detection and identiﬁcation of mycobacterium species ...genomics research institute,1 center for...

JOURNAL OF CLINICAL MICROBIOLOGY, June 2003, p. 2605–2615 Vol. 41, No. 60095-1137/03/$08.00�0 DOI: 10.1128/JCM.41.6.2605–2615.2003Copyright © 2003, American Society for Microbiology. All Rights Reserved.

Detection and Identification of Mycobacterium Species Isolatesby DNA Microarray

Masao Fukushima,1,2 Kenichi Kakinuma,1,2 Hiroshi Hayashi,1 Hiroko Nagai,3Kunihiko Ito,4 and Ryuji Kawaguchi1*

Genomics Research Institute,1 Center for Molecular Biology and Cytogenetics,2 and Laboratory of Infection andImmunology,3 SRL, Inc., 5-6-50 Shinmachi, Hino-shi, Tokyo 191-0002, and Clinical Research Division,

Research Institute of Tuberculosis, Department of Respiratory Disease, Fukujuji Hospital,Japan Anti-Tuberculosis Association, 3-1-24 Matsuyama, Kiyose-shi, Tokyo 204-8533,4 Japan

Received 12 July 2002/Returned for modification 3 October 2002/Accepted 13 February 2003

Rapid identification of Mycobacterium species isolates is necessary for the effective management of tuber-culosis. Recently, analysis of DNA gyrase B subunit (gyrB) genes has been identified as a suitable means forthe identification of bacterial species. We describe a microarray assay based on gyrB gene sequences that canbe used for the identification of Mycobacteria species. Primers specific for a gyrB gene region common to allmycobacteria were synthesized and used for PCR amplification of DNA purified from clinical samples. A setof oligonucleotide probes for specific gyrB gene regions was developed for the identification of 14 Mycobacteriumspecies. Each probe was spotted onto a silylated glass slide with an arrayer and used for hybridization withfluorescently labeled RNA derived from amplified sample DNA to yield a pattern of positive spots. Thismicroarray produced unique hybridization patterns for each species of mycobacteria and could differentiateclosely related bacterial species. Moreover, the results corresponded well with those obtained by the conven-tional culture method for the detection of mycobacteria. We conclude that a gyrB-based microarray can rapidlydetect and identify closely related mycobacterial species and may be useful in the diagnosis and effectivemanagement of tuberculosis.

Tuberculosis is a disease with worldwide significance (7).Effective treatment of tuberculosis requires the rapid detectionand identification of Mycobacterium tuberculosis. Culture of theisolates is the traditional method used to confirm a diagnosis oftuberculosis, but culture is time-consuming because M. tuber-culosis isolates can take 4 to 8 weeks to grow in culture. Adiagnosis can also be made by biochemical or immunologicaltesting, but this can take even longer. Direct staining andmicroscopic examination of clinical specimens can produceresults more quickly, but this methodology lacks sensitivity andspecificity.

On the other hand, the AccuProbe system (Gen-Probe, SanDiego, Calif.) has been the “gold standard” among the com-mercial systems that identify mycobacteria by means of DNAprobes. However, hybridization with the AccuProbe systemthat was commercially available at that time was found to failwith a number of strains displaying the phenotypic features ofthe species Mycobacterium kansasii (34, 41).

PCR, which permits the amplification of specific DNA se-quences and multiplies even a single copy of a given DNAsequence by a factor of 1012 (31), has been applied to variousfields of diagnosis and has proved to be a most useful tool forthe rapid diagnosis of infectious diseases (13, 20, 28). PCR hasbeen used to analyze various mycobacterial genes for diagnos-tic purposes, including 16S and 23S rRNA genes, genus- andspecies-specific fragments in the chromosome (8, 11, 16, 26),genes coding for the 65-kDa heat shock protein (2, 15, 24) and

the 38-kDa protein B antigen (38), the dnaJ gene (39), andinsertion sequences such as IS6110 (9, 14, 30, 37, 40). 16SrRNA has been reported to be a suitable target for use in PCRamplification assays for the detection of Mycobacterium spp. ina variety of clinical samples (21) and has frequently been usedto identify various specific microorganisms because 16S rRNAgenes show species-specific polymorphisms (5, 18, 22, 25).However, because of the extremely slow speed of the molecu-lar evolution of 16S rRNA, the number of substituted basesbetween the 16S rRNA genes of closely related bacterialstrains, such as those belonging to the M. tuberculosis complex,is either nonexistent or too small to differentiate between thesespecies.

