Inhibition of West Nile virus entry by using a
recombinant domain III from the envelope
J. J. H. Chu,13 R. Rajamanonmani,23 J. Li,1R. Bhuvanakantham,1J. Lescar2
and M.-L. Ng1
1Flavivirology Laboratory, Department of Microbiology, 5 Science Drive 2, National University of
Singapore, Singapore 117597, Singapore
2School of Biological Sciences, Nanyang Technological University, 1, Nanyang Walk, Singapore
Received 1 July 2004
Accepted 6 October 2004
The envelope glycoprotein located at the outermost surface of the flavivirus particle mediates
entry of virus into host cells. In this study, the involvement of domain III of West Nile virus
(WNV-DIII) envelope protein in binding to host cell surface was investigated. WNV-DIII was first
expressed as a recombinant protein and purified after a solubilization and refolding procedure.
The refolded WNV-DIII protein displays a content of b-sheets consistent with known homologous
structures of other flavivirus envelope DIII, shown by using circular dichroism analysis. Purified
recombinant WNV-DIII protein was able to inhibit WNV entry into Vero cells and C6/36
mosquito cells. Recombinant WNV-DIII only partially blocked the entry of dengue-2 (Den 2)
virus into Vero cells. However, entry of Den 2 virus into C6/36 was blocked effectively by
recombinant WNV-DIII. Murine polyclonal serum produced against recombinant WNV-DIII protein
inhibited infection with WNV and to a much lesser extent with Den 2 virus, as demonstrated
by plaque neutralization assays. Together these results provided strong evidence that
immunoglobulin-like DIII of WNV envelope protein is responsible for binding to receptor on
the surface of host cells. The data also suggest that similar attachment molecule(s) or receptor(s)
were used by WNV and Den 2 virus for entry into C6/36 mosquito cells.
West Nile virus (WNV), a single-stranded positive sense
RNA envelope virus, was first isolated and identified in the
West Nile region of Uganda in 1937 from a febrile female
adult (Smithburn et al., 1954). It has been classified as a
tests with polyclonal antisera (Boctor et al., 1989). Being
neuroinvasive (George et al., 1984), severe human menin-
goencephalitis might occur as seen clearly in the outbreaks
in North America (CDC, 1999, 2002). During 1999–2002,
the virus extended its range throughout much of the eastern
parts of the USA, and its range within the western hemis-
phere is expected to continue to expand. Birds are the
natural reservoir hosts, and WNV is maintained in nature
in a mosquito-bird-mosquito transmission cycle primarily
involving Culex species mosquitoes.
The WNV, like all flavivirus, is made up of three structural
proteins, the large envelope protein (E), a single nucleo-
capsid protein (C) and a lipid membrane protein (M).
Together with the C protein, the genome RNA forms an
isometric nucleocapsid, while the M and E proteins, toge-
ther with the host’s membrane, form the envelope that
surrounds the nucleocapsids (Murphy, 1980).
The flavivirus major envelope glycoprotein has a dual func-
tion: as a receptor binding protein, it is the primary deter-
minant of host range, cell tropism, virulence and is a major
antigen in eliciting neutralizing antibodies during the
glycoprotein is responsible for fusing the virus and host
membranes in the acidic milieu of the late endosomes (Chu
& Ng, 2004; Heinz & Allison, 2003). The crystallographic
structures of the tick-borne encephalitis (TBE) virus E pro-
tein and that of dengue-2 (Den 2) virus have been deter-
mined (Rey et al., 1995; Modis et al., 2003). Together with
the E1 protein from alphaviruses, they have common struc-
tural and functional features which set them apart from the
well-studied influenza haemaglutinin and have thus, been
named class II fusion glycoproteins (Lescar et al., 2001).
In their pre-fusion conformation, the glycoprotein E forms
homodimers that lie flat on the outer surface of the virus
particles (Rey et al., 1995; Kuhn et al., 2002; Mukhopadhyay
3Both J. J. H. Chu and R. Rajamanonmani contributed equally to this
0008-0411 G 2005 SGMPrinted in Great Britain405
Journal of General Virology (2005), 86, 405–412
et al., 2003). Each E protein monomer folds into three
structural domains predominantly composed of b-strands.
