Journal of Immunological Methods 242 (2000) 91–100www.elsevier.nl / locate / jim

A multi-antigen print immunoassay for the development ofserological diagnosis of infectious diseases

a,b b a ,1Konstantin P. Lyashchenko , Mewa Singh , Roberto Colangeli ,a ,*Maria Laura Gennaro

aPublic Health Research Institute, 455 First Avenue, New York, NY 10016, USAbChemBio Diagnostic Systems, Inc., 3661 Horseblock Road, Medford, New York, NY 11763, USA

Received 29 February 2000; received in revised form 8 May 2000; accepted 12 May 2000

Abstract

Serological diagnosis of infectious diseases that generate a highly heterogeneous antibody repertoire, such as tuberculosis,requires tests based on cocktails of antigens. We describe a new method called multi-antigen print immunoassay (MAPIA)for cocktail-based serological diagnosis. The assay entails the application of antigen to nitrocellulose membranes bymicro-aerosolization (printing), followed by antibody detection using standard chromogenic immunodevelopment. Cocktailsof protein antigens of Mycobacterium tuberculosis tested by MAPIA were found to maintain the serological activity of eachof their components. In contrast, the same cocktails tested by enzyme-linked immunosorbent assay (ELISA) had aserological activity that was lower than the sum of the activities of their components. Consequently, cocktail-based MAPIAattained the diagnostic sensitivity expected on the basis of single antigen results, while a significant loss of diagnosticsensitivity was observed with cocktail-based ELISA. Thus, the MAPIA format is superior to conventional ELISA for theserological diagnosis of infectious diseases characterized by heterogeneous antibody responses. 2000 Elsevier ScienceB.V. All rights reserved.

Keywords: Rapid diagnosis; ELISA; Solid-phase immunoassay; Antigen cocktail; Serodiagnosis; Tuberculosis

1. Introduction

Detection of antibodies is one of the approachesmost commonly used to diagnose infectious diseases(Rose, 1998). Serological techniques are simple,Abbreviations: ELISA, enzyme-linked immunosorbent assay;usually rapid, relatively non-invasive, and do notkDa, kilodalton; MAPIA, multi-antigen print immunoassay; PBS,

phosphate-buffered saline; OD, optical density require isolation or culturing of the pathogen. In*Corresponding author. Tel.: 11-212-578-0844; fax: 11-212- many infections, most, if not all, infected persons

578-0804. produce antibodies against the same, immuno-E-mail address: gennaro@phri.nyu.edu (M.L. Gennaro).1 dominant antigen(s). In other cases, most notably inPresent address: Division of Infectious Diseases, Department

infections caused by certain intracellular pathogens,of Medicine, Montefiore Medical Center, Bronx, New York, NY10467, USA. the antibody repertoire is highly diverse, with sera

0022-1759/00/$ – see front matter 2000 Elsevier Science B.V. All rights reserved.PI I : S0022-1759( 00 )00241-6

92 K.P. Lyashchenko et al. / Journal of Immunological Methods 242 (2000) 91 –100

Table 1from different persons reacting with different an-Recombinant antigens of M. tuberculosistigens. Examples include tuberculosis (LyashchenkoProtein Molecular Referenceet al., 1998), Crohn’s disease (Elsaghier et al., 1992),

mass (kDa)chlamydial infections (Essig et al., 1999), andcryptococcosis (Chen et al., 1999). Accurate ESAT-6 6 (Sørensen et al., 1995)

14 kDa 14 (Verbon et al., 1992)serodiagnosis of such infectious diseases requiresMPT63 16 (Manca et al., 1997b)cocktails of multiple antigens to cover the breadth of19 kDa 19 (Ashbridge et al., 1989)

the antibody response. MPT70 22 (Matsumoto et al., 1995)To develop a serodiagnostic test that utilizes a MPT64 23 (Yamaguchi et al., 1989)

multi-antigen cocktail, it is necessary to have an MPT51 27 (Ohara et al., 1995)MTC28 28 (Manca et al., 1997a)immunoassay that gives optimal cocktail perform-Ag85B 30 (Matsuo et al., 1988)ance. Below, we describe a new, cocktail-based38 kDa 38 (Andersen and Hansen, 1989)

