JOURNAL OF CLINICAL MICROBIOLOGY, Dec. 1990, p. 2668-2673 Vol. 28, No. 120095-1137/90/122668-06$02.00/0Copyright C 1990, American Society for Microbiology

Characterization of a Mycobacterium tuberculosis InsertionSequence, IS6110, and Its Application in Diagnosis

DOMINIQUE THIERRY,' ANNE BRISSON-NOËL,2 VERONIQUE VINCENT-LÉVY-FRÉBAULT,SIMON NGUYEN,2 JEAN-LUC GUESDON,' AND BRIGITTE GICQUEL4*

Laboratoire des Sondes Froides,' Unité de la Tuberculose et des Mycobactéries,3 and Unité de Génie Microbiologique,4Centre National de la Recherche Scientifique Unité de Recherche Associée No. 1300, Institut Pasteur, 28 rue du Docteur

Roux, Paris 75015, and Diagnostics Pasteur, Marnes-la-Coquette,2 France

Received 11 June 1990/Accepted 14 September 1990

An insertion sequence-like element, IS6110, was isolated from a Mycobacterium tuberculosis cosmid libraryas a repetitive sequence. IS6110 shows similarities with elements of the IS3 family. This insertion sequence wasfound to be specific to mycobacteria belonging to the M. tuberculosis complex. For detection and identificationofM. tuberculosis bacilli in uncultured specimens, oligonucleotides derived from the IS6110 sequence were usedas primers and probes in polymerase chain reaction studies. The results obtained were consistent with resultsof classical identification procedures, bacteriological data, and clinical criteria.

At present, diagnosis of tuberculosis and related diseasesis dependent on isolation of the pathogenic bacteria and theirsubsequent identification by biochemical testing. Such pro-cedures usually require 4 to 8 weeks. More rapid techniquessuch as the BACTEC detection system (18), which is basedon measurement of the radioactive C02 released by metab-olism of radioactive palmitate added to cultures, and theGen-Probe identification system (8), which is based onhybridization of 1251I-labeled DNA fragments correspondingto 16S RNA sequences, have been described previously.Although these techniques reduce the identification time,they still require at least a 1-week culturing period.The technique of in vitro amplification of specific DNA

sequences is a sensitive method for the detection of virusesand bacteria in pathological samples (10, 16, 20). We haverecently adapted this technique to the detection and identi-fication of mycobacteria in uncultured clinical specimens (3,12) by making use of polymorphism in the gene coding forthe 65-kDa heat shock protein in bacteria.

In several cases it has been shown that repetitive DNAsequences display species specificity. For example, theinsertion sequence IS900 was shown to be specific to Myco-bacterium paratuberculosis (9). In this report, we describethe identification of a repetitive sequence from Mycobacte-rium tuberculosis. This element, IS6110, possesses similar-ities to insertion sequences of the IS3 family (15, 19, 22).IS6110 is 1,361 bp long and contains 28-bp, imperfect in-verted repeats at its extremities with three mismatches and3-bp direct repeats that probably result from repetition of thetarget sequence. Among the various mycobacterial speciesexamined, IS6110 was detected only in species belonging tothe M. tuberculosis complex. Oligonucleotides derived fromthis sequence were used to detect M. tuberculosis in clinicalspecimens following in vitro DNA amplification.

MATERIALS AND METHODSConstruction of an M. tuberculosis cosmid library. M.

tuberculosis H37rv genomic DNA was partially digested for1 h at 37°C with 0.03 U of SalI per ,ug of total DNA. DNAfragments were fractionated by electrophoresis on a 0.6%

* Corresponding author.

agarose gel in TAE (0.04 M Tris acetate, 0.001 M EDTA).The 30- to 40-kb fragments were electroeluted in TBE (0.089M Tris borate, 0.089 M boric acid, 0.002 M EDTA) indialysis membranes (4). Fragments (1.5 ,ug) were ligated with100 ng of the pHC79 cosmid (13), digested with SalI, anddephosphorylated with alkaline phosphatase. Reaction wascarried out at 14°C for 16 to 18 h with 2.5 U of T4 ligase in0.06 M Tris hydrochloride (pH 7.5) buffer containing 5 mMMgCl2, 5 mM dithiothreitol, and 1 mM ATP. Recombinantcosmids were packaged (14) and used to infect Escherichiacoli HB101. Ampicillin-resistant, tetracycline-susceptiblecolonies were picked and grown in LB broth supplementedwith ampicillin, and cosmid DNAs were extracted (2).