Domain I is centrally located in the structure and carries
the N-glycosylation site. Structural domain II of the E pro-
tein promotes dimerization and bears the fusion loop that
inserts into the target host membrane during the pH-
dependent fusion of thevirus (Modiset al., 2004; Bressanelli
et al., 2004). Located at the C-terminal end of the molecule,
domain III, which was named domain B in earlier anti-
genic mapping studies, carries several epitopes able to elicit
virus-neutralizing antibodies (Roehrig, 2003). In addition,
studies with several flaviviruses including TBE (Mandl et al.,
2001) have also suggested that domain III, which has a
fold typical of an immunoglobulin constant domain, could
also mediate flavivirus attachment to host cells (Anderson,
Since the C-terminal domain can fold independently into a
stable conformation, it is therefore an attractive target to
block infectivity either through the design of molecules
that would compete with the whole virus to enter cells or
by eliciting neutralizing antibodies. In addition, it should
also be useful to help identify receptors and co-receptors
located at the cell surface. In the present study, the putative
and purified. The recombinant protein was subjected to
circular dichroism (CD) spectroscopy to estimate its secon-
dary structure, and the antigenic and antagonist properties
Cells and viruses. Vero cells (ATCC) were maintained in Medium
199 (M199; Gibco) containing 10% inactivated fetal calf serum
(FCS) while C6/36 cells (ATCC) were grown in L15 (Gibco) con-
taining 10% FCS. Flaviviruses, West Nile virus [Sarafend, WNV(S)]
and Den 2 (New Guinea) (gifts from E. G. Westaway, Sir Albert
Sakzewski Virus Research Laboratory, Queensland, Australia) were
propagated in Vero cells throughout this study.
Cloning and expression of recombinant WNV-DIII protein.
The sequence corresponding to aa 299–401 of the C-terminal
domain III of the WNV(S) envelope protein was amplified by PCR
using Advantage II polymerase enzyme (Clontech) and the following
setof primers:forward primer,
CATATGGTGTATGCTC-39 and reverse primer, 59-ACCGGATCC-
restriction sites are underlined. The PCR product of 440 bp was
digested with the restriction enzymes XhoI and BamHI and ligated
into the BamHI/XhoI sites of the pET16b vector (Novagen). The
nucleotide sequence was confirmed by using DNA sequence analysis.
E. coli cells [strain BL21(DE3)] with the pET16b vector containing
the inserted fragment were grown until OD600of the culture reached
0?6, and protein expression was induced by adding IPTG to a final
concentration of 0?5 mM at 30uC for 4–6 h. Cells were harvested
by centrifugation at 5000 g for 20 min at 4uC and washed twice
with a buffer containing 1% NP40, 20 mg ml21DNase and 1 mM
PMSF. Cell lysis was performed in a French pressure cell at
1200 p.s.i. (8280 kPa), followed by centrifugation at 48000 g
for 30 min at 4uC (Beckman). The recombinant protein was
expressed in the insoluble fraction as inclusion bodies as shown by
G-39. The XhoI andBamHI
Purification of recombinant WNV-DIII protein. Inclusion bodies
were solubilized in 8 M urea and passed through a Ni-NTA agarose
column (Qiagen) for affinity purification. The protein was eluted at
pH 2?4. The pH was then adjusted to 7?9 and the protein diluted
to a concentration of 5–10 mg ml21. Refolding was carried out
through extensive dialysis at 4uC against a buffer containing 50 mM
Tris/HCl, 1 mM EDTA, 20% glycerol, 100 mM NaCl and 3 mM
DTT at pH 8?0. After concentration by ultrafiltration using a mole-
cular mass cut-off of 5 kDa (Amicon), the protein was purified by
size exclusion chromatography (Superdex 75; Amersham) in a
buffer containing 12 mM Tris/HCl, 200 mM NaCl and 5 mM
DTT at pH 8?0. Western blot analysis was performed using an
anti-penta-histidine antibody (Qiagen). In order to confirm the
identity of the protein cleavage of the His tag was carried out using
Factor Xa (Qiagen) following the manufacturer’s recommenda-
tions, and sequencing of the first 10 aa from the cleaved product
were obtained using Edman degradation on an Applied Biosystems
Analysis by CD. The CD spectra were recorded on a Jasco J810
spectro-polarimeter by using three accumulations of data at 0?1 nm
intervals and were smoothed using the noise reduction routines
provided with the instrument, including solvent background sub-
traction. The buffer used was 12 mM Tris/HCl, 10 mM sodium
phosphate, 2 mM DTT at pH 8?0 and the protein concentration was
0?1 mg ml21. Deconvolution of the CD far UV-spectrum was car-
ried out with the CDNN and CONTIN software (Sreerama & Woody,
2000). Mass spectrometry analysis using a matrix-assisted laser desor-
ption ionization time of flight mass spectrometer (MALDI-TOF)
was used to determine the molecular mass of the sample.
Generation of murine polyclonal antibodies against soluble
WNV-DIII protein. Soluble WNV-DIII protein was incubated with
ImmunEasy mouse adjuvant (Qiagen) at a concentration recom-
mended by the manufacturer. The antigen–adjuvant mixture was
used to immunize BALB/c mice five times subcutaneously at 14-day
intervals. Mouse sera were collected 12 days after the last booster.
Mouse sera were purified using Econo-Pac serum IgG purification
kits (Bio-Rad) and dialysed overnight with PBS. The purified immuno-
globulins were stored at 220uC. Sera were tested by Western blot
detection for the presence and specificity of antibodies against the
WNV-DIII protein as described previously in Chu & Ng (2002).
Inhibition of WNV infection by soluble WNV-DIII protein.