As an alternative to 16S rRNA analysis, Yamamoto andHarayama (44, 45, 46) designed a set of PCR primers thatallowed both the amplification of the gyrB gene, which encodesthe subunit B protein of DNA gyrase (topoisomerase type II),and the rapid nucleotide sequencing of the amplified gyrBfragments from a wide variety of bacteria. They used these gyrBgenes in the taxonomic classification of Pseudomonas putidaand Acinetobacter strains. We have reported that such closelyrelated bacteria, for example, Shigella and Escherichia coli,might be classified by gyrB analysis (12). The rate of molecularevolution inferred from gyrB gene sequences is faster than thatinferred from 16S rRNA gene sequences. For detection ofMycobacterium species, Kasai et al. (19) have determined thegyrB gene sequences of 43 slowly growing strains belonging to15 species in the genus Mycobacterium and developed a meth-od of PCR and PCR-restriction fragment length polymorphismanalysis to differentiate these species.

The identification of bacteria by molecular genetics can beadvanced further by DNA microarray technology (23, 27, 35).

* Corresponding author. Mailing address: Genomics Research In-stitute, SRL, Inc., 5-6-50 Shinmachi, Hino-shi, Tokyo 191-0002, Japan.Phone: 81-426-48-3873. Fax: 81-426-48-4054. E-mail: [email protected].

2605

on March 6, 2021 by guest

http://jcm.asm

.org/D

ownloaded from

http://jcm.asm.org/

The DNA microarray or DNA chip generally comprises a glasssurface on which multiple DNA probes with known identitiesare fixed for molecular hybridization with DNA samples, whichallows the examination of parallel gene expression or genotyp-ing. This method allows the simultaneous analysis of thousandsof genes in a short assay time and so is useful for phylogeneticanalysis and species identification. For the identification ofbacteria, this method may involve the labeling of in vitro RNAtranscribed from a target gene from bacteria in specimens,subsequent hybridization of the labeled in vitro transcribedRNA to species-specific oligonucleotide probes on a micro-array, and detection of the label, usually by fluorescence. Forexample, the Affymetrix Genechip, which uses large sets ofoligonucleotides that are synthesized rather than spotted ontoa glass substrate, has been successfully applied by using 16SrRNA genes as a target for the identification of Mycobacteriumspecies isolates (42).

In the present study, we have investigated the use of amicroarray technology with gyrB-derived DNA probes to dif-ferentiate Mycobacterium isolates at the species level. Usingthe nucleotide sequence data in GenBank (Bethesda, Md.), wedesigned specific probes for the identification of the Myco-bacterium species M. tuberculosis, M. bovis, M. africanum,M. avium, M. intracellulare, M. kansasii, M. gordonae, M. asi-aticum, M. gastri, M. malmoense, M. marinum, M. scrofulaceum,M. simiae, and M. szulgai. We show that species-specific hy-bridization patterns on a microarray containing these probescan differentiate and identify these mycobacteria to the specieslevel.

MATERIALS AND METHODS

Sample preparation for bacterial strain identification. The strains used in thisstudy originated from reference collections (the American Type Culture Collec-tion) or were clinical isolates. One or two freshly grown colonies of bacteria werescraped into a 1.5-ml Eppendorf tube and resuspended in 500 �l of sterile water.The bacterial suspension was then boiled (at 100°C for 10 min) to release theDNA.

Sample preparation for clinical application. Clinical sputum samples wereobtained from the Japan Anti-Tuberculosis Association in Fukujuji Hospital.Some standard strains were obtained from the American Type Culture Collec-tion and used as control bacteria. They were processed by the N-acetyl-L-cysteine(NALC)–NaOH method (29) and used for direct identification assays with themicroarray or by the AMPLICOR MTB-PCR (Roche Diagnostics, Inc., Tokyo,Japan). An equal volume of the NALC-NaOH solution (2% NaOH, 1.45%sodium citrate, 0.5% NALC) was mixed with the processed samples, and themixture was incubated at room temperature for 20 min. Phosphate buffer (67mM; pH 6.8) was added, and the mixture was centrifuged (3,500 � g) for 25 min.The excess fluid was poured off, and the sediment was resuspended in 1.0 ml ofphosphate buffer; 0.8 ml of this suspension was used for culture by standardmethods (29), and 0.2 ml was used to isolate chromosomal DNA, as follows. Thecells (0.2 ml) were added to 2� TES buffer (Tris-HCl [pH 8.5], 20 mM; EDTA,2 mM; NaCl, 300 mM) containing 100 �g of proteinase K (Roche Diagnostics,Basel, Switzerland) and 10% sodium dodecyl sulfate (SDS; final concentration,1%), and the mixture was heated at 65°C for 1 h. The lysates were extracted withan equal volume of phenol-chloroform and precipitated with ethanol. The pelletwas resuspended in 10 �l of TE buffer (10 mM Tris-HCl [pH 7.5], 1 mM EDTA).