Library screening and isolation of a fragment hybridizing toM. tuberculosis DNA. Cosmid DNA samples were digestedwith restriction enzyme SalI, electrophoresed through a0.6% agarose gel, and transferred onto Hybond N filters(Amersham) by the Southern blot method (17). Filters wereprobed with M. tuberculosis and/or Mycobacterium bovisBCG DNA labeled with [a-32P]dCTP by using the mul-tiprime random-labeling system (7) (Amersham). Hybridiza-tions were performed at 68°C for 16 to 18 h in 6x SSC (lxSSC is 0.15 M NaCl plus 0.015 M sodium citrate)-10%dextran sulfate-Sx Denhardt-10 mM EDTA-0.5% sodiumdodecyl sulfate (SDS)-100 ,ug of salmon sperm denaturedDNA per ml-106 cpm of probe.

Filters were washed twice with 2x SSC at 65°C for 10 min,once with 2x SSC-0.1% SDS at 65°C for 30 min, and oncewith O.lx SSC at 65°C for 10 min. Filters were briefly airdried and exposed to Kodak XAR5 film with intensifyingscreen at -80°C.A SalI fragment which strongly hybridized with M. tuber-

culosis DNA was eluted from the agarose gel, purified, andsequenced by the chain termination method of Sanger et al.(20a) by using Taq polymerase and 7-deaza-dGTP instead ofdGTP, to prevent band compression in the GC-rich regions(TaqTrack sequencing system from Promega Corp.). Tocomplete the IS6110 sequence, a larger fragment (3.5 kb)was cloned after HindIII digestion of the original cosmid,I21.

Preparation of an IS6100 internal fragment as probe. Fromthe 3.5-kb fragment described above, a 902-bp HindIII-BamHI fragment was subcloned into pUC18 (26); the result-

2668

on May 21, 2018 by guest

http://jcm.asm

.org/D

ownloaded from

CHARACTERIZATION OF M. TUBERCULOSIS IS6110 2669

ing plasmid was named pMT02. The 902-bp fragment waslabeled with [c_-32P]dCTP by using the multiprime random-labeling system (7) (Amersham) and was used as a probe inSouthern hybridization analysis.Southern hybridization analysis of genomic DNAs. Total

DNA (1.2 ,ug) from species of the M. tuberculosis complex(M. tuberculosis, M. bovis BCG, M. bovis, and M. microti)and DNAs from other mycobacteria (M. avium, M. paratu-berculosis, M. intracellulare, and M. scrofulaceum) (seelegend to Fig. 4) were digested with 70 U of the restrictionenzymes PstI, EcoRI, BamHI, and SalI in appropriatebuffers for 16 to 18 h at 37°C. DNA fragments were electro-phoresed through a 0.6% agarose gel. Gels were incubatedfor 15 min in 0.25 M HCI, twice for 15 min each time in 1.5M NaCl-0.5M NaOH, and twice for 15 min each time in 1 MCH3COONH4. DNAs were then transferred onto a HybondN filter (Amersham) and hybridized with the denatured32P-radiolabeled 902-bp fragment.PCR experiments. (i) Synthetic oligonucleotides. Oligonu-

cleotides were synthesized on an Applied Biosystems syn-thesizer by the automated phosphoramidite coupling method(1). Oligonucleotides used as primers and probes for poly-merase chain reaction (PCR) experiments are described inFrench patent application no. 90.02676 (A. Brisson-Noël).Oligonucleotides used in hybridization assays were radiola-beled at their 5' ends with T4 polynucleotide kinase (Boehr-inger) and [32P]ATP (Amersham). Labeled oligonucleotideswere separated from unincorporated [32P]ATP by chroma-tography on Sephadex G-50.

(ii) DNA preparation. For amplification experiments, my-cobacteria from 1-ml cultures were pelleted, resuspended in200 ,uI of lysis buffer (0.1 M NaOH, 2M NaCl, 0.5% SDS),and incubated at 95°C for 15 min. DNAs from the resultinglysates were extracted twice with phenol-chloroform, pre-cipitated with ethanol, and resuspended in 100 ,ul of sterilewater.DNA was extracted from 0.2 to 1 ml of crude or sodium

hydroxide-decontaminated clinical specimens by the sameprocedure described above. Ten-microliter aliquots wereused for amplification.

