977

Polymerase Chain Reaction Amplification of a Repetitive DNA Sequence Specific for Mycobacterium tuberculosis Kathleen D. Eisenach, M. Donald Cave, Joseph H. Bates, and Jack T. Crawford

From the Medical Research Service, McClellan Memorial I4?terans Hospital, and the Departments of Pathology, Anatomy, Medicine, and Microbiology and Immunology. University of Arkansas for Medical Sciences, Liuie Rock

An absolute diagnosis of tuberculosis continues to depend on culture of Mycobacterium tuberculosis from secretions or tissue from the infected host together with a compatible clinical picture of the disease. Because of the long time required for isolation of M. tuberculosis, the development of a rapid, sensitive, and specific test for detection of the organism in clinicalspecimens has been a long-standing goal. DNA probes for direct detection of mycobacterial sequences in clinical specimens appear promising. Although a probe system is commercially available for identification of several species of mycobacteria in culture (Gen-Probe, San Diego), no direct probe for detecting mycobacteria in clinical specimens is currently available. The primary limitation of direct tests using DNA probes is sensitivity. The recent development of DNA amplification by polymerase chain reaction (PCR) provides an alternative to simple probe assays with a dramatic improvement in sensitivity of detection of specific DNA sequences [I]. This technique is being applied for the detection of a number of infectious agents [2-5]. We have described three cloned sequences of M. tuberculosis that are promising as clinical probes [6]. These probes are specific for the M. tuberculosis complex (M. tuberculosis and the subspecies Mycobacterium bovis) and are of special

Received 12 May 1989; revised 16 November 1989. Supported by funds from General Medical Research of the Department of Veterans Affairs. Reprints and correspondence: Dr. Jack T. Crawford. Medical Research Service-ISt. McClellan Memorial Veterans Hospital, 4300 W. 7th sr.. Little Rock. AR 72205. The Journal of Infectious Diseases 1990;161:971-981 © 1990 by The University of Chicago. All rights reserved. 0022-1899/90/6105-0025$01.00

interest because they are repeated several (10-16) times in the chromosome. Here we report use of a segment of one of these clones as a target for amplification by PCR.

Materials and Methods Mycobacterial strains. The strains used are listed in table I. Clinical isolates of M. tuberculosis and M. kansasii were obtained locally. M. avium strains LR541, 542. 551, 552, 588, and 593 were AIDS-associated isolates obtained from the Centers for Disease Control, Atlanta. All other strains were obtained from the reference laboratories at National Jewish Center for Immunology and Respiratory Medicine, Denver. DNA isolation and cloning. Mycobacterial DNA was isolated as described previously [7]. The concentration of DNA was determined by measuring the optical density at 260 nm. Cloning of segments of M. tuberculosis DNA into bacteriophage Ml3 was described previously [6]. A single fragment of clone M13KE39 was obtained by digestion with endonuclease Sau3A and subcloned into M13mpl8 and MI3mpl9 for sequencing using the dideoxy chain termination technique (8]. Both strands were sequenced. Synthesis ofprimers. The oligomers were synthesized on a DNA synthesizer (Model 381A; Applied Biosystems, Foster City, CA) using l3-eyanoethyl phosphoramidite chemistry. DNA was deblocked using ammonium hydroxide and purified by passage through a Sephadex G-25 column (NAP-25; Pharmacia, Piscataway. NJ) and ethanol precipitation. The sequences of the primers (5' to 3') were CCTGCGAGCGTAGGCGTCGG and CTCGTCCAGCGCCGCTTCGG. peR. Amplification reactions were performed using Thermus aquaticus (Taq) polymerase and reagents according to the manufacturer's instructions (GeneAmp kit; Perkin-Elmer Cetus, Norwalk. CT). Buffer. nucleotides, primers (I p.M each), and enzyme were mixed and then dispensed in 45-p.1 aliquots. DNA was added in a volume of 5 J.tl, and the reaction mixture was overlaid with mineral oil. A control tube containing no DNA was included with each set of reactions.

