RNA-dependent RNA polymerase activity encodedby GB virus-B non-structural protein 5BW. Zhong, P. Ingravallo, J. Wright-Minogue, A. S. Uss, A. Skelton, E. Ferrari, J. Y. N. Lauand Z. Hong Department of Antiviral Therapy, Schering-Plough Research Institute, Kenilworth, NJ, USA

Received September 1999; accepted for publication January 2000

INTRODUCTION

GB virus-B (GBV-B) is a single-stranded positive-sense RNA

virus [1,2]. It is the causative agent of the acute GB

hepatitis in tamarins (Saguinus species). GB agent hepatitis

was originally described by Deinhardt et al. [3], who

inoculated tamarins with the serum obtained from a sur-

geon (having the initials G. B.) with acute hepatitis. All

tamarins inoculated with the GB serum developed hepa-

titis, and serum from these animals obtained during the

acute-phase hepatitis also induced hepatitis in naive ani-

mals, con®rming the infective nature of this agent. Results

from later studies further indicated that the aetiological

agent for GB hepatitis had physical and chemical char-

acteristics of a virus, although its classi®cation as a hu-

man virus was unclear [4,5]. Recently, two positive-sense

RNA viruses, GB virus-A (GBV-A) and GBV-B, were iso-

lated from the serum of infected tamarins using repre-

sentational difference analysis (RDA) [1,2]. Subsequent

studies further showed that GBV-B is the aetiological agent

of GB agent-induced hepatitis [6,7].

Based on sequence analysis, GBV-B has been characterized

as a member of the Flaviviridae family [1,2]. Characteristic of

this virus family, the viral genomic RNA contains a single,

long open reading frame (ORF), which in GBV-B consists of

» 2860 amino acids [1]. The 5¢ and 3¢ non-translated re-

gions (NTRs) that ¯ank the ORF are required for translation

of the viral proteins as well as for viral RNA replication,

based on analogy with other positive-sense RNA viruses. The

polyprotein translated from the viral ORF is processed co-

and post-translationally into several structural (Core, E1, E2

and p7) and non-structural (NS2, NS3, NS4A, NS4B, NS5A

and NS5B) proteins. Processing of the structural proteins is

mediated by cellular signal peptidases, whereas cleavages of

the non-structural region (NS) are mediated by viral prote-

ases (NS2±3 and NS3/4A) [reviewed in 8]. The non-struc-

tural proteins, possibly in conjunction with certain host

Abbreviations: GBV-B, GB virus-B; HCV, hepatitis C virus; NTR, non-

translated regions; NS5B, non-structural protein 5B; ORF, open

reading frame; RdRp, RNA-dependent RNA polymerase.

Correspondence: Weidong Zhong, Antiviral Therapy, K-15-4/4945,

Schering-Plough Research Institute, 2015 Galloping Hill Road,

Kenilworth, NJ 07033-0539, USA.

Journal of Viral Hepatitis, 2000, 7, 335±342

Ó 2000 Blackwell Science Ltd

SUMMARY. Phylogenetic analysis and polyprotein organi-

zation comparison have shown that GB virus-B (GBV-B) is

closely related to hepatitis C virus (HCV). In this study, the

coding region for GBV-B non-structural protein 5B (NS5B)

was isolated by reverse transcription±polymerase chain

reaction (RT±PCR) from pooled serum of GBV-B-infected

tamarins. Expression of soluble GBV-B NS5B protein in Esc-

herichia coli was achieved by removal of a 19-amino acid

hydrophobic domain at the C-terminus of the protein. The

truncated GBV-B NS5B (NS5BDCT19) was puri®ed to

homogeneity and shown to possess an RNA-dependent RNA

polymerase (RdRp) activity in both gel-based and scintillation

proximity assays. NS5BDCT19 required the divalent cation

Mn2+ for enzymatic activity, at an optimal concentration of

15 mM. Interestingly, Mg2+, at concentrations up to 20 mM,

did not support the GBV-B NS5B activity. This differs from

HCV NS5B where both Mn2+ and Mg2+ can support RdRp

activity. Zn2+ was found to inhibit the activity of GBV-B NS5B,

with a 50% inhibitory concentration (IC50) of 5±10 lM.

Higher concentrations of monovalent salts (NaCl or

KCl > 100 mM) and glycerol (> 3%) were also inhibitory.