(iii) DNA amplification. Amplification reactions were per-formed in a total volume of 100 ,ul in 10 mM Tris hydrochlo-ride (pH 8.3)-1.5 mM MgCl2-50 mM KCl-12.5 pmol of eachprimer-125 ,uM (each) deoxynucleoside triphosphates(dATP, dGTP, dTTP, and dCTP)-2 U of Thermus aquaticusDNA polymerase (Perkin Elmer Corp., Norwalk, Conn.).The amplification mixture was overlaid with 50 ,ul of mineralou and subjected to 40 cycles of amplification as follows.Samples were incubated for 2 min at 94°C to denature theDNA, for 2 min at 60°C to anneal the primers, and for 2 minat 72°C to extend the annealed primers. Thermal cycling wasperformed in a programmable heat block (Perkin Elmer).Samples were then electrophoresed through a 1.5% agarosegel; amplification products were visualized by ethidiumbromide staining.

(iv) Hybridization with labeled oligonucleotides. AmplifiedDNAs were resolved on agarose gels and transferred ontoHybond N nylon filters (Amersham) by Southern blotting(17). After a 15-min prehybridization in Rapid HybridizationBuffer (Amersham) at 65°C, membranes were hybridizedwith 106 cpm of 32P-labeled probe in the same buffer for atleast 2 h at 65°C. Filters were then washed twice in 2xSSC-0.1% SDS at 200C for 10 min and once in lx SSC-0.1%SDS at 650C for 15 min. Autoradiography was performed for18 h at -80°C with an intensifying screen.

A B

23- f9.6- &6.4-2 -

0.6-

FIG. 1. Southern hybridization analysis of cosmid 121. 121 DNA(15 ,ul, corresponding to about 3 ,ug) was digested with 10 U ofrestriction enzyme SalI. After agarose electrophoresis, DNA wastransferred to filters by Southern blotting and hybridized withradiolabeled probes: M. bovis BCG total DNA (A) and M. tubercu-losis total DNA (B). The arrow indicates the 1-kb fragment ofinterest.

RESULTS AND DISCUSSIONConstruction of an M. tuberculosis cosmid library and

isolation of IS6110. With the aim of detecting M. tuberculo-sis-specific repeated genomic DNA sequences, we con-structed a cosmid library. About 2 x 103 recombinantcosmids were obtained. Since in vitro packaging ensuresselective pressure for insertion of large fragments (20- to40-kb fragments), and since M. tuberculosis is believed tocontain about 4 x 106 bp, this cosmid library was expectedto be representative of the complete M. tuberculosis ge-nome.A set of 150 recombinant clones was screened by dot blot

analysis (4) by using total M. tuberculosis DNA as a probe.Cosmids giving strong hybridization signals were furtheranalyzed by Southern blotting by using M. tuberculosis, M.bovis BCG, or M. avium total DNA as probes. Afterrestriction enzyme digestion, several clones were found tocontain fragments that hybridized strongly with M. tubercu-losis DNA. Typical results for clone 121 are shown in Fig. 1.Cosmid I21 contains a 1-kb fragment that gives a strongpositive signal with M. tuberculosis DNA after a 24-hexposure and a weak signal with M. bovis BCG DNA after a48-h exposure (see arrow in Fig. 1). This fragment waspurified and cloned in phages M13mpl8 and M13mpl9. Thenucleotide sequences of this fragment and its adjacent re-gions were determined and shown to be similar to insertionsequences of the IS3 family. The complete nucleotide se-quence of this insertion sequence-like element IS6110 hasbeen described previously (21).

Specificity of IS6110. Total DNA from mycobacteria of theM. tuberculosis complex and of closely related mycobacteriawas digested with the restriction enzyme BamHI and ana-lyzed in Southern blots by using the HindIII-BamHI 902-bpfragment of PMT02 as a probe (see Materials and Methods).Results are shown in Fig. 2. Significant hybridization wasobserved only in mycobacteria belonging to the M. tubercu-losis complex, i.e., M. tuberculosis, M. bovis, and M.microti. No hybridization was observed with M. avium, M.paratuberculosis, or M. intracellulare. The weak signalobserved with M. scrofulaceum DNA may have corre-sponded to the presence of sequences sharing similaritieswith IS6110. However, with IS6110-derived primers, noamplification was detected with M. scrofulaceum DNA (seeFig. 4), which suggests that M. scrofulaceum does not

