In situ reactions of monoclonal antibodies with a viable mutant of Murray Valley encephalitis virus reveal an absence of dimeric NS1 protein

Murray Valley encephalitis virus (MVEV) is a mosquito-borne flavivirus endemic to tropical regions of Australia and Papua New Guinea (Mackenzie et al., 1994). It is a member of the Japanese encephalitis virus (JEV) complex and is closely related antigenically to Alfuy virus (ALFV), JEV and Kunjin virus (KUNV), which co-circulate with MVEV in regions of northern Australia and Papua New Guinea (Hall et al., 2002; Mackenzie et al., 2002; May et al., 2006). MVEV is the main aetiological agent of arboviral encephalitis in Australia, occurring as isolated cases or sporadic outbreaks mainly in the north-western part of the continent (Mackenzie et al., 1994).

The ∼11 kb single-stranded, positive-sense RNA genome of flaviviruses is translated from a single open reading frame in the order NH₂-C-prM-E-NS1-NS2A-NS2B-NS3-NS4A-NS4B-NS5-COOH (Lindenbach & Rice, 2001). Cleavage of the polyprotein by host and viral proteases occurs both co- and post-translationally to yield three structural and seven non-structural proteins. Following translation, the first non-structural protein (NS1) translocates to the lumen of the endoplasmic reticulum, where dimerization of NS1 occurs, making the protein amphipathic (Winkler et al., 1989). This event was proposed to be required for the interaction of NS1 with the membrane-bound virus replication complex with which it plays an essential, yet undefined role (Mackenzie et al., 1996; Lindenbach & Rice, 1997; Muylaert et al., 1997). NS1 is also secreted from infected mammalian cells inducing a strong, non-neutralizing antibody response, which has been shown to convey protection against subsequent challenge with a homologous virus (Schlesinger et al., 1985, 1986; Hall et al., 1996). Interestingly, secreted NS1 has been implicated in dengue virus pathogenesis (Jacobs et al., 2000); however, a similar role for NS1 in the induction of pathogenesis in other flaviviral infections remains to be demonstrated. The secreted forms of tick-borne encephalitis virus and dengue virus 1 NS1 have been shown to be hexameric (Crooks et al,. 1994; Flamand et al., 1999), the later dissociating into dimeric subunits by the disruption of weak hydrophobic bonds (Flamand et al., 1999). The predominance of oligomeric NS1 in infected cells during infection in vitro and in vivo coupled with the proposed role of NS1 dimerization in facilitating membrane association and the formation of hexamers have suggested that this event is critical for NS1 function.

Previously, we reported the identification of a single amino acid substitution in KUNV NS1 that prevented NS1 dimerization yet allowed virus replication (Hall et al., 1999). The substitution occurred at a Pro²⁵⁰ of NS1, which is conserved throughout the flavivirus genus. In the present study, we have shown that a homologous substitution at residue 250 of MVEV NS1 also abolished NS1 dimerization. We also report the generation of a panel of 27 hybridomas to MVEV NS1 and the selection of a monoclonal antibody (mAb), 2E3, that selectively binds dimeric but not monomeric NS1. To determine whether replication of mutant MVEV was due to the presence of residual or unstable NS1 dimer, we fixed viral proteins in situ and probed with mAb 2E3. The results address the significance of NS1 dimerization in virus replication and suggest that mutations at residue 250 of NS1 may be useful attenuation markers for multiple flaviviruses.

Cell culture.
Mouse myeloma cells (X63-Ag8.653) and hybridoma cell lines were cultured in RPMI 1640 (Gibco) supplemented with 20 % fetal bovine serum (FBS) and incubated at 37 °C with 5 % CO₂ in a humidified atmosphere. African green monkey kidney (Vero) cells were propagated in DMEM containing 10 % FBS and incubated at 37 °C with 5 % CO₂ in a humidified environment. Aedes albopictus (C6/36) cells were grown in RPMI 1640 containing 10 % FBS and incubated at 28 °C. All media were supplemented with penicillin and streptomycin.