method for antibody detection. The method, which MPT32 45/47 (Laqueyrerie et al., 1995)we call multi-antigen print immunoassay (MAPIA), KatG 80 (Heym et al., 1993)is based on immobilization of antigens onto nitro-cellulose membranes by semi-automated micro-spraying, followed by standard chromogenic im- polyhistidine-tagged protein fusions by using meth-munodevelopment. As a test of the method, we used ods described previously (Colangeli et al., 1998).tuberculosis. The serological activity in MAPIA of acocktail containing multiple antigens of Mycobac- 2.3. ELISAterium tuberculosis was equal to that expected fromsingle antigen results and it was much greater than Polystyrene 96-well microtiter plates (Costar)that observed with an enzyme-linked immunosorbent were coated with 100 ml of single antigens or multi-assay (ELISA) in polystyrene microtiter plates. Thus, antigen cocktails at concentrations ranging betweenMAPIA constitutes a significant step towards de- 0.1 and 2 mg/ml (see below, section 2.5) in 0.05 Mveloping a cocktail-based, serodiagnostic test for carbonate–bicarbonate buffer, pH 9.6, at 48C over-tuberculosis and other infectious diseases. night. After coating with antigen, plates were

blocked with 1% non-fat skim milk (Sigma) in 0.15M phosphate-buffered saline plus 0.05% Tween 20(PBS-T) for 1 h at room temperature. A 100-ml2. Materials and methodsvolume of serum was added to each well at a 1:50dilution in PBS-T containing 1% non-fat skim milk

2.1. Sera and was incubated for 1 h at room temperature. Afterwashing with PBS-T, plates were incubated with

Sera were obtained from 75 patients with culture- alkaline-phosphatase-conjugated goat anti-humanconfirmed active pulmonary tuberculosis and from IgG antibody (Sigma), diluted 1:2000 in PBS-T, for67 healthy, control individuals. All study subjects 1 h at room temperature. Plates were washed withwere negative for infection with human immuno- PBS-T, and enzyme activity was assayed by incuba-deficiency virus. tion for 15 min at room temperature with 0.1 ml per

well of p-nitrophenylphosphate (Bio-Rad). After2.2. Antigens addition to each well of 0.1 ml 0.4 N NaOH to stop

the reaction, optical density was measured at 405 nmIn this study, we used twelve proteins of M. (OD ) by using an automatic microtiter plate405

tuberculosis (Table 1) that had been previously reader (Spectra Shell, Tecan).shown to generate antibody responses during tuber-culosis (Lyashchenko et al., 1998; Colangeli et al., 2.4. MAPIA1999). Proteins were purified to near-homogeneityfrom cultures of recombinant Escherichia coli as Antigens were immobilized on nitrocellulose

K.P. Lyashchenko et al. / Journal of Immunological Methods 242 (2000) 91 –100 93

membranes (Protran BA83, Schleicher and Schuell) When specified, semi-quantitative analyses ofas narrow bands by using a semi-automatic air-brush MAPIA data were performed by scanning nitro-printing device (Linomat IV, Camag). Antigens were cellulose strips using a Hewlett Packard ScanJetprinted at concentrations ranging from 50 to 200 IIcx /T and by densitometric evaluationmg/ml in PBS (see below, section 2.5). A 6-ml (ScanAnalysis software, version 2.56). Test sensitivi-volume of antigen solution was sufficient to print a ty was calculated as the percentage of positive testband that was 15 cm long. After antigen printing, the results obtained with sera of tuberculosis patients.membrane was cut perpendicular to the antigen Test specificity was calculated as the percentage ofbands into strips that were 4 mm wide. A detailed negative test results obtained with sera of negativedescription of MAPIA is presented in Results (sec- control individuals.tion 3.2) and in Fig. 1. Nitrocellulose strips wereblocked for 1 h with 1% non-fat skim milk in PBS-Tand were incubated for 1 h with serum samples that 3. Resultswere diluted 1:50 in blocking buffer. After washingwith PBS-T, strips were incubated for 1 h with 3.1. Choice of solid phasealkaline-phosphatase-labeled anti-human IgG anti-body (Sigma), diluted 1:2000 in PBS-T, and were The formulation of a multi-antigen cocktail forwashed again with PBS-T. Enzyme activity was serodiagnosis is a two-stage process. First, antigensvisualized by incubating the strips for 10 min with are selected on the basis of seroreactivity. Second,5-bromo-4-chloro-3-indolyl-phosphate /nitroblue tet- the serological activity of a cocktail of the selectedrazolium phosphatase substrate (Kirkegaard and antigens is assessed. We found that ELISA usingPerry Laboratories). Strips were extensively rinsed in polystyrene microtiter plates was inadequate becausedistilled water to stop the reaction. The entire multi-antigen cocktails tended to perform subopti-procedure was performed at room temperature. mally: for a given serum, absorbance values mea-

