rna-dependent rna polymerase activity encoded by gb virus-b non-structural protein 5b
TRANSCRIPT
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.