JOURNAL OF CLINICAL MICROBIOLOGY, Oct. 1990, p. 2200-2204 Vol. 28, No. 10 0095-1137/90/102200-05$02.00/0 Copyright © 1990, American Society for Microbiology Polymerase Chain Reaction for Detection of Mycobacterium tuberculosis ULF SJOBRING,l MICHAEL MECKLENBURG,"2 ASE BENGÂRD ANDERSEN,3 AND HÂKAN MIORNERl* Department of Medical Microbiologyl and Department ofPure and Applied Biochemistry,2 University ofLund, Lund, Sweden, and Mycobacteria Department, Statens Seruminstitut, Copenhagen, Denmark3 Received 6 March 1990/Accepted 25 June 1990 A polymerase chain reaction for the specific detection of mycobacteria belonging to the Mycobacterium tuberculosis complex was developed. Using a single primer pair derived from the nucleotide sequence of protein antigen b of M. tuberculosis, we achieved specific amplification of a 419-base-pair DNA fragment in M. tuberculosis and M. bovis. After DNA was extracted from mycobacteria by using a simple, safe lysis procedure, we detected the 419-base-pair sequence in samples containing few mycobacteria. Preliminary data suggested that this technique could be applied to clinical specimens for early and specific diagnosis of tuberculosis.

Several mycobacterial species are associated with dis- to detect DNAs extracted from various mycobacteria (3, 7). eases in humans. Tuberculosis is caused by Mycobacterium The latter application is based on the sequences of the heat tuberculosis, Mycobacterium africanum, or Mycobacterium shock proteins of M. bovis bacillus Calmette-Guérin (BCG) bovis and is a major health problem, particularly in devel- (21) and M. tuberculosis (18). These sequences are shared by oping countries. The number of cases of tuberculosis world- many other species, and labeled probes are required to wide is estimated to be some 30 million, with an annual identify amplified DNA specific for M. tuberculosis. incidence of 10 million (9). Mycobacteria other than M. The nucleotide sequence of the 38-kilodalton protein an- tuberculosis are widely distributed in nature, and infections tigen b (Pab) of M. tuberculosis was recently published (2). with mycobacteria other than M. tuberculosis are a medi- When serological methods were used, this protein could be cally important problem (4, 22). The emergence of acquired detected only in M. tuberculosis and M. bovis. Thus, hyper- immunodeficiency syndrome has also resulted in an in- immune sera from rabbits immunized with other mycobac- creased incidence of mycobacterial infections (20). terial species did not bind affinity-purified Pab (10, 23). This Bacteriological diagnosis of mycobacterial infections re- species restriction makes Pab a suitable candidate for diag- mains a problem, and there is an urgent need for a rapid and nostic purposes. In this paper we describe a PCR which was reliable diagnostic test. Culturing of organisms has a speci- used for detection of M. tuberculosis and which was based ficity that approaches 100% and permits susceptibility test- on the nucleotide sequence of the Pab gene. ing of the isolates, but the slow growth of most pathogenic mycobacteria (3 to 6 weeks) results in delays in diagnosis. MATERIALS AND METHODS When few mycobacteria are present at the site of infection, Bacterial strains and growth conditions. The bacterial multiple cultures may also be necessary to ensure a positive strains which we used are listed in Table 1. The mycobac- result. Identification of acid-fast organisms in stained smears terial strains were kindly provided by Veronique Lévy- may provide rapid diagnosis, but this method does not allow Frébault, Institut Pasteur, Paris, France, and by J. Stanford. identification at the species level and requires that a rela- M. bovis BCG was obtained from the Institut Pasteur, Paris, tively large number of bacteria (>104/ml) be present in the France. Mycobacteria were grown in Middlebrook 7H9 sample (6). Serological techniques may be useful in some broth (Difco Laboratories, Detroit, Mich.) supplemented clinical situations, but both the sensitivity and the specificity with OADC enrichment or on Middlebrook 7H10 