Viruses.
The prototype MVEV strain (MVE-1-51) used in this study was isolated originally from the brain of a fatal human case of encephalitis during the 1951 outbreak in the Murray Valley of south-eastern Australia. An infectious clone (vM212) derived from the prototype MVEV (Lee & Lobigs, 2000) was also included. Other flaviviruses used in this study were as follows: ALFV, MRM 3929 strain; dengue virus 2 (DEN2), New Guinea C strain; JEV, Nakayama strain; Kokobera virus (KOKV), MRM 32 strain; KUNV, MRM 61C strain; West Nile virus (WNV) Sarafend strain; and yellow fever virus (YFV), 17D 204 vaccine strain. Virus stocks were prepared as 20 % suckling mouse brain suspensions or as infected Vero cell culture supernatants and viral titres were determined by TCID₅₀ assay.

Production of hybridoma cell lines.
Six-week-old BALB/c mice were immunized with 5 µg immunoaffinity-purified MVEV-1-51 NS1 diluted in PBS with 50 % Freunds incomplete adjuvant by the intraperitoneal (i.p.) route. The procedures for immunoaffinity purification of MVEV NS1 and mouse immunization have been described elsewhere (Hall et al., 1991, 1996). Mice were boosted at 2 weeks with the same preparation i.p. and at 8 weeks with 5 µg pure NS1 without adjuvant by the intravenous (i.v.) route. Five days after the final boost, mice were culled and the spleens were harvested for fusions. Hybridomas were produced as described previously (Hall et al., 1988). Hybridoma cell lines that secreted antibodies specific to MVEV NS1 were identified by screening culture supernatants on C6/36 cells infected with MVEV-1-51 by ELISA as described previously (Hall et al., 1988). Monoclonal antibodies were isotyped using the Mouse Typer Panel (Bio-Rad) according to the manufacturers instructions.

Determination of epitope conformation of MVEV NS1 by Western blot.
NS1 antigens were prepared by infection of C6/36 or Vero cells with MVEV at an m.o.i. of 1. At 96 (C6/36) or 60 (Vero) h post-infection (p.i.), cells were harvested with a cell scraper and washed three times with BS9 buffer (120 mM NaCl, 50 mM H₃BO₃, pH 9.0). The cells were resuspended in BS9 buffer containing 1 % Triton X-100 and 0.1 % sodium lauryl sulfate and sonicated. The lysate was clarified by centrifugation at 12 000 g for 10 min at 4 °C and stored at 70 °C. Secreted NS1 was harvested in the cell culture supernatants from Vero cells infected for 60 h and clarified by centrifugation at 3000 r.p.m. for 10 min at 4 °C.

For reduction and carboxymethylation of NS1, lysate antigen was buffer exchanged into 0.1 M Tris/HCl (pH 8.1) and treated with dithiothreitol to a final concentration of 0.1 M. This preparation was flushed with nitrogen and heated at 95 °C for 5 min. Reduced protein was carboxymethylated by the addition of iodoacetic acid to a final concentration of 0.3 M. The antigen was again flushed with nitrogen gas and incubated at 37 °C for 1 h. Complete carboxymethylation of the protein sample was demonstrated with Ellmans reagent, as described elsewhere (Ellman, 1959).

