Journal of Virological Methods 99 (2002) 41–51

A rapid immune plaque assay for the detection of Hendraand Nipah viruses and anti-virus antibodies

Gary Crameri, Lin-Fa Wang, Christopher Morrissy, John White,Bryan T. Eaton *

CSIRO, Australian Animal Health Laboratory, Pri�ate Bag 24, Geelong, Vic. 3220, Australia

Received 21 May 2001; received in revised form 1 August 2001; accepted 1 August 2001

Abstract

Rapid immune plaque assays have been developed to quantify biohazard level 4 agents Hendra and Nipah virusesand detect neutralising antibodies to both viruses. The methods rely on the fact that both viruses rapidly generatelarge syncytia in monolayers of Vero cells within 24 h and that monospecific antiserum to the Hendra virusphosphoprotein (P) detects that protein in both Hendra and Nipah viruse-induced syncytia after methanol fixation ofvirus-infected cells. The P protein is a constituent of the ribonucleoprotein core of the viruses and a component ofthe viral RNA-dependent RNA polymerase and is made in significant amounts in infected cells. In the immune plaqueassay, anti-P antibody is localised by an alkaline phosphatase-linked second antibody and the Western blot substrates5-bromo-4-chloro-3-indolyl phosphate and p-nitro blue tetrazolium. A modification of the rapid immune plaque assaywas also used to detect antibodies to Nipah virus in a panel of porcine field sera from Malaysia and the resultsshowed good agreement between the immune plaque assay and a traditional serum neutralisation test. After methanolfixation, plates can be stored for up to 7 months and may be used in the immune plaque assay to complement theenzyme-linked immunosorbent assay screening of sera for antibodies to Nipah virus. At present, all enzyme-linkedimmunosorbent assay positive sera are subject to confirmatory serum neutralisation tests. Use of the immune plaqueassay may reduce the number of sera requiring confirmatory neutralisation testing for Nipah virus antibodies underbiohazard level 4 conditions by identifying those that generate false positive in the enzyme-linked immunosorbentassay. © 2002 Elsevier Science B.V. All rights reserved.

Keywords: Rapid immune plaque assay; Hendra virus; Anti-virus antibody; Nipah virus

www.elsevier.com/locate/jviromet

1. Introduction

Hendra and Nipah viruses are related zoonoticmembers of the Paramyxo�iridae family (Wang et

al., 2001), whose natural hosts are most probablyfruit bats in the genus Pteropus. Anti-Hendravirus antibodies have been detected in a highproportion of the four species of Pteropid batsfound in Australia (Young et al., 1996) and pre-liminary data indicate two different members ofthe same genus in Malaysia have antibodies to

* Corresponding author. Tel.: +61-3-5227-5020; fax: +61-3-5227-5532.

E-mail address: bryan.eaton@li.csiro.au (B.T. Eaton).

0166-0934/02/$ - see front matter © 2002 Elsevier Science B.V. All rights reserved.

PII: S 0166 -0934 (01 )00377 -9

G. Crameri et al. / Journal of Virological Methods 99 (2002) 41–5142

Nipah virus (Field et al., 2001). Both viruses arehighly virulent. Hendra virus was responsible forthe death of 17 horses and two humans in threeseparate outbreaks in Queensland since 1995(Murray et al., 1995; Hooper et al., 1996;Rogers et al., 1996; Field et al., 2000; Hooper etal., 2000). In Malaysia although Nipah viruscaused relatively few mortalities during itsspread through the pig population, there havebeen over 100 fatal human cases of encephalitiscaused by the virus (Chua et al., 2000). Theextreme pathogenicity of Hendra and Nipahviruses requires that investigations using livevirus be conducted under the highest levels ofmicobiological security (biohazard level 4). Atthis level, virus is contained within flexible filmisolators or in environments, where workerswear protective plastic suits supplied with air forbreathing and maintenance of a positive pres-sure compared to the laboratory.

Comparison of nucleic acid and deducedamino acid sequences with other members of thefamily confirms that Hendra and Nipah virusesare members of the sub-family Paramyxo�iridae,but the limited homology with members of theMorbilli�irus, Rubula�irus and Respiro�irus gen-era suggest that they should be classified in anew genus within the sub-family (Chua et al.,2000; Harcourt et al., 2000; Wang et al., 2001).The name Henipa�irus has been suggested forthe new genus (Wang and Eaton, in press).

Current tests for Hendra and Nipah virusesinclude polymerase chain reaction, indirect im-munofluorescence and peroxidase tests to detectantigen in infected cells and tissue, a microtitreassay to quantify virus and a serum neutralisa-tion test and enzyme-linked immunosorbent as-say for antibody detection (Murray et al., 1995).The microtitre assay relies on the generation ofa cytopathic effect in Vero cells. Although syn-cytia caused by Hendra and Nipah viruses arevisible within 24 h, the traditional microtitre as-say requires up to 3 days, because it depends onthe presence of microscopically visible cell de-struction. A microtitre test to detect anti-Hendraand anti-Nipah virus antibodies also requires aminimum of 3 days. In recent years, many pro-cedures have been developed that significantly

reduce virus assay times, often to �24 h. Suchtests rely on detection of antigen in infectedcells by monoclonal or polyclonal antibodiesthat are localised by either immunofluorescenceor immunoperoxidase staining methods.

