431–451 replaced by amino acids S–D–G–K) were PCR amplified from pJC84 in two pieces. The DNA sequences encoding for residues 181–430 and 327–430 were both amplified with forward primers that included an NdeI site and addition of a GSHM sequence at the N-terminus of the constructs (PD2F: 5'-GGGAATTCGCATATGGGCTCCGTCACTGTGCCC-3' for d2{Delta}-HCVh and PFF for d1,2{Delta}-HCVh as above for d1-HCVh and d1,2,3-HCVh) and a reverse primer that included the nucleotides encoding for the S–D–G–K sequence (PD2MB: 5'-TTTGCCATCGCTGGTGTTGCAGTCTATCACAG-3'). For both constructs, the DNA sequence that encodes for residues 452–481 was amplified with the nucleotides encoding S–D–G–K in the forward primer (PD2MT: 5'-CACCAGCGATGGCAAACCCCAGGATGCTGTC-3') and a HindIII site in the reverse primer (PD2R: 5'-GGGCCAAGCTTAGCGCTCCCCCGGTGCCAC-3'). After all of the fragments had been synthesized, purified by gel electrophoresis and isolated using Geneclean or MERmaid, the fragments encoding for residues 181–430–S–D–G–K and S–D–G–K–452–481 were combined in an aliquot as were the fragments encoding for residues 327–430–S–D–G–K and S–D–G–K–452–481. The mixtures were subject to a final round of PCR using the same forward primers used in previously amplifying each of the upstream regions of the final gene constructs and the same reverse primer used in amplifying the DNA fragment encoding for 452–481. The resulting PCR products were prepared substantially as described above for pNS3d1 and pNS3d1,2,3. The final vector constructs, pNS3d2{Delta} and pNS3d1,2{Delta}, encoded fusion proteins of HCV NS3 residues 327–430–SDGK–452–481 and 181–430–SDGK–452–481, respectively, C-terminal to a polyhistidine-tag and thrombin cleavage site.

The gene sequences of pNS3d1, pNS3d1,2{Delta}, pNS3d2{Delta} and pNS3d1,2,3 were checked by DNA sequencing (Sanger et al., 1977Go) to ensure that no mutations occurred during the PCR and cloning procedures. Any unanticipated sequences were corrected using Quickchange following the manufacturer's directions. The final constructs were confirmed with DNA sequencing.

Protein expression of isotope-labeled HCV NS3 helicase constructs

The respective plasmid was transformed into E.coli BL21(DE3) for production of a particular HCV NS3 helicase construct. A single colony was used to initiate growth in LB medium supplemented with 30 µg/ml kanamycin. When the cell density reached an OD600 of 1–2, the culture was used to inoculate a growth in M9 media with 1g/l 15NH4Cl as the sole nitrogen source (Lech and Brent, 1998Go) supplemented with 30 µg/ml kanamycin and 17 µg/ml thiamine. When the cell density reached an OD600 of 1–2 it was used to inoculate an M9 culture supplemented with 30 µg/ml kanamycin and 17 µg/ml thiamine which contained 1g/l 15NH4Cl as the sole nitrogen source. For perdeuterated protein samples 99% D2O instead of H2O was used to prepare the M9 medium. For d1,2{Delta}-HCVh and d2{Delta}-HCVh protein expression was induced with addition of 1 mM isopropyl thiogalactopyranoside (IPTG) at an OD of 1.5. For d1-HCVh and d1,2,3-HCVh the cell culture was cooled to 16°C when reaching an OD600 of 0.7–1.0 before inducing recombinant expression with 1 mM IPTG. At an OD of 3.0 or after 3 h at 37°C (for d2{Delta}-HCVh and d1,2{Delta}-HCVh) or 16 h at 16°C (for d1-HCVh and full-length d1,2,3-HCVh), the cells were harvested by centrifugation and the pellets were stored at –20°C.

Protein purification of HCV NS3 helicase constructs

The cell pellets were subsequently thawed on ice and resuspended in 50–100 ml/l culture of lysis buffer (Table IIIGo). The suspension was homogenized using a glass dounce homogenizer and incubated at room temperature for 20 min with gentle stirring. After cell lysis, the lysate was cleared by centrifugation at 186 000 g for 20 min. The supernatant was added to 4 ml/l culture of Ni2+ resin which had previously been equilibrated in lysis buffer without DTT. The lysate and resin mixture were incubated for 1 h in a cold-room (~4°C) on a rotator. The resin was pelleted by centrifugation, packed into a column and washed with wash buffer (Table IIIGo) until the {lambda}max 280 nm stabilized at a value close to zero. The bound recombinant protein was then eluted with elution buffer (Table IIIGo). Next, the isolated protein product was digested with thrombin to remove a leader His-tag from the N-terminus of the target. Ten NIH units of thrombin were added per milligram of fusion protein and the sample was dialyzed in a cold-room (~4°C) for 16 h against 75 mM potassium phosphate, 1 mM DTT, pH 8.0 for d1-HCVh and d1,2,3-HCVh or pH 6.5 for d2{Delta}-HCVh or against 50 mM potassium phosphate, pH 7.0, 300 mM NaCl, 1 mM DTT and 20% glycerol for d1,2{Delta}-HCVh. The sample was then dialyzed against gel filtration buffer (Table IIIGo) for 4–16 h. After dialysis the sample was concentrated and applied to a Superdex-200 size-exclusion column (26x60 cm) equilibrated in gel filtration buffer (Table IIIGo). Fractions containing the protein of interest, as judged by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS–PAGE), were pooled and concentrated for NMR sample preparation.


