Vol. 30, No. 12
JOURNAL OF CLINICAL MICROBIOLOGY, Dec. 1992, p. 3065-3069
0095-1137/92/123065-05$02.00/0
Copyright X) 1992, American Society for Microbiology
Production of Monoclonal Antibody to a Phenolic Glycolipid of Mycobacterium tuberculosis and Its Use in Detection of the Antigen in Clinical Isolates SANG-NAE CHO,`* JEON-SOO SHIN,1 MAMADOU DAFFE,2 YUNSOP CHONG,3 SUNG-KYU KIM,4 AND JOO-DEUK KIM1 1 Departments of Microbiology, Clinical Pathology,3 and Medicine,4 Yonsei University College of Medicine, Seoul 120-752, Republic of Korea, and Centre de Recherche de Biochimie et Genetique Cellulaires du Centre National de la Recherche Scientifique, 31062 Toulouse Cedex, France2 Received 29 May 1992/Accepted 4 September 1992
A monoclonal antibody (MAbIII604) specific to phenolic glycolipid Th (PGL-Th), a Mycobacterium tuberculosis-specific antigen, was produced and used in the detection of the antigen. MAbI604 reacted with the PGL-Th antigen but not with other phenolic glycolipids from Mycobacterium leprae, M. bovis, and M. kansasii, thus indicating the specificity of the monoclonal antibody to PGL-Th. A dot enzyme-linked immunosorbent assay with MAbUI604 was employed to detect the PGL-Th antigen in lipids purified from M. tuberculosis clinical isolates. Of 50 isolates, 32 (64.0%o) showed clear evidence of the PGL-Th antigen by the dot enzyme-linked immunosorbent assay, but there were marked variations in the intensities and sizes of spots. This suggests differences in PGL-Th antigen production among M. tuberculosis strains even when they are grown in the same culture media and conditions. This was most evident from the fact that in only eight (16.0%o) of the isolates examined was the PGL-Th antigen detectable by thin-layer chromatography, which is much less sensitive for the detection of glycolipid antigens. This study shows that monoclonal antibodies specific to PGL-Th are useful in detecting the antigen in lipid extracts and that there is a marked variation in the PGL-Th production among M. tuberculosis clinical isolates.
Isolation and characterization of phenolic glycolipid Th (PGL-Th), a Mycobacterium tuberculosis-specific antigen (10), has quickly attracted attention for use of the antigen in the serodiagnosis of tuberculosis and for understanding the pathogenesis of tuberculosis. However, the rates of seropositive response to the antigen among tuberculosis patients varied from about 20% to more than 95% in different studies (5, 9, 27, 28). There were also conflicting reports on PGL-Th production in clinical isolates of M. tuberculosis. In one study, PGL-Th was detectable in all clinical isolates by an enzyme immunoassay (24); in another study, however, only about 9% of isolates produced an antigen that was detectable by thin-layer chromatography (TLC) (9). Since the studies examined only a few clinical isolates and employed different tools for detecting the PGL-Th antigen, we were interested in investigating a larger number of M. tuberculosis clinical isolates for the production of the antigen by an enzyme immunoassay and TLC. For enzyme immunoassays, monoclonal antibodies to PGL-Th provide a great advantage in the specific detection of the antigen. Monoclonal antibodies against numerous antigens of M. tuberculosis have been produced (11, 17, 20, 23) and used for the cloning and characterization of the genes responsible for expression of the antigens (1, 2, 18, 26, 29). Some of the M. tuberculosis-specific monoclonal antibodies were to detect antibodies specific to epitopes present only in tubercle bacilli for the serodiagnosis of tuberculosis by competitive immunoassays (14, 15). In addition, monoclonal antibodies to phenolic glycolipid I (PGL-I) of Mycobacterium leprae were used to detect PGL-I antigen (31, 32) and to study the role of the antigen in the pathogenesis of M. leprae infection (22). *
Corresponding author.
Monoclonal antibodies to the PGL-Th antigen thus could be used in various studies related to the antigen. In this study, a monoclonal antibody specific to the PGL-Th antigen was produced and used to detect the antigen in lipids purified from M. tuberculosis clinical isolates.
MATERIALS AND METHODS Mycobacterial antigens and antibodies. M. tuberculosis Canetti and PGL-Th antigen were prepared as previously described (10). Phenolic glycolipids (3) from M. leprae (PGL-I), Mycobacterium bovis (mycoside B), and M. kansasii (mycoside A) were provided by P. J. Brennan (Colorado State University, Fort Collins). The glycopeptidolipid (GPL) core (3) of Mycobacterium avium complex and lipoarabinomannan (LAM) of M. tuberculosis antigens and monoclonal antibody to PGL-I were also provided by P. J. Brennan. A monoclonal antibody to LAM produced previously (7) was also used in this study. Production of monoclonal antibodies to PGL-Th. For the production of monoclonal antibodies, BALB/c mice were immunized with the M. tuberculosis Canetti mixed with Freund's incomplete adjuvant (Sigma Chemical Co., St. Louis, Mo.) intraperitoneally and given booster injections after 3 weeks. Four days after a second intravenous booster dose, the spleen was removed and the spleen cells were fused with Sp2/O-Ag-14 mouse myeloma cells as described by Kohler and Milstein (19). The culture soup from the well with growing cells was examined for the presence of antibodies to PGL-Th as described below. The colonies producing anti-PGL-Th antibodies were cloned twice at a density of 0.3 cell per well by limiting dilution. The culture soup from the cloned cells was used to characterize the monoclonal antibodies and to detect PGL-Th from clinical isolates. The 3065
3066
CHO ET AL.
isotypes of the monoclonal antibodies determined with a commercially available isotyping kit (Hyclone Laboratories, Inc., Logan, Utah). ELISA. The enzyme-linked immunosorbent assay (ELISA) described by Voller et al. (30), with the minor modifications reported previously (8, 9), was used. Briefly, 50 ,ul of PGL-Th antigen dissolved (5 ,ug/ml) in ethanol was added to wells of 96-well U-bottom microtiter plates (Dynatech Laboratories, Inc., Alexandria, Va.) and incubated at room temperature until complete evaporation. Peroxidase-conjugated anti-mouse immunoglobulins (immunoglobulin G [IgG] plus IgA plus IgM or IgG alone) and a substrate solution (o-phenylenediamine-H202) were employed in the assay, and the A490 was read. Each test was performed in duplicate, and the A490 of the wells without antigen were subtracted from those of wells with PGL-Th before analysis. The concentrations of other antigens were 5.0 ,g/ml for mycoside A, mycoside B, and GPL core and 0.2 ug/ml for LAM. Purification of lipid from M. tuberculosis and TLC analysis. M. tuberculosis clinical isolates from tuberculosis patients at the chest clinic at Yonsei Medical Center were grown on the surface of Sauton medium (50 ml in a 75-cm2 tissue culture flask) for 4 to 6 weeks, heat killed, and harvested by centrifugation. The whole cells were washed twice by centrifugation with phosphate-buffered saline solution (pH 7.2) and lyophilized. Lipids were extracted from about 100 mg of each M. tuberculosis isolate with the CHCl3-CH30H (2:1) (30 ml/g of dry weight) at 50°C for 48 h and dried under an N2 flow. The lipid extracts were washed with 3.5 ml of CHCl3CH30H-H20 (4:2:1), and the organic phase was collected and dried under N2. The mean weight of lipids at this stage was 11.9 mg (standard deviation, 1.7 mg). The washed lipid samples were divided in half and processed separately for TLC and dot ELISA. For purification, each lipid dissolved in CHCl3 was applied to a florisil (60- to 100-mesh) column (Sigma Chemical Co., St. Louis, Mo.) prepared in a pasteur pipette (bed volume, 2 ml) and eluted with 2 bed volumes of chloroform, chloroform-methanol (19:1, vol/vol), and then chloroform-methanol (9:1, vol/vol). The lipid fraction eluted with 5% (vol/vol) CH30H in CHCl3 was then examined by TLC and dot ELISA as described below. The glycolipid spot was visualized with a 10% H2SO4 spray. The residual lipids after TLC analysis were hydrolyzed and examined for the presence of the specific sugars of the PGL-Th antigen as described previously (10). Dot ELISA with monoclonal antibodies. The dot ELISA described by Hawkes et al. (12), with the minor modification reported previously (6), was employed. The purified lipid fractions from each M. tuberculosis isolate were dissolved in 100 pl of hexane, and 5-pl portion were applied to a Tuifryn (polysulfone) membrane (HT-200; Gelman Sciences, Inc., Ann Arbor, Mich.). A monoclonal antibody specific to PGL-Th produced in this study was used as the primary antibody, and peroxidase-conjugated goat anti-mouse IgG (Cooper Biomedical, Inc., Malvern, Pa.) was used as the secondary antibody. The results were read visually. RESULTS Characterization of the monoclonal antibody to PGL-Th. From several fusion attempts, one clone, MAbIII604, that produced antibodies (IgG3) to the PGL-Th antigen was obtained. To examine the reactivity of the PGL-Th antigen with the monoclonal antibodies, serial twofold dilutions of monoclonal antibodies to PGL-I, LAM, and normal mouse