sured with an antigen cocktail were often 10 to 50%2.5. Antigen concentrations in the immunoassays lower than those obtained with single antigens

(examples are shown in Table 2). This phenomenonOptimal concentration of antigen for coating was is likely to reduce the diagnostic sensitivity of

determined in preliminary ELISA and MAPIA ex- cocktail-based tests.periments. Each antigen was used at the same Nitrocellulose membranes have a higher protein-concentration both in single-antigen- and in cocktail- binding capacity than polystyrene microtiter platesbased assays. ELISA: ESAT-6, 1 mg/ml; 14 kDa, (Kemeny and Challacombe, 1988). Moreover, nitro-0.3 mg/ml; MPT63, 0.2 mg/ml; 19 kDa, 1 mg/ml; cellulose is amenable to simultaneous testing ofMPT70, 1 mg/ml; MPT64, 1 mg/ml; MTC28, 0.1 many antigens and/or many sera (Alric et al., 1986;mg/ml, and 38 kDa, 2 mg/ml. MAPIA: ESAT-6, 0.2 Kazemi and Finkelstein, 1990; Noya and Noya,mg/ml; 14 kDa, 0.05 mg/ml; MPT63, 0.2 mg/ml; 1998). Thus, nitrocellulose was examined as a19 kDa, 0.05 mg/ml; MPT70, 0.2 mg/ml; MPT64, possible solid phase.0.2 mg/ml; MPT51, 0.2 mg/ml; MTC28, 0.1 mg/ml; Ag85B, 0.1 mg/ml; 38 kDa, 0.1 mg/ml; 3.2. Multi-antigen print immunoassayMPT32, 0.1 mg/ml, and KatG, 0.1 mg/ml.

We took advantage of techniques and equipment2.6. Data analysis employed for thin-layer chromatography (Touch-

stone, 1992) to develop an assay on nitrocelluloseFor ELISA, cut-off values were calculated as membranes that allowed efficient antigen screening

mean OD values obtained with 67 control sera and cocktail testing. Multi-antigen print immuno-405

plus two standard deviations. MAPIA results were assay (MAPIA) involves spraying antigens (printing)evaluated visually, with the presence of a band of onto nitrocellulose membranes. The key steps ofany intensity being considered a positive result. MAPIA are outlined in Fig. 1. By using a semi-

94 K.P. Lyashchenko et al. / Journal of Immunological Methods 242 (2000) 91 –100

Fig. 1. Multi-antigen print immunoassay. The four key steps of MAPIA are illustrated. From top to bottom: (1) printing antigens onto anitrocellulose sheet to generate narrow, parallel bands (represented as gray lines). A 15315 cm nitrocellulose sheet can contain as many as50 antigen bands, 3 mm apart; (2) cutting membrane strips that were 4 mm wide, each carrying all antigens; (3) incubation of each stripwith a test serum followed by standard antibody detection; (4) data analysis. Images of strips printed with ten single antigens and aten-antigen cocktail (Combi-10) and developed after incubation with sera from four healthy individuals (negative controls) and fourtuberculosis patients. Strip images were obtained using a Hewlett Packard ScanJet IIcx /T scanner.