Soluble WNV-DIII protein or BSA (in the concentration range of
5–100 mg ml21) was incubated with either Vero or C6/36 cells
(16106cells) in cell culture medium at 4uC for 1 h. Unbound pro-
tein molecules were removed by washing the cells three times with
PBS. This was followed by incubating with 16104p.f.u. ml21of
35S-methionine labelled WNV or Den 2 virus for 1 h at 37uC. The
ratio of WNV-DIII molecules to WNV/Den 2 virus particles is 5:1.
After the incubation period, excess or unbound virus were inacti-
vated with acid citrate buffer, pH 2?8 (Chu & Ng, 2003) and
removed by washing three times with PBS. The cells were then
lysed with 1% SDS and the specific radioactivity was determined.
Three independent experiments were carried out.
Plaque neutralization assay. Purified murine polyclonal anti-
bodies generated against soluble WNV-DIII protein or pre-immune
sera (control) were evaluated using plaque reduction assay to deter-
mine the neutralization ability of these antibodies on WNV or Den
2 virus infectivity. Fifty microlitres of antibodies (500 mg ml21) in
twofold serial dilutions (from 1:2 to 1:8192) were prepared in
microcentrifuge tubes. The WNV or Den 2 virus was adjusted to
500 p.f.u. in 50 ml of virus diluent (10% concentrated Hanks’
balanced salt solution, 0?1% BSA; pH 7?2–7?4) was added to the
tube containing serial diluted antibodies. The antibody and virus
was mixed, pulse centrifuged and then incubated at 37uC for 1 h. A
406Journal of General Virology 86
J. J. H. Chu and others
24-well plate with confluent monolayer of Vero cells was used for
virus infection. Before inoculation of the antibody–virus mixture,
the cell monolayer was rinsed once with virus diluent, after which
100 ml of the antibody–virus mixture was added to the appropriate
wells. The plates were left at 37uC for 1 h, and rocked at 15 min
intervals. After incubation, the inocula were removed and the cell
monolayer was rinsed once with virus diluent. Overlay medium
(2% carboxymethly cellulose in M199 containing 2% FCS) was
added and incubated further at 37uC for 4 days, and virus plaques
were stained with 0?5% crystal violet. Three independent experi-
ments were carried out.
Purification and characterization of the
solubilized recombinant WNV-DIII protein
The products from the various purification steps were
analysed in a 15% SDS-PAGE (Fig. 1). The total protein
profile of E. coli BL21 was shown in lane 1. Lane 2 shows the
total protein profile after induction of protein expression
with IPTG (the arrow shows the expressed DIII protein).
The supernatant from the inclusion bodies (lane 3) after
solubilization showed the presence of the WNV-DIII pro-
tein (arrow). The protein sample was refolded (lane 4,
arrow) and purified (lane 5, arrow).
The homogeneity of WNV-DIII protein was assessed using
MALDI-TOF and showed the presence of a single peak with
the expected molecular mass of 13?7 kDa (data not shown).
The protein could be concentrated up to about 15 mg ml21
in Tris/HCl buffer (pH 7?4). At higher concentration, the
protein started to form aggregates, which are consistent
with the presence of several additional exposed hydro-
phobic residues. Purification by size-exclusion chromato-
graphy demonstrated that the protein eluted as a monomer.
The amino acid sequences of the recombinant WNV-DIII
protein [WNV(S) and WNV (NY)] were aligned with the
sequences of other flaviviruses namely: Yellow fever virus
[(vaccine strain 17D) (17D YF)], TBE virus, Den 2 virus,
Japanese encephalitis virus (JE), Kunjin (Kun) virus, Murray
Valley encephalitis (MVE) virus and Langat virus (Lgt)
(Fig. 2). Secondary structure elements (b-strands 1–7) are
represented above the sequences as reported by Rey et al.
(1995). The single conserved disulphide bond is indicated
by a dotted line. The RGD/RGE integrin-binding motif
present in YF, JE and MVE viruses is located in the loop
between b6 and b7 strands of DIII. The known antigenic
regions 307, 330–332 and 390 are indicated in bold
The CD spectrum of the recombinant protein revealed
a secondary structure predominantly composed of anti-
parallel b-sheets as shown by the presence of a large positive
maximum at approximately 200 nm and a shallow nega-
tive peak at 217 nm (Fig. 3). This pattern which is found in
immunoglobulin structures (Tetin et al., 2003) allows for an
Deconvolution of the CD far UV-spectrum using several
publicly available programs (Sreerama & Woody, 2000)
revealsthe presence of about 45%of b-sheetsin the protein.
This result is consistent with a content of 45?3% of residues
in a b-strand conformation as observed in the crystallo-
graphic 3D structure of the C-terminal domain of TBE
virus (residues 301–395) (Rey et al., 1995; PDB code 1svb).