Horse radish peroxidase-linked antibodies(Nakane and Pierce, 1967) have been used ex-tensively to detect viral antigen directly (Uber-tini et al., 1971; Borovec and Uren, 1997) andas secondary antibodies (Terpstra et al., 1984;Swenson and Kaplan, 1985; Lipson et al., 1992;Harder et al., 1993). Primary anti-viral antibodyhas been localised by biotinylated secondary an-tibody and avidin-peroxidase (Chou and Scott,1988), horseradish peroxidase-conjugated proteinA (Belanger et al., 1988) and non-enzyme-linkedsecondary antibody and peroxidase–anti-perox-idase complex (Tanishita et al., 1984). Immunos-taining techniques have also been used inconjunction with conventional plaque assays toquantify viruses that grow slowly or form ill-defined plaques and to detect antibody to suchviruses (Tanishita et al., 1984; Harder et al.,1993). We describe a method, the immuneplaque assay, for rapid detection and quantifica-tion of Hendra and Nipah viruses and anti-viralantibody. The method makes use of the factthat these viruses rapidly generate large syncytiain cell monolayers and that after methanol fixa-tion the viral phosphoprotein (P protein), a con-stituent of the ribonucleoprotein core of thevirus and a component of the viral transcriptase,can be detected in syncytia using an antibodygenerated to a bacterial-expressed portion of theHendra virus P protein, alkaline phosphatase-linked second antibody and the Western blotsubstrates 5-bromo-4-chloro-3-indolyl phosphateand p-nitro blue tetrazolium. The immuneplaque assay has also been used to detect anti-viral antibodies in sera by virtue of their abilityto bind to viral antigen in methanol-fixed syncy-tia in infected cell monolayers. This variation ofthe immune plaque assay may have applicationas a secondary serological test for anti-Hendraand anti-Nipah antibodies in laboratories thatdo not have the biohazard level 4 facilities re-quired for the safe handling of these pathogens.

G. Crameri et al. / Journal of Virological Methods 99 (2002) 41–51 43

2. Materials and methods

2.1. Virus and cells

Continuous cells lines Vero, MARC and 143Bwere maintained in the absence of antibiotics inMinimal Essential Medium containing Earle’ssalts and 10% fetal calf serum. Primary lambtestes (LT) cells were grown in the samemedium. MDCK and L929 cells were grown inMinimal Essential Medium containing Earle’ssalts containing 10% fetal calf serum and 10mM N-2-hydroxyethylpiperazine-N �-2-ethanesul-phonic acid. Swine testes (ST) cells were grownin Minimal Essential Medium containing Earle’ssalts and 10% fetal calf serum, 10 mM N-2-hy-droxyethylpiperazine-N �-2-ethanesulphonic acid,1% sodium pyruvate and 0.25% lactalbumin hy-drolysate. BHK cells were grown in Basal Me-dia Eagle supplemented with 10% fetal calfserum.

Hendra virus stock virus (titre 2–4×107

TCID50/ml) was prepared as described (Wang etal., 1998). Nipah virus was obtained from Dr K.Lam, University of Malaya and plaque purifiedtwice in Vero cells. Stock virus was generatedby passage three times in Vero cells at a multi-plicity of 0.05 TCID50/cell.

2.2. Con�entional microtitre assay

Replicates of 10-fold dilutions of Hendra andNipah viruses in Minimum Essential Mediumcontaining Earle’s salts were either added towells of a 96 well plate containing monolayersof Vero cells seeded the previous day or mixedwith a suspension of 2.5×104 Vero cells in thewells of the microtitre plate. After 3–4 days in-cubation in Minimal Essential Medium contain-ing Earle’s salts and 10% fetal calf serum at37 °C, cells were examined microscopically forthe presence of a cyptopathic effect. Virus titre,expressed in TCID50/ml, was the reciprocal ofthe dilution that generated a cytopathic effect in50% of replicate wells and was calculated by themethod of Reed and Muench (Reed andMuench, 1938).

2.3. Immune plaque assay for �iral antigen

Three methods were used to infect cells withvirus. In method 1, 2×104 Vero cells wereseeded into individual wells of 96-well microtitreplates and incubated at 37 °C overnight in 150�l Minimal Essential Medium containing Earle’ssalts and 10% fetal calf serum. Under biohazardlevel 4 conditions, medium was removed fromthe plates and serial dilutions of virus in Mini-mal Essential Medium containing Earle’s saltsand 10% fetal calf serum added to replicatewells in volumes of 100 �l. Cells were incubatedovernight at 37 °C. Method 2 is identical to thefirst method except that serial dilutions of virusin Minimal Essential Medium containing Earle’ssalts and 10% fetal calf serum were added toreplicate wells of Vero cells in volumes of 50 �l.After incubation at 37 °C for 30 min, virus wasremoved and 100 �l of Minimal EssentialMedium containing Earle’s salts and 10% fetalcalf serum was added prior to incubationovernight at 37 °C. In method 3, 2×104 Verocells were added to dilutions of virus in microt-itre wells and incubated overnight at 37 °C. Inall three methods, the culture medium was dis-carded after overnight incubation and plateswere immersed in absolute methanol in heatsealed plastic bags at room temperature and thebags’ surface sterilised with lysol during removalfrom the biohazard level 4 laboratory.