Human control DNA was also extracted from peripheral blood leukocytesfrom healthy volunteers by standard techniques (33).

Preparation of DNA microarray. The oligonucleotides used to prepare theDNA microarray were synthesized at Sawady Technology Co., Ltd. (Tokyo,Japan). Each 10 �l of the 14- to 15-mer oligonucleotides at a concentration of200 �M was dispensed into a 96-well microplate with 10 �l of 2� ArrayItMicro-Spotting solution (TeleChem International, Inc., Sunnyvale, Calif.) perwell. The amino acid-modified DNA was printed onto silylated microscope slideswith an arrayer (SPBIO 2000; Hitachi Software, Tokyo, Japan). Following print-ing of the slides, the slides were left at 65°C for 18 h to permit thorough drying

of the DNA onto the surface of the silylated slides. After the slides had dried,they were washed in 0.2% SDS at 25°C for 5 min each, twice in distilled H2O(dH2O) at 25°C for 2 min each time, and once in dH2O at 95°C for 2 min; cooledto 25°C for 5 min; washed once in sodium borohydride solution (1.3 g of Na2BH4

dissolved in 375 ml of phosphate-buffered saline and 125 ml of pure ethanol) at25°C for 5 min, twice in 0.2% SDS for 3 min each time, and twice in dH2O at25°C for 2 min each time; and then left to air dry.

PCR and in vitro RNA transcription. The gyrB region was amplified by PCRwith Mycobacterium genus-specific primers (nucleotide positions 794 to 818 and894 to 910 in the reference M. tuberculosis sequence in GenBank [accessionnumber AB014241]; the M. tuberculosis amplicon size is 184 bp). The Mycobac-terium species-specific primers were derived from regions of the gyrB gene thatare conserved among all mycobacterial species. The primers that we designedwere F119 (5�-TGGGCAACACCGAAGTGAAGTCGTT-3�) and R184T7 (5�-GTAATACGACTCACTATAGGGCCGCACCARYTCWCGYGCYTT-3�),which contained a bacteriophage T7 promoter sequence at the 5� ends. Chro-mosomal DNAs were amplified by PCR in a thermocycler 480 (Perkin-ElmerCo., Norwalk, Conn.). PCR was performed in a total volume of 100 �l with 5 Uof Taq DNA polymerase (AmpliTaq; Perkin-Elmer Co.), 50 mM KCl, 10 mMTris-HCl (pH 8.3), 1.5 mM MgCl2, 0.001% (wt/vol) gelatin, 200 mM each de-oxynucleoside triphosphate (dATP, dCTP, dGTP, and dTTP), 10 �M primerF119, and 10 �M primer R184T7. A 5-�l bacterial sample was added to the PCRsolution, which underwent an initial denaturation step of 95°C for 5 min before30 cycles of 96°C for 1 min, 63°C for 1 min, and 72°C 1 min and then a final stepof 72°C for 7 min for the last cycle. The PCR products were analyzed byelectrophoresis on a 3% agarose gel. The promoter-tagged PCR amplicons wereused to generate fluorescently labeled single-stranded RNA targets by in vitrotranscription. Each 50-�l reaction mixture contained approximately 50 ng ofPCR product; 20 U of T7 RNA polymerase (Promega, Madison, Wis.); 40 mMTris-HCl (pH 8.1); 6 mM MgCl2; 2 mM spermidine; 10 mM NaCl; 10 mMdithiothreitol; 2 mM (each) ATP, CTP, and GTP; 0.04 mM UTP; and 0.2 mMFluoorlink cyanine 5 (Cy-5)–UTP (Amersham Pharmacia Biotech, Piscataway,N.J.). The reaction was carried out at 37°C for 1 h, and then the template DNAwas removed by adding 3 �l of DNase I (GIBCO BRL, Grand Island, N.Y.) at37°C for 15 min. The RNA transcribed in vitro was fragmented by incubationwith 30 mM MgCl2 at 94°C for 30 min (42).