NS5BDCT19 was able to bind to RNA homopolymers, but

utilized most ef®ciently poly(C), the one with the lowest

binding af®nity for RNA synthesis. Mutational analysis of

GBV-B NS5B demonstrated the importance of several

conserved sequence motifs for enzymatic activity. Based on

sequence homology (» 37% identity and 52% similarity)

between GBV-B and HCV NS5B proteins, the active GBV-B

RdRp provides a good surrogate assay system for HCV

polymerase studies.

Keywords: copy-back RNA replication, GBV-B, RNA-

dependent RNA polymerase.

protein(s), are believed to form a replication complex

responsible for viral RNA replication.

Located at the C-terminus of the polyprotein, NS5B con-

tains sequence motifs characteristic of an RNA-dependent

RNA polymerase (RdRp).2 Recombinant NS5B of both

hepatitis C virus (HCV) and bovine viral diarrhoea virus

(BVDV) has been previously reported to be able to catalyse

nucleotidyl transfer by extending from the 3¢ hydroxyl of a

template (`copy-back'), or from an RNA or DNA primer

[9±14]. The BVDV NS5B was also shown to be capable of

initiating RNA synthesis in vitro in a primer-independent

fashion [15]. As RdRp encoded by positive-sense RNA

viruses functions as the catalytic subunit in concert with

other viral proteins, as well as host proteins, in the replica-

tion of the viral genome, characterization of active viral

RdRps represents the ®rst step towards understanding and

elucidating the mechanisms underlying viral RNA replica-

tion. Additionally, owing to its role in virus replication, RdRp

has been considered an important target, in the case of HCV,

for antiviral therapy.

In this work we have isolated and characterized a

recombinant NS5B protein of GBV-B. The GBV-B NS5B was

shown to be active in RdRp assays in vitro. Optimal reaction

conditions and sequence motifs required for its enzymatic

activity have been identi®ed. Because of its homology to HCV

NS5B, GBV-B RdRp may serve as a surrogate assay system

for studying HCV polymerase and for identi®cation of effect-

ive anti-HCV therapy.

MATERIALS AND METHODS

Cells and plasmid

Bacterial strains used in this study were JM109 (DE3) (Pro-

mega3;4 , Madison, WI), DH10B (GIBCO-BRL3;4 , Rockville, MD) and

XL-1 Blue (Stratagene5 , La Jolla, CA). The expression vector,

pET-28a, was purchased from Novagen6 (Madison, WI).

Cloning, expression and puri®cationof GBV-B NS5B protein

cDNA fragments containing the GBV-B NS5B were isolated

by reverse transcription-coupled polymerase chain reaction

(RT±PCR) from a GBV-B-infected tamarin serum pool (kindly

provided by Drs Jens Bukh and Robert Purcell at the

National Institute of Allergy and Infectious Diseases (NIAID)

of the National Institutes of Health [NIH], Bethesda, MD8 ).

The primers for RT±PCR were designed according to the

GBV-B sequence published previously [1]. Additional se-

quences (coding for a methionine at the N-terminus and a

polyhistidine tag, GSHHHHHH, at the C-terminus) were engi-

neered to facilitate cloning, expression and puri®cation. Viral

RNA was extracted from the serum using Trizol (GIBCO-

BRL) according to the manufacturer's instructions. RT±PCR

reactions were performed using the SuperScript one-step

system (GIBCO-BRL). The ampli®ed RT±PCR products were

cloned into pET-28a between the NcoI and BamHI restriction

sites. A total of six clones were isolated and sequenced. One

clone representing the consensus sequence was selected for

expression and further characterization. Production of sol-

uble GBV-B NS5B protein was induced in freshly trans-

formed Escherichia coli JM109 (DE3) cells, at an OD value of

0.6, by isopropylthio-b-D-galactoside (IPTG) used at a ®nal

concentration of 0.2 mM. After 4 h of induction at 24 °C,

the cells were harvested and NS5B protein expression was

con®rmed by Western blot analysis. Soluble cell lysates were

batch-adsorbed onto a nickel-chelated (Ni-NTA) column.

After extensive washing (10 column volumes) with a high

concentration of salt (1 M NaCl), the protein was eluted off

the column with a buffer containing 0.3 M imidazole. The

protein was further puri®ed using a Superdex-200 gel-

®ltration column (Pharmacia Biotech9 , Piscataway, NJ).

Fractions from the gel-®ltration column were subjected to

sodium dodecyl sulphate±polyacrylamide gel electrophoresis

(SDS±PAGE). The ®nal protein concentration was determined

by using the Bradford protein assay (Bio-Rad Laboratories10 ,

Melville, NY) according to the manufacturer's instructions.