J. CLIN. MICROBIOL.
1.2
-
1.0
-
E 0.8-
c
0
0.6 -
1:1
1:2
1:4 1:8 MAb dilution
1:16
1:32
FIG. 1. Anti-PGL-Th reactivity of monoclonal antibodies to PGL-Th (MAbIII604) (0), PGL-I (A), and LAM (A) and of normal mouse serum (0) in the ELISA. Each monoclonal antibody was diluted to give an optical density of 1.0 to 1.2 against the corresponding antigen before the twofold dilution was started.
serum were examined for reactivity with the antigen by ELISA. Only anti-PGL-Th monoclonal antibody showed a dose response-curve of seroreactivity with the PGL-Th antigen (Fig. 1), thus indicating that the culture soup contained monoclonal antibodies to the antigen. In the next experiment, the MAbIII604 was investigated for reactivity with phenolic glycolipids from M. leprae (PGL-I), M. bovis (mycoside B), M. kansasii (mycoside A), and with the GPL core of M. avium complex to examine the specificity of the monoclonal antibodies by ELISA. MAbIII604 reacted mainly with the PGL-Th antigen and showed little reactivity with other phenolic glycolipids and the GPL core (Fig. 2). This suggests that MAbIII604 is specific to the PGL-Th antigen. The specificity of MAbIII604 to the PGL-Th antigen also became clear in the dot ELISA, in which reactivity with the antigen was shown only with PGL-Tb (Fig. 3). Detection of the PGL-Th antigen by using MAbM6O4. A dot ELISA with MAbIII604 was employed to detect the PGL-Th antigen in lipid fractions of M. tuberculosis clinical 1.0
* PGL-Th E PGL-1 O Mycoside A 0 Mycoside B
0.8
E o 0.6 0) a
o C-Mycoside
0.4 0.2
0.0
Antigens FIG. 2. Seroreactivity of MAbIII604 to phenolic glycolipids from M. tuberculosis (PGL-Th), M. leprae (PGL-I), M. kansasii (mycoside A), and M. bovis (mycoside B) and to the mycoside C core of M. avium complex in the ELISA.
VOL. 30, 1992
MONOCLONAL ANTIBODY FOR DETECTION OF PGL-TB ANTIGEN
3067
TABLE 1. Detection of the PGL-Th antigen by dot ELISA with MAbIII604 and by TLC Dot ELISA results
Reactivityb
No. (%)
samplesa
+++ ++ +
10 (20.0) 12 (24.0) 10 (20.0) 18 (36.0) 50 (100)
8 0 0 0 8
+
FIG. 3. Dot ELISA results showing the specific reactivity of MAbIII604 with the PGL-Th antigen. A 5-pl sample of hexane solution containing 50 ng of glycolipid was applied to each spot. Spots: A, no antigen; B, PGL of M. leprae; C, PGL of M. kansasii; D, PGL of M. tuberculosis; E, GPL of M. avium; F, PGL of M. bovis.
isolates. To determine the sensitivity of the dot ELISA with MAbIII604, purified PGL-Tb was diluted from 10 ,ug/ml to 100 ng/ml in hexane containing 100 p,g of cholesterol per ml as an accompanying nonspecific lipid, because it was assumed that the lipid eluents from the florisil column contained at least 100 pg of other lipids besides PGL-Th. PGL-Th at a concentration of 500 ng/ml was detected by the dot ELISA in repeated tests; 250 ng of PGL-Th per ml gave a doubtful result (Fig. 4A). Since only 5 ,ul per dot was applied, the absolute amount of the PGL-Th antigen in a spot was 2.5 ng. Without the cholesterol supplement in the hexane diluent, a higher concentration of PGL-Th antigens (about 10 times) was required to test positive in the dot ELISA (data not shown). Figure 4B shows the dot-ELISA results for lipids eluted with 5% CH30H-CHCl3 from 50 M. tuberculosis isolates. Interestingly, there was a marked variation in the intensities and sizes of spots, indicating differences in PGL-Th antigen 1
q A
9
[i A 7 A Q I
No. of TLC-positive
Total
a Based on visual observation of the yellowish spot with the same Rf value with the standard PGL-Th antigen upon spraying with 10% H2SO4 in ethanol. b Based on visual observation of the intensities and sizes of spots in the dot ELISA.
production among M. tuberculosis clinical isolates despite growth in the same culture media and conditions. However, because lipids eluted with 5% CH30H-CHCl3 contained M. tuberculosis lipids, although in amounts too small to be accurately measured, it was difficult to determine the concentration of PGL-Th in each lipid simply by comparison with the results obtained from the standard PGL-Th antigen. Therefore, dot ELISA results were graded visually as +, +, + +, and + + + on the basis of intensities and sizes of colored spots (Fig. 4B). Of 50 M. tuberculosis isolates, 32 (64%) gave definite evidence of PGL-Th antigen and 18 (36.0%) showed doubtful (±) results (Table 1); in contrast, PGL-Th was detectable only in 8 (16.0%) of the clinical isolates by TLC analysis. As expected, all of the isolates that gave a positive TLC result showed a strong (+ + +) reactivity in the dot ELISA (Table 1). Some of the TLC patterns are shown in Fig. 5. Therefore, the dot ELISA with MAbIII604 was much more sensitive than the conventional TLC method in detecting PGL-Th antigen from lipid extracts. In an experiment to determine the sensitivity, TLC gave a positive result with a sample 250 ng (data not shown), thus indicating that the dot ELISA with MAbIII604 is about 100 times more sensitive than TLC in detecting the PGL-Th antigen.
A 1
2
3
4
5
6
7
8
9 10
I
B
2
3 4
FIG. 4. Reactivity of standard PGL-Th (A) and lipid fractions purified from 50 M. tuberculosis clinical isolates (B) with MAbIII604 in the dot ELISA. (A) Duplicate 5-p1 samples of serially diluted standard PGL-Th were applied; samples of 50, 37.5, 25, 12.5, 5, 3.8, 2.5, 1.25, 0.5, and 0 ng, respectively, were applied in columns 1 through 10. (B) Samples of 5 p1 of lipid fractions (dissolved in 100 p.1 hexane) eluted with 5% CH30H-CHCl3 from the florisil column applied with lipid extracts from each isolate were applied. Dot ELISA results were graded visually as , +, ++, and + ++, respectively, based on the intensities and sizes of colored spots as shown in columns 7, 9, 3, and 8 in row 1.