K.P. Lyashchenko et al. / Journal of Immunological Methods 242 (2000) 91 –100 95

Table 2Reactivity in ELISA of selected sera from tuberculosis patients with antigens of M. tuberculosis, both singly and in combination

aSerum OD Loss of signal405bwith cocktail , (%)

Single antigens Combi-4

ESAT-6 14 kDa 14 kDa 38 kDa

B51 1.26 0.10 0.21 0.06 0.72 43U9 1.07 0.03 0.04 0.03 0.53 50B123 0.01 0.53 0.06 0.03 0.32 40C8 0.01 0.95 0.30 0.09 0.68 28B56 0.03 0.05 0.82 0.05 0.74 10C6 0.01 0.05 0.50 0.01 0.44 12B15 0.01 0.06 0.20 0.73 0.66 10U8 0.03 0.05 0.14 0.60 0.50 17Cut-off 0.12 0.17 0.22 0.16 0.21

a Values represent absorbance at 405 nm (OD ) obtained with sera tested by ELISA as described in Materials and Methods. Shown in405

bold are OD values above the cut-off for the corresponding antigen. Cut-off (bottom line) was calculated as the mean OD readings405 405

obtained with 67 negative control sera plus two standard deviations.b Calculated as [1002(OD /OD 3100)].cocktail single Ag

automatic micro-aerosolization device, antigens are serum (1.5%) reacted with one antigen (MPT32)sprayed through a syringe needle onto a nitrocellu- (data not shown).lose sheet to generate parallel bands. Multi-antigen We next selected antigens for inclusion in acocktails are prepared in one of two ways: (i) cocktail (Fig. 2). The first choice was the 38 kDaproteins are mixed prior to printing or (ii) individual protein, since it was the most seroreactive antigen inantigens are printed one after the other on the same, the panel (recognized by 30.7% of tuberculosissingle track of the membrane. Since cocktails have patients; Fig. 2, top line). The next antigen selectedthe same serological activity irrespective of the was the 14 kDa protein, because it contributed thecocktail preparation method (data not shown), se- largest number of additional reactors to those alreadyquential antigen printing is preferred because it obtained with the 38 kDa antigen. The proportion ofutilizes the same protein dilution steps as single reactors to 38 kDa plus 14 kDa antigens was 42.7%,antigen printing. After antigen printing, the mem-

Table 3brane is cut into strips perpendicular to the antigenReactivity in MAPIA of sera from TB patients and healthybands. Next, each strip is incubated with a testcontrols to antigens of M. tuberculosis

serum. Bound antibody is then detected by standardAntigen Reactors amongchromogenic development. MAPIA is a very ver-

75 TB patientssatile assay format, since it does not entail the use of No. Namea fixed template, such as an immunoblotter (Noya n %and Noya, 1998), for antigen alignment and im-

1 38 kDa 23 30.7mobilization. 2 14 kDa 15 20

3 ESAT-6 10 13.34 19 kDa 13 17.33.3. Selection of antigens for cocktail formulation5 MPT63 11 14.76 MPT64 5 6.7We tested twelve antigens of M. tuberculosis by7 MTC28 14 18.7

MAPIA (Table 1) with sera from 75 tuberculosis 8 MPT70 7 9.3patients and 67 healthy, control individuals. Results 9 KatG 17 22.7

10 MPT32 12 16are shown in Table 3. We found that 58 sera from11 MPT51 2 2.7tuberculosis patients (77%) reacted with at least one12 Ag85B 9 12antigen in the panel. Only one negative control

96 K.P. Lyashchenko et al. / Journal of Immunological Methods 242 (2000) 91 –100

to those obtained with the first eight antigens (Fig. 2,last two lines). Moreover, KatG tended to elicit veryweak antibody responses, while, as mentioned above,MPT32 contributed false-positive results (data notshown).

3.4. MAPIA analysis of multi-antigen cocktails

We next evaluated the serological activity of afour-antigen cocktail (ESAT-6, 14 kDa, 19 kDa, and38 kDa antigens), referred to as Combi-4, and of aneight-antigen cocktail (the same four antigens as inCombi-4 plus MPT63, MPT64, MPT70, andMTC28), Combi-8. We used the same 75 sera frompatients and 67 control sera that were utilized in theanalysis of single antigens. Serological testing of

Fig. 2. Selection of antigens for cocktail formulation. The fraction Combi-4 by MAPIA showed that the number of(%) of tuberculosis patient sera expected to react with antigen reactors to each cocktail among tuberculosis patientscocktails was determined as described in the text. Antigen