agar (Dif- of these tests are unsatisfying (5). Gas chromatography with co) for 1 to 3 weeks at 37°C. During the first week the flame ionization detection has been shown to be useful for cultures were incubated in an atmosphere containing 5% the identification of cultured mycobacteria, but this method C02, and thereafter the cultures were incubated in air. To is not sensitive enough to be used for clinical specimens (12). obtain samples containing a known number of viable cells, Mass spectrometric detection has been used successfully to M. bovis BCG was cultured in Middlebrook 7H9 medium detect tuberculostearic acid in clinical samples but is an containing 0.05% Tween 80 and harvested at the exponential elaborate technique which requires expensive equipment growth phase. The suspended bacteria were mixed with a (13). The DNA or RNA hybridization tests with labeled vortex mixer for 10 min and then were allowed to sediment specific probes which have been described so far are not for 1 h. The supernatant was decanted and briefly sonicated sensitive enough to be used for clinical specimens without in a water bath, and then serial 10-fold dilutions were made prior culturing (15). in phosphate-buffered saline containing 0.05% Tween 80. The advent of the polymerase chain reaction (PCR) (16, Samples (100 pl) from the dilutions were inoculated onto 17) has opened new possibilities for diagnosis of microbial Middlebrook 7H10 agar, and the preparations were incu- infections. This technique has already been applied to a bated for 3 weeks at 37°C to determine the number of viable number of microorganisms. In the field of mycobacteria the organisms present. PCR has been used to identify Mycobacterium leprae (8) and Isolation of DNA. Mycobacterial DNA was prepared es- sentially as described by Baess (1). Bacteria grown in liquid cultures were lysed with sodium dodecyl sulfate, and the * Corresponding author. proteins were removed with proteinase K. Cell debris, 2200 VOL. 28, 1990 PCR FOR DETECTION OF M. TUBERCULOSIS 2201

TABLE 1. Specificity of the PCR based on the DNA sequence in 50 mM Tris hydrochloride (pH 8.0), and boiled for 5 min. of Pab of M. tuberculosis A 100-pul portion of each cell suspension was mixed with 25 pul of acid-washed glass beads (diameter, <150 ,um; Sigma SpeciesSpeciesor prepn (type culture) Presence~~~~~~~PCRofproductthe 419-bp Chemical Co., St. Louis, Mo.) suspended in 50 mM Tris hydrochloride. The mixture was then subjected to ultrason- Mycobacterium tuberculosis (H37Rv) ...... + M. bovis (ATCC 19210) ...... + ication at 20 kHz in a water bath (Schoeller & Co., Frank- M. bovis BCG ...... + furt, Federal Republic of Germany) for 15 min at 50°C; 10 pul M. kansasii (HB4962). of the supernatant was used as the target DNA in the M. marinum (ATCC 927) . amplification assay. M. scrofulaceum (HB1565). Clinical samples. Seven clinical samples were selected at M. szulgai (NCTC 10831). random from among smear-positive and smear-negative M. gordonae (ATCC 14470). samples. The sputum samples were treated with dithiothre- M. flavescens (NCTC 10271) . M. xenopi (HB1895). itol (Sputolysin; Behring Diagnostics, La Jolla, Calif.), M. avium (NCTC 8551) . whereas the samples from bronchial and gastric lavages were M. intracellulare (ATCC 13950) . not treated. The specimens were centrifuged at 2,300 x g for M. gastri (W471) . 20 min, the supernatants were discarded, and the pellets M. malmoense (ATCC 29571). were suspended in approximately 250 pul of 50 mM Tris M. nonchromogenicum (NCTC 10424). hydrochloride (pH 8.0). A 100-,ul portion of each of the M. terrae (W45) . samples was boiled. DNA was extracted as described above. M. triviale (T-254-3) . Selection of primers. The sequences of oligonucleotides M. paratuberculosis (ATCC 43015). MT1 and MT2, which were used as primers in the PCR, and M. fortuitum (HB1792) . the sequence of oligonucleotide MT3, which was used as a M. chelonae subsp. abscessus . M. phlei (NCTC 10266). probe to identify the amplified DNA in DNA hybridization M. smegmatis (ATCC 14468). experiments, were homologous to