1
2
3
4
S
5
6
7
FIG. 5. TLC of lipids (lanes 1 through 7) eluted with 5% CH30H from a florisil column to which lipid extracts of M. tuberculosis isolates were applied. Lane S contains the standard PGL-Th. Chromatogram was developed in CHC13-CH30H (96:4), and glycolipids were located with 10% H2SO4 in ethanol. The arrow indicates a PGL-Th antigen spot.
3068
J. CLIN. MICROBIOL.
CHO ET AL.
DISCUSSION
Phenolic glycolipids consist of phenol phthiocerol as a structure and a carbohydrate that determines the specificity to each mycobacterial species (4). For example, the carbohydrate determinant of the PGL-Th antigen is 2,3,4-Me3Fuc-Rha-2-O-MeRha (10), and that of 3,6-0-Me2 Glc-2,3-O-Me2Rha-3-O-MeRha (13). The nature of the terminal sugar residues and the positions of 0 methylation at rhamnose residues determine the species specificities of the two phenolic glycolipids. This was supported by our findings that PGL-Th antigen reacted with MAbIII604 but not with monoclonal antibodies to PGL-I. The specific reactivity of MAbIII604 with PGL-Th and no significant reactivity with phenolic glycolipids from other Mycobacterium species confirmed that MAbIII604 was reactive to the carbohydrate immunodeterminant of the PGL-Th antigen. Therefore, it may be safe to assume that glycolipids reacting with MAbIII604 in the dot ELISA contain the sugar determinant of PGL-Th. Interestingly, there were marked variations in the PGL-Th contents in lipid fractions from different isolates. The significance of the differences among the PGL-Th antigen production levels of M. tuberculosis clinical isolates is not known. The PGL-Th antigen was originally isolated from the Canetti strain of M. tuberculosis (10) and was not detectable in lipids from M. tuberculosis H37Rv (24). When rabbit polyclonal anti-PGL-Tb antibodies were used to detect the antigen by ELISA, all of the clinical isolates of M. tuberculosis produced the antigen, although in different relative quantities (24). However, the PGL-Th antigen was detectable in only 1 (9.1%) of 11 isolates by TLC, although the oligoside was present in two other isolates (9). The different results in the two studies might be due mainly to differences in the sensitivities of the enzyme immunoassays and TLC in detecting the PGL-Th antigen. In this study, only 8 (16%) of the 50 clinical isolates of M. tuberculosis showed evidence of PGL-Th by TLC; this roughly matched the proportion (20%) of isolates whose lipids showed strong (+ + +) reactivity with MAbIII604 in the dot ELISA (Table 1). In another 40% of the isolates, the PGL-Th antigen was positively identified by the dot ELISA with the monoclonal antibodies but not by TLC. From this study with a larger number of isolates, therefore, it becomes more evident that there are marked differences in PGL-Th production levels among M. tuberculosis strains. Interestingly, the production of the PGL-Th antigen at a relatively higher level in about 20% of M. tuberculosis clinical isolates coincided with the prevalence (about 20%) of anti-PGL-Th IgM antibodies among tuberculosis patients (9). The correlation between anti-PGL-Th antibody response and the PGL-Th antigen production level of the isolate from the same patient is currently under investigation. In nearly 40% of clinical isolates in this study, the dot ELISA results were ambiguous, mainly because of the subjective visual reading of the results. Most of these isolates seemed to have some deposit of blue color but were not dark enough to be considered a definite positive reaction. This might imply that the dot ELISA using MAbIII604 was not sensitive enough to detect the PGL-Th antigen in the lipids, because the lipids were purified only from 100 mg (dry weight) of whole cells for each isolate. Recently, it was reported that all 14 clinical isolates had PGL-Th detectable by dot-ELISA when rabbit polyclonal antibodies to the antigen were used (25). In that study, however, 10 g (wet weight) of whole cells from each strain was used and about common
10 ng of purified PGL-Th antigen per spot was required to give a positive result in the dot ELISA with polyclonal antibodies. Since 2.5 to 5.0 ng of PGL-Th per spot was detectable in our study, MAbIII604 was comparable to rabbit polyclonal antibodies in detecting the PGL-Th antigen in the dot ELISA. Therefore, to detect the PGL-Th antigen from all clinical isolates, it may be necessary to purify lipids from larger amounts of whole cells than those used in this study. Although the dot ELISA with MAbIII604 provided a qualitative analysis, this study showed clear evidence of marked variations in the ability to produce the PGL-Th antigen. Differences in the amounts of PGL-Th produced by different clinical isolates were also found with an ELISA (24) and a dot ELISA (25) when rabbit polyclonal antibodies to the antigen were used. To date, little is known about the exact mechanism of the variation in PGL-Th production level among M. tuberculosis clinical isolates. It is not fully understood how the PGL-Th antigen production is stable in different culture media and conditions and is reproducible after repeated subculture. To minimize such confounding factors in this study, only clinical isolates that were passed in an egg-based medium a maximum of two times were grown on the surface of the Sauton medium. The variation in PGL-Th antigen levels among clinical isolates of M. tuberculosis might be due to differences in efficiency of lipid extraction from lyophilized whole cells. In this study, the mean weight of lipid extracts after two-phase washing was 11.9 mg and ranged from 9 to 15 mg. This difference in lipid amount was not remarkable given the dot ELISA results between isolates, thus indicating that the ability to produce PGL-Th varies among clinical isolates. In a previous study, mycoside B was detectable in 35 (92.1%) of 38 field isolates of M. bovis, and three strains did not produce the antigen when TLC was used to detect the antigen (16); no explanation was given for the absence of mycoside B in three isolates. The permanent loss of PGL-Th production in M. tuberculosis H37Rv might be due to genetic variation. This was partially supported by our previous study (21), in which it was shown that a 13-kb DNA fragment was present in clinical isolates producing the PGL-Th antigen but not in M. tuberculosis H37Rv and M. bovis. Cloning and expression of the genes responsible for the biosynthesis of the PGL-Th antigen may help explain the differences in antigen production level among M. tuberculosis clinical isolates. Therefore, the monoclonal antibody to the PGL-Th antigen produced in this study will be useful in related studies in the future. ACKNOWLEDGMENTS We thank E. 0. Shin and M. K. Lee for their excellent technical assistance. Special thanks are given to P. J. Brennan (Colorado State University) for providing various mycobacterial antigens and monoclonal antibodies and for helpful discussion. This work was supported in part by a Yonsei University Faculty Research Grant for 1991, a grant from the Korean Research Foundation, a grant from the IMMTUB Program of the World Health Organization, and a grant from the Fonds special des Comites departmentaux contre les maladies respiratoires et la tuberculose (contract 90 T/2).
REFERENCES 1. Andersen, A. B., and E. B. Hansen. 1989. Structure and mapping of antigenic domains of protein antigen b, a 38,000-molecularweight protein of Mycobacterinum tuberculosis. Infect. Immun. 57:2481-2488. 2. Baird, P. N., L. M. C. Hall, and A. R. M. Coates. 1989. Cloning
VOL. 30, 1992
3. 4.
5.
6.
7.
8.
9.
10.
11.
12. 13. 14.
15.
16. 17.