Hybridization control. A hybridization probe (5�-GATCAGACACTTCAAGGTCTAG-3�) was printed onto silylated microscope slides with an arrayer. ADNA probe (5�-CTAGACCTTGAAGTGTCTGATC-3�) labeled withFluoorlink Cy-5–CTP (Amersham Pharmacia Biotech), together with labeledsample DNA, was allowed to hybridize to the microarray. The control probe andthe complementary target were made such that, ideally, they had similar meltingtemperatures and did not have consensus sequences that were the same as thesequence of the other probe. The hybridization signals for the control probeswere used as hybridization controls.

Hybridization and analysis. The fluorescently labeled RNA was resuspendedin 2.0 �l of sterile water and then in 8.0 �l of prewarmed 1.25� UniHybhybridization solution (TeleChem International, Inc.). The microarray was in-cubated in the presence of the fragmented labeled RNA solution for 30 min at30°C and then washed in 2� SSC (0.3 M NaCl plus 30 mM sodium citrate)–0.02% SDS at 25°C for 3 min and in 0.2� SSC at 25°C for 30 s. The fluorescentsignal emitted by a target bound to the microarray was detected at a pixelresolution of 10 �m by using the ScanArray Lite instrument (GSI Lumomics,Northville, Mass.). Sixteen-bit TIFF images of 10-�m resolution were importedinto QuantArray software (GSI Lumomics). After subtraction of the backgroundintensity (by a fixed circle-based quantification method), the mean intensities ofthe individual spots were used to calculate match-to-mismatch signal intensityratios for pairs of spots corresponding to different alleles.

Effect of target sequence on signal intensity. To determine whether the oli-gonucleotide DNA targets arrayed retained their expected hybridization prop-erties, we first tested the hybridization signal intensities on the microarray bycomparing the differences in fluorescence intensities between spots encodinghomologous targets and those encoding nonhomologous targets. A perfectlymatched oligonucleotide probe (M1-1 [5�-ACCGACGCGAAAGT-3�]) and amismatched oligonucleotide probe (M1-2 [5�-ACCGACTCGAAAGT-3�) (wherethe underscores indicate the mismatched nucleotide]) were printed onto silylatedmicroscope slides (n � 5) with an arrayer. The Fluoorlink Cy-5–CTP (AmershamPharmacia Biotech)-labeled DNA probe (5�-ACTTTCGCGTCGGT-3�) was al-lowed to hybridize to the microarray, and then the hybridization signals wereanalyzed.

Direct M. tuberculosis amplification tests. PCRs for M. tuberculosis amplifica-tion (MTB-PCR; AMPLICOR) were performed according to the instructions ofthe manufacturer (1).

2606 FUKUSHIMA ET AL. J. CLIN. MICROBIOL.


http://jcm.asm

.org/D

ownloaded from

http://jcm.asm.org/

Nucleotide sequence accession numbers. The nucleotide sequence data re-ported in this paper appear in the GenBank nucleotide sequence database underthe following accession numbers: AB014192, AB014206, AB014189, AB014184,AB014294, AB014191, AB014188, AB014302, AB014187, AB014203, AB014205,AB014027, AB014182, AB014185, and AB014242.

RESULTS

Amplification of mycobacterial species. The DNAs of themycobacterial species from clinical specimens were amplifiedand analyzed by 3% agarose gel electrophoresis to confirm that

the primers were specific for the gyrB genes of all Mycobacte-rium species. A representative example of the mycobacterialDNA amplification efficiencies is shown in Fig. 1. We obtainedspecific amplification of a 184-bp DNA fragment by PCR withprimers F119 and R184T7. No amplification products wereobserved from human genomic DNA (Fig. 1, lane 10), indicat-ing that there is no similar or homologous region in humanDNA. These results show that only a single band was amplifiedfrom the clinical specimens and that the primers used wereappropriate for amplification of the gyrB region of mycobacte-ria at the genus level.

Effect of target sequence on signal intensity. A DNA probe(5�-ACTTTCGCGTCGGT-3�) labeled with Fluoorlink Cy-5–CTP (Amersham Pharmacia Biotech) was allowed to hybridizeto the microarray, and then the hybridization signals wereanalyzed. Figure 2 shows that the signal intensities varied from85,000 to 200 fluorescence units. Quantification of the fluores-cence signals showed that the relative intensity ratio of thehomologous target to the nonhomologous target was about 8.0for three different concentrations. These data show that theseprobes can differentiate between signal intensities arising fromhomologous and nonhomologous targets.