VOL. 28, 1990

on May 21, 2018 by guest

http://jcm.asm

.org/D

ownloaded from

2670 THIERRY ET AL.

1 2 3 4 5 6 7 8

2

1 2 34 5 6 7 8

23-

9.6- M-«6.4- 8

4.3- s

2.3-2.0- le

1.3-

0.9-

0.6-

FIG. 2. Southern hybridization analysis of total mycobacterialDNAs probed with an IS6110 internal fragment, the 902-bp fragmentfrom pMT02, as the probe. A total of 1.2 ,ug of genomic DNA fromM. tuberculosis (lane 1), M. bovis BCG (lane 2), M. bovis (lane 3),M. microti (lane 4), M. paratuberculosis (lane 5), M. intracellulare(lane 6), M. scrofulaceum (lane 7), and M. avium (lane 8) wasdigested with 70 U of BamHI. After agarose electrophoresis, DNAwas transferred onto filters by Southern blotting and hybridized withthe radiolabeled 902-bp fragment from pMT02.

contain IS6110 sequences. Similar results have already beendescribed for IS900, an insertion sequence-like element fromM. paratuberculosis. Cross-hybridization was observed be-tween IS900 and chromosomal fragments of several M.avium strains. When IS900-derived oligonucleotides wereused as primers for PCR, no amplification was obtained withDNA extracted from the M. avium strains that cross-hybrid-ized with IS900 DNA (24).With IS6110, different hybridization patterns were ob-

served for M. tuberculosis, M. bovis, and M. microti,suggesting differences in copy number and genomic locationof the element. We estimated that M. tuberculosis and M.microti harbor 10 to 20 copies of IS6110, while M. bovis andM. bovis BCG (Pasteur strain) appear to contain singlecopies. Southern blot experiments with M. bovis DNAcleaved with different restriction enzymes (PstI, EcoRI,BamHI, or SalI) are in agreement with the presence of asingle IS6110 sequence in this species (Fig. 3). Preliminaryexperiments with pulsed-field electrophoresis followed bySouthern blotting showed that IS6110 sequences are scat-tered throughout the M. tuberculosis chromosome (data notshown). A search for the presence of IS6110 in M. tubercu-losis strains isolated from various patients revealed a dif-ferent arrangement of these elements in the genomes of thesestrains (6). These findings suggest that IS6110 could be used

FIG. 3. Restriction fragment length polymorphism analysis ofmycobacterial genomic DNAs with an IS6110 internal fragment usedas the probe. Genomic DNAs from M. tuberculosis (lanes 1, 3, 5,and 7) and M. bovis BCG (lanes 2, 4, 6, and 8) were digested withPstI (lanes 1 and 2), EcoRI (lanes 3 and 4), BamHI (lanes 5 and 6),and SalI (lanes 7 and 8). After agarose electrophoresis, DNAs weretransferred onto filters by Southern blotting and hybridized with theradiolabeled 902-bp fragment from pMT02.

for epidemiologic studies of M. tuberculosis. In addition, ananalysis of restriction fragment length polymorphism withIS6110 used as a probe could differentiate M. bovis fromother members of the M. tuberculosis complex.

Specificity of IS6110-derived primers and probes in PCRexperiments. In order to detect mycobacteria of the M.tuberculosis complex directly in uncultured specimens, oli-gonucleotides derived from IS6110 were used as primers andprobes for in vitro amplification. Two 20-mer primers,ISTB2 and ISTB7, were selected for amplification experi-ments. A 30-mer designated IS-2 was chosen as a probe forthe detection of the 325-bp fragment amplified with ISTB2and ISTB7.The amplification specificity was tested on 30 mycobacte-

rial strains, including 19 reference mycobacterial species and11 M. tuberculosis strains isolated from patients. One nano-gram of purified total DNA was amplified simultaneouslywith primers ISTB2 and ISTB7, corresponding to IS6110,and TB1 and TB2, corresponding to the 65-kDa antigen (11).TB1 and TB2 amplified a 383-bp fragment in all mycobacte-rial species tested. Amplification of both 325- and 383-bpfragments was observed for all species belonging to the M.tuberculosis complex, i.e., M. tuberculosis, M. africanum,M. bovis, M. bovis BCG, and M. microti reference strains(Fig. 4), and for the 11 clinical M. tuberculosis strains (datanot shown). In contrast, DNAs from all other mycobacterialspecies were amplified with TB1 and TB2 but not withISTB2 and ISTB7 (Fig. 4). These results were confirmed byhybridization of the 325-bp amplified fragment with the IS-2

Kb

21 -

5-4.2-3.6-

1.9-1.3_-

0.9-

Û.3-

-

-il...,:F

te

lll

J. CLIN. MICROBIOL.

on May 21, 2018 by guest

http://jcm.asm

.org/D

ownloaded from

CHARACTERIZATION OF M. TUBERCULOSIS IS6110 2671

rUt, i Il ;2 1

K~8' Un.)