MONOCLONAL ANTIBODY FOR DETECTION OF PGL-TB ANTIGEN
and sequence analysis of the 10 kDa antigen gene of Mycobacterium tuberculosis. J. Gen. Microbiol. 135:931-939. Brennan, P. J. 1988. Mycobacterium and other actinomycetes, p. 203-298. In C. Ratledge and S. G. Wilkinson (ed.), Microbial lipids, vol. 1. Academic Press Ltd., London. Brennan, P. J., S. W. Hunter, M. McNeil, D. Chatterjee, and M. Daffe. 1990. Reappraisal of the chemistry of mycobacterial cell walls, with a view to understanding the roles of individual entities in disease processes, p. 55-75. In E. M. Ayoub, G. H. Cassell, W. C. Branche, Jr., and T. J. Henry (ed.), Microbial determinants of virulence and host response. American Society for Microbiology, Washington, D.C. Casabona, N. M., T. G. Fuente, L. A. Arce, J. 0. Entraigas, and R. V. Pla. 1989. Evaluation of a phenolglycolipid antigen (PGL-Thl) from M. tuberculosis in the serodiagnosis of tuberculosis: comparison with PPD antigen. Acta Leprol. 7(Suppl. 1):89-93. Cho, S. N., R. V. Cellona, T. T. Fajardo, Jr., R. M. Abalos, E. C. Dela Cruz, G. P. Walsh, J. D. Kim, and P. J. Brennan. 1991. Detection of phenolic glycolipid-I antigen and antibody in sera from new and relapsed lepromatous patients treated with various drug regimens. Int. J. Lepr. 59:25-31. Cho, S. N., J. S. Shin, J. D. Kim, and Y. Chong. 1990. Production of monoclonal antibodies to lipoarabinomannan-B and use in the detection of mycobacterial antigens in sputum. Yonsei Med. J. 31:333-338. Cho, S. N., D. L. Yanagihara, S. W. Hunter, R. H. Gelber, and P. J. Brennan. 1983. Serological specificity of phenolic glycolipid I from Mycobacterium leprae and use in serodiagnosis of leprosy. Infect. Immun. 41:1077-1083. Daffe, M., S. N. Cho, D. Chatterjee, and P. J. Brennan. 1991. Chemical synthesis and seroreactivity of neoantigen containing the oligosaccharide hapten of the Mycobacterium tuberculosisspecific phenolic glycolipid. J. Infect. Dis. 163:161-168. Daffe, M., C. Lacave, M. A. Laneelle, and G. Laneelle. 1987. Structure of the major triglycosyl phenol-phthiocerol of Mycobacterium tuberculosis (strain Canetti). Eur. J. Biochem. 167: 155-160. Engers, H. D., B. R. Bloom, and T. Godal. 1985. Monoclonal antibodies against mycobacterial antigens. Immunol. Today 6:345-348. Hawkes, R., E. L. Niday, and J. A. Gordon. 1982. A dotimmunobinding assay for monoclonal and other antibodies. Anal. Biochem. 119:142-147. Hunter, S. W., and P. J. Brennan. 1981. A novel phenolic glycolipid from Mycobacterium leprae possibly involved in immunogenicity and pathogenicity. J. Bacteriol. 147:728-735. Ivanyi, J., E. Krambovitis, and M. Keen. 1983. Evaluation of a monoclonal antibody (TB72) based serological test for tuberculosis. Clin. Exp. Immunol. 54:337-345. Jackett, P. S., G. H. Bothamley, H. V. Batra, A. Mistry, D. B. Young, and J. Ivanyi. 1988. Specificity of antibodies to immunodominant mycobacterial antigens in pulmonary tuberculosis. J. Clin. Microbiol. 26:2313-2318. Jarnagin, J. L., P. J. Brennan, and S. K. Harris. 1983. Rapid identification of Mycobacterinum bovis by a thin-layer chromatographic technique. Am. J. Vet. Res. 44:1920-1922. Kadival, G. V., and S. D. Chaparas. 1987. Production, charac-
18.
19. 20.
21.
22; 23.
24. 25.
26.
27.
28.
29.
30. 31. 32.
3069
terization, and species specificity of five monoclonal antibodies to Mycobacterium tuberculosis. J. Clin. Microbiol. 25:76-80. Kingston, A. E., P. R. Salgame, N. A. Mitchison, and M. J. Colston. 1987. Immunological activity of a 14-kilodalton recombinant protein of Mycobacteinum tuberculosis H37Rv. Infect. Immun. 55:3149-3154. Kohler, G., and C. Milstein. 1975. Continuous cultures of fused cells secreting antibody of predefined specificity. Nature (London) 256:495-497. Kolk, A. H. J., M. L. Ho, P. R. Kiatser, T. A. Eggelte, S. Kuiper, S. De Jonge, and J. Van Leeuwen. 1984. Production and characterization of monoclonal antibodies to Mycobacterium tuberculosis, M. bovis (BCG) and M. leprae. Clin. Exp. Immunol. 58:511-521. Lee, T. Y., S. N. Cho, K. H. Yoon, J. S. Shin, and J. D. Kim. 1991. Comparison of DNA fragment patterns between the phenolic glycolipid-Th producers and non-producers of Mycobacterium tuberculosis. Yonsei Med. J. 32:243-249. Neill, M. A., and S. J. Klebanoff. 1988. The effect of phenolic glycolipid-I from Mycobacterium leprae on the antimicrobial activity of human macrophages. J. Exp. Med. 167:30-42. Olds, G. R., A. J. Sanson, and T. M. Daniel. 1987. Characterization of Mycobacterium tuberculosis antigen 5 epitopes by using a panel of 19 monoclonal antibodies. J. Clin. Microbiol. 25:471-475. Papa, F., A. Laszlo, H. L. David, and M. Daffe. 1989. Serological specificity of Mycobacterium tuberculosis glycolipids. Acta Leprol. 7(Suppl. 1):98-101. Papa, F., M. Luquin, and H. L. David. 1992. DOT-ELISA for detection of phenolic glycolipid PGL-Thl and diacyl-trehalose antigens of Mycobacterium tuberculosis. Res. Microbiol. 143: 327-331. Shinnick, T. M. 1987. The 65-kilodalton antigen of Mycobacterium tuberculosis. J. Bacteriol. 169:1080-1088. Torgal-Garcia, J., H. L. David, and F. Papa. 1988. Preliminary evaluation of a Mycobactenium tuberculosis phenolglycolipid antigen in the serologic diagnosis of tuberculosis. Ann. Inst. Pasteur/Microbiol. 139:289-294. Torgal-Garcia, J., F. Papa, and H. L. David. 1989. The use of PGL-Thl from M. tuberculosis in the serological diagnosis of tuberculous meningitis: a preliminary report. Acta Leprol. 7(Suppl. 1):136-137. Vismara, D., M. F. Mezzopreti, M. S. Gilardini, P. Del Porto, G. Lombardi, E. Piccolella, G. Damiani, R. Rappuoli, and V. Colizzi. 1990. Identification of a 35-kilodalton Mycobactenium tuberculosis protein containing B- and T-cell epitopes. Infect. Immun. 58:245-251. Voller, A., D. E. Bidwell, and A. Bartlette. 1979. The enzymelinked immunosorbent assay (ELISA). Dynatech Laboratories, Alexandria, Va. Young, D. B., J. P. Harnisch, J. Knight, and T. M. Buchanan. 1985. Detection of phenolic glycolipid I in sera from patients with lepromatous leprosy. J. Infect. Dis. 152:1978-1081. Young, D. B., S. R. Khanolkar, L. L. Barg, and T. M. Buchanan. 1984. Generation and characterization of monoclonal antibodies to the phenolic glycolipid of Mycobacterium leprae. Infect. Immun. 43:183-188.