Methanol-fixed plates were air dried at roomtemperature and stored (Section 2.10). Prior touse, plates were blocked for 15 min with 100 �l0.01 M phosphate-buffered saline, pH 7.2,0.05% Tween 20 containing 2% skim milk pow-der and incubated for 30 min at 37 °C with 50�l rabbit anti-Hendra virus P antiserum diluted1:200 in 0.01 M phosphate-buffered saline, pH7.2, 0.05% Tween 20 containing 2% skim milkpowder. Plates were washed three times with0.01 M phosphate-buffered saline, pH 7.2,0.05% Tween 20 and incubated for 30 min at37 °C with 50 ul Promega anti-rabbit alkalinephosphatase conjugate (Catalog number S3731)diluted 1:2000 in 0.01 M phosphate-buffered sa-line, pH 7.2, 0.05% Tween 20 containing 2%skim milk powder. Plates were washed three

G. Crameri et al. / Journal of Virological Methods 99 (2002) 41–5144

times with 0.01 M phosphate-buffered saline, pH7.2, 0.05% Tween 20 containing 2% skim milkpowder and incubated with 50 �l Promega 5-bromo-4-chloro-3-indolyl phosphate and p-nitroblue tetrazolium substrate (Catalog numberS3771) for 10–30 min at room temperature oruntil purple coloured plaques were obviousagainst a clear background. The substrate wasremoved and plates rinsed with distilled water andair-dried. The plaques were counted using a mag-nifying glass.

2.4. Antibody detection using IPA

In a modification of the procedure, air-driedplates blocked with 0.01 M phosphate-bufferedsaline, pH 7.2, 0.05% Tween 20 containing 2%skim milk powder were incubated for 30 min at37 °C with 50 �l Malaysian porcine anti-Nipahvirus antisera diluted 1/10 and 1/100 with 0.01 Mphosphate-buffered saline, pH 7.2, 0.05% Tween20 containing 2% skim milk powder. After wash-ing as described above, plates were incubated for30 min at 37 °C with protein A-conjugated alka-line phosphatase (ICN, Catalog number 55965)diluted 1:500 in 0.01 M phosphate-buffered saline,pH 7.2, 0.05% Tween 20 containing 2% skim milkpowder. Plates were washed, incubated with 50 �lPromega 5-bromo-4-chloro-3-indolyl phosphateand p-nitro blue tetrazolium substrate and theplaques stained and counted as described above.In experiments to compare the reactivity ofMalaysian porcine field sera in the serum neutral-isation and immune plaque assays, the followingmethod was used to score the immune plaqueassay results. Plaques exhibiting maximum inten-sity with no reduction in number compared withvirus control wells (i.e. incubated without anti-serum) were scored as 2+ . Plaques exhibitingreduced staining intensity with no reduction innumber compared with control wells received a1+ score. In situations, where both the numberand intensity of plaques detected were reduced,the sera were given a +/− rating.

2.5. Con�entional serum neutralisation assay

Replicates of 2-fold dilutions of porcine sera in

growth medium were mixed in the wells of a 96well microtitre plate with an equal volume (50 �l)of medium containing �200 TCID50 Hendra orNipah virus and incubated for 1 h at 37 °C. Verocells (100 �l) in growth medium (2.5×105 per ml)were subsequently added to each well. After 3–4days at 37 °C in an atmosphere of 5% CO2,individual wells were scored for the presence orabsence of a cytopathic effect and the serum titreexpressed as the reciprocal of the highest dilutionthat prevented growth in 50% of replicate wells.End points were calculated by the method ofReed and Muench (Reed and Muench, 1938).

2.6. Neutralisation using immune plaque assay

Sera were diluted 2-fold in Minimal EssentialMedium containing Earle’s salts and 10% fetalcalf serum. Under biohazard level 4 conditions,antiserum in 50 �l duplicates or quadruplicateswas mixed with 50 �l of diluted virus, incubatedat 37 °C for 30 min and added to Vero cellmonolayers in 96 well plates and incubated at37 °C. Virus dilutions were chosen to generate25–40 plaques following adsorption of 50 �l ofvirus for 30 min at 37 °C to Vero cell monolayersin 96 well plates. After 30 min, the virus/anti-serum inocula were removed and 100 �l MinimalEssential Medium containing Earle’s salts and10% fetal calf serum was added to each well andplates incubated overnight at 37 °C. The culturemedium was discarded the next day and platesimmersed in absolute methanol for at least 10 minat room temperature prior to air-drying outsidethe biohazard level 4 facility. Plaques were de-tected by the immune plaque assay as described inSection 2.3 and the titre expressed as the recipro-cal of the serum dilution that reduced the numberof plaques to 75% of that of control untreatedvirus.