Design of species-specific oligonucleotides for microarrayanalysis. The sequence alignment in Fig. 3 was used to identifyregions that were both unique to one particular mycobacterialspecies and sufficiently different from all other species to avoidcross-hybridization (17, 32). In addition, we designed eachprobe of 15 bases in such a way that the species-specific basesequence was located in the center of the probe. The probes

FIG. 1. Amplification of mycobacterial DNAs with primers F119and R184T7. Lanes: 1 to 8, amplification of gyrB fragments fromclinical samples; 1, M. scrofulaceum; 2, M. tuberculosis; 3, M. kansaii; 4,M. intracelluare; 5, M. gordonae; 6, M. avium; 7, M. bovis; 8, M. simiai;9, positive control with M. tuberculosis (GenBank accession numberAB014241); 10, human genomic DNA; 11, negative control; M, mo-lecular size markers (�X174-digested HindIII). The arrow indicatesthe position of the expected amplification product.

FIG. 2. Quantification for the hybridization signals. (a) Perfectly matched oligo-probe (PM) (M1-1; 5� ACCGACGCGAAAGT 3�) andmismatched oligo-probe (MM) (M1-2; 5� ACCGACTCGAAAGT 3�) were printed onto silylated microscope slides (n � 5) by an arrayer. Theseprobes in serial 10-fold dilutions were 4 to 6 log10 units (10�4 M, 10�5 M, and 10�6 M). Fluoorlink Cy-5-CTP (Amersham Pharmacia Biotech.,N.J.)-labeled DNA probe (M [5� ACTTTCGCGTCGGT 3�]) was allowed to hybridize to the microarray. (b) Hybridization signals were analyzed.The fluorescent signal emitted by a target bound to the microarray was detected at a pixel resolution of 10 �m by using ScanArray Lite. Sixteen-bitTIFF images of 10 �m resolution were imported into QuantArray software.

VOL. 41, 2003 MYCOBACTERIUM DETECTION AND IDENTIFICATION 2607


http://jcm.asm

.org/D

ownloaded from

http://jcm.asm.org/

chosen for each mycobacterial species sequenced are shown inTable 1.

Because most Mycobacterium species cannot be identifiedwith a single probe, we used a combination of probes to iden-tify individual species. Figure 4 indicates how a set of 28 probescan be used to differentiate 14 Mycobacterium species. EachMycobacterium species was expected to show a unique pattern

of reactivity to this set of probes. For example, whereas asample with M. tuberculosis is expected to hybridize with theM1-1, M3-1, M4-1, and M5-1 probes, no other Mycobacteriumspecies is able to react with the exact same four probes.

The probes were printed onto the microarray as shown inFig. 4. Figure 5 shows the actual scanning image obtained onthe microarray, in which the colors of the spots, which are

FIG. 3. Nucleotide sequence alignment of the gyrB genes from the 14 Mycobacterium species strains. Nucleotides identical to those inM. tuberculosis gyrB are indicated with dots, and PCR primers are underlined.



http://jcm.asm

.org/D

ownloaded from

http://jcm.asm.org/

pseudo-colored from yellow (highest) to blue (backgroundlevel) according to the fluorescence intensity, represent theintensities of the reactions with the probes. Positive spots couldbe differentiated from negative spots, and a clear, specific pat-tern of reactivity was observed for each six Mycobacteriumspecies which had been identified by the culture method. Theculture results were consistent with the predictions in Fig. 5.The specific set of probes complementary to M. tuberculosis,for example, hybridized only to nucleic acids from the corre-sponding species and not to nucleic acids from any other my-cobacterial species examined here.

Microarray assay with labeled RNA from DNA samplesfrom cultured specimens. We analyzed in a blinded fashion 68cultured specimens, including specimens with mycobacteriaand nonmycobacteria, by both the conventional culture meth-od and the microarray method, and the data were coded at alater time (Table 2). Both methods gave the same results. The28 specimens identified as nonmycobacteria by the culturemethod were classified as negative controls and were not rec-ognized by the microarray, as expected. In the culture assay,five specimens (specimens 6 to 10) were identified at best to bemembers of the M. avium-M. intracellulare complex (MAC). Bycontrast, the microarray method was able to identify the indi-vidual species in these five specimens, that is, to specify eitherM. avium or M. intracellulare. For the isolate in one specimen(specimen 15) that was identified as MAC by the culture meth-od, the scanning image obtained on the microarray showed acomposite of the pattern specific for M. avium and that specificfor M. intracellulare, suggesting that this patient had a dualinfection.