This is slightly higher than the 35% (residues 301–401)
reported for the closely related recombinant domain III
protein of JE virus determined by NMR (Wu et al., 2003;
PDB code 1pjw). The effect of increasing concentrations of
urea on the CD spectrum of the recombinant WNV-DIII
protein showed that the protein is completely devoid of
regular secondary structures at a concentration of 2 M urea.
Competitive inhibition of WNV entry with
soluble recombinant WNV-DIII protein
By using Western blotting, the expressed soluble recombi-
nant WNV-DIII was detected by monoclonal antibodies
against the WNV envelope protein as well as anti-His anti-
bodies (not shown), hence indicating the antigenicity of the
expressed protein. Since WNV-DIII has been proposed to
be the receptor-binding domain, the ability of recom-
binant WNV-DIII binding to the cell and blocking the entry
of WNV infection was investigated. Vero cells were first
incubated with a range of concentrations (5–100 mg ml21)
of soluble recombinant WNV-DIII protein or BSA (to rule
out any possibility of steric hindrance that block virus
binding) for 1 h at 4uC. Radiolabelled WNV or Den 2
virus was added to the pre-treated cells and quantified for
Fig. 1. Analyses of expressed and purified WNV-DIII using
15% SDS-PAGE. M, Molecular mass markers (New England
Biolabs). Lanes: 1, total protein from E. coli BL 21; 2, total
proteins after induction of protein expression with IPTG; 3,
supernatant from inclusion bodies solubilized in urea; 4, sample
after the refolding step and 5, purified fraction from the gel
filtration column. The arrow indicates the WNV-DIII protein.
Antagonist and antigenic activity of WNV-DIII
virus entry (Fig. 4). Fig. 4(a) shows that pre-treatment of
Vero cells at a concentration of 100 mg ml21of soluble
WNV-DIII protein resulted in more than 60% inhibition
ing to 20% when the Vero cells were pre-treated with low
concentration (5 mg ml21) of WNV-DIII protein. Although
the inhibition of WNV entry was not complete, it was still
very significant when compared to cells that were pre-
treated with BSA, at high concentration of 100 mg ml21.
There was a baseline of 5–10% non-specific inhibition of
WNV entry. The solubilized WNV-DIII protein was only
able to partially inhibit Den 2 virus entry (30% inhibition)
into the pre-treated Vero cells. Again the inhibition was
dose-dependent, decreasing to about 1% inhibition in cells
that were pre-treated with WNV-DIII protein concentra-
tions ranging from 5 to 10 mg ml21.
Fig. 2. Alignment of the amino acid sequences of domain III of flaviviruses E protein. Amino acid numbering refers to WNV.
Evolutionary conserved residues are shaded. The disulphide bond between Cys305 and Cys336 is indicated by a dotted line.
Secondary structure elements are labelled from b1 to b7 based on the WNV domain III structure determined using NMR (Volk
et al., 2004). Abbreviations and amino acid sequence accession numbers are WNV(S), West Nile virus (strain Sarafend;
AY688948); Lgt, Langat virus (strain TP21; AAF75259); Kun, Kunjin virus (strain MRM61C; P14335); MVE, Murray Valley
encephalitis virus (NP_722531). Sequences for WNV(NY), West Nile virus (strain New York 385–399; PDB code 1s6n); JE,
Japanese encephalitis virus (strain CH2195 LA; PDB code 1pjw); Den 2 virus (strain Pr159 s1; PDB code 1oke); TBE, tick-
borne encephalitis virus (strain Neudorfl; PDB code 1svb) and YF, Yellow fever virus (vaccine strain 17D; PDB code 1na4)
were obtained from the protein database. The RGD integrin-binding motif present in some flaviviruses (e.g. MVE) is located in
the loop connecting strands b6 and b7. The three known antigenic regions located around residues 307, 330–332 and 390,
which are likely to be involved in receptor binding, are indicated in bold characters.
Fig. 3. CD spectrum of recombinant WNV-DIII protein. A large
maximum at 200 nm and a shallow minimum at approximately
217 nm which are indicative of the presence of anti-parallel
b-strands in the structure are visible.
408 Journal of General Virology 86
J. J. H. Chu and others
Similar experimental procedures were carried out to deter-
mine if WNV-DIII protein was responsible for binding
to the cell surface of mosquito cells, C6/36. Entry of
WNV can be effectively blocked by the recombinant WNV-
DIII (>70%, Fig. 4b) and, interestingly, the entry of Den
2 virus into C6/36 cells was also significantly inhibited
(>60%, Fig. 4b). Again, BSA has minimal effect in
blocking the entry of WNV and Den 2 virus. Studies are
also currently being carried out to determine the ability
of recombinant Den 2 virus E DIII protein in blocking
the binding of WNV to vertebrate and mosquito cells.
These will provide further information as to whether simi-
lar attachment or receptor molecules are utilized by WNV
and Den 2 virus in different cell types.