The protein was stored in small aliquots at ±70 °C in the

presence of 10% glycerol and 5 mM dithiothreitol (DTT).

Gel-based RdRp assay

The standard activity assay was performed in a reaction

mixture (40 ll) containing: 20 mM HEPES, pH 7.3; 15 mM

MnCl2; 100 lM of ATP, UTP and GTP; 10 lCi of [a-33P]CTP;

300 ng of puri®ed protein; and 0.2 lg of a 36-base synthetic

RNA (5¢-GGA25UAUAUAUAU-3¢). The sequence of the RNA

was designed to form a stem±loop at the 3¢ end to support

the `copy-back' priming of RNA synthesis. The polymeriza-

tion reaction was performed at 30 °C for 1 h and terminated

by phenol and chloroform extraction. RNA was ethanol

precipitated and separated on a 15% TBE (100 mM Tris

(pH8)/90 mM Boric acid/1 mM EDTA)11 /6 M urea polyacryl-

amide gel (Novex12 , San Diego, CA). The gel was ®xed, vacuum

dried and subjected to autoradiography.

RNA-binding assay

The RNA-binding assay was performed in a reaction mixture

(10 ll) containing: 20 mM HEPES, pH 7.3; 7.5 mM MnCl2;

7.5 mM DTT; 5% glycerol; 125 mM NaCl; 100 lg ml±1 of

bovine serum albumin (BSA); 1 unit of RNase inhibitor;

100 ng of GBV-B NS5B protein; 1 pmol of an end-labelled

RNA probe with the sequence 5¢-U28GGACUUCGGUCC-3¢;and different amounts of homopolymeric RNA, as indicated.

The probe was found to bind most tightly to the GBV-B NS5B

among all the RNA probes of similar length tested (data not

shown). Following incubation at room temperature for

30 min, the reaction mixture was separated on a non-dena-

turing 6% polyacrylamide gel, at 180 V for approximately

Ó 2000 Blackwell Science Ltd, Journal of Viral Hepatitis, 7, 335±342

336 W. Zhong et al.

1h13 , in 0.5 ´ TBE buffer. The gel was dried and then analysed

using autoradiography.

RESULTS

Expression and puri®cation of GBV-B NS5B

The cDNA of the GBV-B NS5B region was isolated (using

RT±PCR) from a pooled serum sample of GBV-infected tam-

arins. The PCR primers were designed based on the pub-

lished GBV-B sequence (GenBank acc. no.: U22304) [1]. To

facilitate protein expression and puri®cation, a methionine

codon was introduced at the N-terminus for initiation of

translation and a six-histidine epitope tag (GSHHHHHH) was

introduced at the C-terminus for af®nity puri®cation. The

cDNA was cloned into the bacterial expression vector

pET-28a (Novagen). A number of independent clones were

isolated and sequenced. Based on sequence alignment, a

consensus NS5B clone, with the coding sequence identical

to the published GBV-B NS5B sequence [1], was selected

for subsequent protein expression and enzymatic character-

ization.

The initial attempt to express full-length GBV-B NS5B

protein in E. coli proved problematic owing to poor solubility

of the protein in the host (data not shown). Prompted by our

recent observation that removal of the C-terminal 21-amino

acid hydrophobic domain of HCV NS5B signi®cantly im-

proved its solubility in E. coli [9], a similar approach was

adopted. As shown in Fig. 1(a), the hydropathy pro®le of

GBV-B NS5B revealed the existence of a similar hydrophobic

region (19 amino acids) at the C-terminus of the protein.

Removal of the C-terminal 19 amino acids (NS5BDCT19)

rendered the protein soluble and readily puri®able through a

Ni-NTA column (Fig. 1b, lanes 3 and 4). The ®nal GBV-B

NS5BDCT19 eluate was » 90% pure and was used for en-

zymatic characterization.

GBV-B NS5BDCT19 possesses RdRp activity

To demonstrate that the puri®ed NS5BDCT19 protein can

direct RNA synthesis in vitro, a gel-based RdRp assay was

established in which a 36-base synthetic RNA with a short

stem±loop at the 3¢ terminus was used as template. As

shown in Fig. 1(c), NS5BDCT19 was able to extend the input

RNA template and produced a near dimer-size product

(lane 3). This result indicated that NS5BDCT19 was able to

initiate RNA synthesis via a `copy-back' priming mechanism

in which the 3¢ terminal sequence of the RNA template

folded back intramolecularly to form a stem±loop. This is

similar to the in vitro `copy-back' activity observed for HCV

[9±13] and BVDV NS5B [14]. Figure 1(c) also showed that

Zn2+ metal ion was inhibitory to the GBV-B RdRp activity at

low lM concentrations (lanes 4±8). Further characterization

of this inhibition is described below.