Blind test of microarray and AMPLICOR assays for directidentification of clinical specimens. To determine whether thegyrB fragment could be directly detected in clinical samples, weanalyzed 122 sputum samples by both the microarray assay andAMPLICOR assays (for M. tuberuclosis, M. avium, and M. in-tracellulare). Both methods produced identical results for thesemycobacteria (Table 3). Most specimens were both AMPLI-COR MTB-PCR and microarray assay negative. Ten of 122specimens were positive by the microarray assay and theAMPLICOR assay for M. tuberculosis, and 6 and 5 of 122specimens were positive by both methods for M. avium andM. intracellulare, respectively (Table 3). Moreover, of six spec-imens that were positive for M. avium by both methods, themicroarray assay identified two cases of dual infections withM. avium-M. intracelluare and M. avium-M. kansasii. Likewise,the microarray method detected a case of dual infection withM. avium-M. intracellulare, for which the AMPLICOR methoddetected only M. intracelluare.

DISCUSSION

Mycobacterial species are usually identified by time-consum-ing culture methods. Recently, the development of rapid diag-nostic tests that use molecular genetic methods, such as PCRamplification, has been reported. The microarray has provedto be a valuable tool for the specific detection of microorgan-isms directly from clinical samples. In particular, it has severaladvantages over classical detection methods: first, the numberof organisms in a clinical sample is not always large enough forthe organisms to be detected by microscopic methods; second,the period required for culture of these organisms is long;finally, not all acid-fast bacilli are M. tuberculosis, so an iden-tification test must be used to differentiate Mycobacterium spe-cies.

Among several DNA regions that have been targeted fordiagnostic detection of Mycobacterium species, the 16S rRNAgene has been used the most frequently. DNA detection hasbeen further advanced by exploiting the DNA microarray tech-nology (27), which can simultaneously detect hybridization tomultiple DNA probes arranged in an array.

In the study described here we have developed a microarrayassay for the detection of Mycobacterium species that uses thegyrB gene as the target. Our study shows that a microarrayassay targeting the gyrB gene can identify mycobacteria at thespecies level and can even differentiate among closely relatedspecies. A previous study (42) used 16S rRNA sequence datato construct a DNA probe array for the detection of mycobac-teria, but that array could not distinguish among closely relatedspecies of mycobacteria. In contrast, our microarray analysiswith the gyrB gene was able to classify the closely relatedspecies M. tuberculosis and M. bovis. Of possible therapeuticrelevance was the fact that closely related species of very dif-ferent clinical importance were clearly differentiated by thistechnique, as M. tuberculosis complex species could be distin-guished from M. avium, M. marinum, M. asiaticum, and M.intracellulare. These results confirm those of previous studiesreported by Yamamoto and Harayama (44, 45, 46) and Kasaiet al. (19) that analysis of gyrB gene sequences is a rapid andeffective method for the identification of bacterial species.

We have also shown that the microarray can readily differ-

TABLE 1. Mycobacteria-specific probes

Probe Sequence Probe locationa

M1-1 ACCGACGCGAAAGT 871–884M1-2 ACCGACTCGAAAGT 871–884M1-3 GCAGACGCCAAAGT 871–884M1-4 CGGACGCCAAGGT 872–884M1-5 GCTGACGCTAAAAC 871–884M1-6 TCGGAAGCCAAAAC 871–884M1-7 GCTGATGCTAAAACC 871–884M1-8 TCGGAAGCTAAAACC 871–884M1-9 CGGATGCCAAAACT 872–884M1-10 TCCGAGGCGAAAAC 872–884M2-1 CCAATCCGTCGGA 863–875M2-2 GCCAATCCTTCGGA 862–875M2-3 CCAACCCGTCGGA 863–875M2-4 CCAACCCATCGGA 863–875M2-5 CCAACCGCACGGA 863–875M3-1 TCTGTAACGAACAGC 830–844M3-2 TCTGTAATGAACAGC 830–844M4-1 GCGAAAGTCGTTGT 877–890M4-2 CCAAGGTGGTGGT 878–890M4-3 GCCAAAACCGTTGT 877–890M4-4 GCTAAAACCGTTGTG 877–891M4-5 GCCAAAACTGTTGTA 877–891M4-6 GCCAAAACCATTGTG 877–891M4-7 CGAAAACGGTGGTG 888–891M4-8 GCTAAAACCGTTGTA 888–891M4-9 CCAAAACCGTCGTG 888–891M5-1 AGGTCTGTAACGAAC 827–841M5-2 AGGTGTGCAATGAAC 827–841

a The locations of probes are shown by the nucleotide number of the gyrBsequence of M. tuberculosis (GenBank accession number AB014241).