To rule out the presence of other bacterial proteins that
may co-purify with WNV-DIII protein and participate
in blocking virus attachment, supernatant of bacterial cell
lysate not expressing WNV-DIII (purified in the same
manner as for recombinant WNV-DIII protein) was used
to assess virus binding. The E. coli cell lysate inhibited
WNV and Den 2 virus binding to Vero and C6/36 cells at a
low level of not more than 2%. This was less than that
observed using BSA as a competitor (approx. 5%) (data not
shown). Therefore, these results confirm that WNV-DIII
protein is responsible for binding to the cell surface of
both vertebrate and invertebrate cells and the plausibility
of WNV and Den 2 virus sharing the same attachment/
receptor molecule(s) in C6/36 cells.
Murine polyclonal antibodies to recombinant
WNV-DIII protein neutralized WNV
To affirm further that WNV-DIII protein is indeed the
receptor-binding domain, murine polyclonal antibodies
against recombinant WNV-DIII were produced and used
for plaque neutralization assay of WNV. The murine poly-
clonal was specific in detecting WNV-DIII protein in a
Western blot (data not shown). Plaque neutralization
assays were then carried out for both WNV and Den 2
virus with a dilution series of antibodies against WNV-DIII
protein. At dilution of 1:64, there was at least 90% neutra-
lization of the WNV (Fig. 5a). High neutralization (80%)
of the WNV was maintained until the murine polyclonal
antibodies were diluted to 1:256. At higher dilutions, there
was an exponential decrease in the degree of neutraliza-
tion reaching 0% when the dilution factor was 1:8192.
When the WNV was reacted with the pre-immunized sera,
no neutralization effect was observed. In contrast, Den 2
virus was only partially inhibited (50%) at low dilution
(<1:64) of the antibodies and minimal neutralization
effect was observed with higher dilution of the antibodies
(>1:256) (Fig. 5b).
Among the three structural domains of the flavivirus major
envelope glycoprotein E, DIII forms a continuous polypep-
tide segment that can fold independently (Bhardwaj et al.,
100 µg ml
50 µg ml
Concentration of WNV-Dlll or BSA
25 µg ml
10 µg ml
5 µg ml
Inhibition of virus entry (%)
Concentration of WNV-Dlll or BSA
Inhibition of virus entry (%)
100 µg ml
50 µg ml
25 µg ml
10 µg ml
5 µg ml
Fig. 4. Competitive inhibition of WNV entry with soluble
recombinant WNV-DIII protein. (a) Vero and (b) C6/36 cells
were first incubated with different concentrations of soluble
WNV-DIII protein or BSA (negative control). Radiolabelled
WNV or Den 2 virus was added and assayed for virus entry.
Entry of WNV is significantly blocked in the presence of WNV-
DIII protein in (a) Vero and (b) C6/36 cells. Recombinant
WNV-DIII protein is able to partially block the entry of Den 2
(>100 mg ml”1)] into (a) Vero cells and drastically inhibit Den
2 virus entry into (b) C6/36 cells. No effect of BSA on the
entry of WNV and Den 2 virus into (a) Vero and (b) C6/36
cells is seen.
Antagonist and antigenic activity of WNV-DIII
2001). It was earlier identified as a stable tryptic fragment
of about 10 kDa, which required formation of a native
intramolecular disulfide bond to correctly present B-cell
epitopes (Roehrig et al., 1998). Due to the absence of the
interacting central domain I, the recombinant truncated
protein is poorly soluble, presumably because of the
exposure of additional hydrophobic residues which are
buried in the context of the native homodimer. To
overcome solubility problems, several investigators have
expressed the DIII of flaviviruses as a recombinant pro-
tein fused either with thioredoxin as for JE virus (Wu et al.,
2003) or with GST for Langat virus (Bhardwaj et al., 2001).
In the protocol used in this study, the truncated protein
was obtained within inclusion bodies and a refolding pro-
cedure was performed. Despite the relative high yield and
ease of purification, the procedure introduced an uncer-
tainty about the exact proportion of protein molecules,
which have the native fold. Indeed, some proteins asso-
ciated as dimers or trimers via the formation of non-native
inter-monomer disulfide bonds. Most of these misfolded
proteins were eliminated during the size-exclusion chro-
matography step. Proper folding of the expressed recombi-
nant WNV-DIII protein is of paramount importance since
it has been shown that a reduced and denatured WNV E
glycoprotein was unable to elicit neutralizing antibodies
in mice (Wengler & Wengler, 1989). The secondary struc-
ture composition of the recombinant WNV-DIII protein
derived from CD indicates an anti-parallel b-strands con-
formation consistent with the experimentally determined
structures of other flaviviruses envelope protein ecto-
domains (Rey et al., 1995; Modis et al., 2003) or DIII
fragments (Wu et al., 2003; Volk et al., 2004).