Reaction conditions for NS5BDCT19 RdRp activity

To further characterize the RdRp activity associated with the

GBV-B NS5B protein, a scintillation proximity assay (SPA)

was developed using polycytidylic acid (polyC) as the tem-

plate and biotinylated oligoguanylic acid (oligoG12) as the

primer. Incorporation of 3H-GMP was measured after the

polymerization products were captured onto the streptavidin-

coated SPA beads [9]. Optimization experiments showed

that the optimal reaction time and temperature for the

GBV-B RdRp were » 3 h and 22±30 °C, respectively

(Fig. 2a, 2b). The enzyme was active in the pH range be-

tween 6.5 and 8, with the peak activity at » pH 7.3

(Fig. 2c). Glycerol was inhibitory at concentrations higher

than 3% (Fig. 2d). This ®nding was different from that ob-

served for a full-length HCV NS5B protein expressed in insect

cells, in which the presence of glycerol at higher concen-

trations (> 10%) was required for maximum activity [13]. It

is probable that full-length recombinant NS5B protein re-

quired higher concentrations of glycerol to remain soluble

owing to the hydrophobic domain at the C-terminus.

To determine the optimal concentrations of monovalent

and divalent cations required for the NS5BDCT19 enzymatic

activity, the SPA-based RdRp assay was performed under the

standard conditions (see the legend to Fig. 2) with increasing

concentrations of KCl, NaCl, MgCl2, MnCl2 or ZnCl2,

respectively (Fig. 3). The RdRp activity was highest at low

concentrations (0±25 mM) of the monovalent salts (Fig. 3a).

The polymerase required the divalent cation Mn2+ at an

optimal concentration of 15 mM (Fig. 3b). Surprisingly,

Mg2+ could not replace Mn2+, even at concentrations as

high as 20 mM (Fig. 3b). This differs from what is known for

HCV NS5B where either Mn2+ or Mg2+ can be used to

achieve comparable activity [13]. In contrast to previous

reports that Zn2+ did not affect the activity of baculovirus-

expressed HCV NS5B at concentrations up to 100 lM [12],

or modestly inhibited an E. coli-expressed, C-terminally

truncated HCV NS5B (50% inhibitory concentration

[IC50] � 60 lM) [9], Zn2+ was quite inhibitory to GBV-B

NS5BDCT19, with an IC50 between 5 and 10 lM (Fig. 3c).

Further analysis will be required to characterize the struc-

tural base responsible for the Zn2+ inhibition.

Comparison of RNA-binding and RdRp activities

In addition to polymerase activity, viral RdRps must be able

to bind to RNA as the ®rst step of the replication process. To

demonstrate the RNA-binding activity associated with the

GBV-B NS5B protein, a gel-shift binding assay was estab-

lished using a radiolabelled 40-base synthetic RNA as a

probe (5¢-U28GGACUUCGGUCC-3¢). This RNA probe was

chosen for its high binding af®nity to the GBV-B NS5B (data

not shown). As shown in Fig. 4(a), NS5BDCT19 was able to

bind to the RNA, resulting in a bandshift of the probe

(lane 2). The RNA±protein complex was not detected in

Ó 2000 Blackwell Science Ltd, Journal of Viral Hepatitis, 7, 335±342

RdRp activity encoded1 by GBV-B NS5B 337

lanes either with no enzyme (lane 1) or after proteinase K

treatment (lane 3). To assess the binding af®nity of

NS5BDCT19 to various homopolymeric RNAs, competition

experiments were performed with increasing concentrations

of non-radiolabelled RNA homopolymers (´ 0.1, ´ 1, ´ 10,

´ 100 probe, respectively). As shown in Fig. 4(a), poly(U)

competed most ef®ciently for binding (lanes 5±8), followed

by poly(G) (lanes 20±23), poly(A) (lanes 10±13) and poly(C)

Fig. 1 Expression and puri®cation of an enzymatically active GB virus-B (GBV-B) non-structural protein 5B (NS5B).