Samples prepared as above were diluted in an equal volume of non-reducing Laemmli sample buffer [62.5 mM Tris/HCl (pH 6.8), 10 % glycerol, 0.025 % bromophenol blue] and electrophoresed at 150 V through 420 % gradient polyacrylamide gels (iGels; Life Gels). Separated proteins were transferred to a Hybond C nitrocellulose membrane (Amersham) using a semi-dry transfer apparatus (Bio-Rad). Nitrocellulose membranes were incubated for 1 h in blocking buffer [0.05 M Tris/HCl (pH 8.0), 1 mM EDTA, 0.15 M NaCl, 0.05 % (v/v) Tween 20, 0.2 % w/v casein] prior to probing with mAbs diluted in blocking buffer for a further 1 h. Membranes were washed three times with PBST wash buffer [PBS containing 0.05 % (v/v) Tween 20, pH 7.2] and incubated for 1 h in horseradish peroxidase (HRP)-conjugated goat anti-mouse IgG (H+L) (Dako) diluted 1 : 2000 in blocking buffer. Membranes were again washed three times and specifically bound protein bands were stained with substrate solution [PBS containing 1.5 mM diaminobenzidine, 0.06 % (v/v) H₂O₂, pH 7.2].

Fixed-cell ELISA.
Fixed-cell ELISA was performed as described previously (Hall et al., 1988). Briefly, C6/36 or Vero cells were infected with the appropriate virus at an m.o.i. of 1. At 96 (C6/36) or 60 (Vero) h p.i., the cells were fixed with PBS containing 20 % (v/v) acetone and 0.2 % (w/v) BSA or with PBS containing 4 % formaldehyde. The plates were blocked with 200 µl blocking buffer for 1 h prior to probing of fixed antigen with mAbs at a pre-determined optimum dilution or serially diluted 2-fold across the plate. Wells were washed four times with wash buffer and bound antibodies were detected with HRP-conjugated goat anti-mouse IgG (H+L) diluted 1 : 2000 in blocking buffer. The plates were washed six times and enzyme activity was visualized by the addition 100 µl substrate solution [1 mM 2,2'-azino-bis(3-ethylbenzthiazoline-6-sulfonic acid) (ABTS), 3 mM H₂O₂ in a buffer prepared by mixing approximately 6 vols 0.1 M citric acid with 5 vols 0.2 M Na₂HPO₄ to give a pH of 4.2]. Absorbance was measured at 405 nm. The criterion for specific recognition of antigen was defined as an A₄₀₅ value of at least 0.4 and at least 2-fold greater than that generated by probing uninfected C6/36 cells with the corresponding antibody dilutions.

In vitro viral growth kinetics.
Vero cells (4x10⁵) in 35 mm diameter tissue culture dishes were infected with virus at an m.o.i. of 1. Infected cells were washed twice with Hanks balanced salts solution containing 20 mM HEPES buffer (pH 8.0) and 0.2 % BSA at 1 h p.i. and fresh growth medium was added and the incubation continued. At 16, 20, 24 and 28 h p.i., 0.2 ml samples were taken and frozen at 70 °C. Virus titres were determined by plaque assay on Vero cell monolayers, as described previously (Licon Luna et al., 2002).

Mouse virulence studies.
Groups of 3-week-old Swiss outbred mice (n=5) were inoculated with tenfold serial dilutions of each virus by the i.p. route to assess the efficiency of neuroinvasion or by the intracranial (i.c.) route to assess the level of neurovirulence, as described previously (Lee & Lobigs, 2000).

Interferon-α receptor deficient (IFN-α-R^/) mice (Müller et al., 1994) were obtained from the Animal Breeding Facility at the John Curtin School of Medical Research, Canberra, Australia. Groups of 6-week-old male or female IFN-α-R^/ mice were infected with 10² p.f.u. MVEV_NS1-250Leu i.v., and morbidity and mortality were monitored daily over a period of 4 weeks. Alternatively, mice were euthanized at 7 days p.i. and brains were harvested for determination of virus titres, as described previously (Lobigs et al., 2003).

Statistics.
Differences in survival ratios for mouse virulence experiments were assessed using Fishers exact test and differences in mean survival time were analysed for significance using the MannWhitney test.

Selection of mAbs that distinguish between dimeric and monomeric NS1
Previously, we demonstrated that substitution of the conserved proline at residue 250 with leucine in KUNV NS1 prevented dimerization, but did not abolish virus replication (Hall et al., 1999). To gain more conclusive evidence that this mutation in NS1 completely prevents dimer formation within the infected cell and to test the generality of the effect of the Pro²⁵⁰→Leu mutation on NS1 dimer formation and other biological properties of flaviviruses belonging to the JEV serotype, we generated a NS1 dimer-specific mAb to be used in situ on cells infected with a mutant of MVEV encoding the Pro²⁵⁰→Leu mutation.