(42 of 75; 56%) and controls (0 of 67; 0%) was thenumbering (1 through 12) is as shown in Table 3.same as that predicted on the basis of single antigen

i.e., 12% of tuberculosis patients reacted with the 14 results (Table 4). With Combi-8, we found that onlykDa protein and not with the 38 kDa protein (Fig. 2, one serum of the 55 expected reactors was negative,second line). The same type of analysis was per- with a 1% loss of test sensitivity, and one controlformed for each antigen in the panel. Reactors to serum reacted with the cocktail, with a 1.5% loss offour antigens (38 kDa antigen, 14 kDa antigen, test specificity (Table 4).ESAT-6 and 19 kDa antigen) accounted for 56% ofthe tuberculosis patients (Fig. 2, third line). The 3.5. ELISA analysis of multi-antigen cocktailsproportion of reactors in the tuberculosis groupreached 73.3% as four more antigens (MPT63, We also tested cocktail activity by ELISA. Two ofMPT64, MPT70, and MTC28) were added to the 67 healthy controls gave false-positive results withanalysis (Fig. 2, fourth line). The remaining four both cocktails, resulting in a 3% loss of test spe-antigens in the panel (KatG, MPT32, MPT51, and cificity. The serological activity of both cocktailsAg85B) were excluded from subsequent studies was lower than expected on the basis of singlebecause they failed to contribute additional reactors antigen results. Only 37 of the expected 42 reactors

Table 4Serodiagnostic performance of multi-antigen cocktails in MAPIA and ELISA

Cocktail Sensitivity, (%) Specificity, (%)a b a bExpected Observed Expected Observed

MAPIA ELISA MAPIA ELISA

Combi-4 56 56 49 100 100 97Combi-8 73 72 56 100 98.5 97

a Calculated on the basis of results obtained with single antigens (see text and Fig. 2).b Experimentally obtained by testing multi-antigen cocktails.Test sensitivity and specificity were calculated as described in Materials and methods by using results obtained with sera from 75

tuberculosis patients and 67 healthy control individuals. Combi-4, ESAT-6114 kDa119 kDa138 kDa. Combi-8, Combi-41MPT631

MPT641MPT701MTC28.

K.P. Lyashchenko et al. / Journal of Immunological Methods 242 (2000) 91 –100 97

were positive with Combi-4, constituting a loss of technology countries where most tuberculosis cases7% of test sensitivity (Table 2). Loss of sensitivity are found. MAPIA results can also be evaluatedwith Combi-8 was even greater (17%), as only 42 of semi-quantitatively by scanning and densitometrythe 55 expected reactors gave positive results (Table (Fig. 3). While a negative serum gave only little, if4). The loss of sensitivity of the multi-antigen any, background reading (Fig. 3, panel A), positiveELISA test was generally caused by lower absor- signals obtained with single or multiple antigens

`bance values obtained with antigen cocktails vis-a- gave rise to peaks. Peak areas were measured byvis those obtained with single antigens (refer to densitometry, and numerical data were obtained (Fig.Table 4). 3, panels B and C).

Semi-quantitative MAPIA may be useful during3.6. Semi-quantitative MAPIA test development, because it allows measurement of

`the analytical sensitivity of a cocktail vis-a-vis thatVisual evaluation of MAPIA results is simple and of its individual components by comparing peak

rapid. A ‘yes-or-no’ result is suitable for efficient areas. For example, when only one antigen wasantigen screening and evaluation of cocktails, and is recognized by a given serum, the signal obtainedhighly desirable for use in the low-income, low- with that antigen alone had the same intensity as

Fig. 3. Semi-quantitative analysis of MAPIA data. Panels A, B and C represent analysis of three different sera. Panel A, normal, negativecontrol serum; panels B and C, sera from two tuberculosis patients. Left: images obtained with a Hewlett Packard ScanJet IIcx /T scanner ofnitrocellulose strips printed with ten single antigens and a ten-antigen cocktail (Combi-10), incubated with serum and immunodeveloped(see Materials and Methods). Center: plots obtained by densitometric image analysis performed using ScanAnalysis software, version 2.56.Right: bar graph representation of peak area measurements of densitometry plots.