JOURNAL OF CLINICAL MICROBIOLOGY, Dec. 1992, p. 3065-3069
0095-1137/92/123065-05$02.00/0
Copyright X) 1992, American Society for Microbiology
Production of Monoclonal Antibody to a Phenolic Glycolipid of Mycobacterium tuberculosis and Its Use in Detection of the Antigen in Clinical Isolates SANG-NAE CHO,`* JEON-SOO SHIN,1 MAMADOU DAFFE,2 YUNSOP CHONG,3 SUNG-KYU KIM,4 AND JOO-DEUK KIM1 1 Departments of Microbiology, Clinical Pathology,3 and Medicine,4 Yonsei University College of Medicine, Seoul 120-752, Republic of Korea, and Centre de Recherche de Biochimie et Genetique Cellulaires du Centre National de la Recherche Scientifique, 31062 Toulouse Cedex, France2 Received 29 May 1992/Accepted 4 September 1992
A monoclonal antibody (MAbIII604) specific to phenolic glycolipid Th (PGL-Th), a Mycobacterium tuberculosis-specific antigen, was produced and used in the detection of the antigen. MAbI604 reacted with the PGL-Th antigen but not with other phenolic glycolipids from Mycobacterium leprae, M. bovis, and M. kansasii, thus indicating the specificity of the monoclonal antibody to PGL-Th. A dot enzyme-linked immunosorbent assay with MAbUI604 was employed to detect the PGL-Th antigen in lipids purified from M. tuberculosis clinical isolates. Of 50 isolates, 32 (64.0%o) showed clear evidence of the PGL-Th antigen by the dot enzyme-linked immunosorbent assay, but there were marked variations in the intensities and sizes of spots. This suggests differences in PGL-Th antigen production among M. tuberculosis strains even when they are grown in the same culture media and conditions. This was most evident from the fact that in only eight (16.0%o) of the isolates examined was the PGL-Th antigen detectable by thin-layer chromatography, which is much less sensitive for the detection of glycolipid antigens. This study shows that monoclonal antibodies specific to PGL-Th are useful in detecting the antigen in lipid extracts and that there is a marked variation in the PGL-Th production among M. tuberculosis clinical isolates.
Isolation and characterization of phenolic glycolipid Th (PGL-Th), a Mycobacterium tuberculosis-specific antigen (10), has quickly attracted attention for use of the antigen in the serodiagnosis of tuberculosis and for understanding the pathogenesis of tuberculosis. However, the rates of seropositive response to the antigen among tuberculosis patients varied from about 20% to more than 95% in different studies (5, 9, 27, 28). There were also conflicting reports on PGL-Th production in clinical isolates of M. tuberculosis. In one study, PGL-Th was detectable in all clinical isolates by an enzyme immunoassay (24); in another study, however, only about 9% of isolates produced an antigen that was detectable by thin-layer chromatography (TLC) (9). Since the studies examined only a few clinical isolates and employed different tools for detecting the PGL-Th antigen, we were interested in investigating a larger number of M. tuberculosis clinical isolates for the production of the antigen by an enzyme immunoassay and TLC. For enzyme immunoassays, monoclonal antibodies to PGL-Th provide a great advantage in the specific detection of the antigen. Monoclonal antibodies against numerous antigens of M. tuberculosis have been produced (11, 17, 20, 23) and used for the cloning and characterization of the genes responsible for expression of the antigens (1, 2, 18, 26, 29). Some of the M. tuberculosis-specific monoclonal antibodies were to detect antibodies specific to epitopes present only in tubercle bacilli for the serodiagnosis of tuberculosis by competitive immunoassays (14, 15). In addition, monoclonal antibodies to phenolic glycolipid I (PGL-I) of Mycobacterium leprae were used to detect PGL-I antigen (31, 32) and to study the role of the antigen in the pathogenesis of M. leprae infection (22). *
Corresponding author.
Monoclonal antibodies to the PGL-Th antigen thus could be used in various studies related to the antigen. In this study, a monoclonal antibody specific to the PGL-Th antigen was produced and used to detect the antigen in lipids purified from M. tuberculosis clinical isolates.
MATERIALS AND METHODS Mycobacterial antigens and antibodies. M. tuberculosis Canetti and PGL-Th antigen were prepared as previously described (10). Phenolic glycolipids (3) from M. leprae (PGL-I), Mycobacterium bovis (mycoside B), and M. kansasii (mycoside A) were provided by P. J. Brennan (Colorado State University, Fort Collins). The glycopeptidolipid (GPL) core (3) of Mycobacterium avium complex and lipoarabinomannan (LAM) of M. tuberculosis antigens and monoclonal antibody to PGL-I were also provided by P. J. Brennan. A monoclonal antibody to LAM produced previously (7) was also used in this study. Production of monoclonal antibodies to PGL-Th. For the production of monoclonal antibodies, BALB/c mice were immunized with the M. tuberculosis Canetti mixed with Freund's incomplete adjuvant (Sigma Chemical Co., St. Louis, Mo.) intraperitoneally and given booster injections after 3 weeks. Four days after a second intravenous booster dose, the spleen was removed and the spleen cells were fused with Sp2/O-Ag-14 mouse myeloma cells as described by Kohler and Milstein (19). The culture soup from the well with growing cells was examined for the presence of antibodies to PGL-Th as described below. The colonies producing anti-PGL-Th antibodies were cloned twice at a density of 0.3 cell per well by limiting dilution. The culture soup from the cloned cells was used to characterize the monoclonal antibodies and to detect PGL-Th from clinical isolates. The 3065
3066
CHO ET AL.
isotypes of the monoclonal antibodies determined with a commercially available isotyping kit (Hyclone Laboratories, Inc., Logan, Utah). ELISA. The enzyme-linked immunosorbent assay (ELISA) described by Voller et al. (30), with the minor modifications reported previously (8, 9), was used. Briefly, 50 ,ul of PGL-Th antigen dissolved (5 ,ug/ml) in ethanol was added to wells of 96-well U-bottom microtiter plates (Dynatech Laboratories, Inc., Alexandria, Va.) and incubated at room temperature until complete evaporation. Peroxidase-conjugated anti-mouse immunoglobulins (immunoglobulin G [IgG] plus IgA plus IgM or IgG alone) and a substrate solution (o-phenylenediamine-H202) were employed in the assay, and the A490 was read. Each test was performed in duplicate, and the A490 of the wells without antigen were subtracted from those of wells with PGL-Th before analysis. The concentrations of other antigens were 5.0 ,g/ml for mycoside A, mycoside B, and GPL core and 0.2 ug/ml for LAM. Purification of lipid from M. tuberculosis and TLC analysis. M. tuberculosis clinical isolates from tuberculosis patients at the chest clinic at Yonsei Medical Center were grown on the surface of Sauton medium (50 ml in a 75-cm2 tissue culture flask) for 4 to 6 weeks, heat killed, and harvested by centrifugation. The whole cells were washed twice by centrifugation with phosphate-buffered saline solution (pH 7.2) and lyophilized. Lipids were extracted from about 100 mg of each M. tuberculosis isolate with the CHCl3-CH30H (2:1) (30 ml/g of dry weight) at 50°C for 48 h and dried under an N2 flow. The lipid extracts were washed with 3.5 ml of CHCl3CH30H-H20 (4:2:1), and the organic phase was collected and dried under N2. The mean weight of lipids at this stage was 11.9 mg (standard deviation, 1.7 mg). The washed lipid samples were divided in half and processed separately for TLC and dot ELISA. For purification, each lipid dissolved in CHCl3 was applied to a florisil (60- to 100-mesh) column (Sigma Chemical Co., St. Louis, Mo.) prepared in a pasteur pipette (bed volume, 2 ml) and eluted with 2 bed volumes of chloroform, chloroform-methanol (19:1, vol/vol), and then chloroform-methanol (9:1, vol/vol). The lipid fraction eluted with 5% (vol/vol) CH30H in CHCl3 was then examined by TLC and dot ELISA as described below. The glycolipid spot was visualized with a 10% H2SO4 spray. The residual lipids after TLC analysis were hydrolyzed and examined for the presence of the specific sugars of the PGL-Th antigen as described previously (10). Dot ELISA with monoclonal antibodies. The dot ELISA described by Hawkes et al. (12), with the minor modification reported previously (6), was employed. The purified lipid fractions from each M. tuberculosis isolate were dissolved in 100 pl of hexane, and 5-pl portion were applied to a Tuifryn (polysulfone) membrane (HT-200; Gelman Sciences, Inc., Ann Arbor, Mich.). A monoclonal antibody specific to PGL-Th produced in this study was used as the primary antibody, and peroxidase-conjugated goat anti-mouse IgG (Cooper Biomedical, Inc., Malvern, Pa.) was used as the secondary antibody. The results were read visually. RESULTS Characterization of the monoclonal antibody to PGL-Th. From several fusion attempts, one clone, MAbIII604, that produced antibodies (IgG3) to the PGL-Th antigen was obtained. To examine the reactivity of the PGL-Th antigen with the monoclonal antibodies, serial twofold dilutions of monoclonal antibodies to PGL-I, LAM, and normal mouse