VS-.' bp l1

i 2 4 tj I é8 I

6

06mwl I 12 I3 ]14 1 5 16h i 1 8 1 ( 20 2? 1 Il I2 I 1s 4 15 16( i lU l 2is1, :

FIG. 4. Specificity of DNA amplification. Mycobacterial DNA (1 ng) was amplified simultaneously with ISTB2 and ISTB7 (325 bp) andTB1 and TB2 (383 bp) primers. (A) Ethidium bromide-stained agarose gel containing a 1/lOth volume of amplification reactions; (B) DNAsof panel A transferred to a nylon membrane and hybridized with 32P-labeled IS-2 probe. Lanes 1, M. asiaticum ATCC 25276 (American TypeCulture Collection); lanes 2, M. avium ATCC 25291; lanes 3, M. chelonae ATCC 19977; lanes 4, M. flavescens ATCC 14474; lanes 5, M.gordonae ATCC 12478; lanes 6, M. kansasii 890493 (clinical isolates, references of the Centre National de Reference des Mycobactéries,Paris, France); lanes 7, M. malmoense ATCC 29571; lanes 8, M. marinum ATCC 927; lanes 9, M. tuberculosis ATCC 27294 H37rV; lanes10, control without DNA; lanes 11, M. paratuberculosis 1120006; lanes 12, M. scrofulaceum 0220031; lanes 13, M. simiae ATCC 25275; lanes14, M. szulgai NCTC 10831 (National Collection of Type Cultures); lanes 15, M. terrae ATCC 15755; lanes 16, M. xenopi NCTC 10042; lanes17, M. bovis ATCC 19210; lanes 18, M. bovis BCG (Institut Pasteur); lanes 19, M. africanum ATCC 25420; lanes 20, M. microti NCTC 8710;lanes 21, 4X174 DNA hydrolyzed by HaeIII.

probe. The amplification specificity was further verified byusing DNA from other bacterial species, i.e., Staphylococ-cus aureus, E. coli, Micromonospora chalcea, Nocardiaasteroides, and several streptomycetes (S. lividans, S. hy-groscopicus, S. viridochromogenes, S. antibioticus, and S.fradiae). None of these species was amplified with theISTB2 and ISTB7 set of primers (data not shown).

In order to test the sensitivity of in vitro amplification ofmycobacterial DNA with ISTB2 and ISTB7 and TB1 andTB2 primers, different concentrations of M. tuberculosisDNA (10 ng to 1 fg) were amplified. The results presented inFig. 5 show that the ISTB2 and ISTB7 primers enabledetection of less than 10 fg of DNA. This DNA amountcorresponds to about three genome equivalents (one genomebeing approximately 2.5 x 109 Da mol-'). For the lowestDNA amounts, the signals obtained with ISTB2 and ISTB7were stronger than those obtained with TB1 and TB2. Infact, amplification with ISTB2 and ISTB7 primers alloweddetection ofDNA from mycobacteria at a 10-fold-lower levelthan amplification with TB1 and TB2 did. This result is inagreement with the presence of 10 to 20 copies of IS6110 inthe M. tuberculosis genome. The gene coding for the 65-kDaprotein was believed to be present as a single copy, whilemore recent studies with E. coli strains have suggested thepresence of another gene similar to groEL encoding the65-kDa heat shock protein (11). An IS6110-derived PCRappears to be more specific, since both amplification andhybridization are specific for bacilli of the M. tuberculosiscomplex, whereas the 65-kDa-based PCR uses nonspecific

A

383 bp x

325 bp r

B

FIG. 5. Sensitivity of M. tuberculosis detection by PCR. Dilu-tions of M. tuberculosis DNA were coamplified with ISTB2 andISTB7 and TB1 and TB2. (A) A 1.5% Agarose gel stained withethidium bromide; (B) hybridization with 32P-labeled IS-2 probe.Lanes 1 to 8, 10 ng, 1 ng, 100 pg, 10 pg, 1 pg, 100 fg, 10 fg, and 1 fgof M. tuberculosis DNA, respectively; lane 9, control withoutmycobacterial DNA; lane 10, bX174 DNA hydrolyzed with HaeIII.