98 K.P. Lyashchenko et al. / Journal of Immunological Methods 242 (2000) 91 –100

when the same antigen was present in a ten-antigen tive accumulation of antibodies at the site of antigencocktail (Fig. 3, panel B). Moreover, when several immobilization on nitrocellulose. In contrast, ELISAantigens in a cocktail were recognized by a given readings for cocktails were often lower than thoseserum, the peak area obtained with the cocktail was generated with single antigens. This phenomenon isusually greater than any peak area obtained with presumably due to antigens interfering with eachsingle antigens (Fig. 3, panel C). other in the microtiter plate for binding to poly-

styrene.MAPIA offers several additional advantages over

4. Discussion ELISA. First, MAPIA constitutes a more efficientand cost-effective method for screening many an-

In the present work, we describe MAPIA, a novel tigens. A single 15315 cm sheet of nitrocelluloseassay for antibody detection that entails the applica- can carry as many as 50 antigen bands and can betion of antigen onto nitrocellulose membranes by used to assay 50 serum samples. To test 50 sera andmicro-aerosolization. In MAPIA, a cocktail of multi- 50 antigens by ELISA, one would need approximate-ple antigens retains the serological activity of its ly 50 microtiter plates, ten-times more serum than incomponents. Thus, MAPIA is well suited for de- MAPIA (if used at the same dilution), about 40 timesveloping cocktail-based serological assays for the more antigen (200–400 ng of antigen per serumdiagnosis of tuberculosis and other infectious dis- sample in ELISA versus 5–10 ng in MAPIA), andeases that are characterized by a heterogeneous about ten-times more conjugate and substrate. Fur-antibody repertoire (Lyashchenko et al., 1998; Brank thermore, antigen screening is much more time-et al., 1999; Chen et al., 1999; Essig et al., 1999). efficient with MAPIA than with ELISA. The second

The present report describes a method for multi- advantage of MAPIA over ELISA is that an assayantigen test development rather than the composition that utilizes antigen bound to nitrocellulose mem-of a cocktail for tuberculosis serodiagnosis. The branes is inherently more reproducible than ELISA,choice of antigens and multi-antigen cocktails made a method characterized by significant test-to-testin the present work was dictated by the pattern of variation (Engvall, 1980; Kemeny and Challacombe,antigen recognition expressed by the particular set of 1988). Third, the nitrocellulose membrane is moresera used. Selection of antigens and cocktail formu- versatile than the ELISA microtiter plate. For exam-lation to attain a highly accurate serodiagnostic test ple, membranes other than nitrocellulose can befor tuberculosis is still in progress, as it requires utilized in MAPIA to achieve the best binding andanalysis of a larger pool of M. tuberculosis antigens retention of certain protein (our unpublished observa-and much broader sets of sera than those used in the tions) and non-protein antigens (Vera-Cabrera et al.,present, methodological study. 1999). Fourth, unlike ELISA, a membrane-based

We have demonstrated that MAPIA provides a assay can easily evolve into a rapid test that utilizesbetter format than ELISA for a multi-antigen assay. thin-layer immunochromatography. This type of testCocktails tested in ELISA lacked the serological is used for the rapid diagnosis of many infectiousactivity expected on the basis of single antigen and non-infectious diseases (Bhaskar et al., 1996;results, with the loss of test sensitivity increasing Zhou et al., 1996; Valecha et al., 1998). Thus,with the number of antigens in the cocktail (refer to MAPIA constitutes a step toward developing a rapid,Table 4). This finding is consistent with the notion point-of-care diagnostic test for tuberculosis andthat the protein binding capacity of the solid phase, other infectious diseases that is most desirable inwhich is higher for nitrocellulose than for poly- low-technology settings.styrene (Kemeny and Challacombe, 1988), is crucialfor a cocktail to express its full serological activity.Furthermore, MAPIA signals, as evaluated by scan-ning and densitometry, were often of greater intensi- 5. Human subjectsty for a cocktail than for any single antigen (Fig. 3).This observation can be explained by non-competi- Research described in this manuscript is IRB-

K.P. Lyashchenko et al. / Journal of Immunological Methods 242 (2000) 91 –100 99

the humoral immune response to Chlamydia pneumoniae byexempt, because human sera were obtained through aimmunoblotting and immunoprecipitation. Clin. Diagn. Lab.commercial supplier and had no personal identifiers.Immunol. 6, 819.