Downloaded from http://jid.oxfordjournals.org/ at Russian Archive on December 23, 2013

A segment of DNA repeated in the chromosome of Mycobacterium tuberculosis was sequenced and used as a target for amplification using polymerase chain reaction (PeR). The sequences of the primers (5' to 3') wereCCIGCGAGCGTAGGCGTCGG and crcGTCCAGCGCCGCITCGG, and a temperature of 68°C was used for annealing the primers in the reaction. Amplification produced a 123-base-pair fragment with an internal Sail site. The specific PCR product was obtained with input DNA from 11 different strains of M. tuberculosis and Mycobacterium bovis and one strain of Mycobacterium simiae. No product was detected with DNA from 28 strains of the Mycobacterium avium complex, Mycobacterium scrofulaceum, Mycobacterium kansasii, Mycobacteriumjortuitum, Mycobacterium chelonei, and Mycobacteriumgordonae. The PeR product was detected by gel electrophoresis after 30 cycles using 1 fg of input DNA. Amplification of this sequence may provide the basis for an assay to detect M. tuberculosis directly in clinical material.

Eisenach et al.

978

JID 1990;161 (May)

Table 1. Mycobacterial strains used in this study. Organism, strain, serotype

Organism, strain

Organism, strain, serotype

The reaction was performed using an automated thermal cycler (DNA Thermal Cycler; Perkin-Elmer Cetus). The samples were denatured at 94°C for 5 min, and then 25 amplification cycles were performed. The cycle consisted of denaturation at 94°C for 2 min, annealing of primers at 68°C for 2 min, and primer extension at noc for 2 min. The extension time was increased by 5 s with each subsequent cycle. The product was analyzed by electrophoresis on 0.8% agarose gels or 12% acrylamide gels. The DNA was stained with ethidium bromide and photographed on a 302-nm ultraviolet transilluminator. Demonstration of repetitive DNA. DNA from M. tuberculosis strain T2 was digested with endonuclease BamHI and aliquots containing 1.45 p.g and 145 ng were run on an 0.8% agarose gel (20 x 25 em). The lane containing 1.45 ,.,.g of DNA was transferred to a nylon membrane (GeneScreen Plus; NEN-Du Pont, Boston) using the alkaline procedure [9]. The blot was probed by hybridization with the 123-base-pair fragment produced by the PCR reaction. For the hybridization, the PCR product was harvested from an agarose gel, and one strand was labeled by incorporation of [J'P]dCTP by extension of one of the primers using large-fragment DNA polymerase I at a temperature of 37°C. The reaction was cycled three times with addition of enzyme at each step. Hybridization conditions were as described previously [6]. The lane containing 145 ng of DNA was sliced into 0.5- or l-cm fractions. The agarose slices were placed in water (0.5 or 0.8 ml, respectively), melted in a boiling water bath, and then diluted 1000-fold in water. Aliquots (5 ,.,.1) of fractions 3, 5-22, 24, 26, and 28 were then amplified by PCR. The samples were run on a 12% acrylamide gel.

Results We previously described cloning of repetitive DNA sequences from M. tuberculosis. A segment of one repeat was selected as a possible target for amplification by PCR. The sequence of this segment is shown in figure I. Oligonucleo-

60

80

100

CCCAGGTCGACACATAGGTGAGGTCTGCTACCCACAGCCGGTTAGGTGCT Sail

120

GGTGGTCCGAAGCGGCGCTGGACGAGATC

Figure 1. Sequence of the repetitive DNA segment from Mycobacterium tuberculosis used for polymerase chain reaction. The 20 base regions used for priming the reaction and the internal Sail site are underlined.