(a) Hydropathy pro®le of GBV-B NS5B. Sequence of the C-terminal hydrophobic region consisting of 19 amino acids is

indicated. This region was deleted to improve solubility of the protein (GBV-B NS5BDCT19) in Escherichia coli. (b) Puri®cation of

NS5BDCT19 expressed in E. coli. NS5BDCT19 was cloned in expression vector pET21 and protein expression was carried out in

bacterial host JM109. Protein puri®cation was performed under conditions similar to those described previously [9]. Lane 1,

uninduced total cell lysate (UN); lane 2, total cell lysate after induction (IND) with 0.2 mM isopropylthio-b-D-galactoside

(IPTG); lane 3, soluble fraction (SOL) from the induced lysate; lane 4, eluate (EL) from the Ni-NTA af®nity column. The upper

panel was stained using Coomassie Brilliant Blue and the lower panel represents Western blotting analysis of the same samples

using an antipenta-histidine monoclonal antibody (Qiagen16 , Washington DC, USA). Mr, molecular mass. (c) RNA-dependent

RNA polymerase (RdRp) activity of GBV-B NS5BDCT19. The standard activity assay was performed as described in the

Materials and methods. Lane 1, RNA size marker; lane 2, input RNA template end-labelled using T4 polynucleotide kinase and

[c-33P]ATP. Lanes 3±8 represent RdRp reactions with different concentrations of ZnCl2 (0, 1, 2.5, 5, 7.5, 10 lM), respectively.

Ó 2000 Blackwell Science Ltd, Journal of Viral Hepatitis, 7, 335±342

338 W. Zhong et al.

(lanes 15±18). These results demonstrated that NS5BDCT19

binds to RNA homopolymers with the af®nity order of

poly(U) > poly(G) > poly(A) > poly(C).

The abilities of NS5BDCT19 to use these RNA homo-

polymers as templates for RNA synthesis were also com-

pared. As shown in Fig. 4(b), only poly(C)-oligo(G) was able

to support the polymerase activity, while poly(A)-oligo(dT),

poly(U)-oligo(dA) and poly(G)-oligo(dC) were inactive. This

result was similar to those observed for HCV RdRp [12],

suggesting an inverse correlation between the RNA-binding

af®nity and the RdRp activity for both polymerases. It is

consistent with the hypothesis that tight binding to an RNA

template will prevent the polymerase from translocating

ef®ciently along the template and thus interfere with its

elongation. As a weak binder, poly(A) was inactive for

GBV-B NS5B (Fig. 4b) but was modestly active for HCV

NS5B [9,12]. This suggests that the GBV-B NS5B may bind

more tightly than HCV NS5B to poly(A).

Mutational analysis of conserved sequence motifs

Amino acid sequence alignment of GBV-B NS5B and other

¯aviviral RdRps revealed six conserved sequence motifs, four

10

20

30

Time (h)

10

20

30

0 20 40 60 80T (ºC)

10

20

30

8.0 7.5 6.5 6.0 5.5pH

5

10

15

20

25

0 10 20 30 40 50 60Glycerol (%)

0 2 4 6 8

7.0

3 H in

corp

orat

ion

(× 1

03 c

.p.m

.)

3 H in

corp

orat

ion

(× 1

03 c

.p.m

.)

3 H in

corp

orat

ion

(× 1

03 c

.p.m

.)

3 H in

corp

orat

ion

(× 1

03 c

.p.m

.)

(a) (b)

(c) (d)

Fig. 2 Effects of (a) time, (b) temperature, (c) pH and

(d) glycerol on polymerase activity of GB virus-B (GBV-B)

NS5BDCT19. The scintillation proximity assay (SPA) was

performed under the standard conditions described previ-

ously [9] with indicated modi®cations. Brie¯y, reaction

mixtures (50 ll total volume) containing (in addition to

other standard reagents) 250 ng of poly(C), 25 ng of bioti-

nylated oligo(G)12, 5 lM GTP/0.1 lCi of 3H-GTP and 175 ng

of GBV-B non-structural protein 5B (NS5B) were incubated

in a 96-well plate at room temperature for 3 h. The reac-

tions were terminated by addition of 50 ll of 100 mM EDTA,

the products were captured by SPA beads coated with

streptavidin and 3H-GMP incorporation was determined

using a scintillation counter.

Fig. 3 Requirements of monovalent and divalent cations for

GB virus-B (GBV-B) non-structural protein 5B (NS5B)

activity. (a) Effects of monovalent cations, KCl and NaCl, at

concentrations between 0 and 200 mM. (b) Effects of di-

valent cations, MnCl2 and MgCl2, at concentrations between

0 and 20 mM. (c) Inhibition of polymerase activity by Zn2+

at concentrations between 0 and 200 lM. The scintillation

proximity assay (SPA)-based RNA-dependent RNA poly-

merase assay was performed as described for Fig. 2 using

poly(C) template and biotinylated oligo(G)12 primer.