http://jcm.asm

.org/D

ownloaded from

http://jcm.asm.org/



http://jcm.asm

.org/D

ownloaded from

http://jcm.asm.org/

FIG

.4.

Posi

tions

ofM

ycob

acte

rium

-spe

cific

olig

o-pr

obes

onth

eD

NA

mic

roar

ray.

Eac

hsp

ecie

sof

Myc

obac

teriu

mw

asex

pect

edto

show

aun

ique

patt

ern

ofre

activ

ityto

this

set

ofpr

obes

.(A

)M

.tub

ercu

losi

s;(B

)M

.bov

is;(

C)

M.a

vium

;(D

)M

.int

race

llula

re;(

E)

M.k

ansa

sii;

(F)

M.g

ordo

nae;

(G)

M.g

astr

i;(H

)M

.afr

ican

um;(

I)M

.asi

atic

um;(

J)M

.m

alm

oens

e;(K

)M

.mar

inum

;(L

)M

.scr

oful

aceu

m;(

M)

M.s

imia

e;(N

)M

.szu

lgai

.NC

,Non

bact

eria

lhyb

ridi

zatio

nco

ntro

l.



http://jcm.asm

.org/D

ownloaded from

http://jcm.asm.org/

entiate between correctly matched and mismatched sequences(Fig. 2). The hybridization signals arising from perfectlymatched oligonucleotide probes and mismatched oligonucleo-tide probes gave a signal-to-background ratio of 8.0 for threeconcentrations of arrayed DNA (10�4, 10�5, and 10�6 M).

We compared the clinical performance of the microarrayassay with that of the traditional culture method. The resultsfor isolates from 68 clinical specimens, including mycobacteriaand nonmycobacteria, show that the overall performance ofthe microarray was comparable to that of the culture method.In cases in which the culture assay could identify specimensonly as members of MAC, however, the microarray methodwas able to identify the individual species, that is, either M.avium or M. intracellulare. The microarray method was alsoable to identify both types of bacteria in cases of dual infection.For example, one specimen identified as MAC by the culturemethod was identified to include both M. avium and M. intra-cellulare through its composite pattern on the microarray (Ta-ble 3). Our assay also performed well when its performance

was compared with that of the AMPLICOR MTB-PCR assay.Studies on the use of PCR for the detection of M. tuberculosisorganisms show that overall it has good sensitivity and speci-ficity, although the results for sensitivity vary from approxi-mately 50 to 100% (3, 4, 6, 10, 24, 36, 43). Cartuyvels et al. (4)have reported that the AMPLICOR PCR cannot yet replaceculture as a first-line screening method for the detection of M.tuberculosis isolates, but it can be used as a rapid confirmatorytest for smear-positive specimens or in the case of a stronglysuspected M. tuberculosis infection (36). Rapid identification ofMycobacterium species is an important factor for the successfuldiagnosis of mycobacteriosis. However, because of the variablenature of the sputum specimens and the low sensitivities ofthese tests, there is a risk of false-negative results. No false-negative results were obtained by the microarray assay in thisstudy.

Our assay offers several advantages over other assays de-scribed in the literature. In addition, these initial studies sug-gest that our microarray method is at least as sensitive as and

FIG. 5. Mycobacterium microarray. Colors represent the various intensities of the Mycobacterium-specific oligonucleotide probes. Microarrayanalysis was performed after amplification of DNAs from cultured clinical specimens with primers F99 and R184T7, immobilization of specificprobes on a glass slide, and hybridization on the glass slide with Cy-5-labeled clinical specimens. (A) M. tuberculosis; (B) M. bovis; (C) M. avium;(D) M. intracellulare; (E) M. kansaii; (F) M. gordonae.