sequences in a 419-base- M. vaccae (ATCC 15483). pair (bp) region of the gene encoding Pab of M. tuberculosis M. gilvum (NCTC 10742). (2) (Fig. 1). The oligonucleotides were synthesized by using M. chitae (NCTC 10485). a model 381A DNA synthesizer (Applied Biosystems, Foster M. aurum. City, Calif.) according to the instructions of the manufac- M. thermoresistibile (NCTC 10409). turer. The oligonucleotides were purified by fast-protein M. diernhoferi (ATCC 19340). liquid chromatography on a Mono-Q column (Pharmacia M. rhodesiae (ATCC 27024) . M. gadium. Fine Chemicals, Uppsala, Sweden). PCR. A PCR in which we used heat-stable DNA polymer- Nocardia asteroides. ase from Thermus aquaticus (AmpliTaq; Perkin Elmer Ce- N. brasiliensis. tus, Norwalk, Conn.) was performed by using an Intelligent N. farcinica. heating block (Hybaid Ltd., Teddington, United Kingdom) N. otitidiscaviarum. essentially as described by Saiki et al. (16); 100 pul of a reaction mixture containing 50 mM KCl, 10 mM Tris hydro- Staphylococcus aureus . chloride (pH 8.3), 1.5 mM MgCl2, 100 ,ug of gelatin per ml, Staphylococcus epidermidis. each of the deoxynucleotides (Pharmacia) at a concentration Streptococcus pyogenes. Group G streptococcus. of 300 ,uM, 20 pmol of oligonucleotide MT1, 20 pmol of Human leukocytes. oligonucleotide MT2, and 10 pli of a target DNA solution was overlaid with 100 pul of light mineral oil (Sigma) and dena- a Target DNA was extracted and the PCR was performed as described in tured at 98°C for 10 min. AmpliTaq enzyme (2.5 U) was the legend to Fig. 3. The PCR products were analyzed on 3% agarose gels. added, and the PCR was performed. The thermocycling profile consisted of 34 cycles with denaturation at 92°C for 1 polysaccharides, and the remaining proteins were removed min, annealing at 58°C for 1 min, and polymerization at 72°C by precipitation with cetyltrimethyl-ammonium bromide, for 1 min. which was followed by extraction with chloroform-isoamyl Analysis of DNA. Bacterial chromosomal DNA was ana- alcohol (24:1); then high-molecular-weight DNA was recov- lyzed on 0.8% agarose (SeaKem ME agarose; FMC Bioprod- ered from the resulting aqueous solution by isopropanol ucts, Rockland, Maine) gels run in TAE buffer (40 mM Tris, precipitation. DNAs from other bacterial species were iso- 20 mM acetate, 2 mM EDTA, pH 8.0). The PCR product was lated as described by Maniatis et al. (14). The bacteria were separated on 3% agarose gels (NuSieve GTG agarose; FMC digested with lysozyme, and this was followed by lysis with Bioproducts) in TAE buffer. pBR322 digested with Hinfl sodium dodecyl sulfate in the presence of proteinase K. The (New England BioLabs, Inc., Beverly, Mass.) was run as a DNA was extracted with phenol and chloroform and recov- molecular weight marker. Bands were visualized by staining ered by ethanol precipitation. Human high-molecular-weight with ethidium bromide (1 ,ugIml) and destaining with distilled DNA prepared from human leukocytes was a kind gift from water. Gels were blotted onto nitrocellulose filters (Schlei- L. Truedsson, Department of Medical Microbiology, Uni- cher & Schuell, Dassel, Federal Republic of Germany) as versity of Lund, Lund, Sweden. Plasmid pAA26 is a deriv- described by Southern (19). Dot blot experiments were done ative of pBR322 with a 7.8-kilobase HindIII-SmaI DNA by using a manifold apparatus (Schleicher & Schuell). The fragment containing the entire gene for Pab of M. tubercu- DNA samples were denatured in 0.4 M NaOH at 80°C and losis. The construction of this plasmid has been described neutralized before they were applied to nitrocellulose filters. previously (2). The preparation of target DNA by boiling and The filters were baked for 2 h at 80°C and prehybridized for ultrasonication is described below. 30 min at room temperature in a solution containing 50% Extraction of target DNA from mycobacteria. Cultured formamide, 25% 20x SSC (lx SSC is 0.15 M NaCl plus mycobacteria were harvested by centrifugation, suspended 0.015 M sodium citrate, pH 7.0), 5% NaPO4 (pH 6.6), 5% 2202 SJOBRING ET AL. J. CLIN. MICROBIOL.