Ó 2000 Blackwell Science Ltd, Journal of Viral Hepatitis, 7, 335±342

RdRp activity encoded1 by GBV-B NS5B 339

of which are characteristic of polymerase (A, B, C and D)

[16]. The other two conserved motifs, initially named as `nc'

and `cc', based on their locations relative to motif A and motif

D, are shared among ¯aviviral RdRps (Fig. 5a) [17]. The

motif `nc' is rich in positively charged residues (lysine and

arginine) and thus may play a role in template and/or tem-

plate/primer binding. The motif `cc' consists of a conserved

`cysteine±serine' pair. This motif was subsequently identi®ed

as motif E based on a computer modelling analysis that it is

located at the junction between the palm and thumb sub-

domains (data not shown). Of the known motifs, motif A is

involved in the co-ordination of divalent cations and possibly

sugar selection. Motif B is probably involved in template and/

or primer positioning as well as selection of NTP vs dNTP.

Motif C, the signature motif of most polymerases, is impli-

cated in binding to a second divalent cation that is catalyti-

cally essential to the nucleotidyl transferring reaction. The

function of motif D is unclear; a positively charged residue is

conserved in this motif among all RdRps. Motif E is found

only in RNA-dependent polymerases, and residues from this

motif interact with the nascent primer strand (Fig. 5a).

To determine the importance of those conserved sequence

motifs for GBV-B NS5B polymerase activity, a panel of ala-

nine substitutions (Ala-scanning) were introduced by site-

directed mutagenesis, focusing on the conserved residues in

each motif (Fig. 5a, bold text). All mutant proteins were

expressed and puri®ed using the same protocol as for the

wild-type protein and comparable protein yields were ob-

tained (Fig. 5b). The enzymatic activities of the mutant

proteins were then compared with those of the wild-type

protein in the gel-based RdRp assay. As shown in Fig. 5(c),

substitutions of the positively charged residues in motif `nc'

(K155A, R158A and R168A) showed different effects. The

mutant NS5B with the K155A substitution had an activity

similar to that of the wild-type protein (Fig. 5c, compare

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23

No

enzy

me

Enz

yme

Enz

yme+

PK

Poly (U) Poly (A) Poly (C) Poly(G)

No RNA

Poly(

G)-o

ligo(

dC)

Poly(

A)-olig

o(dT

)Pol

y(C)-o

ligo(

G)

5

10

15

20

25

Probe

Complex

Poly(

U)-olig

o(dA

)

3 H-N

TP

inco

rpor

atio

n(×

103

c.p

.m.)

(a)

(b)

Fig. 4 Template preference for RNA binding and RNA

synthesis. (a) RNA-binding activity of GB virus-B (GBV-B)

NS5BDCT19. Lane 1, no enzyme; lane 2, 100 ng of enzyme;

lane 3, treatment with proteinase K (2.8 mg ml±1) at 37 °C

for 15 min; lanes 5±8, 10±13, 15±18 and 20±23 are

competition experiments with increasing concentrations of

poly(U), poly(A), poly(C) and poly(G), respectively. The

amounts of competing RNA homopolymers used were:

0.1 pmol (lanes 5, 10, 15 and 20), 1.0 pmol (lanes 6, 11, 16

and 21), 10 pmol (lanes 7, 12, 17 and 22) and 100 pmol

(lanes 8, 13, 18 and 23). (b) Activity of RNA homopolymers

as template for GBV-B non-structural protein 5B (NS5B).

The scintillation proximity assay (SPA) was performed as

described for Fig. 2 except that different types of homo-

polymer±primer pairs were tested.

A B C D ccnc

1 590

KPPRLISYPHLEMR

168

158

DTVCFD

222

217

RSSGVYTTSSSN

288

279

GDD

316

SWMK

341

CS

363

155

WT

K15

5A

R15

8A

R16

8A

D21

7A

D22

2A

S27

9A

N28

8A

D31

6A

K34

1A

C36

3A

1 2 3 4 5 6 7 8 9 10 11

68 kDa

wt (

100)

R15

8A

(0.

3)

R16

8A

(0.

9)

K15

5A

(63

.9)

D21

7A

(1.

0)

D22

2A

(1.

8)

S27

9A

(3.

5)

N28

8A

(0.

6)

D31

6A

(1.