The sites of mutations that affect host range or virulence has
been mapped on the major flavivirus envelope protein (Rey
et al., 1995). They cluster at the hinge junction between
domain I and III next to the fusion loop, which is buried in
the dimer interface and in domain III. In the virus struc-
ture (Kuhn et al., 2002; Mukhopadhyay et al., 2003) the
lateral face of domain III is largely exposed. In particular,
residues located around positions 305–308, 330–333 and
384–386 are clustered on the upper surface of WNV-DIII,
which is exposed to the solvent and hence accessible to
antibodies. Residue 307 is a lysine in WNV-DIII protein
and also in TBE virus but is acidic in other flaviviruses (JE
virus and MVE virus). Position 307 is a hot spot for
mutation enabling the generation of flavivirus escape
mutants (Beasley & Barrett, 2002; Chambers et al., 1998).
Residues Ser331 and Asp332 [corresponding respectively
to Lys332 and Asp333 in WNV(S)] belong to a neutralizing
epitope on JE virus (Lin & Wu, 2003). Thus, a number of
these residues which belong to virus-neutralizing epitopes
(Beasley & Barrett, 2002; Lin & Wu, 2003; Wu et al., 2003;
Volk et al., 2004) account for the antigenic fine structure of
the flavivirus E protein, and are likely to be involved in
determining the binding specificity for different cell types
(Hung et al., 2004). In this study, a polyclonal anti-WNV-
DIII protein serum was produced in mice. The serum was
able to neutralize WNV (Fig. 5a). This suggested that the
conformational epitopes on the purified refolded protein
were correctly presented at the external surface and the
WNV-DIII recombinant protein adopts a native fold.
At the surface of the virion, DIII cluster around fivefold
Fig. 5. Plaque neutralization of WNV with murine polyclonal
antibodies against WNV-DIII protein. The polyclonal antibodies
against WNV-DIII protein are diluted in a twofold series. Equal
volume (50 ml) of anti-WNV-DIII antibodies and (a) WNV or (b)
Den 2 virus (500 p.f.u.) were incubated for 1 h before this mix-
ture is overlaid onto Vero cells monolayer. Plaques were
stained with 0?5% crystal violet. Virus diluent was used as a
Greater than 90% of the WNV is neutralized when the anti-
serum used is at a dilution of <1:64. The neutralization capa-
city is maintained at 80% with the antiserum dilution up to
1:256. The neutralization percentages decrease exponentially
after that to 0% at the antiserum dilution of 1:8192. (b) Partial
neutralization of Den 2 virus with polyclonal anti-WNV-DIII is
only observed at low dilutions. Up to 40% neutralization is
observed with dilutions of 1:16. The neutralization capability
is reduced to less than 10% with dilution of 1:256 or higher.
410 Journal of General Virology 86
J. J. H. Chu and others
icosahedral axis (Kuhn et al., 2002; Mukhopadhyay et al.,
2003) and such pentameric clusters might be a more effi-
cient way to elicit neutralizing antibodies that would
possibly interfere with receptor binding. It is interesting
to note that the Langat virus DIII protein crystallizes with
five monomers per asymmetric unit thus possibly mimick-
ing the arrangement found on the viral surface (White et al.,
Domain III was initially postulated to form the receptor-
binding site for the virus particles because of its exposed
location at the surface of the virus (Rey et al., 1995; Kuhn
et al., 2002; Mukhopadhyay et al., 2003). In addition, the
presence of a tripeptide Arg-Gly-Asp motif in the loop
connecting strands b6 and b7 of several mosquito-borne
flaviviruses E protein (e.g. MVE virus and several strains of
JE virus) (Fig. 2) suggested that this motif could bind to
the integrin family of cell-surface matrix receptors. This
hypothesis, however, was not supported by mutagenesis
studies of the YF virus (Van der Most et al., 1999) and
cellular receptors are still not unequivocally identified for
flaviviruses. Indeed a range of different surface molecules
could act as flavivirus receptors on different cell types
(Anderson, 2003). The fact that the recombinant WNV-
DIII protein was able to substantially inhibit infection
with the Den 2 virus in C6/36 cells (Fig. 4b) indicated that
the attachment/receptor molecules situated at the surface
of the mosquito cells may be common for both WNV and
Den 2 virus. Work is in progress to address this interesting
However, the lack of high antagonistic effect of WNV-DIII
on Den 2 virus infection (Fig. 4a) in Vero cells as com-
pared to WNV, may suggest that the two viruses utilized
a different set of attachment or receptor molecules for
binding to Vero cells. This was consistent with the obser-
vation that the murine polyclonal antibodies against
WNV-DIII also failed to exhibit high neutralization effect
on Den 2 virus (Fig. 5b).
Glycosaminoglycans such as heparan sulfate and choindri-
tin sulfate (negatively charged carbohydrates) present on
the surface of many vertebrate cells were proposed to be
involved in the attachment and entry of Den 2 virus (Chen
et al., 1997). However, treatment with heparinase on
vertebrate cells (Chu & Ng, 2003) did not significantly
inhibit infection with WNV. Instead, a 105 kDa protease-
sensitive, N-linked glycoprotein has been implicated to
mediate attachment and entry of WNV into permissive
vertebrate cells (Chu & Ng, 2003). Other studies also
suggested that heparan sulfate may not necessary serve as
the universal flavivirus attachment or receptor molecule on
cell surfaces (Bielefeldt-Ohmann et al., 2001; Kroschewski
et al., 2003).