View this table:
[in this window]
[in a new window]
  		Table III. . Buffers used during purification of different HCV helicase constructs
 
NMR sample preparation

For the NMR experiments the protein samples were either directly concentrated to the appropriate protein concentration (~0.15–1.0 mM) plus adding 5–10% D2O and protease inhibitors or if a buffer other than the gel filtration buffer was used, it was exchanged using a desalting column. Table IIIGo summarizes the final sample buffers and concentrations which were used to acquire the NMR spectra presented in this work.

NMR spectroscopy

Two-dimensional (2D) 1H–15N heteronuclear single quantum coherence (HSQC) NMR spectra were obtained at 25°C (d1-HCVh, d2{Delta}-HCVh and d1,2{Delta}-HCVh) or 30°C (d1,2,3-HCVh) on Varian INOVA 500 and 600 MHz and Bruker AVANCE 500 MHz NMR spectrometers equipped with triple-resonance probes and a z-axis pulse field gradient. A fast HSQC pulse sequence (Mori et al., 1995Go) was used for the HSQC experiments. The mixed States-TPPI method was employed to obtain sign discrimination in t1 (Marion and Wüthrich, 1983Go) and spectral widths of ~13.3 p.p.m. (centered on the water resonance) and ~30.4 p.p.m. (centered at 119 p.p.m.) were used for the 1H and 15N dimensions, respectively. A total of 1024 t2 data points and either 64 (d1-HCVh and d2{Delta}-HCVh) or 128 (d1,2{Delta}-HCVh) or 200 (d1,2,3-HCVh) t1 increments were collected. The signal was averaged over 16 or 32 transients. Time domain data were zero-filled and multiplied with phase-shifted squared sine-bell functions prior to Fourier transformation. Data processing was carried out with the FELIX program on Silicon Graphics workstations. The sizes of the final matrices were 1024x128 (d1-HCVh), 1024x256 (d2{Delta}-HCVh and d1,2{Delta}-HCVh) and 1024x1024 (d1,2,3-HCVh).

Steady-state kinetic analysis

ATPase rates were assayed using the continuous, enzyme coupled spectrophotometric method (Pullman et al., 1960Go). To determine the 2Km for ATP, helicase construct (40 nM) was assayed at 25°C in 0.103 M sodium MOPS buffer, pH 7.2, 2.6 mM MgCl2, 0.28 mg/ml BSA, 0.4 mM DTT, 0.1 mM EDTA, 1 mM Tris–HCl, 1 mM sodium HEPES, 2 mM PEP (phosphoenol pyruvate), 20 U/ml LDH (lactate dehydrogenase), 10 U/ml PK (pyruvate kinase), 0.17 mM NADH, ±525 µM polyU ([U]), plus 0.05, 0.1, 0.2, 0.4, 0.8 1.6 or 3.2 mM Mg-ATP.

To determine the constructs' relative steady-state affinities for RNA, 20 nM of each were assayed as described above for Km, with the following modifications: [Mg-ATP] was 10 mM, [MgCl2] was 5.1 mM and [U] was between 0 and 660 µM.

NMR titration experiments

To study the NTP binding of the HCV NS3 helicase, a non-hydrolyzable ATP analog, ATP-{gamma}-S [adenosine 5'-O-(3-thiotriphosphate)], was used. ATP-{gamma}-S (0.0005, 0.001, 0.0025, 0.005, 0.008 M) was added incrementally to 240 µM 15N-labeled d1,2{Delta}-HCVh in 25 mM Tris buffer pH 6.7, 20 mM DTT, 5% D2O, 0.015% sodium azide. For comparison, ATP-{gamma}-S (0.0002, 0.0005, 0.001, 0.003, 0.006, 0.012 M) was also added incrementally to 230 µM 15N-labeled d1,2,3-HCVh in 50 mM Tris buffer pH 7.5, 5 mM DTT, 5% D2O, 0.015% sodium azide. MgCl2 was also added to the solution at a ratio of 1:1 relative to ATP-{gamma}-S. 2D 1H–15N HSQC spectra were collected after each addition of ATP-{gamma}-S. The dissociation constant of ATP-{gamma}-S was derived from the amide chemical shift changes of protein residues at the binding site as a function of the concentration of ATP-{gamma}-S. For an interaction of a compound C with a protein R:


This equationGo can directly be correlated with chemical shifts as follows:

(1)
where [C]0 and [R]0 are the total concentrations of compound and protein, respectively, and [CR] is the concentration of the complex, {delta} is the chemical shift of the protein measured at each concentration [CR], {delta}f is the chemical shift of the protein in the absence of the compound ([C]0 = 0) and {delta}b is the chemical shift of the protein at saturation with the compound.