tide primers corresponding to the indicated sequences were synthesized for use in PCR. Amplification of the sequence bounded by the primers should produce a I23-base-pair fragment with an internal site for endonuclease SaLI, and preliminary assays demonstrated the appropriate product. In initial assays an annealing temperature of 50°C was used. This produced the expected product with M. tuberculosis DNA but also gave products of various sizes with M. kansasii and M. avium DNA. Increasing the annealing temperature to 68°C eliminated the products produced with these latter two species, and this temperature was used in all subsequent reactions. Although previous hybridization analysis demonstrated that the sequence was repeated in the chromosome of M. tuberculosis, we did not know if each copy of the sequence was perfectly conserved and could be amplified using the primers we had synthesized. To determine this, M. tuberculosis DNA was digested with endonuclease BamHI, which does not cleave within the PCR target, and run on an agarose gel to separate the individual copies (figure 2). The gel was then sliced into fractions and the DNA in the fractions was amplified. The products of the reaction were compared with the distribution of the repeated sequences in the BamHI fragments as determined by hybridization. The hybridization results indicated that there were at least 12 copies of the repeat, and the intensity of the hybridization suggested that some of the bands represented multiple copies. As expected, PCR product was detected in fractions 5-22 due to the presence of small amounts of target DNA throughout the gel. However, the highest amount was obtained with those fractions corresponding to the various BamHI fragments, indicating that most or all of the copies were being amplified. To determine the specificity of the reaction, DNA from 39 additional strains of mycobacteria were assayed (figure 3). Relatively high amounts of input DNA were used to ensure that amplification would be detected if it was occurring. DNA from all 10 strains of M. tuberculosis and M. bovis was amplified to give the expected 123-base-pair product. In addition, DNA from the M. simiae strain tested was amplified. No amplified product was detected with DNA from 4 strains of M. kansasii, 21 strains of the M. avium complex and M.

Downloaded from http://jid.oxfordjournals.org/ at Russian Archive on December 23, 2013

M. kansasii M. avium complex M. tuberculosis LRI47, I TMCI204 Clinical T2 LR588, 1 TMCI217 H37Rv (TMCI02) LR593, I Clinical K4 H37Ra Goldman LRI13,4 Clinical K5 Clinical TI LR541, 4 Other Clinical T3 LR542, 4 M. fortuitum TMCI530 Clinical T4 LRI50,8 Clinical T5 M. simiae TMCI226 M. gordonae TMC 1324 LR551, 8 Clinical T6 M. chelonei TMCI543 LR552,8 M. bovis TMC715,2 TMC401 LR107,3 TMC4IO LRI63,6 BCG Glaxo (TMCI024) TMCI479,9 M. scrofulaceum LR158, 10 TMC1315,41 LRI05, II TMC1309,42 TMCI466, 13 LR121, 43 LRI27, 19 LR141, 26

40

20

GATCCTGCGAGCGTAGGCGTCGGTGACAAAGGCCACGTAGGCGAACCCTG

peR of Repetitive M. tuberculosis DNA

lID 1990;161 (May)

Figure 2 Amplification ofrepetitive DNA. DNA from Mycobacterium tuberculosis strain T2 was digested with endonuclease BamHI, and two aliquots (lane a, 1.45 Jkg; lane b, 145ng) were electrophoresed on an 0.8% agarose gel. Molecular weight standards are indicated at top. Lane a was blotted and probed with labeledPCR product (autoradiograph shown in lane c). Lane b was sliced into 28 fractions as indicated, slices were melted, and DNA was diluted and amplified (product corresponding to each fraction shown in lane d).

1.1 I

2.3

24 6.8 I I

I

0.6 I

a ~_'I' .

b c I

1

d

3

," I I I I I I I I I I I I I I I I I I I

A

M+a bed e f 9 h i j kim no p q r 5 t u

B

M+a bed efg hijabedefgh ijk-M

Figure 3. Amplification of DNA from various strains of mycobacteria by 25 cycles of PCR. Input DNA was 625 pg per reaction; 5 Jkl of each reaction mixture was run on an agarose gel. M = 123-basepair ladder marker; + = positive control, strain M. tuberculosis T2; - = negative control, no DNA. A, lanes a-j, M. tuberculosis and M. bovis strains in the order listed in table I starting with H37Rv; k-n, M. kansasii strains as listed in table I; 0, M. fortuitum; p, M. simiae; q, M. gordonae; r, M. chelonei. B, lanes a-u (all negative), M. avium complex and M. scrofulaceum strains in the order listed in table I. C, top, agarose gel; bottom, auto radiograph of blot probed with labeled PCR product. Lanes a-k, right, reactions with the in-