The entry process of WNV into both vertebrate and
invertebrate is poorly understood and this process is
important in explaining tissue tropism and pathogene-
sis of WNV infection. To our knowledge, this study has
defined direct evidence, for the first time, that DIII of
the WNV envelope protein is involved in binding to cell
surface molecules for both Vero and C6/36 cells. However,
we cannot exclude any possibility that other domains of
the envelope protein may also be necessary for post-binding
process of WNV entry into host cells. Future work is also
necessary to elucidate the molecular mechanism of WNV
entry mechanism. Nevertheless, the identification of the
receptor-binding domain of WNV envelope protein can
serve as a potential target for the design of anti-viral agent
development to eradicate this emerging flavivirus infection.
We thank Loy Boon Pheng for technical assistance and Jacques
d’Alayer (Pasteur Institute) for performing the N-terminal amino acid
sequencing of WNV-DIII protein. M.-L.N group is supported by
grants from Biomedical Research Council, Singapore (BMRC/01/1/21/
18/003), and National University of Singapore (R-182-000-055-112).
J.L. group is supported by grants from Nanyang Technological
University (SUG14/02), Biomedical Research Council (BMRC/02/1/
22/17/043) and National Medical Research Council (NMRC/SRG/001/
Anderson, R. (2003). Manipulation of cell surface macromolecules
by flaviviruses. Adv Virus Res 59, 229–274.
Beasley, D. W. C. & Barrett, A. D. T. (2002). Identification of
neutralizing epitopes within structural domain III of the West Nile
virus envelope protein. J Virol 76, 13097–13100.
Bhardwaj, S., Holbrook, M., Shope, R. E., Barrett, A. D. T. &
Watowich, S. J. (2001). Biophysical characterization and vector-
specific antagonist activity of domain III of the tick-borne flavivirus
envelope protein. J Virol 75, 4002–4007.
Bielefeldt-Ohmann, H., Meyer, M., Fitzpatrick, D. R. & Mackenzie,
J. S. (2001). Dengue virus binding to human leukocyte cell lines:
receptor usage differs between cell types and virus strains. Virus Res
Boctor, F. N., Calisher, C. H. & Peter, J. B. (1989). Dot-ELISA for
serodiagnosis of human infections due to Western equine encepha-
litis virus. J Virol Methods 26, 305–311.
Bressanelli, S., Stiasny, K., Allison, S. L., Stura, E. A., Duquerroy, S.,
Lescar, J., Heinz, F. X. & Rey, F. A. (2004). Structure of a flavivirus
envelope glycoprotein in its low-pH-induced membrane fusion
conformation. EMBO J 23, 728–738.
CDC (1999). Update: West Nile virus encephalitis – New York 1999.
MMWR Morbid Mortal Wkly Rep 48, 944–946.
CDC (2002). Update: West Nile virus encephalitis – New York 1999.
MMWR Morbid Mortal Wkly Rep 51, 1135–1136.
Chambers, T. J., Halevy, M., Nestorowicz, A., Rice, C. M. & Lustig, S.
(1998). West Nile virus envelope proteins: nucleotide sequence
analysis of strains differing in mouse neuroinvasiveness. J Gen Virol
Chen, Y., Maguire, T., Hileman, R. E., Fromm, J. R., Esko, J. D.,
Linhardt, R. J. & Marks, R. M. (1997). Dengue virus infectivity
depends on envelope protein binding to target cell heparan sulfate.
Nat Med 3, 866–871.
Chu, J. J. H. & Ng, M. L. (2002). Trafficking mechanism of West Nile
(Sarafend) virus structural proteins. J Med Virol 67, 127–136.
Antagonist and antigenic activity of WNV-DIII
Chu, J. J. H. & Ng, M. L. (2003). Characterization of a 105-kDa
plasma membrane associated glycoprotein that is involved in West
Nile virus binding and infection. Virology 312, 458–469.
Chu, J. J. H. & Ng, M. L. (2004). Infectious entry of West Nile virus
occurs through a clathrin-mediated endocytic pathway. J Virol 78,
George, S., Gourine-Devi, M., Rao, J. A., Prasad, S. R. & Pavri, K. M.
(1984). Isolation of West Nile virus from the brain of children who
had died of encephalitis. Bull W H O 62, 879–882.
Heinz, F. X. & Allison, S. L. (2003). Flavivirus structure and
membrane fusion. Adv Virus Res 59, 63–97.
Hung, J. J., Hsieh, M. T., Young, M. J., Kao, C. L., King, C. C. &
Chang, W. (2004). An external loop region of domain III of dengue
virus type 2 envelope protein is involved in serotype-specific binding
to mosquito but not mammalian cells. J Virol 78, 378–388.