Non-linear regression methods were used to estimate Kd and {delta}b in the titration experiment. Data from an experiment consist of chemical shift ({delta}) values measured at a number of different compound concentrations. The values of [C]0, [R]0 and {delta}f are known. Estimates of Kd and {delta}b are computed by fitting the data to Equation 1Go using non-linear least squares in the statistical package SAS. From a non-linear fit, estimates of the standard errors were obtained for Kd and {delta}b.


   Results
Top
 Abstract
 Introduction
 Materials and methods
 Results
Discussion
 References
 
Design of truncated HCV NS3 helicase constructs

The 451 amino acid residue HCV NS3 helicase consists of three nearly equal-sized subdomains which form a triangular-shaped molecule ~65 Å on a side and 35 Å thick (Figure 1aGo) (Yao et al., 1997Go). Subdomain 1 of the HCV NS3 helicase contains the highly conserved NTP-binding motifs I and II (Figure 1bGo), often referred to as ‘Walker motifs A and B’ (Walker et al., 1982Go), respectively, which are shared by all helicases and also by a wide variety of other NTP-utilizing enzymes (Gorbalenya and Koonin, 1989Go). Residues in motif I are directly involved in binding of the ß and {gamma} phosphates of NTP, while residues in motif II serve to chelate the Mg2+ of the Mg–NTP complex. Residues within the five additional motifs that are used to classify helicases reside in subdomains 1 and 2 of the HCV NS3 helicase (Figure 1bGo). Structural comparisons suggest that residues in motif III are involved in coupling NTP hydrolysis to subdomain 2 rotations. However, the functional roles of the remaining conserved motifs are less clear. An additional motif, TxGx (Pause and Sonenberg, 1993Go), located in subdomain 1 was shown to be involved in ss-DNA binding (Korolev et al., 1997Go; Kim et al., 1998Go). Subdomain 2 contains conserved sequence motifs IV and VI which are implicated in interacting with nucleic acid and motif V which is involved in coupling NTP hydrolysis to nucleic acid unwinding and translocation (Korolev et al., 1998Go). Subdomain 3 forms an ‘{alpha}-helical subdomain’, part of which interacts with ss-DNA in the crystal structure of an HCV helicase–poly(U) complex (Kim et al., 1998Go). Subdomains 1 and 3 share a more extensive interface than either share with subdomain 2. Therefore, subdomains 1 and 3 form a rigid unit, whereas subdomain 2 is connected to subdomains 1 and 3 by solvent-exposed polypeptide segments capable of supporting large-scale, relative rotations of subdomain 2. In particular, an unusual molecular feature is a long antiparallel ß-loop that extends from the central ß-sheet of subdomain 2 to subdomain 3 where the end of the loop becomes an integral part of the subdomain 3 structure. Thus, similarly to other helicases, subdomain motions are characteristic for the unwinding activity of the HCV RNA helicase (Korolev et al., 1998Go).




View larger version (101K):
[in this window]
[in a new window]
  		Fig. 1. . (a) Ribbon diagram of three-domain HCV NS3 helicase (PDB entry 1HEI). Subdomains 1, 2 and 3 are shown in blue, yellow and red, respectively. (b) Backbone worm representation. Conserved motifs between superfamily I and superfamily II helicases are colored and labeled.

 
Our goal was to engineer truncated constructs of the HCV NS3 helicase which are suitable for NMR-based screening techniques in which most of the binding surfaces that are critical for the function of the enzyme would be preserved. First, we designed three constructs that are derived from subdomains 1 and 2 to obtain proteins with NTP-binding and basal NTPase activities. Figure 2Go summarizes the design of d1,2{Delta}-HCVh, which is composed of the two N-terminal subdomains 1 and 2 of HCV NS3 helicase including an engineered ß- hairpin in subdomain 2 to avoid potential protein instability and aggregation (see below). The corresponding single-domain constructs, d1-HCVh (subdomain 1) and d2{Delta}-HCVh (subdomain 2) were also produced.