2.7. Preparation of anti-P, G, F1 and F2 antisera

For production of mono-specific antisera inrabbits, recombinant Hendra virus structuralproteins were expressed in bacterial systems usingeither the pGD, pTD or pRSET vectors as de-scribed previously (Wang et al., 1997). The fol-

G. Crameri et al. / Journal of Virological Methods 99 (2002) 41–51 45

lowing recombinant proteins were used in thisstudy: truncated P protein corresponding to theP-specific region containing aa 407–614; a por-tion of the attachment (G) protein spanning aa188–459; N-terminus truncated fusion (F1)protein containing aa 220–547 of the F0

protein; a portion of the F2 protein covering aa1–92. Approximately 1.5 mg recombinantproteins were purified using preparative poly-acrylamide gel electrophoresis and bands of in-terest were visualised in ice-cold 0.3 M KCl, cutand sliced into 2×2 mm pieces, followed bypassive elution in 50 mM NH4HCO3 containing0.1% SDS. Eluted proteins (0.5 ml at 0.5 mg/ml)were mixed with an equal volume of Freund’sincomplete adjuvant and antisera raised in rab-bits. The same injection was repeated 4 weekslater, followed by a third injection 6 weeks laterwithout adjuvant. Antibody titres and specificitywere determined by enzyme-linked immunosor-bent assay and Western blot using both recom-binant antigens and purified virus proteins. Todetermine the effectiveness of monospecific anti-sera in detecting Hendra virus-infected cells,Vero cells in 96 well plates were infected with arange of Hendra virus dilutions to generatefrom 20 to 50 microscopically visible foci perwell. After 24 h incubation, culture medium wasremoved and the plates immersed in methanol.Foci of infection were detected using the anti-P,G, F1 and F2 antisera as described in Section2.4.

2.8. Production of mouse anti-Hendra �irusascitic fluid

Plaque purified Hendra virus (20 �l containing105 TCID50) was inoculated intracerebrally intoten 5-day old suckling Balb/c mice. By 7 dayspost-inoculation two mice had died and at thattime brains were harvested from the eight re-maining moribund mice and a 10% suspensiongenerated by homogenisation in PBS. After afurther passage in suckling mouse brain the timeto death decreased to �4 days. Ten adult micewere inoculated intraperitoneally with an equalmixture of incomplete Freund’s adjuvant and a10% brain suspension from the second suckling

mouse passage. The inoculation process was re-peated 2 weeks later and after 4 weeks micewere inoculated intraperitoneally with 106

Ehrlich ascites cells and 50 �l Hendra virus-in-fected suckling mouse brain homogenate. Micewere euthanized after a single tapping.

2.9. Virus adsorption studies

Hendra virus was grown in Vero cells in me-thionine-deficient medium containing 50 �Ci/ml[35S]-methionine and purified as described previ-ously (Wang et al., 1998). [35S]-labelled-Hendravirus was diluted in phosphate-buffered salineand �3000 cpm was added in volumes of 100�l to confluent monolayers of Vero cells in 24well plates at 4 °C. After 1 h, virus was re-moved, monolayers were washed twice with ice-cold phosphate-buffered saline and the cellssolubilised in 2% sodium dodecyl sulphate. Ra-dioactivity in cell lysates was counted in a scin-tillation counter and results expressed as theproportion of added cpm that became cell asso-ciated. In a control experiment, pre-incubationof the virus with ascitic fluid diluted 1:50 com-pletely neutralised 3000 cpm [35S]-labelled-Hen-dra virus and reduced the proportion of virusbound by �60%.

2.10. Storage and testing of methanol-fixed�irus-infected plates

Microtitre plates with monolayers of Verocells containing �25 foci of infection weremethanol-fixed, removed from the biohazardlevel 4 laboratory, air dried, sealed and storedat either room temperature (20–22 °C), 4 or –20 °C. At monthly intervals a plate at each ofthe three temperatures was removed and testedto determine if the foci of virus infection couldbe detected using anti-P antiserum and standardformat as described in Section 2.3.

2.11. Methanol inacti�ation of Hendra �irus

Two experiments were done to ensure thatmethanol treatment of Hendra virus-infectedcells in the immune plaque assay inactivated the

G. Crameri et al. / Journal of Virological Methods 99 (2002) 41–5146

virus. In the first experiment, 100 �l of stockvirus was mixed with 900 �l of either ice-coldmethanol or Minimal Essential Medium contain-ing Earle’s salts and 10% fetal calf serum for 5min at 4 °C. Untreated and methanol-treatedHendra virus samples were subject to serial logdilution and virus titre determined by microtitreassay. To assess its effect on cells, methanol wasmixed in a 9:1 ratio with Minimal EssentialMedium containing Earle’s salts and 10% fetalcalf serum and serial log dilutions of the mix-ture added to Vero cells for 3 days at 37 °C,the length of time required for the microtitreassay. Both the undiluted and 1:10 dilution ofthe methanol/culture medium mix caused celldeath, but dilutions of 1:100 and above did notcause a cytopathic effect. No cytopathic effectwas observed in methanol-treated Hendra virussamples diluted 1:100, indicating a maximumtitre of 5.0×102 TCID50/ml compared with1.4×108 TCID50/ml in the untreated control.This represents a minimum reduction in titre of3×105 in the presence of 90% methanol. In theimmune plaque assay, the concentration ofmethanol added to 96 well plates after removalof the contents exceeds the 90% used in thisexperiment.