In total, 28 hybridoma cell lines expressing mAbs reactive to the NS1 protein of MVEV virus were isolated. Cross-reactivity of these antibodies with NS1 protein from a panel of flaviviruses revealed that all but five also recognized the ALFV subtype of MVEV, two mAbs (4D12 and 10G9) reacted with the closely related JEV, and a single mAb (2H9) cross-reacted with KUNV and WNV (Table 1). Two mAbs (4G4 and 2E5) recognized all flaviviruses tested (Table 1). Each mAb was also tested by Western blotting to define the nature of the epitopes recognized (Table 2). Representative blots probed with 2E3, 4G4 and 4F7 are presented in Fig. 1 to demonstrate the typical band pattern observed for MVEV NS1 in cell lysate. Boiling NS1 generated two prominent bands of 45 and 53 kDa, respectively (Fig. 1b). These bands corresponded to NS1 and NS1' (an elongated form of NS1 containing part of the N terminus of NS2A; Mason, 1989; Blitvich et al., 1999). Two faint bands representing viral envelope (E)NS1 and ENS1' were also seen in this sample (Blitvich et al., 1995). Unboiled, unreduced MVEV NS1 appeared as a cluster of three bands (88, 102 and 110 kDa), which represent the NS1 homodimer, the NS1/NS1' heterodimer and the NS1' homodimer. Three minor, high-molecular-mass bands were also seen in this preparation, which were shown previously to be complexes between the three dimeric forms of NS1/NS1' and the E protein (Blitvich et al., 1995). Eleven mAbs recognized boiled and reduced NS1. Reduction of NS1 was associated with a concomitant reduction in the electrophoretic mobility of both NS1 and NS1' as a result of structural changes imposed by elimination of intramolecular disulfide bonds in these proteins (Fig. 1b). A single mAb (4F7) bound boiled, reduced and carboxymethylated MVEV NS1, indicating that this antibody recognized a linear epitope (Fig. 1c). The other ten mAbs were probably binding to epitopes that had renatured following the removal of reducing agents during the process of electrophoresis and blotting. Whilst the binding of a number of mAbs was adversely affected by boiling NS1, only 2E3 failed to bind to the heat-denatured protein (Fig. 1a). Denaturation of dimeric NS1 with mild acid treatment produced similar results (data not shown). Thus, binding of this mAb to NS1 was exclusively dependent on dimerization of the antigen and provided us with a tool to test the effect of the Pro²⁵⁰→Leu mutation on dimerization of NS1 in situ.

Table 1. Viral specificity of anti-MVEV NS1 mAbs in ELISA mAbs in hybridoma supernatant were serially diluted 2-fold and tested for reactivity with infected C6/36 cells fixed with 20 % acetone in PBS. The criterion for specific reaction was defined as an absorbance value of at least 0.4 and at least twice the absorbance value generated by probing uninfected C6/36 cells with the corresponding antibody dilution.

Table 2. Western blot analysis of epitopes recognized by MVEV NS1 mAbs ND, Not determined.

(32K): Fig. 1. (a, b) MVEV-infected Vero cell lysates were electrophoresed on 420 % gradient polyacrylamide gels, transferred to nitrocellulose membranes and probed with mAb 2E3 (a) or 4G4 (b). (c) Lysates were heated (lane 1), heated and reduced (lane 2) or heated, reduced and carboxymethylated (lane 3) prior to PAGE and Western blotting with mAb 4F7.