View larger version (197K):
[in this window]
[in a new window]
  		Fig. 2. . Design of d1,2{Delta}-HCVh. (a) Three-domain HCV NS3 helicase (PDB entry 1HEI) with subdomains 1 and 2 shown in a molecular surface representation and subdomain 3 as a backbone worm. Conserved motifs between SFI and SFII helicases are colored as in Figure 1bGo. (b) Removal of the C-terminal 148 amino acids corresponding to subdomain 3 reduces the molecular weight from ~49 to ~32 kDa, but would expose a long ß-hairpin including a large hydrophobic surface area (dark brown) which would be mobile and solvent exposed. (c) The latter was replaced by an engineered ß-hairpin in subdomain 2 to avoid potential protein instability and aggregation which further reduces the molecular weight of this two-domain construct to ~30 kDa.

 
Inspection of the three-dimensional structure of the HCV NS3 helicase reveals that the conserved residues responsible for NTP-binding and energy coupling are mainly located within the interface between subdomains 1 and 2. Therefore, removal of the C-terminal 148 amino acids corresponding to subdomain 3, which reduces the molecular weight from ~49 to ~32 kDa, should not interfere with NTP-binding and the basal NTPase activity of the protein. However, as described above the second subdomain of the HCV NS3 helicase has an extended ß-hairpin (Thr430–Pro452) that has extensive contacts with subdomain 3. In constructs that exclude subdomain 3, the nearly 20 residues of this ß-loop would become solvent exposed (Figure 2bGo) and likely adopt a non-native structure. Given that subdomain 2 itself appears to be a relatively rigid structure that is preserved during rotation (Yao et al., 1997Go), it appeared feasible to replace the extended ß-loop with a designed short ß-hairpin.

Even though subdomains 1 and 2 have little sequence identity, they share the same structure. Fifty-four residues in the six central parallel ß-strands and in helix B can be superimposed in the two subdomains with an r.m.s.d. of 1.4 Å (Figure 3Go). Within this overlapped set, seven residues (13%) are identical and 27 (50%) match polar or hydrophobic character. A truncated ß-hairpin was designed for subdomain 2 that would correspond to the size of the hairpin in subdomain 1 and increase protein solubility (Table IGo and Figure 4Go). The optimal designed construct containing subdomains 1 and 2 (d1,2{Delta}-HCVh) is a ~30 kDa protein in which the NTP binding site, part of the nucleic acid binding surface and all the sequence motifs conserved between SFI and SFII helicases are retained (Figure 2Go).



View larger version (61K):
[in this window]
[in a new window]
  		Fig. 3. . HCV NS3 helicase (PDB entry 1HEI). Six ß-strands and one helix (see text) were used to superimpose subdomain 1 (blue) and subdomain 2 (yellow). A designed ß-turn (magenta) replaces the long ß-hairpin, which extends to the right, giving a truncated subdomain 2 that corresponds to subdomain 1 in size.

 


View larger version (61K):
[in this window]
[in a new window]
  		Fig. 4. . For HCV NS3 helicase subdomain 2 (PDB entry 1HEI), junction of extended ß-hairpin with the protein core. Displayed are the six residues (red) that were templates for the ß-turn search, the neighboring backbone (green) and hydrogen bonds (cyan). Four Asx–Gly turns from the PDB are shown in atom colors.

 
Plasmid construction

The coding sequences for the HCV NS3 fragments of 181–324, 327–481 with 431–451 deleted and replaced with S–D–G–K, 181–481 with 431–451 deleted and replaced with S–D–G–K and 181–631 were cloned into pet28b(+). These constructs encode for d1-HCVh, d2{Delta}-HCVh, d1,2{Delta}-HCVh and d1,2,3-HCVh with N-terminal poly-His-tags which are subsequently removed by the addition of thrombin. The plasmids were used to transform E.coli strain BL21(DE3).

Expression and purification of isotope-labeled HCV NS3 helicase constructs

Transformed E.coli were grown on minimal media containing 15NH4Cl. With induction, the fusion proteins were expressed in correctly folded soluble forms as N-terminal His-tag fusion proteins. We noted that with the addition of zinc and iron to the M9 minimal medium d1,2{Delta}-HCVh was expressed, but incorporated into an insoluble cellular fraction.

Our purification strategy involved initial isolation of the N-terminal His-tag fusion proteins using standard Ni-chelate chromatography. The isolated proteins were then proteolytically cleaved with thrombin to remove the histidine tag. After thrombin proteolysis, the protein of interest was separated from the histidine tag fragment by size-exclusion chromatography. Figure 5Go shows a gel analysis of the protein purification steps. The protocol described allowed for efficient isolation and purification of ~16, ~5, ~9 and ~15 mg/l of HCV NS3 helicase fragments d1-HCVh, d2{Delta}-HCVh, d1,2{Delta}-HCVh and d1,2,3-HCVh, respectively. We noted that DTT had to be added periodically to avoid protein precipitation.