J. CLIN. MICROBIOL.
1.2
-
1.0
-
E 0.8-
c
0
0.6 -
1:1
1:2
1:4 1:8 MAb dilution
1:16
1:32
FIG. 1. Anti-PGL-Th reactivity of monoclonal antibodies to PGL-Th (MAbIII604) (0), PGL-I (A), and LAM (A) and of normal mouse serum (0) in the ELISA. Each monoclonal antibody was diluted to give an optical density of 1.0 to 1.2 against the corresponding antigen before the twofold dilution was started.
serum were examined for reactivity with the antigen by ELISA. Only anti-PGL-Th monoclonal antibody showed a dose response-curve of seroreactivity with the PGL-Th antigen (Fig. 1), thus indicating that the culture soup contained monoclonal antibodies to the antigen. In the next experiment, the MAbIII604 was investigated for reactivity with phenolic glycolipids from M. leprae (PGL-I), M. bovis (mycoside B), M. kansasii (mycoside A), and with the GPL core of M. avium complex to examine the specificity of the monoclonal antibodies by ELISA. MAbIII604 reacted mainly with the PGL-Th antigen and showed little reactivity with other phenolic glycolipids and the GPL core (Fig. 2). This suggests that MAbIII604 is specific to the PGL-Th antigen. The specificity of MAbIII604 to the PGL-Th antigen also became clear in the dot ELISA, in which reactivity with the antigen was shown only with PGL-Tb (Fig. 3). Detection of the PGL-Th antigen by using MAbM6O4. A dot ELISA with MAbIII604 was employed to detect the PGL-Th antigen in lipid fractions of M. tuberculosis clinical 1.0
* PGL-Th E PGL-1 O Mycoside A 0 Mycoside B
0.8
E o 0.6 0) a
o C-Mycoside
0.4 0.2
0.0
Antigens FIG. 2. Seroreactivity of MAbIII604 to phenolic glycolipids from M. tuberculosis (PGL-Th), M. leprae (PGL-I), M. kansasii (mycoside A), and M. bovis (mycoside B) and to the mycoside C core of M. avium complex in the ELISA.
VOL. 30, 1992
MONOCLONAL ANTIBODY FOR DETECTION OF PGL-TB ANTIGEN
3067
TABLE 1. Detection of the PGL-Th antigen by dot ELISA with MAbIII604 and by TLC Dot ELISA results
Reactivityb
No. (%)
samplesa
+++ ++ +
10 (20.0) 12 (24.0) 10 (20.0) 18 (36.0) 50 (100)
8 0 0 0 8
+
FIG. 3. Dot ELISA results showing the specific reactivity of MAbIII604 with the PGL-Th antigen. A 5-pl sample of hexane solution containing 50 ng of glycolipid was applied to each spot. Spots: A, no antigen; B, PGL of M. leprae; C, PGL of M. kansasii; D, PGL of M. tuberculosis; E, GPL of M. avium; F, PGL of M. bovis.
isolates. To determine the sensitivity of the dot ELISA with MAbIII604, purified PGL-Tb was diluted from 10 ,ug/ml to 100 ng/ml in hexane containing 100 p,g of cholesterol per ml as an accompanying nonspecific lipid, because it was assumed that the lipid eluents from the florisil column contained at least 100 pg of other lipids besides PGL-Th. PGL-Th at a concentration of 500 ng/ml was detected by the dot ELISA in repeated tests; 250 ng of PGL-Th per ml gave a doubtful result (Fig. 4A). Since only 5 ,ul per dot was applied, the absolute amount of the PGL-Th antigen in a spot was 2.5 ng. Without the cholesterol supplement in the hexane diluent, a higher concentration of PGL-Th antigens (about 10 times) was required to test positive in the dot ELISA (data not shown). Figure 4B shows the dot-ELISA results for lipids eluted with 5% CH30H-CHCl3 from 50 M. tuberculosis isolates. Interestingly, there was a marked variation in the intensities and sizes of spots, indicating differences in PGL-Th antigen 1
q A
9
[i A 7 A Q I
No. of TLC-positive
Total
a Based on visual observation of the yellowish spot with the same Rf value with the standard PGL-Th antigen upon spraying with 10% H2SO4 in ethanol. b Based on visual observation of the intensities and sizes of spots in the dot ELISA.
production among M. tuberculosis clinical isolates despite growth in the same culture media and conditions. However, because lipids eluted with 5% CH30H-CHCl3 contained M. tuberculosis lipids, although in amounts too small to be accurately measured, it was difficult to determine the concentration of PGL-Th in each lipid simply by comparison with the results obtained from the standard PGL-Th antigen. Therefore, dot ELISA results were graded visually as +, +, + +, and + + + on the basis of intensities and sizes of colored spots (Fig. 4B). Of 50 M. tuberculosis isolates, 32 (64%) gave definite evidence of PGL-Th antigen and 18 (36.0%) showed doubtful (±) results (Table 1); in contrast, PGL-Th was detectable only in 8 (16.0%) of the clinical isolates by TLC analysis. As expected, all of the isolates that gave a positive TLC result showed a strong (+ + +) reactivity in the dot ELISA (Table 1). Some of the TLC patterns are shown in Fig. 5. Therefore, the dot ELISA with MAbIII604 was much more sensitive than the conventional TLC method in detecting PGL-Th antigen from lipid extracts. In an experiment to determine the sensitivity, TLC gave a positive result with a sample 250 ng (data not shown), thus indicating that the dot ELISA with MAbIII604 is about 100 times more sensitive than TLC in detecting the PGL-Th antigen.
A 1
2
3
4
5
6
7
8
9 10
I
B
2
3 4
FIG. 4. Reactivity of standard PGL-Th (A) and lipid fractions purified from 50 M. tuberculosis clinical isolates (B) with MAbIII604 in the dot ELISA. (A) Duplicate 5-p1 samples of serially diluted standard PGL-Th were applied; samples of 50, 37.5, 25, 12.5, 5, 3.8, 2.5, 1.25, 0.5, and 0 ng, respectively, were applied in columns 1 through 10. (B) Samples of 5 p1 of lipid fractions (dissolved in 100 p.1 hexane) eluted with 5% CH30H-CHCl3 from the florisil column applied with lipid extracts from each isolate were applied. Dot ELISA results were graded visually as , +, ++, and + ++, respectively, based on the intensities and sizes of colored spots as shown in columns 7, 9, 3, and 8 in row 1.