NUCLEOTIDE NUMBER

o 250 500 750 1000 1250 1500 1

MT1 MT3 5' V//Y/////Y//Y/Y//Y///A/// -3 3' * - 5'

MT1: 5' ACCACCGAGCGGTTCGCCTGA 3'

MT2: 5' 3'

MT3: 5' CGCTGTTCAACCTGTGGGGTCCGGCCTTTC ,3'

FIG. 1. Oligonucleotides used for amplification and identification of M. tuberculosis-specific DNA. The sequences and positions of oligonucleotides MT1 and MT2, which were used as primers, and oligonucleotide MT3, which was used as a probe, are indicated in a schematic representation of the Pab gene of M. tuberculosis.

100x Denhardt solution (1% Ficoll, 1% polyvinylpyrroli- bovis, and M. bovis BCG were as used as target sources done, 1% bovine serum albumin), and 15% TE (10 mM Tris (Table 1). A weak band at 419 bp was observed with one hydrochloride [pH 8.0], 1 mM EDTA). The filters were then strain of Mycobacterium malmoense, but this band could hybridized in a similar solution, in which the TE was easily be distinguished from the clearly visualized bands replaced by an equal volume of a probe solution containing produced by M. tuberculosis. Four other M. malmoense 32P-labeled oligonucleotide MT3 and 0.3 mg of sheared, strains, including an American Type Culture Collection denatured salmon sperm DNA per ml in TE. After overnight reference strain, were negative. When the annealing temper- hybridization at room temperature, the filters were washed ature of the PCR was lowered to 52°C, weak bands at 250 to four times (10 min each) in lx SSC containing 0.1% sodium 300 bp were observed with a few mycobacterial strains dodecyl sulfate at room temperature. After drying, the filters (Mycobacterium fortuitum, Mycobacterium gilvum, Myco- were exposed to type XAR film (Eastman Kodak Co., bacterium gastri), indicating the presence of structures Rochester, N.Y.) at -70°C. The oligonucleotide was labeled homologous to Pab in mycobacterial strains that do not with [,y-32P]ATP (Amersham International plc, Amersham, belong to the M. tuberculosis complex. United Kingdom) by using T4 polynucleotide kinase (New Sensitivity of the PCR performed with lysed mycobacteria. England BioLabs). The sensitivity of the amplification performed with target DNAs extracted by using the lysis procedure described RESULTS above was estimated by using serial dilutions of a culture Cycling parameters of PCR. Conditions for performing the containing approximately 4 x 106 M. bovis BCG cells per ml. PCR were optimized by using linearized plasmid pAA26, The number of viable organisms was determined by cultur- which contains the entire gene for Pab (2), as the target ing samples of each dilution on Middlebrook 7H10 agar. A DNA. Annealing temperatures of 56 to 58°C with denatur- 100-,ul sample of each dilution was boiled and sonicated ation at 92°C and primer extension at 72°C gave the best together with glass beads as described above; 10 pli of the results. At temperatures above 60°C amplification was less resulting sonic extract was used as the target source in effective, and at temperatures below 50°C considerable PCRs. After amplification 20 ,uI of the mixture was dena- mispriming occurred; 20 pmol of each of the two primers was tured and applied to a nitrocellulose membrane, which was found to be adequate. With MgCl2 concentrations below 0.5 subsequently probed with 32P-labeled oligonucleotide MT3 mM and above 3 mM no amplification was achieved. (Fig. 3A). A 10-,ul portion of the PCR product was also Specific amplification of a 419-bp DNA fragment by the separated on an agarose gel (data not shown). Our results PCR. DNAs were extracted and purified from seven myco- indicated that DNA from less than 10 mycobacterial cells bacterial strains, as well as from two other bacterial species. can be detected by