Pro²⁵⁰→Leu mutation inserted into the NS1 protein of MVEV impairs virus growth
The NS1 Pro²⁵⁰→Leu mutation was incorporated into the infectious cDNA clone of MVEV (Lee & Lobigs, 2000) to yield a viable mutant virus (MVEV_NS1-250Leu) following transfection of BHK cells with in vitro-synthesized RNA corresponding to the viral genome. However, the growth efficiency of the mutant in mammalian cells was reduced significantly relative to that of infectious clone-derived wild-type (WT) MVEV (vM212). In a single-step growth experiment on Vero cells, virus titres in the culture supernatant of MVEV_NS1-250Leu-infected cells were 100- to 1000-fold less than those of vM212-infected cells at all time points (Fig. 2). This was consistent with the observation that MVEV_NS1-250Leu produced uniformly smaller plaques on Vero cell monolayers compared with WT virus (results not shown). However, growth of MVEV_NS1-250Leu was less markedly reduced (∼10-fold) in BHK cells compared with WT MVEV, giving rise to virus stocks of 1x10⁶ p.f.u. ml¹ at 48 h p.i. (results not shown).

(11K): Fig. 2. Replication of MVEV infectious clone-derived WT (vM212) and mutant MVEVNS1-250Leu viruses in Vero cells. Vero cells were infected at an m.o.i. of 1 and virus titres in the culture fluid were determined at the indicated times by plaque assay on Vero cells. , WT MVEV; , MVEVNS1-250Leu.

MVEV_NS1-250Leu exhibits reduced neuroinvasiveness in weanling mice and is attenuated in mice defective in IFN-α/β responses
Prototype virus, MVEV-1-51, and the corresponding infectious clone-derived WT virus are lethal when inoculated into 3-week-old mice by the i.p. or i.c. route (Lobigs et al., 1988; Lee & Lobigs, 2000). Given that the i.c. LD₅₀ and i.p. LD₅₀ values do not differ by more than 10-fold, the WT viruses are considered virulent. By this criterion, MVEV_NS1-250Leu is partially attenuated in the 3-week-old mouse model; thus, despite comparable neurovirulence (LD₅₀<1) relative to WT MVEV, infection by the i.p. route gave only 60 % mortality over the range 6x10²6x10⁴ p.f.u. (Table 3).

Table 3. Virulence of MVEVNS1-250Leu in 3-week-old Swiss mice ND, Not done.

To investigate further the mechanism of attenuation, the virulence of MVEV_NS1-250Leu in mice defective in IFN-α/β responses was tested. These mice are highly susceptible to extraneural infection with JEV serotype flaviviruses, even as adults (Lee & Lobigs, 2002; Lobigs et al., 2003; Lee et al., 2004). Whilst infection of 6-week-old IFN-α/β-R^/ mice with 10² p.f.u. WT MVEV i.v. resulted in 100 % mortality by day 6 p.i., infection with MVEV_NS1-250Leu gave <50 % mortality (Fig. 3a). This significant difference in mortality (P=0.045) was accompanied by a significant difference in mean survival time between groups of mice infected with WT (5.6 days) or NS1 mutant MVEV (8.6 days) (P=0.005). Brains were also collected from a group (n=12) of IFN-α/β-R^/ mice infected with 10² p.f.u. MVEV_NS1-250Leu i.v. in order to determine virus load at 7 days p.i. (i.e. ∼2 days preceding the mean time of death) (Fig. 3b). It is noteworthy that these mice uniformly succumb to infection with the same dose of WT MVEV prior to day 7 p.i. and always show detectable virus in the brain at the time of death (Lobigs et al., 2003; M. Lobigs and E. Lee, unpublished results). In comparison, relatively low levels of virus (10³10⁴ p.f.u. g¹) were detected in the brains of only five out of 12 MVEV_NS1-250Leu-infected mice (42 %), which was consistent with the mortality rate and the putative cause of death by encephalitis. In summary, the virulence experiments suggested that the NS1 mutation prevents or delays virus entry into the brain, most likely by reducing virus growth in extraneural tissues, and that an increased sensitivity to type I IFN does not markedly contribute to the mechanism of attenuation of MVEV_NS1-250Leu.