Kroschewski, H., Allison, S. L., Heinz, F. X. & Mandl, C. W. (2003).
Role of heparan sulfate for attachment and entry of tick-borne
encephalitis virus. Virology 308, 92–100.
Kuhn, R. J., Zhang, W., Rossmann, M. G. & 9 other authors (2002).
Structure of dengue virus: implications for flavivirus organization,
maturation, and fusion. Cell 108, 717–725.
Lescar, J., Roussel, A., Wien, M. W., Navaza, J., Fuller, S. D.,
Wengler, G., Wengler, G. & Rey, F. A. (2001). The fusion
glycoprotein shell of Semliki Forest virus: an icosahedral assembly
primed for fusogenic activation at endosomal pH. Cell 105, 137–148.
Lin, C. W. & Wu, S.-C. (2003). A functional epitope determinant on
domain III of the Japanese encepahalitis envelope protein interacted
with neutralizing-antibody combining sites. J Virol 77, 2600–2606.
Mandl, C. W., Kroschewski, H., Allison, S. L., Kofler, R., Holzmann,
H., Meixner, T. & Heinz, F. X. (2001). Adaptation of tick-borne
encephalitis virus to BHK-21 cells results in the formation of
multiple heparan sulfate binding sites in the envelope protein and
attenuation in vivo. J Virol 75, 5627–5637.
Modis, Y., Ogata, S., Clements, D. & Harrison, S. C. (2003). A
ligand-binding pocket in the dengue virus envelope glycoprotein.
Proc Natl Acad Sci U S A 100, 6986–6991.
Modis, Y., Ogata, S., Clements, D. & Harrison, S. C. (2004).
Structure of the dengue virus envelope protein after membrane
fusion. Nature 427, 313–319.
Mukhopadhyay, S., Kim, B. S., Chipman, P. R., Rossmann, M. G. &
Kuhn, R. J. (2003). Structure of West-Nile virus. Science 302, 248.
Murphy, F. A. (1980). Togavirus morphology and morphogenesis. In
The Togavirus. Biology, Structure, Replication, pp. 241–316. Edited by
R. W. Schlesinger. New York: Academic Press.
Rey, F. A., Heinz, F. X., Mandl, C., Kunz, C. & Harrison, S. C. (1995).
The envelope glycoprotein from tick-borne encephalitis virus at 2 A˚
resolution. Nature 375, 291–298.
Roehrig, J. T. (2003). Antigenic structure of flavivirus proteins. Adv
Virus Res 59, 141–175.
Roehrig, J. T., Bolin, R. A. & Kelly, R. G. (1998). Monoclonal antibody
mapping of the envelope glycoprotein of the dengue 2 virus, Jamaica.
Virology 246, 317–328.
Smithburn, K. C., Taylor, R. M., Rizk, F. & Kader, A. (1954).
Immunity to certain arthropod-borne viruses among indigenous
residents of Egypt. Am J Trop Med Hyg 3, 9–18.
Sreerama, N. & Woody, R. W. (2000). Estimation of protein
secondary structure from circular dichroism spectra: comparison of
CONTIN, SELCON, and CDSSTR methods with an expanded reference
set. Anal Biochem 287, 252–260.
Tetin, S. Y., Prendergast, F. G. & Venyaminov, S. Y. (2003).
Accuracy of protein secondary structure determination from circu-
lar dichroism spectra based on immunoglobulin examples. Anal
Biochem 321, 183–187.
van der Most, R. G., Corver, J. & Strauss, J. H. (1999). Mutagenesis
of the RGD motif in the yellow fever virus 17D envelope protein.
Virology 265, 83–95.
Volk, D. E., Beasley, D. W. C., Kallick, D. A., Holbrook, M. R., Barrett,
A. D. T. & Gorenstein, D. G. (2004). Solution structure and anti-
body binding studies of the envelope protein domain III from the
New York strain of West Nile virus. J Biol Chem 279, 38755–38761.
Wengler, G. & Wengler, G. (1989). Cell-associated West Nile
flavivirus is covered with E+pre-M protein heterodimers which
are destroyed and reorganized by proteolytic cleavage during virus
release. J Virol 63, 2521–2526.
White, M. A., Liu, D., Holbrook, M. R., Shope, R. E., Barrett, A. D. T. &
Fox, R. O. (2003). Crystallization and preliminary X-ray diffraction
analysis of Langat virus envelope protein domain III. Acta Crystallogr
D Biol Crystallogr 59, 1049–1051.
Wu, K. P., Wu, C. W., Tsao, Y. P., Kuo, T. W., Lou, Y. C., Lin, C. W., Wu,
S. C. & Cheng, J. W. (2003). Structural basis of a flavivirus
recognized by its neutralizing antibody: solution structure of the
domain III of the Japanese encephalitis virus envelope protein. J Biol
Chem 278, 46007–46013.
412Journal of General Virology 86
J. J. H. Chu and others