5

7

9

11 13

15 17 19

21

I

23

I

24

I

I

26

I

I

28

II scrofulaceum, and single strains of M. fortuitum, M. chelonei, and M. gordonae. The gels shown in figure 3A and 3B were blotted and probed with labeled I23-base-pair peR product to increase the sensitivity of detection. No hybridization was detected in any of the samples judged to be negative by ethidium bromide staining (not shown). The lack of amplification with DNA from the nontuberculous mycobacteria could result from inhibition of the reaction by contaminants in the DNA samples. To rule out this possibility, mixtures of M. tuberculosis DNA and DNA from 10 of the nontuberculous strains were amplified (figure 3C). The amplified product was obtained in each case. Again, no product was obtained with the nontuberculous mycobacterial DNAs alone, and the lack of amplification was confirmed by hybridization of a blot of the gel. The product obtained in the reactions with the 10 M. tuberculosis complex strains and M. simiae (figure 3A)was digested with endonuclease Sail and analyzed on an acrylamide gel (figure 4). Apparently identical Sail fragments were obtained in all cases. The sensitivity of detection was assessed by amplifying samples containing seriallO-fold dilutions of M. tuberculosis DNA (figure 5). The product produced from 100 fg of input DNA was detected after 25 cycles. After 30 cycles, product was detected with 1 fg of input DNA. This is equivalent to about one copy of the M. tuberculosis chromosome ( rv3000 kilobases). Addition of 10 ng of M. avium DNA or human DNA to the above reactions had no effect. No product was detected after 30 cycles with the M. avium DNA or human DNA alone or in the control with no input DNA (data not shown).

dividual DNAs; lanes a-j, left, reactions containing DNA from nontuberculous mycobacteria plus M. tuberculosis TMC102 DNA: a, M. kansasii TMC1204; b, TMCI2l7; c, M. avium LR147; d, LR107, e, LR163; f, TMC1479; g, M. scrofulaceum TMC1315; h, M. fortuitum; i, M. gordonae; j, M. chelonei; k, M. tuberculosis TMC102.

Downloaded from http://jid.oxfordjournals.org/ at Russian Archive on December 23, 2013

I

M+ abedefghi jklmnopqr

c

979

Eisenach et al.

980

Cab c d e f 9 h

j

k

Discussion Direct detection of infectious agents in clinical material using DNA probes provides an exciting possibility for rapid diagnosis of disease, especially diseases caused by agents such as mycobacteria that require long periods of time for culture [10]. However,direct specimen probe tests have suffered from low sensitivity, and at present there is no direct specimen test available for M. tuberculosis. Pao et al. [II] described detection of M. tuberculosis in various clinical samples using a cloned segment of M. tuberculosis DNA as probe. Their procedure involvedpurification of the DNA from the clinical sample, but the details of the technique were not provided. Amplification of target DNA using PCR provides a means for dramatically increasing the sensitivity of detection. The procedure and the variables that affect sensitivity and specificity have been reviewed recently [12]. Our results indicate that the repetitive sequence used in this study is well suited for this purpose. The sequence is fairly short and contains an internal endonuclease site that allows confirmation of the product by digestion with Sail. The primer segments have a 75 % G + C content with the A and T residues