(13K): Fig. 3. Attenuation of MVEVNS1-250Leu in IFN-α/β-R/ mice. (a) Six-week-old IFN-α/β-R/ mice (n=16) were infected with 102 p.f.u. MVEVNS1-250Leu i.v. Mortality was recorded daily and surviving mice were monitored for 28 days. The susceptibility of a group of IFN-α/β-R/ mice (n=5) to infection with WT MVEV is shown for comparison (Lobigs et al., 2003). (b) Virus titres in brains of 6-week-old IFN-α/β-R/ mice at 7 days after infection with 102 p.f.u. MVEVNS1-250Leu i.v. were determined by plaque assay on Vero cells. The lower limit of detection is indicated by the dashed line.

Pro²⁵⁰→Leu mutation abolishes dimerization of intracellular and extracellular NS1 of MVEV
Consistent with our previous study on the Pro²⁵⁰→Leu mutation in NS1 of KUNV, Western blots of cell culture supernatant of MVEV_NS1-250Leu-infected Vero cells revealed that this substitution abolished NS1 dimerization (Fig. 4). To determine more rigorously whether the mutation completely eliminated dimerization or, alternatively, resulted in the formation of an unstable dimer that was denatured during PAGE, in situ binding of the dimer-specific mAb 2E3 was evaluated. Mammalian and arthropod cells infected with either WT MVEV or MVEV_NS1-250Leu were fixed with 4 % formalin in PBS and tested for reaction with mAb 2E3 (hybridoma supernatant diluted 1 : 5 in blocking buffer) in ELISA. Fig. 5(a, b) shows that 2E3 recognized NS1 in both Vero and C6/36 cells infected with WT infectious clone-derived MVEV, but failed to react with cells infected with the MVEV_NS1-250Leu mutant. This experiment was replicated with a PBS/20 % acetone fixative buffer with similar results (Fig. 5c, d). As a further control, mAb 4G4, which binds with high affinity to both monomeric and dimeric NS1, was used as a control to determine whether the differences in 2E3 binding correlated with the absence of dimeric NS1 in MVEV_NS1-250Leu virus-infected cells, rather than a reduction in the amount of NS1 generated by this virus due to attenuation. Fig. 5 clearly demonstrates that comparable amounts of NS1 were present in cells infected with each virus. Together, these results indicated that MVEV_NS1-250Leu fails to produce detectable levels of cell-associated dimeric NS1.

(13K): Fig. 4. Western blot of culture supernatants from Vero cultures infected with WT MVEV or mutant MVEVNS1-250Leu. Supernatants were either boiled (+) or unboiled () in non-reducing sample buffer prior to electrophoresis through 420 % gradient SDS-polyacrylamide gels. Protein bands were electrotransferred from gels to nitrocellulose membranes and probed with NS1-specific mAb 4G4 (a) or NS1 dimer-specific mAb 2E3 (b). Arrows indicate NS1 dimers and monomers.

(16K): Fig. 5. Detection of dimeric NS1 by mAb 2E3 in fixed-cell ELISA. C6/36 cells (a, c) or Vero cells (b, d) were infected with infectious clone-derived WT MVEV or with mutant MVEVNS1-250Leu. Infected cells were fixed with 4 % formaldehyde in PBS (a, b) or 20 % acetone in PBS (c, d) and probed with mAb 4G4 (filled columns) or 2E3 (open columns) followed by goat anti-mouse HRP conjugate and ABTS substrate. Absorbance readings for each sample were normalized by subtracting the background absorbance for the reactions of each mAb reacting with uninfected fixed Vero or C6/36 cells as appropriate.