1
2
3
4
S
5
6
7
FIG. 5. TLC of lipids (lanes 1 through 7) eluted with 5% CH30H from a florisil column to which lipid extracts of M. tuberculosis isolates were applied. Lane S contains the standard PGL-Th. Chromatogram was developed in CHC13-CH30H (96:4), and glycolipids were located with 10% H2SO4 in ethanol. The arrow indicates a PGL-Th antigen spot.
3068
J. CLIN. MICROBIOL.
CHO ET AL.
DISCUSSION
Phenolic glycolipids consist of phenol phthiocerol as a structure and a carbohydrate that determines the specificity to each mycobacterial species (4). For example, the carbohydrate determinant of the PGL-Th antigen is 2,3,4-Me3Fuc-Rha-2-O-MeRha (10), and that of 3,6-0-Me2 Glc-2,3-O-Me2Rha-3-O-MeRha (13). The nature of the terminal sugar residues and the positions of 0 methylation at rhamnose residues determine the species specificities of the two phenolic glycolipids. This was supported by our findings that PGL-Th antigen reacted with MAbIII604 but not with monoclonal antibodies to PGL-I. The specific reactivity of MAbIII604 with PGL-Th and no significant reactivity with phenolic glycolipids from other Mycobacterium species confirmed that MAbIII604 was reactive to the carbohydrate immunodeterminant of the PGL-Th antigen. Therefore, it may be safe to assume that glycolipids reacting with MAbIII604 in the dot ELISA contain the sugar determinant of PGL-Th. Interestingly, there were marked variations in the PGL-Th contents in lipid fractions from different isolates. The significance of the differences among the PGL-Th antigen production levels of M. tuberculosis clinical isolates is not known. The PGL-Th antigen was originally isolated from the Canetti strain of M. tuberculosis (10) and was not detectable in lipids from M. tuberculosis H37Rv (24). When rabbit polyclonal anti-PGL-Tb antibodies were used to detect the antigen by ELISA, all of the clinical isolates of M. tuberculosis produced the antigen, although in different relative quantities (24). However, the PGL-Th antigen was detectable in only 1 (9.1%) of 11 isolates by TLC, although the oligoside was present in two other isolates (9). The different results in the two studies might be due mainly to differences in the sensitivities of the enzyme immunoassays and TLC in detecting the PGL-Th antigen. In this study, only 8 (16%) of the 50 clinical isolates of M. tuberculosis showed evidence of PGL-Th by TLC; this roughly matched the proportion (20%) of isolates whose lipids showed strong (+ + +) reactivity with MAbIII604 in the dot ELISA (Table 1). In another 40% of the isolates, the PGL-Th antigen was positively identified by the dot ELISA with the monoclonal antibodies but not by TLC. From this study with a larger number of isolates, therefore, it becomes more evident that there are marked differences in PGL-Th production levels among M. tuberculosis strains. Interestingly, the production of the PGL-Th antigen at a relatively higher level in about 20% of M. tuberculosis clinical isolates coincided with the prevalence (about 20%) of anti-PGL-Th IgM antibodies among tuberculosis patients (9). The correlation between anti-PGL-Th antibody response and the PGL-Th antigen production level of the isolate from the same patient is currently under investigation. In nearly 40% of clinical isolates in this study, the dot ELISA results were ambiguous, mainly because of the subjective visual reading of the results. Most of these isolates seemed to have some deposit of blue color but were not dark enough to be considered a definite positive reaction. This might imply that the dot ELISA using MAbIII604 was not sensitive enough to detect the PGL-Th antigen in the lipids, because the lipids were purified only from 100 mg (dry weight) of whole cells for each isolate. Recently, it was reported that all 14 clinical isolates had PGL-Th detectable by dot-ELISA when rabbit polyclonal antibodies to the antigen were used (25). In that study, however, 10 g (wet weight) of whole cells from each strain was used and about common
10 ng of purified PGL-Th antigen per spot was required to give a positive result in the dot ELISA with polyclonal antibodies. Since 2.5 to 5.0 ng of PGL-Th per spot was detectable in our study, MAbIII604 was comparable to rabbit polyclonal antibodies in detecting the PGL-Th antigen in the dot ELISA. Therefore, to detect the PGL-Th antigen from all clinical isolates, it may be necessary to purify lipids from larger amounts of whole cells than those used in this study. Although the dot ELISA with MAbIII604 provided a qualitative analysis, this study showed clear evidence of marked variations in the ability to produce the PGL-Th antigen. Differences in the amounts of PGL-Th produced by different clinical isolates were also found with an ELISA (24) and a dot ELISA (25) when rabbit polyclonal antibodies to the antigen were used. To date, little is known about the exact mechanism of the variation in PGL-Th production level among M. tuberculosis clinical isolates. It is not fully understood how the PGL-Th antigen production is stable in different culture media and conditions and is reproducible after repeated subculture. To minimize such confounding factors in this study, only clinical isolates that were passed in an egg-based medium a maximum of two times were grown on the surface of the Sauton medium. The variation in PGL-Th antigen levels among clinical isolates of M. tuberculosis might be due to differences in efficiency of lipid extraction from lyophilized whole cells. In this study, the mean weight of lipid extracts after two-phase washing was 11.9 mg and ranged from 9 to 15 mg. This difference in lipid amount was not remarkable given the dot ELISA results between isolates, thus indicating that the ability to produce PGL-Th varies among clinical isolates. In a previous study, mycoside B was detectable in 35 (92.1%) of 38 field isolates of M. bovis, and three strains did not produce the antigen when TLC was used to detect the antigen (16); no explanation was given for the absence of mycoside B in three isolates. The permanent loss of PGL-Th production in M. tuberculosis H37Rv might be due to genetic variation. This was partially supported by our previous study (21), in which it was shown that a 13-kb DNA fragment was present in clinical isolates producing the PGL-Th antigen but not in M. tuberculosis H37Rv and M. bovis. Cloning and expression of the genes responsible for the biosynthesis of the PGL-Th antigen may help explain the differences in antigen production level among M. tuberculosis clinical isolates. Therefore, the monoclonal antibody to the PGL-Th antigen produced in this study will be useful in related studies in the future. ACKNOWLEDGMENTS We thank E. 0. Shin and M. K. Lee for their excellent technical assistance. Special thanks are given to P. J. Brennan (Colorado State University) for providing various mycobacterial antigens and monoclonal antibodies and for helpful discussion. This work was supported in part by a Yonsei University Faculty Research Grant for 1991, a grant from the Korean Research Foundation, a grant from the IMMTUB Program of the World Health Organization, and a grant from the Fonds special des Comites departmentaux contre les maladies respiratoires et la tuberculose (contract 90 T/2).
REFERENCES 1. Andersen, A. B., and E. B. Hansen. 1989. Structure and mapping of antigenic domains of protein antigen b, a 38,000-molecularweight protein of Mycobacterinum tuberculosis. Infect. Immun. 57:2481-2488. 2. Baird, P. N., L. M. C. Hall, and A. R. M. Coates. 1989. Cloning
VOL. 30, 1992
3. 4.
5.
6.
7.
8.
9.
10.
11.
12. 13. 14.
15.
16. 17.