using the combination of the lysis A 1-,ug portion of purified DNA from each of the strains was procedure described above and the PCR. used as a target DNA in the PCR, and oligonucleotides MT1 Detection of the 419-bp sequence in clinical specimens. and MT2 were used as primers. Amplification of the 419-bp Seven samples from different patients with suspected tuber- fragment was achieved only when DNA isolated from M. culosis were collected. After DNA was extracted, 10 pul of tuberculosis or M. bovis BCG was included as the target each sonic extract was subjected to the PCR. Three of the DNA (Fig. 2). No amplification of human chromosomal samples produced sharp bands at 419 bp when 10 pi of the DNA was observed under these conditions (data not shown). PCR product was separated on a 3% agarose gel, whereas Target DNAs were also extracted from 29 mycobacterial the remaining four samples did not produce any visible band strains, 4 Nocardia strains, and 4 strains of gram-positive on agarose gels (data not shown). A 20-pul portion of the PCR cocci by using a simple lysis procedure. A 10-pul portion of product was also analyzed by dot blot hybridization, using each target DNA was subsequently subjected to the PCR. oligonucleotide MT3 as a probe. The samples which were The 419-bp DNA fragment was clearly visualized on an positive as determined by analysis on agarose gels also agarose gel only when sonic extracts of M. tuberculosis, M. hybridized with the oligonucleotide MT3 probe, whereas the VOL.VOL.28,28,19901990PCR FOR DETECTION 0F M. TUBERCULOSIS 2203

a b c d e -1.6 kb A. 2 e

Ba 2 e -510 bp 3 FIG. 3. Sensitivity of the PCR in which mycobacterial cell sonic extracts were used as DNA target sources. Samples (100 pul) of 10-fold dilutions of M. bovis BCG (A) or clinical specimens (B) were boiled for 5 min and subsequently ultrasonicated at 20 kHz for 15 min in the presence of acid-washed glass beads (diameter, <150 ,um). A 10-pI portion of the resulting supernatant was used as a target source in the PCR, in which 20 pmol of oligonucleotide MT1 and 20 pmol of oligonucleotide MT2 were used as primers. The annealing temperature for the PCR was 58°C, and the reactions were otherwise carried out as described in the text. A 20-pil portion of the PCR product was denatured and applied to a nitrocellulose filter, which was subsequently hybridized with 32P-labeled oligonucleotide MT3. (A) M. bovis BCG. Sample la, 10-1 dilution; sample lb, 10-2 a b c d e f g h i k dilution; sample lc, l0-' dilution; sample ld, lO- dilution; sample le, l0-' dilution. (B) Sample 2a, Gastric lavage sample 68; sample FIG. 2. Amplification of a 419-bp gene fragment in M. tubercu- 2b, gastric lavage sample 73; sample 2c, sputum sample 76; sample losis and M. bovis BCG. The PCR was performed with 20 pmol of 2d, sputum sample 81; sample 2e, sputum sample 85; sample 3a, each of the primers (oligonucleotides MT1 and MT2) at an annealing sputum sample 502; sample 3b, bronchial secretion sample 536. temperature of 58°C under the conditions described in the text. A 1-,ug portion of purified chromosomal DNA from each strain was PCR. Further improvement of the lysis technique, possibly used as target DNA. A 10-pul portion of the resulting product was by including a step which concentrates the mycobacterial analyzed by separation on a 3% (wt/vol) agarose gel. Lane a, M. cells in the sample, may increase the sensitivity. tuberculosis; lane b, M. avium; lane c, M. intracellulare; lane d, M. The oligonucleotide primers (oligonucleotides MT1 and bovis BCG; lane e, M. marinum; lane f, M. kansasùi; lane g, M. scrofulaceum; lane h, Staphylococcus aureus; lane i, group G MT2) derived from the Pab DNA sequence were