VOL. 28, 1990

iw

on May 21, 2018 by guest

http://jcm.asm

.org/D

ownloaded from

2672 THIERRY ET AL.

A ~

$25 bp

FIG. 6. Detection of M. tuberculosis in clinical specimens afteramplification with ISTB2 and ISTB7 primers. (A) A 1.5% agarose

gel containing 1/lOth of the amplification reactions; (B) hybridizationwith 32P-labeled IS-2 probe. Lane 1, Control without mycobacterialDNA; lane 2, sputum which is smear negative and culture positive;lanes 3 and 4, gastric aspirates which are smear and culture positive;lane 5, sputum which is smear and culture negative; lane 6,cerebrospinal fluid which is smear negative and culture positive;lane 7, gastric aspirate which is smear negative and culture positive;lane 8, M. tuberculosis DNA-positive control; lane 9, 4X174hydrolyzed with HaeIII.

primers for amplification followed by hybridization withspecific probes. Another PCR based on IS6110 has also beendescribed by Eisenach and colleagues (5, 21), who obtainedcomparable results with different primers.

Detection of M. tuberculosis strains in uncultured clinicalspecimens. In vitro amplification with IS6110 and the 65-kDaantigen-derived oligonucleotides used as primers and probeswas applied to the detection of M. tuberculosis in 75 clinicalspecimens, including 42 gastric aspirate, 17 sputum, 6 cere-

brospinal fluid, 2 serum, 3 cystotomy, 4 pleural liquid, and 1urine specimen. These specimens were tested in parallel byclassical bacteriological methods, direct microscopic exam-ination after Ziehl-Neelsen staining, and culture on Loewen-stein-Jensen medium. Total DNA from 0.2 to 1 ml of eachsample was extracted and resuspended in 100 ,ul of water.Aliquots (10 pl) were amplified with the ISTB2 and ISTB7primers, and the amplification products were analyzed bySouthern blot hybridization by using IS-2 as a probe. Theresults obtained with several samples are shown in Fig. 6.Conclusions obtained by PCR analyses were based on re-sults obtained from two or three independent experiments.

Results obtained by PCRs were consistent with bacterio-logical data and clinical criteria. All samples positive bydirect examination and/or culture gave a positive result byDNA amplification (Table 1). Among 45 samples foundnegative by bacteriological analysis, 5 were positive byPCRs. These samples were not likely to represent false-positive results, since they were from patients in whomtuberculosis was highly suspected or already diagnosed.Three independent cystotomy specimens were from a pa-tient for whom renal tuberculosis was diagnosed on the basis

TABLE 1. Comparison of DNA amplification results with thedata obtained by standard procedures

No. of results of standard proceduresPCR result and Direct exam Direct exam Direct exam

specimen positive, cul- negative, cul- negative, cul-ture positive ture positive ture negative

Positive for M. tuber-culosis

Serum 1 0 0Sputum 6 1 0Gastric aspirate 14 5 0Cerebrospinal fluid 0 3 0Cystotomy O 0 3Pleural liquid O 0 2

Negative for M. tuber-culosis

Serum O 0 1Sputum O 0 10Gastric aspirate O 0 23Cerebrospinal fluid O 0 3Pleural liquid O 0 2Urine O 0 1

a Direct exam, Microscopic examination after Ziehl-Neelsen staining; cul-ture, culture on Loewenstein-Jensen solid medium.

of positive culture results obtained with other specimens(ureteral samples). Two pleural liquid specimens came froma patient for whom tuberculosis was highly suspected andwhose general state greatly improved under anti-tuberculoustreatment. The presence of M. tuberculosis bacilli was alsodetected in two serum specimens, which were not tested byculture, from patients with tuberculous meningitis and forwhom cerebrospinal fluids were found to be positive both byculture and by PCRs. On the other hand, analysis of samplesknown to contain M. avium bacilli by bacteriological analy-sis were found to be negative by amplification with ISTB2and ISTB7 primers derived from the IS6110 sequence.Two samples found to be positive by direct examination