View larger version (56K):
[in this window]
[in a new window]
  		Fig. 5. . SDS–PAGE analysis of the isolation and purification of d1,2{Delta}-HCVh from E.coli. The position of d1,2{Delta}-HCVh is marked. Lane 1, Mark 12 molecular mass standards (Invitrogen, Carlsbad, CA); lane 2, uninduced BL21(DE3) cells containing pNS3d1,2{Delta}; lane 3, induced BL21(DE3) cells containing pNS3d1,2{Delta}; lane 4, soluble fraction of lysed, induced cells; lane 5, poly(His)-d1,2{Delta}-HCVh after Ni2+ chromatography; lane 6, purified d1,2{Delta}-HCVh following thrombin cleavage of the N-terminal poly(His)-tag and gel filtration chromatography.

 
NMR characterization of the designed constructs

Purified d1-HCVh was stable in 75 mM KiPO4, pH 8.0 with 5 mM DTT, up to a protein concentration of about 0.67 mM. However, the protein fragment was already aggregated at a concentration of 0.5 mM as clearly indicated by the increased peak linewidths in a 2D 1H–15N HSQC NMR spectrum (Figure 6aGo). In contrast, the peak linewidths at 0.15 mM are typical for a monomeric protein of this size (Figure 6bGo). The high dispersion of the peaks in this spectrum is indicative of a well-folded protein. To avoid phosphate in the NMR buffer system, which may have an effect on NTP binding and to improve the sensitivity of NMR triple resonance experiments by reducing the exchange rate of the labile amide protons, non-phosphate buffers at lower pH and temperature (20°C) were tested. An optimized buffer system containing 20 mM Tris–HCl at pH 7.4, 50 mM NaCl, 5% glycerol, 5 mM DTT, 5% D2O and 0.015% NaN3 was found which makes this protein construct amenable to NMR-based drug discovery techniques.




View larger version (47K):
[in this window]
[in a new window]
  		Fig. 6. . 2D 1H–15N HSQC spectra of 15N-labeled d1-HCVh at (a) 500 and (b) 150 µM under NMR buffer conditions described in Table IIIGo.

 
Similarly, high-quality 1H–15N HSQC NMR spectra were obtained in phosphate buffer systems for d2{Delta}-HCVh, d1,2{Delta}-HCVh and d1,2,3-HCVh (Figure 7Go). The high dispersion of the peaks in these spectra is indicative of well-folded proteins and their linewidths are consistent with the protein constructs being monomeric under the conditions tested. A comparison between the 1H–15N HSQC NMR spectra of full-length HCV helicase (d1,2,3-HCVh), d1,2{Delta}-HCVh and d1-HCVh illustrates a dramatic reduction in peak overlap and peak linewidths with decreasing protein size and a corresponding increase in sensitivity. As a result, the two-domain and single-domain HCV helicase constructs are amenable to NMR-based drug discovery methods and possibly solution structure determination. Figure 8Go shows an overlay of 1H–15N HSQC NMR spectra of d1-HCVh, d2{Delta}-HCVh and d1,2{Delta}-HCVh fragments. Most peaks in the NMR spectra of the single-domain constructs are easily recognized in the NMR spectrum of d1,2{Delta}-HCVh. Therefore, the majority of the backbone amide 1H–15N assignments obtained on the single-domain constructs are directly transferable to d1,2{Delta}-HCVh (Liu and Wyss, 2000Go, 2001Go). This strongly suggests that the folds of the individual subdomains are preserved in the single-domain constructs. It also implies that the domain–domain interaction between subdomains 1 and 2 in d1,2{Delta}-HCVh is limited.





View larger version (83K):
[in this window]
[in a new window]
  		Fig. 7. . Comparison of 2D 1H–15N HSQC spectra of (a) 15N,D-labeled d1,2,3-HCVh, (b) 15N-labeled d1,2{Delta}-HCVh and (c) 15N-labeled d2{Delta}-HCVh under the NMR buffer conditions described in Table IIIGo.

 


View larger version (33K):
[in this window]
[in a new window]
  		Fig. 8. . Overlay of 2D 1H–15N HSQC spectra of 15N-labeled d1-HCVh (blue), d2{Delta}-HCVh (red) and d1,2{Delta}-HCVh (black), obtained under the NMR buffer conditions described in Table IIIGo. The boxed region of the spectra is shown expanded at the top to demonstrate that most peaks in the NMR spectra of the single-domain constructs are easily recognized in the NMR spectrum of d1,2{Delta}-HCVh.