success- streptococcus; lane k, pBR322 digested with Hinfl. fully used to amplify a 419-bp DNA sequence in mycobac- teria belonging to the M. tuberculosis complex, but not in other species. We also constructed an oligonucleotide (oli- samples that did not produce any visible band on agarose gonucleotide MT3) to be used in DNA hybridization exper- gels did not hybridize with this probe (Fig. 3B). The three iments. This probe hybridized to some extent also with PCR-positive samples were positive in acid-fast staining chromosomal DNAs extracted from Mycobacterium avium tests and were identified as M. tuberculosis by culturing. Of and Mycobacterium intracellulare under low-stringency the four PCR-negative samples, three were negative when conditions (below 25°C) and therefore may not be optimal. they were cultured, whereas one sample (sputum sample However, in Southern blot experiments only the 419-bp 502) produced a single colony of M. tuberculosis on Lowe- product amplified from M. tuberculosis and M. bovis was nstein-Jensen substrate. identified by oligonucleotide MT3, and therefore this probe can be used in conjunction with analysis of amplified DNAs on agarose gels. DISCUSSION In extrapulmonary tuberculosis cases culturing often pro- Culturing of mycobacteria will remain the basis for diag- duces negative results or causes an unacceptable delay in nosis of mycobacterial infections. However, a rapid, sensi- diagnosis. This is especially true in tuberculous meningitis tive, specific test for M. tuberculosis would be of great cases, in which early diagnosis is essential for the outcome value. The nucleotide sequence of the Pab gene of M. of the disease (11). The high sensitivity of the PCR raises tuberculosis was recently published (2). The aims of this hope for more rapid diagnosis of these infections. A rapid study were to develop a simple, safe method for extraction test for M. tuberculosis would also allow early differentiation of mycobacterial DNAs present in clinical samples and to between M. tuberculosis and mycobacteria other than M. develop a specific, sensitive PCR by using primers based on tuberculosis, which has important implications for preven- the DNA sequence of Pab. tive measures, as well as for treatment of patients. Lysis of mycobacterial cells is difficult and time consum- In this study we tested a small number of clinical samples, ing. A safe, simple lysis procedure for extraction of DNA is and the results which we obtained indicated that our ap- essential if the PCR is to be used in routine diagnostic tests. proach could be used to detect M. tuberculosis in clinical We found that the combination of boiling and ultrasonication specimens. The sensitivity of the PCR is largely dependent adequately extracted mycobacterial DNA for use in the on the efficiency of the DNA extraction procedure, which 2204 SJOBRING ET AL. J. CLIN. MICROBIOL. could probably be improved by modifications of the lysis J. Am. Med. Assoc. 241:264-268. technique. Studies have been initiated to substantiate this. 12. Larsson, L., and G. Odham. 1986. Diagnosis of mycobacterial infections using gas chromatography and gas chromatography/ mass spectrometry, p. 42-50. In M. Casal (ed.), Mycobacteria ACKNOWLEDGMENTS of clinical interest. Elsevier Science Publishers BV, Amster- The excellent technical assistance of Barbro Olsson and Anna dam. Knutsson is gratefully acknowledged. 13. Larsson, L., G. Odham, G. Westerdahl, and B. Olsson. 1987. This work was supported by grants from the Swedish National Diagnosis of pulmonary tuberculosis by selected ion monitor- Association against Heart and Chest Diseases. ing: improved analysis of tuberculostearate in sputum using negative ion mass spectrometry. J. Clin. Microbiol. 