were initially found to be negative by PCR methods. To testthe efficiency of DNA amplification in these samples, 10 ngof pUC19 (26) corresponding to plasmid DNA and primersallowing the amplification of a 517-bp internal fragment ofthis plasmid (a part of the bla gene, which is responsible forampicillin resistance) was added to the reaction. No ampli-fication of the 517-bp fragment was observed when 10 ,u ofthe sample mixture was added in the amplification reaction.However, amplification for both pUC19 and IS6110 se-quences was obtained when 10-fold dilution of the samplepreparation was added (Fig. 7). This result demonstrates thatsome samples may occasionally contain inhibitors of theamplification reaction. Consequently, all specimens shouldbe systematically tested for their amplification capability.The results presented here indicate that greatly improved

tuberculosis diagnosis is possible by the PCR technique,which is more rapid and more sensitive than classical bac-teriological techniques. The PCR might be especially usefulin patients for whom no presumptive diagnosis is available,such as for disseminated forms of tuberculosis that arefrequently encountered in patients with acquired immunedeficiency syndrome (23, 25).

Technical improvements of the method, including the useof nonradioactive probes and hybridization format, are beinginvestigated to simplify the technique for routine use inclinical laboratories.

J. CLIN. MICROBIOL.

.ML,

Mübbl A&

v

on May 21, 2018 by guest

http://jcm.asm

.org/D

ownloaded from

CHARACTERIZATION OF M. TUBERCULOSIS IS6110 2673

2 3 4 5 6 7 8

A

517 bp -_

32S bp _

FIG. 7. Presence of inhibitors of the amplification reaction inclinical specimens. Different volumes of sample DNA preparationswere amplified either with TEMA and TEMB primers in thepresence of pUC19 DNA (A) or with ISTB2 and ISTB7 primers (B).Lane 1, Control without mycobacterial DNA; lanes 2 to 4, 10, 1, and0.1 of the 100-,i DNA preparation from sample 1, respectively;lanes 5 to 7, 10, 1, and 0.1 pil of the 10-pul DNA preparation fromsample 2, respectively; lane 8, positive controls with pUC19 DNA(A) or M. tuberculosis DNA (B).

ACKNOWLEDGMENTS

We are grateful to Julian Davies for interest in this project andhelpful advice. We acknowledge Monica Ranes and Eamon Gorm-ley for critical reading of the manuscript.

This work was supported by the Centre National de la RechercheScientifique, grant 88.M.0887 from the Ministère de la Recherche etde la Technologie, Diagnostics Pasteur, and Institut Pasteur.

LITERATURE CITED1. Beaucage, S. L., and M. H. Caruthers. 1981. Deoxynucleoside

phosphoramidites-a new class of key intermediate for deoxy-nucleotide synthesis. Tetrahedron Lett. 22:1859-1862.

2. Birnboim, H. C., and J. Doly. 1979. A rapid alkaline extractionprocedure for screening recombinant plasmid DNA. NucleicAcids Res. 7:1513-1520.

3. Brisson Noël, A., B. Gicquel, D. Lecossier, V. Vincent Lévy-Frébault, X. Nassif, and A. J. Hance. 1989. Rapid diagnosis oftuberculosis by amplification of mycobacterial DNA in clinicalsamples. Lancet ii:1069-1071.

4. Davis, L. G., M. D. Dibner, and J. F. Battey. 1986. Basicmethods in molecular biology. Elsevier, Amsterdam.

5. Eisenach, K. D., M. D. Cave, J. H. Bates, and J. T. Crawford.1990. Polymerase chain reaction amplification of a repetitivesequence specific for Mycobacterium tuberculosis. J. Infect.Dis. 161:977-981.

6. Eisenach, K. D., J. T. Crawford, and J. H. Bates. 1988.Repetitive DNA sequences as probes for Mycobacterium tuber-culosis. J. Clin. Microbiol. 26:2240-2245.

7. Feinberg, A. P., and B. Vogelstein. 1983. A technique forradiolabelling DNA restriction endonuclease fragments to highspecificity activity. Anal. Biochem. 132:6-13.

8. Gonzalez, R., and B. A. Hanna. 1987. Evaluation of Gen-ProbeDNA hybridization systems for the identification of Mycobac-terium tuberculosis and Mycobacterium avium intracellulare.

Diagn. Microbiol. Infect. Dis. 8:69-77.9. Green, E. P., M. L. V. Tizard, M. T. Moss, J. Thompson, D. J.

Winterbourne, J. L. MacFadden, and J. Hermon-Taylor. 1989.Sequence and characteristics of IS900, an insertion elementidentified in a human Crohn's disease isolate of M. paratuber-culosis. Nucleic Acids Res. 17:9063-9073.