4)

K34

1A

(70

)

C63

1A

(17

2)

MW

68 nt

40 nt

1 2 3 4 5 6 7 8 9 10 11 12

(a)

(b)

(c)

Fig. 5 Mutational analysis of GB virus-B (GBV-B)

NS5BDCT19. (a) Schematic diagram showing sequence

motifs conserved among ¯aviviral RNA-dependent RNA

polymerases (RdRps). The amino acid residues in bold were

replaced individually by alanine to generate the mutant

polymerases. (b) Puri®cation of the wild-type and mutant

proteins. (c) Comparison of the polymerase activity between

the wild-type (lane 2) and the mutant proteins (lanes 3±12)

in the gel-based RdRp assay. The RdRp assay was performed

as described in the Materials and methods. The polymerase

activity of each mutant protein was determined using a

Phosphorimager and expressed as a percentage of wild-type

activity, as indicated.

Ó 2000 Blackwell Science Ltd, Journal of Viral Hepatitis, 7, 335±342

340 W. Zhong et al.

lanes 2 and 3) whereas the R158A and R168A mutations

totally abolished this activity (Fig. 5c, lanes 4 and 5). These

results indicate that the two arginine residues (R158 and

R168) in motif `nc' play important roles in RNA replication.

In motif A, substitution of the ®rst (D217A) or the second

(D222A) aspartate residue reduced the enzymatic activity to

near-background level (Fig. 5c, lanes 6 and 7). In motif B,

substitution of the conserved serine residue (S279A) had sig-

ni®cant impact on, but did not totally eliminate the activity

(Fig. 5c, lane 8), while substitution of the asparagine residue

(N288A) completely abolished the activity (Fig. 5c, lane 9).

This asparagine residue is conserved among all RdRps and is

thought to play a role in the selection of ribose vs deoxyribose

[16,18]. In motif C, substitution of the second aspartate

residue (D316A) also completely abolished the enzymatic

activity, con®rming the absolute requirement for this RdRp

hallmark motif (Fig. 5c, lane 10). In motif D, substitution of

the conserved lysine residue (K341A) had no apparent effect

on the polymerase activity (Fig. 5c, lane 11). This is sur-

prising because this lysine residue is conserved amongst all

reverse transcriptases and nearly all viral RdRps except HCV

NS5B, in which a positively charged arginine residue (R) was

found invariably at this position. In a previous report, Loh-

mann et al. found that an R to K mutation in motif D resulted

a slight increase in RdRp activity [12]. Lastly, to our surprise,

substitution of the conserved cysteine residue in motif E did

not impair the GBV-B NS5B activity. Instead, a twofold

increase in RdRp activity over the wild-type protein was

observed with this mutant protein (Fig. 5c, lane 12).

DISCUSSION

In this study, a soluble recombinant GBV-B NS5B protein

was produced and shown to be active in catalysing RNA

synthesis. However, unlike HCV NS5B, the GBV-B RdRp

could only use divalent cation Mn2+, and not Mg2+, for

enzymatic activity. Other buffer components, such as

monovalent cations (KCl or NaCl) and glycerol were not

required. In fact, higher concentrations of these components

in the RdRp reactions had adverse effects on enzymatic

activity. Zn2+ was rather inhibitory to the GBV-B NS5B

activity with an IC50 of 5±10 lM. Mutational analysis of the

GBV-B RdRp con®rmed the requirement of several conserved

sequence motifs for the polymerase activity. Further char-

acterization of these mutant proteins requires the establish-

ment of an assay system suitable for kinetic and mechanistic

evaluation of GBV-B NS5B-directed RNA replication in vitro.

In the past decade, chronic hepatitis C, along with its

aetiological agent, HCV, has emerged as an important

medical problem worldwide. Like other positive-sense RNA

viruses, replication of HCV requires the virally encoded

RNA-dependent RNA polymerase (NS5B). As RdRp is unique

to virus-infected cells, it represents an attractive target for

the design of antiviral agents to treat HCV-associated

chronic hepatitis. Evaluation of anti-HCV inhibitors that

target NS5B will probably require a cell-based replication

system or a small animal model. Unfortunately, no reliable

cell culture system or small animal model that supports HCV

replication is available at the present time. Currently, the

only accepted animal model that is permissive to HCV in-

fection is chimpanzee. However, the limited availability of

this animal model makes it less amendable to support HCV

drug discovery as well as basic research. As GBV-B is most

closely related to HCV (based on sequence homology, tissue

tropism and polyprotein organization), it is conceivable that

GBV-B/tamarin may serve as a surrogate model for HCV

replication. In the case of NS5B, GBV-B and HCV share 37%

identity and 52% similarity. Our biochemical characteriza-

tion of GBV-B NS5B corroborates this hypothesis and dem-

onstrated similar features between the two polymerases.