In the second experiment to demonstrate Hen-dra virus sensitivity to methanol, Vero cellmonolayers in a 96 well plate were each inocu-lated with 1000 TCID50 in 100 �l Minimal Es-

sential Medium containing Earle’s salts and 10%fetal calf serum per well. After 24 h, the culturemedium was removed and virus titre determined(105 TCID50/ml). Ice-cold methanol was addedto the monolayers, removed after 4 min andcells in two wells scraped into 400 �l of Mini-mal Essential Medium containing Earle’s saltsand 10% fetal calf serum, which was added to asub-confluent monolayer of Vero cells in a 25cm2 flask. After 30 min at 37 °C, 5 ml MinimalEssential Medium containing Earle’s salts and10% fetal calf serum was added and the cellmonolayer observed for 7 days at 37 °C. Nocytopathic effect was observed. Culture medium(2 ml) was removed from the first passage flaskat day 7, added to a second monolayer of Verocells, which was observed for 5 days. No cyto-pathic effect was detected indicating that the ab-sence of detectable infectious virus aftermethanol fixation of Hendra virus-infected cells.

3. Results

3.1. Choice of cell line for the Hendra �irusimmune plaque assay

The criteria for choosing a cell line includedfast virus growth, the release of high titres ofinfectious virus and the rapid generation oflarge foci that could be detected by visual, mi-

Table 1Characteristics of Hendra virus infection of a number cell lines

Species Syncytium formationa titre (×107)b % Virus adsorbed in 1 hCell line

Human +++ 6 50143BMonkey +++MARC 6 50MonkeyVero 454++++Hamster ++BHK 6 40

MDBK Cattle ++ 6 30LT Lamb ++ 8 nd

Dog +++MDCK ndc 17+L929 0.04Murine 18(+) –ST 6Pig

a The relative extent of syncytium formation was estimated 24 h pi by comparing the size of syncytia and the number of nucleitherein. In BHK and Vero cells, syncytia contained up to �10 and 50 nuclei, respectively.

b Titres were estimated by the conventional microneutralisation assay.c Not done.

G. Crameri et al. / Journal of Virological Methods 99 (2002) 41–51 47

croscopic or immunological means. The resultsin Table 1 show that more Hendra virus boundto 143B, Vero and MARC cells than to othercells tested and replicated to high titre in 143B,Vero, MARC, MDBK and LT cells. The cyto-pathic effect in Vero and MARC cells wassevere and characterised by rapid and extensivesyncytia formation, especially in Vero cells. Thiswas in contrast to LT, BHK, MDBK and 143Bcells in which, in spite of generating a high virustitre, the cytopathic effect was mild and associ-ated with fewer and smaller syncytia. On thebasis of these experiments, Vero cells were cho-sen for use in the immune plaque assay. Nipahvirus replicated and formed syncytia in Verocells with the same kinetics as that observedwith Hendra virus (data not shown).

3.2. Choice of primary detecting antibody

In immunofluorescence assays monospecificrabbit antiserum to the Hendra virus P, F1 andF2 proteins detected both Hendra and Nipahviruses-induced foci in Vero cells, but antiserumto the Hendra virus G protein detected onlyHendra virus-induced foci (data not shown).Antiserum to Hendra virus P, F1 and F2

proteins also detected both Hendra and Nipahviruses proteins by Western blotting (data notshown). In the immune plaque assay, Hendravirus antigen was detected using Hendra virusantisera to proteins P, F1, F2 and G with theprimary antibody localised using alkaline phos-phatase-linked secondary antibody as describedin Section 2. The anti-P antiserum was the mosteffective, because it detected foci at a muchhigher antiserum dilution and with less back-ground than other antisera tested (data notshown). Hendra virus anti-P antiserum was alsoeffective for detecting Nipah-induced syncytia.Hendra virus foci detected by anti-P antiserumare shown in Fig. 1.

3.3. Comparison of the con�entional microtitreand immune plaque assays to quantify Hendra�irus and neutralising anti-Hendra �irus antibody

Under conditions, where virus was added to

Fig. 1. Top three rows: foci in Vero cells generated by Hendravirus. Bottom row: control uninfected Vero cells.

preformed monolayers of Vero cells and re-moved after 30 min adsorption, the stock prepa-ration of Hendra virus had titres of 4–7×107

pfu/ml and 2–7×107 TCID50 /ml, when titratedby immune plaque and conventional microtitreassay, respectively. Generation of plaques wasnot dependent on addition of virus to preformedcell monolayers. Plaques were also detected 24 hafter mixing 2.4×103 Vero cells and appropriatevirus dilutions in the wells of a microtitre plate.Under these conditions, the titres obtained forstock preparations of Hendra virus were �5×107 pfu/ml.