well spaced, which allows the use of a high primer annealing temperature. This contributes to the specificity of the reaction and may also increase the sensitivity [12]. In recent work we have shortened the cycle by using a single incubation at 68°C for 2 min for the annealing and extension reactions (data not shown). The sequence appears to be specific for members of the M. tuberculosis complex except that it may be present in M. simiae. Additional testing will be required to confirm its presence in this species. Even if found it will present little problem since M. simiae is rarely isolated. The M. avium complex is quite heterogeneous; thus we tested a number of representative strains including strains of serotype 1,4, and 8 most commonly isolated from patients with AIDS. The results shown in figure 2 demonstrate that the target sequence was repeated in the M. tuberculosis chromosome and the individual copies could be amplified using the same primers. The fact that there are multiple copies of the target sequence should increase the sensitivity of detection by a factor of 10-16 compared with any target that occurs only once in the chromosome. More important, we believe the repeated nature of the target will preclude its loss from any strain of M. tuberculosis since either a deletion or a point mutation would affect only one copy of the sequence, leaving the others available for amplification. The sequence shown in figure 1 is part of a larger repeated segment. We have not completed analysis of this segment, but we anticipate that it is an insertion sequence. Hybridization of the segment with BamHI digests of DNA from the various strains of the M. tuberculosis complex showed considerable heterogeneity in the band patterns. The number of copies of the repeat varies, with 2 or 3 copies present in M. bovis and M. bovis BCG and 10-16 copies present in the M. tuberculosis strains (data not shown). The number of copies is similar in strain H37Rv (isolated in 1905) and in recent clinical isolates, indicating that it is quite stable. The target sequence appears to be specific, and we can detect the PCR product from purified DNA equivalent to about one cell if the reaction is extended to 30 cycles. These results were obtained using the PCR kit as supplied by the manufac-

abc d e abc d e abc d e abc d e

25

+ avium

+ human

30

Figure 5. Sensitivity of detection of input DNA by PCR. Tenfold serial dilutions of M. tuberculosis strain T2 DNA were amplified; 25 ill of each reaction was run on a 12% acrylamide gel. a, 10 pg; b, I pg; c, 100 fg; d, 10 fg; e, 1 fg. As indicated, one set of samples was amplified for 25 cycles and another for 30 cycles. The other sets contained 10 ng of M. avium LRI13 DNA or 10 ng of human skin fibroblast DNA in addition to M. tuberculosis DNA and were amplified for 25 cycles.

Downloaded from http://jid.oxfordjournals.org/ at Russian Archive on December 23, 2013

Figure 4. Analysis of PCR product. DNA from peR reactions shown in figure 3 was digested with endonuclease Sail and electrophoresed on a 12% acrylamide gel. Lane C, undigested PCR product from control strain M. tuberculosis T2; lanes a-j, M. tuberculosis complex strains as in figure 3; lane k, M. simiae.

1ID 1990;161 (May)

JID 1990;161 (May)

PCR of Repetitive M. tuberculosis DNA

Acknowledgment We thank Donald Cunningham and Jane Voyles for technical assistance.

References I. Mullis KB, Faloona FA. Specific synthesis of DNA in vitro via a polymerase catalyzed chain reaction. Methods EnzymoI1987;155:335-350

2. au C, Kwok S, Mitchell SW, Mack DH. Sninsky 11,Krebs JW, Feorino P, Warfield 0, Schochetman G. DNA amplification for direct detection of HIV-I in ON A of peripheral blood mononuclear cells. Science 1988;239:295-297 3. Demmler GJ, Buffone GJ, Schimbor CM, May RA. Detection of cytomegalovirus in urine from newborns by using polymerase chain reaction DNA amplification. J Infect Dis 1988;158:1177-1184 4. Shibata 0, Martin WJ, Appleman MD, Causey OM, Leedom JM, Amheim N. Detection of cytomegalovirus DNA in peripheral blood of patients infected with human immunodeficiency virus. J Infect Dis 1988;158:1185-1192 5. Olive OM. Detection of enterotoxigenic Escherichia coli after polymerase chain reaction amplification with a thermostable DNA polymerase. J Clin Microbiol 1989;27:261-265 6. Eisenach KD, Crawford JT, Bates JH. Repetitive DNA sequences as probes for Mycobacterium tuberculosis. J Clin Microbiol 1988; 26:2240-2245 7. Eisenach KD, Crawford JT, Bates JH. Genetic relatedness among strains of the Mycobacterium tuberculosis complex: analysis of restriction fragment heterogeneity using cloned DNA probes. Am Rev Respir Dis 1986;133:1065-1068 8. Sanger }