 
ATPase activity and ATP-{gamma}-S binding of d1,2{Delta}-HCVh

The Km values for ATP and the apparent steady-state affinities for single-stranded RNA of d1,2,3-HCVh and d1,2{Delta}-HCVh were compared using the continuous, enzyme coupled spectrophotometric method (Pullman et al., 1960Go). The results are summarized in Table IVGo. d1,2{Delta}-HCVh retained both nucleic acid-independent and nucleic acid-stimulated ATPase activities. d1,2{Delta}-HCVh exhibited over 60% of the nucleic acid-stimulated ATP turnover rate (kcat) of d1,2,3-HCVh and had a 3-fold higher nucleic acid-independent turnover rate. This accounts for the lower fold stimulation of d1,2{Delta}-HCVh ATPase activity by saturating polyU. The apparent affinities of d1,2{Delta}-HCVh and d1,2,32-HCVh for polyU for ATPase stimulation (KpolyU) were also similar. The major difference between the constructs was in their Kms for ATP. The nucleic acid-stimulated Km-ATP of d1,2{Delta}-HCVh was 12-fold higher than that of d1,2,3-HCVh. The nucleic acid-independent Km-ATP of d1,2{Delta}-HCVh was over 400-fold higher than the published value for full-length helicase (Preugschat et al., 1996Go).


View this table:
[in this window]
[in a new window]
  		Table IV. . ATPase activity parameters and dissociation constants for ATP-{gamma}-S binding
 
NMR titration experiments were performed to determine if and with what affinity d1,2{Delta}-HCVh is able to bind a non-hydrolyzable ATP analog, ATP-{gamma}-S, relative to full-length HCV helicase (d1,2,3-HCVh). The dissociation constant (Kd) of ATP-{gamma}-S was derived from an analysis of the changes in amide chemical shifts of residues in the binding site of the protein as a function of the concentration of ATP-{gamma}-S. Kd values of 0.8 ± 0.1 and 1.3 ± 0.2 mM were derived from the NMR data for d1,2{Delta}-HCVh and d1,2,3-HCVh, respectively (Table IVGo). The data supports the binding of a nucleotide as is expected for a protein with NTPase activity.


   Discussion
Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
References
 
This paper describes the high-level expression, purification and characterization of the three-domain HCV NS3 helicase and of designed truncated subdomains of this 451-residue (49 kDa) enzyme. The three-domain HCV helicase (d1,2,3-HCVh) can be prepared 15N-labeled in high yields (~15 mg/l) as a pure, soluble, active species. After optimization of solution conditions, high-quality 1H–15N HSQC NMR spectra were obtained whose characteristics are indicative of a well-folded monomeric protein. The high quality of the NMR spectra makes it possible to evaluate the direct binding of potential HCV helicase inhibitors by NMR. In addition, we have designed truncated versions of d1,2,3-HCVh with molecular weights <30 kDa which can be prepared in high yields as pure, soluble and folded isotope-labeled proteins for NMR studies. In our first attempt, we designed three constructs that are derived from the two N-terminal subdomains of d1,2,3-HCVh. These two subdomains contain all motifs which are conserved between SFI and SFII helicases and have been implicated in the function of the enzyme, as well as the NTP binding site and part of a single-stranded oligonucleotide binding site (Kim et al., 1998Go). A comparison between 1H–15N HSQC NMR spectra of d1,2{Delta}-HCVh and d1,2,3-HCVh reveals that many of the resolved peaks in the spectrum of d1,2,3-HCVh can be recognized in the spectrum of d1,2{Delta}-HCVh. Similarly, most peaks in the NMR spectra of the single-domain constructs are easily recognized in the NMR spectrum of d1,2{Delta}-HCVh (Figure 8Go). Owing to the sensitivity of 1H–15N HSQC NMR spectra to protein conformational changes, this strongly suggests that the folds of the individual subdomains are preserved in the truncated versions of the HCV helicase. The quality of the NMR spectra of the two-domain HCV helicase, d1,2{Delta}-HCVh (~30 kDa) and the two single-domain HCV helicases, d1-HCVh and d2{Delta}-HCVh (~15 kDa), make these proteins amenable to 1H–15N HSQC-based screening techniques and possibly solution structure determination.

The optimal designed two-domain construct (d1,2{Delta}-HCVh) is a ~30 kDa protein in which the NTP binding site, part of the ss-RNA-binding surface as well as all the sequence motifs conserved between SFI and SFII helicases are retained (Figure 2Go). Therefore, we expected that this protein could potentially retain NTP-binding and basal NTPase activities. We should note, however, that for the NMR-based drug discovery methods it was more important that the binding surfaces of the active sites were retained in this engineered protein. Steady-state kinetic analysis of d1,2{Delta}-HCVh revealed that in fact, this engineered protein retains both its ATPase activity and its sensitivity to nucleic acid stimulation of that activity (Table IVGo). Its steady-state affinity for polyU during ATPase stimulation is only 2.3-fold lower than that of the full-length helicase, d1,2,3-HCVh. This does not indicate, however, that subdomain 3 is not essential for high-affinity binding of nucleic acid or for procession along the nucleic acid during helix unwinding. Recent studies indicate that maximum (20-fold) stimulation of the basal ATPase rate of full-length HCV helicase may not be directly related to helicase procession along the nucleic acid (Hesson et al., 2000Go; Paolini et al., 2000Go). There are marginal differences between the ATPase kcats of d1,2{Delta}-HCVh and d1,2,3-HCVh, but large differences between their respective Km-ATP values (Table IVGo). This suggests that the 2–3 mM Km-ATP values of d1,2{Delta}-HCVh are good approximations of the affinity of this construct for ATP (Preugschat et al., 1996Go). This is supported by NMR titration experiments in which Kd values of 0.8 ± 0.1 and 1.3 ± 0.2 mM for ATP-{gamma}-S binding were determined for d1,2{Delta}-HCVh and d1,2,3-HCVh, respectively (Table IVGo), showing that d1,2{Delta}-HCVh binds this non-hydrolyzable ATP analog with a very similar affinity than full-length HCV helicase.