In assays modified to detect anti-virus anti-body, the immune plaque and traditional mi-crotitre assay yielded titres of 160 and 128,respectively for anti-Hendra virus ascitic fluid.The amount of virus used in the immune plaqueand microtitre assays was �40 pfu and 200TCID50, respectively.

3.4. The effect of neutralising anti-Hendra �irusantiserum on the size and number of plaques

The Hendra virus immune plaque assay pro-vided an accurate estimate of titre only if theplaques detected were generated as a result of

G. Crameri et al. / Journal of Virological Methods 99 (2002) 41–5148

the spread of the infecting virus to neighbouringcells via syncytia and not due to infection ofdistal cells by progeny virus released into themedium. To determine if secondary foci of in-fection were generated in the assay, Hendravirus was adsorbed to Vero cell monolayers for30 min and removed prior to the addition ofmedium containing anti-Hendra virus asciticfluid. A control contained medium withoutascitic fluid. Hendra virus plaques were detectedby immune plaque assay 24 h later. The resultsshowed that the presence of ascitic fluid in theculture medium at dilutions as low as 1:20 didnot reduce either the number or the size ofplaques compared with that generated in the ab-sence of antiserum (data not shown). This indi-cated that under test conditions no detectablesecondary plaques were generated. The failure ofthe antiserum at the concentrations used to ab-rogate syncytium formation may be due to therelative concentrations of antibodies to the Hen-dra virus attachment (G, for glycoprotein) andfusion (F) glycoproteins. Homopolymers ofthese proteins form the spikes on the surface ofHendra virus and other members of the familyParamyxo�iridae. In the case of members of theRespiro�irus and Rubula�irus genera, the attach-ment glycoprotein is designated HN, because ithas the capacity to bind to and remove sialicacid from glycoprotein substrates, propertiesmanifested as haemagglutination (H) and neu-raminidase (N) activities. The attachmentprotein (G or HN) is responsible for binding thevirus to specific receptors on the surface of sus-ceptible cells and the F protein facilitates fusionof the virus and cell membranes, a process nec-essary for viral infection. Antibodies to bothHN (or G) and F proteins may neutralise virusinfectivity with anti-HN antibodies usually moreeffective in that regard. Merz et al. (1980)showed that in CV-1 cells, infection by therubulavirus SV5 can spread by either releasedprogeny virus adsorbing to and infecting othercells, or by fusion of an infected cell with anadjacent cell as a result of the cell-fusing activityof the F glycoprotein. Antibodies specific for theHN glycoprotein prevented the dissemination ofinfection by released infectious virus, but did

not abrogate spread by cell fusion. In contrast,antibodies to the F glycoprotein completely pre-vented the spread of infection by fusion.

3.5. Use of methanol-fixed, Nipah �irus-infectedcells to detect anti-�irus antibody in sera

A number of Malaysian anti-Nipah virus serawas tested by serum neutralisation test and byimmune plaque assay using a protein A alkalinephosphatase detection system. The panel con-tained both serum neutralisation test positiveand negative sera. The results are shown inTable 2. Of 23 serum neutralisation test positivesera (titres 8–32), 22 (95.7%) were positive byimmune plaque assay with values of either 2+(19 sera) or 1+ (3 sera). Only one serum thatwas positive by serum neutralisation test with atitre of eight gave an ambivalent +/− result inthe immune plaque assay. A single serum with aserum neutralisation titre of four also generateda +/− immune plaque assay. Of 73 serumneutralisation test negative sera, 69 (94.5%) wereunreactive by immune plaque assay. The fourserum neutralisation test negative sera that re-acted in the immune plaque assay gave values of1+ (1 serum) and +/− (3 sera).

3.6. Storage of methanol-fixed Hendra�irus-infected cells

To determine if viral antigen in Hendra virus-infected foci retained a capacity to bind anti-Hendra virus P antibody after prolongedstorage, methanol-fixed Hendra virus-infected

Table 2Comparison of the immune plaque assay and serum neutrali-sation test using a panel of Malaysian porcine sera

04�8Serum neutralisation test result123 73Number of sera

Immune plaque assaya

0 02+ 19101+ 3

1+/− 310 0 69–

a The method used to quantify the immune plaque assay isdescribed in Section 2.

G. Crameri et al. / Journal of Virological Methods 99 (2002) 41–51 49

cells were kept at room temperature, 4 and −20 °C for up to 7 months before analysis byimmune plaque assay. Little change was ob-served in either the number of plaques detectedor the intensity of their staining by the immuneplaque assay after storage for 3 months at roomtemperature, 4 or −20 °C. Thereafter, therewas a decline in the reactivity of plates stored atroom temperature. At 7 months post-storage at−20 °C, there was no decrease in activity asmeasured by the number of plaques and theirintensity of staining. Plates stored at 4 °C for 7months showed only a small decrease in the in-tensity with which the foci were stained.