Dimerization of NS1 was shown previously to be associated with a change in its physical characteristics, from a hydrophilic subunit to an amphipathic protein that pelleted with the membranes of infected cells (Winkler et al., 1989; Mason, 1989). Based on this observation it was proposed that dimerization was required for membrane association of NS1 (Winkler et al., 1989; Fan & Mason, 1990). As NS1 localizes with the viral replication complex to membranes associated with RNA replication (Mackenzie et al., 1996) and NS1 has been shown to interact with membranes independent of other viral proteins (Fan & Mason, 1990), it was a natural corollary that dimerization was a pre-requisite of its role in RNA replication. However, we subsequently reported that a Pro²⁵⁰→Leu substitution at residue 250 of NS1 in a KUNV infectious clone (FLSD) resulted in loss of NS1 dimerization, whilst permitting virus replication and the correct trafficking of this protein in infected cells (Hall et al., 1999). In the present study, we were also unable to detect dimeric NS1 when a homologous substitution was introduced into an MVEV infectious cDNA clone (vM212). To ascertain whether this mutation truly eliminated dimerization, we fixed the protein in situ in cells infected with the mutant virus and probed them with mAb 2E3 specific for the dimer form of NS1 of MVEV. We failed to detect even trace levels of dimeric NS1 in cells infected with MVEV_NS1-250Leu. However, 2E3 clearly reacted with cells infected with WT MVEV. Although we have not mapped the contact residues associated with the binding of mAb 2E3 to WT NS1, it is unlikely that a local structural change due to the mutation at residue 250 in mutant NS1 results in the lack of 2E3 binding. This mAb also failed to recognize the monomeric form of WT NS1, even residual monomers observed in unheated cell lysates from WT MVEV infected cultures (see Fig. 1a).

The above data strongly suggest that dimerization is not essential for NS1 function in virus replication. However, we cannot rule out the possibility that trace amounts of dimeric NS1, below the detection threshold of the assays used in this study, are present in cells infected with MVEV_NS1-250Leu. Whether such low levels of functional NS1 are sufficient to sustain productive infection is debatable.

In this study, we also observed that loss of NS1 dimerization resulted in significant attenuation of MVEV in weanling mice by the peripheral route of infection. Reduced mouse neuroinvasiveness was also observed with the corresponding KUNV mutant, suggesting that substitution of the highly conserved Pro²⁵⁰ in NS1 represents a marker of attenuation in flaviviruses of the JEV antigenic complex. This attenuation is most likely associated with retarded growth in extraneural tissues, as indicated by the slower growth kinetics observed with both MVEV and KUNV mutants in Vero cells and extended time to death of the MVEV mutant in IFN-α/β-R^/ mice.

A panel of 28 mAbs reactive to MVEV NS1 protein were produced in the course of this study. This suite of antibodies complements the six MVEV NS1 mAbs that we identified previously (Hall et al., 1990). Two of the antibodies identified in the present study were flavivirus group reactive. Indeed these mAbs have proved useful in our laboratory because of their ability to bind with high affinity to NS1 from a range of flaviviruses. These antibodies have been employed for immunoaffinity chromatography purification of NS1 from several flaviviruses and recombinant NS1 constructs (D. C. Clark & R. A. Hall, unpublished data), for the sensitive detection of NS1 in clinical samples (Macdonald et al., 2005) and as markers of infection and surface expression of NS1 in inoculated cell cultures (Arnold et al., 2004).

The authors thank Megan Pavy for assistance with mouse virulence experiments, Therese Brand for assistance with NS1 preparation and mouse immunizations, Fiona May for assistance with mAb characterization, Steve Matarello for assistance with NS1 dimer denaturation experiments, Nicola Armstrong for assistance with hybridoma culture supernatant production and Wai Yuen Cheah for assistance with manuscript preparation. This study was funded by the National Health and Medical Research Council of Australia.

Footnotes

^,†, Bradley J. Blitvich³ †Present address: Centre for Immunology and Cancer Research, Department of Medicine, University of Queensland, Princess Alexandra Hospital, Brisbane, Australia. ‡Present address: Department of Veterinary Microbiology and Preventive Medicine, Iowa State University, Ames, IA, USA.