Heym, B., Zhang, Y., Pulet, S., Young, D., Cole, S.T., 1993.Characterization of the katG gene encoding a catalase–per-oxidase required for the isoniazid susceptibility of Mycobac-Acknowledgementsterium tuberculosis. J. Bacteriol. 175, 4255.

Kazemi, M., Finkelstein, R.A., 1990. Checkerboard immuno-We thank Karl Drlica, Sam Kayman and Tomas blotting (CBIB): an efficient, rapid, and sensitive method of

assaying multiple antigen/antibody cross-reactivities. J. Im-Haendler for comments on the manuscript. Thismunol. Methods 128, 143.work was supported by NIH grants AI-43781

Kemeny, D.M., Challacombe, S.J., 1988. Microtitre plates and(K.P.L.) and AI-36989 (M.L.G.). other solid phase support. In: Kemeny, D.M., Challacombe,

S.J. (Eds.), Elisa and Other Solid Phase Immunoassays:Theoretical and Practical Aspects. John Wiley and Sons, NewYork, p. 31.References Laqueyrerie, A., Militzer, P., Romain, F., Eiglmeier, K., Cole, S.,Marchal, G., 1995. Cloning, sequencing and expression of the

Alric, M., Cheyvialle, D., Renaud, M., 1986. Cross-blot and apa gene coding for the Mycobacterium tuberculosis 45/47-cross-dot system: a high-performance system for the detection kilodalton secreted antigen complex. Infect. Immun. 63, 4003.of antigen–antibody complexes on nitrocellulose. Anal. Bio- Lyashchenko, K., Colangeli, R., Houde, M., Jahdali, H.A., Men-chem. 155, 328. zies, D., Gennaro, M.L., 1998. Heterogeneous antibody re-

˚Andersen, A.B., Hansen, E.B., 1989. Structure and mapping of sponses in tuberculosis. Infect. Immun. 66, 3936.antigenic domains of protein antigen b, a 38,000-molecular- Manca, C., Lyashchenko, K., Colangeli, R., Gennaro, M.L.,weight protein of Mycobacterium tuberculosis. Infect. Immun. 1997a. MTC28, a novel 28-kilodalton proline-rich secreted57, 2481. antigen specific for the Mycobacterium tuberculosis complex.

Ashbridge, K.R.A., Booth, R.J., Watson, J.D., Lathigra, R.B., Infect. Immun. 65, 4951.1989. Nucleotide sequence of the 19kDa antigen gene from Manca, C., Lyashchenko, K., Wiker, H.G., Usai, D., Colangeli, R.,Mycobacterium tuberculosis. Nucleic Acids Res. 17, 1249. Gennaro, M.L., 1997b. Molecular cloning, purification and

Bhaskar, S., Singh, S., Sharma, M., 1996. A single-step immuno- serological characterization of MPT63, a novel antigen se-chromatographic test for the detection of Entamoeba his- creted by Mycobacterium tuberculosis. Infect. Immun. 65, 16.tolytica antigen in stool samples. J. Immunol. Methods 196, Matsumoto, S., Matsuo, T., Ohara, N., Hotokezaka, H., Naito, M.,193. Minami, J., Yamada, T., 1995. Cloning and sequencing of a

Brank, M., Le Grand, D., Poumarat, F., Bezille, P., Rosengarten, unique antigen MPT70 from Mycobacterium tuberculosisR., Citti, C., 1999. Development of a recombinant antigen for H37Rv and expression in BCG using E. coli–mycobacteriaantibody-based diagnosis of Mycoplasma bovis infection in shuttle vector. Scand. J. Immunol. 41, 281.cattle. Clin. Diagn. Lab. Immunol. 6, 861. Matsuo, K., Yamaguchi, R., Yamazaki, A., Tasaka, H., Yamada, T.,