4. Discussion

We describe the development of an immune-plaque assay for Hendra and Nipah viruses anda modification of the test to detect and quantifyanti-virus antibody. The method has been usedto serologically compare Hendra and Nipahviruses show that the viruses cross neutralise(Chua et al., 2000). The immune plaque assayaddresses a number of problems inherent inworking at biohazard level 4. The reduction intime from up to 3 days in the traditional micro-neutralisation assay to 1 day in the immuneplaque assay ensures that less virus will be pro-duced during the assay. This is an advantage,when the virus concerned is highly pathogenic.In addition, the process of methanol fixation toprepare infected cell monolayers for immunolog-ical screening inactivates the virus and permitsdetection of foci of infection at lower biohazardlevels, thus reducing the time workers have toremain in the stressful environment of the bio-hazard level 4 laboratory to read test results.

Although antibodies to Hendra and Nipahviruses have been found in Australian andMalaysian flying foxes, respectively (Field et al.,2001), whether or not these viruses naturally in-fect other animals remains uncertain. The cur-rent definitive serological test for anti-Hendraand anti-Nipah virus antibodies is a serum neu-tralisation test that is based on a traditional mi-crotitre assay. The test requires comparatively

large volumes of sera that may not be readilyobtained from small native animals. For exam-ple, the question of whether Hendra virus in-fects Australian microbats remains unanswered,because only microlitre amounts of serum canbe obtained from these animals. An advantageof the immune plaque assay is that volumes ofserum as small as 10 �l can be tested in dupli-cate at a dilution of 1/10.

Tests used currently to detect antibodies indomestic and native animals to Hendra or Ni-pah viruses include an indirect enzyme-linkedimmunosorbent assay using antigen from virus-infected cells and a microtitre neutralisation test.Currently, the presence of anti-viral antibodiesin sera that generate positive or ambivalent re-sults in the enzyme-linked immunosorbent assayis confirmed using the serum neutralisation test.The latter test can be done only in a biohazardlevel 4 laboratory and that frequently is locatedoutside the country from which the suspect posi-tive samples are derived. Transport to the level4 laboratory and the time taken to complete theserum neutralisation test may delay the processof collecting serological data and in emergencysituations impede the application of measures tocontrol putative outbreaks.

A third type of serological test that can beapplied locally without the need to use live virusmay facilitate the rapid detection of anti-Hendraand anti-Nipah virus antibodies. To this end, wedetermined if there was a positive correlationbetween the serum neutralisation test and abilityof anti-Nipah virus antibodies in sera to bindspecifically to the discrete virus-induced foci inmethanol-fixed Vero cell monolayers. The strongcorrelation between serum neutralisation testand immune plaque assay suggest thatmethanol-fixed Hendra or Nipah virus-infectedcell monolayers may be used to detect anti-virusantibodies that bind preferentially to virus-in-duced syncytia. The maintenance of antigenic re-activity after storage for 7 months at 4 and−20 °C suggest that such methanol-fixedmonolayers, prepared at biohazard level 4, maybe transported to, and stored and used in thoselaboratories currently testing sera by enzyme-linked immunosorbent assay. The immune

G. Crameri et al. / Journal of Virological Methods 99 (2002) 41–5150

plaque assay may thus be a useful additional testto corroborate the presence of anti-virus antibod-ies in enzyme-linked immunosorbent assay posi-tive sera. By identifying sera that generate falsepositives in the enzyme-linked immunosorbent as-say, use of the immune plaque assay may reducethe number of sera that have to be subject toconfirmatory serum neutralisation testing.

Acknowledgements

We would like to thank Dr M.N. Mohd. Nor,Director General, Department of Veterinary Ser-vices, Malaysia and Dr J. Aziz, Veterinary Re-search Institute, Ipoh, Malaysia for providingporcine field sera.

References

Belanger, F., Alain, R., Payment, P., Lecomte, J., Trudel,M., 1988. Rapid titration of bovine, caprine and humanRS virus by a micro-immunoperoxidase assay using amonoclonal antibody and a permissive ovine kidney cellline. J. Virol. Meth. 20, 101–107.

Borovec, S., Uren, E., 1997. Single-antibody in situ enzymeimmunoassay for infectivity titration of hepatitis A virus.J. Virol. Meth. 68, 81–87.

Chou, S.W., Scott, K.M., 1988. Rapid quantitation of cy-tomegalovirus and assay of neutralizing antibody by us-ing monoclonal antibody to the major immediate-earlyviral protein. J. Clin. Microbiol. 26, 504–507.

Chua, K.B., Bellini, W.J., Rota, P.A., Harcourt, B.H.,Tamin, A., Lam, S.K., Ksiazek, T.G., Rollin, P.E., Zaki,S.R., Shieh, W.J., Goldsmith, C.S., Gubler, D.J.,Roehrig, J.T., Eaton, B., Gould, A.R., Olson, J., Field,H., Daniels, P., Ling, A.E., Peters, C.J., Anderson, L.J.,Mahy, B.W.J., 2000. Nipah virus: a recently emergentdeadly paramyxovirus. Science 288, 1432–1435.

Field, H.E., Barratt, P.C., Hughes, R.J., Shield, J., Sullivan,N.D., 2000. A fatal case of Hendra virus infection in ahorse in north Queensland: clinical and epidemiologicalfeatures. Aust. Vet. J. 78, 279–280.