25:893-896. LITERATURE CITED 14. Maniatis, T., E. Fritsch, and J. Sambrook. 1982. Molecular manual. Harbor Laboratory, 1. Baess, I. 1974. Isolation and purification of deoxyribonucleic cloning: a laboratory Cold Spring acid from mycobacteria. Acta Pathol. Microbiol. Scand. Sect. B Cold Spring Harbor, N.Y. 82:780-784. 15. Roberts, M. C., C. McMillan, and M. B. Coyle. 1987. Whole 2. Bengàrd Andersen, Â., and E. B. Hansen. 1989. Structure and chromosomal DNA probes for rapid identification of Mycobac- mapping of antigenic domains of protein antigen b, a 38,000- terium tuberculosis and Mycobacterium avium complex. J. molecular-weight protein of Mycobacterium tuberculosis. In- Clin. Microbiol. 25:1239-1243. fect. Immun. 57:2481-2488. 16. Saiki, R. K., D. H. Gelfand, S. Stoffel, S. J. Scharf, R. Higuchi, 3. Brisson-Noël, A., D. Lecossier, X. Nassif, B. Gicquel, V. Lévy- G. T. Horn, K. B. Mullis, and H. A. Erlich. 1988. Primer- Frébault, and A. J. Hance. 1989. Rapid diagnosis of tuberculosis directed enzymatic amplification of DNA with a thermostable by amplification of mycobacterial DNA in clinical samples. DNA polymerase. Science 239:487-491. Lancet ii:1069-1071. 17. Saiki, R. K., S. Scharf, F. Faloona, K. B. Mullis, G. T. Horn, 4. Chapman, J. S. 1977. The atypical mycobacteria and human H. A. Erlich, and N. Arnheim. 1985. Enzymatic amplification of mycobacterioses. Plenum Publishing Corp., New York. beta-globin genomic sequences and restriction site analysis for 5. Daniel, T. M., and S. M. Debanne. 1987. The serodiagnosis of diagnosis of sickle cell anemia. Science 230:1350-1354. tuberculosis and other mycobacterial diseases by enzyme- 18. Shinnick, T. 1987. The 65-kilodalton antigen of Mycobacterium linked immunosorbent assay. Am. Rev. Respir. Dis. 135:1137- tuberculosis. J. Bacteriol. 169:1080-1088. 1151. 19. Southern, E. M. 1975. Detection of specific sequences among 6. David, H. L. 1976. Bacteriology of the mycobacterioses. U.S. DNA fragments separated by gel electrophoresis. J. Mol. Biol. Government Printing Office, Washington, D.C. 98:503-517. 7. Hance, A. J., B. Grandchamp, V. Lévy-Frébault, D. Lecossier, J. 20. Sunderam, G., R. J. McDonald, T. Maniatis, J. Oleske, R. Rauzier, D. Bocart, and B. Gicquel. 1989. Detection and identi- Kapila, and L. B. Reichman. 1986. Tuberculosis as a manifes- fication of mycobacteria by amplification of mycobacterial tation of the acquired immunodeficiency syndrome (AIDS). J. DNA. Mol. Microbiol. 3:843-849. Am. Med. Assoc. 256:362-366. 8. Hartskeerl, R. A., M. Y. L. De Wit, and P. R. Klatser. 1989. 21. Thole, J. E. R., W. J. Keulen, A. H. J. Kolk, D. G. Groothius, Polymerase chain reaction for the detection of Mycobacterium L. G. Berwald, R. H. Tiesjema, and J. D. A. van Embden. 1987. leprae. J. Gen. Microbiol. 135:2357-2364. Characterization, sequence determination, and immunogenicity 9. Joint International Union Against Tuberculosis and World Health of a 64-kilodalton protein of Mycobacterium bovis BCG ex- Organization. 1982. Tuberculosis control. Tubercle 63:157-169. pressed in Escherichia coli K-12. Infect. Immun. 55:1466-1475. 10. Kadival, G., S. Chaparas, and D. Hussong. 1987. Characteriza- 22. Wolinsky, E. 1979. Non-tuberculous mycobacteria and associ- tion of serologic and cell-mediated reactivity of a 38-kDa antigen ated diseases. Am. Rev. Respir. Dis. 119:107-159. isolated from Mycobacterium tuberculosis. J. Immunol. 139: 23. Young, D., L. Kent, A. Rees, J. Lamb, and J. Ivanyi. 1986. 2447-2451. Immunological activity of a 38-kilodalton protein purified from 11. Kennedy, D. H., and R. J. Fallon. 1979. Tuberculous meningitis. Mycobacterium tuberculosis. Infect. Immun. 54:177-183.