In summary, this paper describes the engineering, preparation and characterization of HCV NS3 helicase and several truncated subdomains of this enzyme which retain native protein structure. They can be expressed and purified in high yields as isotope-labeled proteins for NMR. A two-domain construct and two single-domain constructs were derived from the two N-terminal subdomains of the HCV helicase which contain most of the catalytic sites of the enzyme. An engineered ß-hairpin was built into subdomain 2 to avoid potential protein instability and aggregation. Kinetic analysis and characterization by NMR strongly suggest that the HCV helicase proteins produced here are correctly folded. They are suitable for NMR-based drug discovery techniques and studying structure–function relationships using NMR and crystallography together with mechanistic studies.


   Notes
 
2 To whom correspondence should be addressed. E-mail: daniel.wyss{at}spcorp.com Back


   Acknowledgments
 
We thank Drs Brian Beyer and Mary Senior for providing NMR samples of d1,2{Delta}-HCVh for the binding studies with ATP and Drs Anita Howe, Bruce Malcolm and Nanhua Yao for helpful discussions.


   References
Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Ago,H., Adachi,T., Yoshida,A., Yamamoto,M., Habuka,N., Yatsunami,K. and Miyano,M. (1999) Struct. Fold. Des., 7, 1417–1426.[Medline]

Ahmed,A. and Keefee,E.B. (1999) J. Gastroenterol. Hepatol., 14, S12–S18.

Barbato,G., Cicero,D.O., Nardi,M.C., Steinkühler,C., Cortese,R., De Francesco,R. and Bazzo,R. (1999) J. Mol. Biol., 289, 371–384.[ISI][Medline]

Bartenschlager,R. (1997) Intervirology, 40, 378–393.[ISI][Medline]

Bartenschlager,R., Ahlborn-Laake,L., Mous,J. and Jacobsen,H. (1993) J. Virol., 67, 3835–3844.[Abstract/Free Full Text]

Bartenschlager,R., Ahlborn-Laake,L., Mous,J. and Jacobsen,H. (1994) J. Virol., 68, 5045–5055.[Abstract/Free Full Text]

Bernstein,F.C., Koetzle,T.F., Williams,G.J.B., Meyer,E.F.,Jr, Brice,M.D., Rogers,J.R., Kennard,O., Shimanouchi,T. and Tasumi,M. (1977) J. Mol. Biol., 112, 535–542.[ISI][Medline]

Bressanelli,S., Tomei,L., Roussel,A., Incitti,I., Vitale,R.L., Mathieu,M., De Francesco,R. and Rey,F.A. (1999) Proc. Natl Acad. Sci. USA, 96, 13034–13039.[Abstract/Free Full Text]

Cho,H.-S., Ha,N.-C., Kang,L.-W., Chung,K.M., Back,S.H., Jang,S.K. and Oh,B.-H. (1998) J. Biol. Chem., 273, 15045–15052.[Abstract/Free Full Text]

Choo,Q.-L., Kuo,G., Weiner,A.J., Overby,L.R., Bradley,D.W. and Houghton,M. (1989) Science, 244, 359–362.[Abstract/Free Full Text]

Cohen,J. (1999) Science, 285, 26–30.[Free Full Text]

Darke,P.L., Jacobs,A.R., Waxman,L. and Kuo,L.C. (1999) J. Biol. Chem., 274, 34511–34514.[Abstract/Free Full Text]

De la Cruz,J., Kressler,D. and Liner,P. (1999) Trends Biochem. Sci., 24, 192–198.[ISI][Medline]

Failla,C., Tomei,L. and De Francesco,R. (1994) J. Virol., 68, 3753–3760.[Abstract/Free Full Text]

Gorbalenya,A.E. and Koonin,E.V. (1989) Nucleic Acids Res., 17, 8413–8440.[Abstract/Free Full Text]

Gorbalenya,A.E. and Koonin,E.V. (1993) Curr. Opin. Struct. Biol., 3, 419–429.