MONOCLONAL ANTIBODY FOR DETECTION OF PGL-TB ANTIGEN
and sequence analysis of the 10 kDa antigen gene of Mycobacterium tuberculosis. J. Gen. Microbiol. 135:931-939. Brennan, P. J. 1988. Mycobacterium and other actinomycetes, p. 203-298. In C. Ratledge and S. G. Wilkinson (ed.), Microbial lipids, vol. 1. Academic Press Ltd., London. Brennan, P. J., S. W. Hunter, M. McNeil, D. Chatterjee, and M. Daffe. 1990. Reappraisal of the chemistry of mycobacterial cell walls, with a view to understanding the roles of individual entities in disease processes, p. 55-75. In E. M. Ayoub, G. H. Cassell, W. C. Branche, Jr., and T. J. Henry (ed.), Microbial determinants of virulence and host response. American Society for Microbiology, Washington, D.C. Casabona, N. M., T. G. Fuente, L. A. Arce, J. 0. Entraigas, and R. V. Pla. 1989. Evaluation of a phenolglycolipid antigen (PGL-Thl) from M. tuberculosis in the serodiagnosis of tuberculosis: comparison with PPD antigen. Acta Leprol. 7(Suppl. 1):89-93. Cho, S. N., R. V. Cellona, T. T. Fajardo, Jr., R. M. Abalos, E. C. Dela Cruz, G. P. Walsh, J. D. Kim, and P. J. Brennan. 1991. Detection of phenolic glycolipid-I antigen and antibody in sera from new and relapsed lepromatous patients treated with various drug regimens. Int. J. Lepr. 59:25-31. Cho, S. N., J. S. Shin, J. D. Kim, and Y. Chong. 1990. Production of monoclonal antibodies to lipoarabinomannan-B and use in the detection of mycobacterial antigens in sputum. Yonsei Med. J. 31:333-338. Cho, S. N., D. L. Yanagihara, S. W. Hunter, R. H. Gelber, and P. J. Brennan. 1983. Serological specificity of phenolic glycolipid I from Mycobacterium leprae and use in serodiagnosis of leprosy. Infect. Immun. 41:1077-1083. Daffe, M., S. N. Cho, D. Chatterjee, and P. J. Brennan. 1991. Chemical synthesis and seroreactivity of neoantigen containing the oligosaccharide hapten of the Mycobacterium tuberculosisspecific phenolic glycolipid. J. Infect. Dis. 163:161-168. Daffe, M., C. Lacave, M. A. Laneelle, and G. Laneelle. 1987. Structure of the major triglycosyl phenol-phthiocerol of Mycobacterium tuberculosis (strain Canetti). Eur. J. Biochem. 167: 155-160. Engers, H. D., B. R. Bloom, and T. Godal. 1985. Monoclonal antibodies against mycobacterial antigens. Immunol. Today 6:345-348. Hawkes, R., E. L. Niday, and J. A. Gordon. 1982. A dotimmunobinding assay for monoclonal and other antibodies. Anal. Biochem. 119:142-147. Hunter, S. W., and P. J. Brennan. 1981. A novel phenolic glycolipid from Mycobacterium leprae possibly involved in immunogenicity and pathogenicity. J. Bacteriol. 147:728-735. Ivanyi, J., E. Krambovitis, and M. Keen. 1983. Evaluation of a monoclonal antibody (TB72) based serological test for tuberculosis. Clin. Exp. Immunol. 54:337-345. Jackett, P. S., G. H. Bothamley, H. V. Batra, A. Mistry, D. B. Young, and J. Ivanyi. 1988. Specificity of antibodies to immunodominant mycobacterial antigens in pulmonary tuberculosis. J. Clin. Microbiol. 26:2313-2318. Jarnagin, J. L., P. J. Brennan, and S. K. Harris. 1983. Rapid identification of Mycobacterinum bovis by a thin-layer chromatographic technique. Am. J. Vet. Res. 44:1920-1922. Kadival, G. V., and S. D. Chaparas. 1987. Production, charac-
18.
19. 20.
21.
22; 23.
24. 25.
26.
27.
28.
29.
30. 31. 32.
3069
terization, and species specificity of five monoclonal antibodies to Mycobacterium tuberculosis. J. Clin. Microbiol. 25:76-80. Kingston, A. E., P. R. Salgame, N. A. Mitchison, and M. J. Colston. 1987. Immunological activity of a 14-kilodalton recombinant protein of Mycobacteinum tuberculosis H37Rv. Infect. Immun. 55:3149-3154. Kohler, G., and C. Milstein. 1975. Continuous cultures of fused cells secreting antibody of predefined specificity. Nature (London) 256:495-497. Kolk, A. H. J., M. L. Ho, P. R. Kiatser, T. A. Eggelte, S. Kuiper, S. De Jonge, and J. Van Leeuwen. 1984. Production and characterization of monoclonal antibodies to Mycobacterium tuberculosis, M. bovis (BCG) and M. leprae. Clin. Exp. Immunol. 58:511-521. Lee, T. Y., S. N. Cho, K. H. Yoon, J. S. Shin, and J. D. Kim. 1991. Comparison of DNA fragment patterns between the phenolic glycolipid-Th producers and non-producers of Mycobacterium tuberculosis. Yonsei Med. J. 32:243-249. Neill, M. A., and S. J. Klebanoff. 1988. The effect of phenolic glycolipid-I from Mycobacterium leprae on the antimicrobial activity of human macrophages. J. Exp. Med. 167:30-42. Olds, G. R., A. J. Sanson, and T. M. Daniel. 1987. Characterization of Mycobacterium tuberculosis antigen 5 epitopes by using a panel of 19 monoclonal antibodies. J. Clin. Microbiol. 25:471-475. Papa, F., A. Laszlo, H. L. David, and M. Daffe. 1989. Serological specificity of Mycobacterium tuberculosis glycolipids. Acta Leprol. 7(Suppl. 1):98-101. Papa, F., M. Luquin, and H. L. David. 1992. DOT-ELISA for detection of phenolic glycolipid PGL-Thl and diacyl-trehalose antigens of Mycobacterium tuberculosis. Res. Microbiol. 143: 327-331. Shinnick, T. M. 1987. The 65-kilodalton antigen of Mycobacterium tuberculosis. J. Bacteriol. 169:1080-1088. Torgal-Garcia, J., H. L. David, and F. Papa. 1988. Preliminary evaluation of a Mycobactenium tuberculosis phenolglycolipid antigen in the serologic diagnosis of tuberculosis. Ann. Inst. Pasteur/Microbiol. 139:289-294. Torgal-Garcia, J., F. Papa, and H. L. David. 1989. The use of PGL-Thl from M. tuberculosis in the serological diagnosis of tuberculous meningitis: a preliminary report. Acta Leprol. 7(Suppl. 1):136-137. Vismara, D., M. F. Mezzopreti, M. S. Gilardini, P. Del Porto, G. Lombardi, E. Piccolella, G. Damiani, R. Rappuoli, and V. Colizzi. 1990. Identification of a 35-kilodalton Mycobactenium tuberculosis protein containing B- and T-cell epitopes. Infect. Immun. 58:245-251. Voller, A., D. E. Bidwell, and A. Bartlette. 1979. The enzymelinked immunosorbent assay (ELISA). Dynatech Laboratories, Alexandria, Va. Young, D. B., J. P. Harnisch, J. Knight, and T. M. Buchanan. 1985. Detection of phenolic glycolipid I in sera from patients with lepromatous leprosy. J. Infect. Dis. 152:1978-1081. Young, D. B., S. R. Khanolkar, L. L. Barg, and T. M. Buchanan. 1984. Generation and characterization of monoclonal antibodies to the phenolic glycolipid of Mycobacterium leprae. Infect. Immun. 43:183-188.