Chen, L.C., Goldman, D.L., Doering, T.L., Pirofski, L., 1988. Cloning and expression of the Mycobacterium bovisCasadevall, A., 1999. Antibody response to Cryptococcus BCG gene for extracellular a-antigen. J. Bacteriol. 170, 3847.neoformans proteins in rodents and humans. Infect. Immun. Noya, O., Noya, B.A.d., 1998. The multiple antigen blot assay67, 2218. (MABA): a simple immunoenzymatic technique for simulta-

Colangeli, R., Antinori, A., Cingolani, A., Ortona, L., neous screening of multiple antigens. Immunol. Lett. 63, 53.Lyashchenko, K., Fadda, G., Gennaro, M.L., 1999. Humoral Ohara, N., Kitaura, H., Hotokezaka, H., Nishiyama, T., Wada, N.,immune responses to multiple antigens of Mycobacterium Matsumoto, S., Matsuo, T., Naito, M., Yamada, T., 1995.tuberculosis in tuberculosis patients co-infected with human Characterization of the gene encoding the MPB51, one of theimmunodeficiency virus. Int. J. Tuberc. Lung Dis. 3, 1127. major secreted protein antigens of Mycobacterium bovis BCG,

Colangeli, R., Heijbel, A., Williams, A., Manca, C., Chan, J., and identification of the secreted protein closely related to theLyashchenko, K., Gennaro, M.L., 1998. Three-step purifica- fibronectin binding 85 complex. Scand. J. Immunol. 41, 433.tion of lipopolysaccharide-free, polyhistidine-tagged recombi- Rose, N.R., 1998. Immunodiagnosis. In: Gorbach, S.L., Bartlett,nant antigens of Mycobacterium tuberculosis. J. Chromatogr. J.G., Blacklow, N.R. (Eds.), Infectious Diseases, second ed..B 714, 223. W.B. Saunders, Philadelphia, p. 145.

˚Elsaghier, A., Prantera, C., Moreno, C., Ivanyi, J., 1992. Anti- Sørensen, A.L., Nagai, S., Houen, G., Andersen, P., Andersen,bodies to Mycobacterium paratuberculosis-specific protein A.B., 1995. Purification and characterization of a low-molecu-antigens in Crohn’s disease. Clin. Exp. Immunol. 90, 503. lar-mass T-cell antigen secreted by Mycobacterium tuber-

Engvall, E., 1980. Enzyme immunoassay ELISA and EMIT. In: culosis. Infect. Immun. 63, 1710.Vukanis, H.V., Langone, J.J. (Eds.). Methods in Enzymology, Touchstone, J., 1992. Practice of Thin Layer Chromatography. J.Vol. 70. Academic Press, New York. Wiley and Sons, New York.

Essig, A., Simnacher, U., Susa, M., Marre, R., 1999. Analysis of Valecha, N., Sharma, V.P., Devi, C.U., 1998. A rapid immuno-

100 K.P. Lyashchenko et al. / Journal of Immunological Methods 242 (2000) 91 –100

chromatographic test (ICT) for diagnosis of Plasmodium Yamaguchi, R., Matsuo, K., Yamazaki, A., Abe, C., Nagai, S.,falciparum. Diagn. Microbiol. Infect. Dis. 30, 257. Terasaka, K., Yamada, T., 1989. Cloning and characterization

Vera-Cabrera, L., Rendon, A., Diaz-Rodriguez, M., Handzel, V., of the gene for immunogenic protein MPB64 of Mycobac-Laszlo, A., 1999. Dot blot assay for detection of an- terium bovis BCG. Infect. Immun. 57, 283.tidiacyltrehalose antibodies in tuberculous patients. Clin. Zhou, A.T., Ma, W.-L., Zhang, P.-Y., Cole, R.A., 1996. DetectionDiagn. Lab. Immunol. 6, 686. of pulmonary and extrapulmonary tuberculosis patients with

Verbon, A., Hartskeerl, R.A., Schuitema, A., Kolk, A.H., Young, the 38-kilodalton antigen membrane-based assay. Clin. Diagn.D.B., Lathigra, R., 1992. The 14,000-molecular-weight antigen Lab. Immunol. 3, 337.of Mycobacterium tuberculosis is related to the alpha-crys-tallin family of low-molecular-weight heat shock proteins. J.Bacteriol. 174, 1352.