Therefore, an active GBV-B NS5B protein will be very helpful

in further characterization of this class of polymerases and

for evaluating inhibitors against viral RdRp.

REFERENCES

1 Muerhoff AS, Leary TP, Simons JN et al. Genomic organiza-

tion of GB viruses A and B, two new members of the Flavi-

viridae associated with GB agent hepatitis. J Virol 1995; 69:

5621±5630.

2 Simons JN, Pilot-Matias TJ, Leary TP et al. Identi®cation of

two ¯avivirus-like genomes in the GB hepatitis agent. Proc

Natl Acad Sci USA 1995; 92: 3401±3405.

3 Deinhardt F, Holmes AW, Capps RB, Popper H. Studies on

the transmission of human viral hepatitis to marmoset

monkeys. I. Transmission of disease, serial passages, and

description of liver lesions. J Exp Med 1967; 125: 673±688.

4 Parks WP, Melnick JL, Voss WR et al. Characterization of

marmoset hepatitis virus. J Infect Dis 1969; 120: 548±559.

5 Parks WP, Melnick JL. Attempted isolation of hepatitis

viruses in marmosets. J Infect Dis 1969; 120: 539±547.

6 Schaluder GG, Dawson GJ, Simons JN et al. Molecular and

serologic analysis in the transmission of the GB hepatitis

agents. J Med Virol 1995; 46: 81±90.

7 Schlauder GG, Pilot-Matias TJ, Gabriel GS et al. Origin of

GB-hepatitis viruses. Lancet 1995; 346: 447±448.

8 Rice CM. Flaviviridae: the viruses and their replication. In:

Fields BN, Knipe DM, Howley PM, eds. Virology, 3rd edn.

New York: Raven Press, 1996: 931±960.

9 Ferrari E, Wright-Minogue J, Fang JWS et al. Characterization

of soluble hepatitis C virus RNA-dependent RNA polymerase

expressed in Escherichia coli. J Virol 1999; 73: 1649±1654.

10 Behrens S-E, Tomei L, De Francesco R. Identi®cation and

properties of the RNA-dependent RNA polymerase of hepa-

titis C virus. EMBO J 1996; 15: 12±22.

11 De Francesco R, Behrens SE, Tomei L, Altamura S, Jiricny J.

RNA-dependent RNA polymerase of hepatitis C virus.

Methods Enzymol 1996; 275: 58±67.

12 Lohmann V, Korner F, Herian U, Bartenschlager R. Bio-

chemical properties of hepatitis C virus NS5B RNA-depend-

ent RNA polymerase and identi®cation of amino acid

sequence motifs essential for enzymatic activity. J Virol 1997;

71: 8416±8428.

Ó 2000 Blackwell Science Ltd, Journal of Viral Hepatitis, 7, 335±342

RdRp activity encoded1 by GBV-B NS5B 341

13 Lohmann V, Roos A, Korner F, Koch JO, Bartenschlager R.

Biochemical and kinetic analyses of NS5B RNA-dependent

RNA polymerase of the hepatitis C virus. Virology14 1998; 249:

108±118.

14 Zhong W, Gutshall L, Del Vecchio AM. Identi®cation and

characterization of an RNA-dependent RNA polymerase

activity within the nonstructural protein 5B of bovine viral

diarrhea virus. J Virol 1998; 72: 9365±9369.

15 Kao CC, Del Vecchio AM, Zhong W. De novo initiation of RNA

synthesis by a recombinant Flavivirus RNA-dependent RNA

polymerase. Virology 1999; 253: 1±7.

16 Poch O, Sauvaget I, Delarue M, Tordo N. Identi®cation of

four conserved motifs among the RNA-dependent RNA

polymerase encoding elements. EMBO J 1989; 12: 3867±

3874.

17 Lai VCH, Kao CC, Ferrari E et al. Mutational analysis of

bovine viral diarrhea virus RNA-dependent RNA polymerase.

J Virol 1999, 73: 10129±1013615 .

18 Koonin EV, Dolja VV. Evolution and taxonomy of positive-

strand RNA viruses: implications of comparative analysis of

amino acid sequences. Crit Rev Biochem Mol Biol 1993; 28:

375±430.

Ó 2000 Blackwell Science Ltd, Journal of Viral Hepatitis, 7, 335±342

342 W. Zhong et al.