Field, H., Young, P., Yob, J.M., Mills, J., Hall, L., Macken-zie, J., 2001. The natural history of Hendra and Nipahviruses. Microb. Infect. 3, 307–314.

Harcourt, B.H., Tamin, A., Ksiazek, T.G., Rollin, P.E., An-derson, L.J., Bellini, W.J., Rota, P.A., 2000. Molecularcharacterization of Nipah virus, a newly emergentparamyxovirus. Virology 271, 334–349.

Harder, T.C., Klusmeyer, K., Frey, H.R., Orvell, C., Liess,B., 1993. Intertypic differentiation and detection of in-

tratypic variants among canine and phocid morbillivirusisolates by kinetic neutralization using a novel im-munoplaque assay. J. Virol. Meth. 41, 77–92.

Hooper, P.T., Gould, A.R., Hyatt, A.D., Braun, M.A., Kat-tenbelt, J.A., Hengstberger, S.G., Westbury, H.A., 2000.Identification and molecular characterization of Hendravirus in a horse in Queensland. Aust. Vet. J. 78, 281–282.

Hooper, P.T., Gould, A.R., Russell, G.M., Kattenbelt, J.A.,Mitchell, G., 1996. The retrospective diagnosis of a sec-ond outbreak of equine morbillivirus infection. Aust. Vet.J. 74, 244–245.

Lipson, S.M., Kaplan, M.H., Simon, J.K., Ciamician, Z.,Tseng, L.F., 1992. Improved detection of cy-tomegalovirus viremia in AIDS patients using shell vialand indirect immunoperoxidase methodologies. J. Med.Virol. 38, 36–43.

Merz, D.C., Scheid, A., Choppin, P.W., 1980. Importance ofantibodies to the fusion glycoprotein of paramyxovirusesin the prevention of spread of infection. J. Exp. Med.151, 275–288.

Murray, K., Selleck, P., Hooper, P., Hyatt, A., Gould, A.,Gleeson, L., Westbury, H., Hiley, L., Selvey, L., Rod-well, B., 1995. A morbillivirus that caused fatal disease inhorses and humans. Science 268, 94–97.

Nakane, P.K., Pierce, G.B., 1967. Enzyme-labeled antibodiesfor the light and electron microscopic localization of tis-sue antigens. J. Cell Biol. 33, 307–318.

Reed, L.J., Muench, H., 1938. A simple method of estimat-ing fifty per cent end points. Am. J. Hygiene 27, 493–497.

Rogers, R.J., Douglas, I.C., Baldock, F.C., Glanville, R.J.,Seppanen, K.T., Gleeson, L.J., Selleck, P.N., Dunn, K.J.,1996. Investigation of a second focus of equine morbil-livirus infection in coastal Queensland. Aust. Vet. J. 74,243–244.

Swenson, P.D., Kaplan, M.H., 1985. Rapid detection of cy-tomegalovirus in cell culture by indirect immunoperoxi-dase staining with monoclonal antibody to an earlynuclear antigen. J. Clin. Microbiol. 21, 669–673.

Tanishita, O., Takahashi, Y., Okuno, Y., Yamanishi, K.,Takahashi, M., 1984. Evaluation of focus reduction neu-tralization test with peroxidase–antiperoxidase stainingtechnique for hemorrhagic fever with renal syndromevirus. J. Clin. Microbiol. 20, 1213–1214.

Terpstra, C., Bloemraad, M., Gielkens, A.L., 1984. The neu-tralizing peroxidase-linked assay for detection of antibodyagainst swine fever virus. Vet. Microbiol. 9, 113–120.

Ubertini, T., Wilkie, B.N., Noronha, F., 1971. Use ofhorseradish peroxidase-labeled antibody for light andelectron microscope localization of reovirus antigen.Appl. Microbiol. 21, 534–538.

Wang, L.F., Eaton, B.T., in press. Henipavirus (Paramyx-o�iridae). In: Tidona, C.A., Daria, G. (Eds.), TheSpringer Index of Viruses. Berlin: Springer.

Wang, L.F., Gould, A.R., Selleck, P.W., 1997. Expression ofequine morbillivirus (EMV) matrix and fusion proteinsand their evaluation as diagnostic reagents. Arch. Virol.142, 2269–2279.

G. Crameri et al. / Journal of Virological Methods 99 (2002) 41–51 51

Wang, L.F., Harcourt, B.H., Yu, M., Tamin, mA., Rota,P.A., Bellini, W.J., Eaton, B.T., 2001. Molecular biologyof Hendra and Nipah viruses. Microb. Infect. 3, 279–287.

Wang, L.F., Michalski, W.P., Yu, M., Pritchard, L.I.,Crameri, G., Shiell, B., Eaton, B.T., 1998. A novel P/V/C gene in a new member of the Paramyxo�iridae family,

which causes lethal infection in humans, horses, andother animals. J. Virol. 72, 1482–1490.

Young, P.L., Halpin, K., Selleck, P.W., Field, H., Gravel,J.L., Kelly, M.A., Mackenzie, J.S., 1996. Serologic evi-dence for the presence in Pteropus bats of a paramyx-ovirus related to equine morbillivirus. Emerg. Infect. Dis.2, 239–240.