Grakoui,A., Wychowski,C., Lin,C., Feinstone,S.M. and Rice,C.M. (1993a) J. Virol., 67, 1385–1395.[Abstract/Free Full Text]

Grakoui,A., McCourt,D.W., Wychowski,C., Feinstone,S.M. and Rice,C.M. (1993b) Proc. Natl Acad. Sci. USA, 90, 10583–10587.[Abstract/Free Full Text]

Grakoui,A., McCourt,D.W., Wychowski,C., Feinstone,S.M. and Rice,C.M. (1993c) J. Virol., 67, 2832–2843.[Abstract/Free Full Text]

Hajduk,P.J., Meadows,R.P. and Fesik,S.W. (1997a) Science, 278, 497–498.[Free Full Text]

Hajduk,P.J. et al. (1997b) J. Am. Chem. Soc., 119, 5818–5827.

Hajduk,P.J. et al. (1997c) J. Med. Chem., 40, 3144–3150.[ISI][Medline]

Hajduk,P.J. et al. (1999a) J. Med. Chem., 42, 3852–3859.[ISI][Medline]

Hajduk,P.J., Zhou,M.-M. and Fesik,S.W. (1999b) Bioorg. Med. Chem. Lett., 9, 2403–2406.[Medline]

Hesson,T., Mannarino,A. and Cable,M. (2000) Biochemistry, 39, 2619–2625.[Medline]

Hijikata,M., Mizushima,H., Akagi,T., Mori,S., Kakiuchi,N., Kato,N., Tanaka,T., Kimura,K. and Shimotohno,K. (1993) J. Virol., 67, 4665–4675.[Abstract/Free Full Text]

Houghton,M. (1996) In Fields,B.N., Knipe,D.M. and Howley,P.M. (eds), Hepatitis C Viruses. Lippincott-Raven, Philadelphia, pp. 1035–1058.

Jankowsky,E., Gross,C.H., Shuman,S. and Pyle,A.M. (2000) Nature, 403, 447–451.[Medline]

Jin,L. and Peterson,D.L. (1995) Arch. Biochem. Biophys., 323, 47–53.[ISI][Medline]

Kadaré,G. and Haenni,A.-L. (1997) J. Virol., 71, 2583–2590.[ISI][Medline]

Kim,D.W., Gwack,Y., Han,J.H. and Choe,J.C. (1995) Biochem. Biophys. Res. Commun., 215, 160–166.[ISI][Medline]

Kim,J.L. et al. (1996) Cell, 87, 343–355.[ISI][Medline]

Kim,J.L., Morgenstern,K.A., Griffith,J.P., Dwyer,M.D., Thomson,J.A., Murcko,M.A., Lin,C. and Caron,P.R. (1998) Structure, 6, 89–100.[Medline]

Korolev,S., Hsieh,J., Gauss,G.H., Lohman,T.M. and Waksman,G. (1997) Cell, 90, 635–647.[ISI][Medline]

Korolev,S., Yao,N., Lohman,T.M., Weber,P.C. and Waksman,G. (1998) Protein Sci., 7, 605–610.[Abstract]

Kuo,G. et al. (1989) Science, 244, 362–364.[Abstract/Free Full Text]

Lech,K. and Brent,R. (1998) In Ausubel,F.M., Brent,R., Kingston,R.E., Moore,D.D., Seidman,J.G., Smith,J.A. and Struhl,K. (eds), Current Protocols in Molecular Biology, Vol. 1. Wiley, New York, Unit 1.1.

Lesburg,C.A., Cable,M.B., Ferrari,E., Hong,Z., Mannarino,A.F. and Weber,P.C. (1999) Nature Struct. Biol., 6, 937–943.[ISI][Medline]

Lin,C., Lindenbach,B.D., Prágai,B.M., McCourt,D.W. and Rice,C.M. (1994a) J. Virol., 68, 5063–5073.[Abstract/Free Full Text]

Lin,C., Prágai,B.M., Grakoui,A., Xu,J. and Rice,C.M. (1994b) J. Virol., 68, 8147–8157.[Abstract/Free Full Text]

Lin,C., Thomson,J.A. and Rice,C.M. (1995) J. Virol., 69, 4373–4380.[Abstract]

Liu,D. and Wyss,D.F. (2000) J. Biomol. NMR, 18, 279–280.[ISI][Medline]

Liu,D. and Wyss,D.F. (2001) J. Biomol. NMR, 19, 283–284.[ISI][Medline]

Love,R.A., Parge,H.E., Wickersham,J.A., Hostomsky,Z., Habuka,N. Moomaw,E.W., Adachi,T. and Hostomska,Z. (1996) Cell, 87, 331–342.[ISI][Medline]

Marion,D. and Wüthrich,K. (1983) Biochem. Biophys. Res. Commun., 113, 967–974.[ISI][Medline]

Mori,S., Abeygunawardana,C., Johnson,M.O. and van Zijl,P.C.M. (1995)