10. Guatelli, J. C., T. R. Gingeras, and D. D. Richman. 1989.Nucleic acid amplification in vitro: detection of sequences withlow copy numbers and application to diagnosis of humanimmunodeficiency virus type 1 infection. Clin. Microbiol. Rev.2:217-226.

11. Hallett, P., A. Mehlert, and A. Maxwell. 1990. Escherichia colicells resistant to the DNA gyrase inhibitor, ciprofloxacin, over-produce a 60 kD protein homologous to GroEL. Mol. Microbiol.4:345-353.

12. Hance, A. J., B. Grandchamp, V. Vincent Lévy-Frébault, D.Lecossier, J. Rauzier, D. Bocart, and B. Gicquel. 1989. Detectionand identification of mycobacteria by amplification of mycobac-terial DNA. Mol. Microbiol. 3:843-849.

13. Hohn, B., and J. Collins. 1980. A small cosmid for efficientcloning of large DNA fragments. Gene 11:291-298.

14. Hohn, B., and K. Murray. 1977. Packaging recombinant mole-cules into bacteriophage particles in vitro. Proc. Natl. Acad.Sci. USA 74:3259-3263.

15. Ishiguro, N., and G. Sato. 1988. Nucleotide sequence of inser-tion sequence IS3411 which flanks the citrate utilization deter-minant of transposon Tn3411. J. Bacteriol. 170:1902-1906.

16. Larzul, D., D. Chevrier, and J.-L. Guesdon. 1989. A non-radioactive diagnostic test for the detection of HBV DNAsequences in serum at the single molecule level. Mol. Cell.Probes 3:45-57.

17. Maniatis, T., E. F. Fritsch, and J. Sambrook. 1982. Molecularcloning: a laboratory manual. Cold Spring Harbor Laboratory,Cold Spring Harbor, N.Y.

18. Peterson, E. M., R. Lu, C. Floyd, A. Nakasone, G. Friedly, andL. M. de la Maza. 1989. Direct identification of Mycobacteriumtuberculosis, Mycobacterium avium, and Mycobacterium intra-cellulare from amplified primary cultures in BACTEC mediausing DNA probes. J. Clin. Microbiol. 27:1543-1547.

19. Prère, M.-F., M. Chandler, and O. Fayet. 1990. Transposition inShigella: isolation and analysis of IS911, a new member of theIS3 group of insertion sequences. J. Bacteriol. 172:4090-4099.

20. Saiki, R. K., D. H. Gelfand, S. Stoffel, S. J. Scharf, R. Higuchi,G. T. Horn, K. B. Mullis, and H. A. Erlich. 1988. Primer-directed enzymatic amplification of DNA with a thermostableDNA polymerase. Science 239:487-491.

20a.Sanger, F., S. Nicklen, and A. R. Coulson. 1977. DNA sequenc-ing with chain-terminating inhibitors. Proc. Natl. Acad. Sci.USA 74:5463-5467.

21. Thierry, D., M. D. Caves, K. D. Eisenach, J. T. Crawford, J. H.Bates, B. Gicquel, and J.-L. Guesdon. 1990. IS6110, an IS-likeelement ofM. tuberculosis complex. Nucleic Acids Res. 18:188.

22. Timmerman, K. P., and C. P. D. Tu. 1985. Complete sequenceof IS3. Nucleic Acids Res. 13:2127-2139.

23. Truffot-Pernot, C., H. F. Lecoeur, L. Maury, B. Dautzenberg,and J. Grosset. 1989. Results of blood cultures for detection ofmycobacteria in AIDS patients. Tubercule 70:187-191.

24. Vary, P. H., P. R. Andersen, E. Green, J. Hermon-Taylor, andJ. J. McFadden. 1990. Use of highly specific DNA probes andthe polymerase chain reaction to detect Mycobacterium paratu-berculosis in Johne's disease. J. Clin. Microbiol. 28:933-937.

25. Yakrus, M. A., and R. Good. 1990. Geographic distribution,frequency, and specimen source of Mycobacterium avium com-plex serotypes isolated from patients with acquired immunode-ficiency syndrome. J. Clin. Microbiol. 28:926-929.

26. Yanisch-Perron, C., J. Vieira, and J. Messing. 1985. ImprovedM13 phage cloning vectors and host strains: nucleotide se-quences of the M13mpl8 and pUC19 vectors. Gene 33:103-119.

VOL. 28, 1990

on May 21, 2018 by guest

http://jcm.asm

.org/D

ownloaded from