http://jcm.asm

.org/D

ownloaded from

http://jcm.asm.org/

TABLE 2. Comparison of culture and microarray results with clinical outcome

Sample no. SourceResults by:

Culture Microarray analysis

1 Sputum M. scrofulaceum M. scrofulaceum2 Sputum M. tuberculosis M. tuberculosis3 Sputum M. kansasii M. kansaii4 Sputum M. tuberculosis M. tuberculosis5 Sputum M. intracelluare M. intracelluare6 Sputum MAC M. intracelluare7 Sputum MAC M. avium8 Sputum M. gordonae M. gordonae9 Sputum M. kansasii M. kansasii10 Sputum MAC M. intracelluare11 Sputum M. gordonae M. gordonae12 Sputum M. scrofulaceum M. scrofulaceum13 Sputum MAC M. avium14 Sputum MAC M. avium15 Sputum MAC M. avium-M. intracelluare16 Sputum M. kansasii M. kansasii17 Sputum M. kansasii M. kansasii18 Sputum M. kansasii M. kansasii19 Sputum M. kansasii M. kansasii20 Sputum M. kansasii M. kansasii21 Sputum M. kansasii M. kansasii22 Sputum M. kansasii M. kansasii23 Sputum M. kansasii M. kansasii24 Sputum M. kansasii M. kansasii25 Sputum M. kansasii M. kansasii26 Sputum M. kansasii M. kansasii27 Sputum M. kansasii M. kansasii28 Sputum M. kansasii M. kansasii29 Sputum M. kansasii M. kansasii30 Sputum M. kansasii M. kansasii31 Sputum M. kansasii M. kansasii32 Sputum M. kansasii M. kansasii33 Sputum M. kansasii M. kansasii34 Sputum M. avium M. avium35 Sputum M. gordonae M. gordonae36 Sputum M. intracelluare M. intracelluare37 Sputum M. gordonae M. gordonae38 Sputum M. szulgai M. szulgai39 Sputum M. gordonae M. gordonae40 Sputum M. gordonae M. gordonae41 ATCC 29903 Shigella flexneri —a

42 ATCC 25931 Shigella sonnei —43 ATCC 9207 Shigella boydii —44 ATCC 54388 Salmonella entericab serovar Parathyphi A —45 ATCC 8759 Salmonella enterica serovar Parathyphi B —46 ATCC 19430 Salmonella enterica serovar Typhi —47 ES 22 Salmonella enterica serovar Enteritidis —48 Clinical isolate Salmonella enterica serovar Oranienburg —49 Clinical isolate Salmonella enterica serovar Chester —50 ATCC 14028 Salmonella enterica serovar Typhimurium —51 ATCC 13883 Klebsiella pneumoniae —52 ATCC 23715 Yersinia enterocolitica —53 ATCC 12453 Proteus mirabilis —54 ATCC 11638 Helicobactor pylori —55 ATCC 33560 Campylobacter jejunii —56 ATCC 25922 Escherichia coli —57 ATCC 2171 Vibrio alginolyticus —58 ATCC 17802 Vibrio parahaemolyticus —59 ATCC 49226 Neisseria gonorrhoeae —60 ATCC 13077 Neisseria meningitisdis —61 ATCC 23355 Enterobacter cloacae —62 ATCC 19433 Enterococcus feacalis —63 ATCC 65389 Staphylococcus aureus —64 ATCC 6303 Streptococcus pnemoniae —65 ATCC 19615 Streptococcus pyogenes —66 ATCC 13813 Streptococcus agalactiae —67 ATCC 16145 Pseudomonas aeruginosa —68 ATCC 29342 Mycoplasma pneumoniae —

a —, absence of amplification products.b Salmonella enterica subsp. enterica.



http://jcm.asm

.org/D

ownloaded from

http://jcm.asm.org/

may be less subject to error than methods based only on PCR.Most importantly, the microarray assay can analyze a samplefor several kinds of bacteria at the same time. Dual infections,such as those caused by M. avium and M. intracellulare or M.avium and M. kansasii, could also be identified by microarrayanalysis. Thus, we have shown that the microarray assay de-scribed here has high levels of analytical sensitivity and speci-ficity as a clinical test.

In summary, the potential of the microarray strategy for theparallel testing of different targets has been demonstrated. Ithas also been shown that gyrB gene-based microarrays have thepotential to be used for the direct testing of samples to providerapid results for species identification. Future studies will focuson defining both the identification of bacterial species and thedrug resistance genotyping features of this technology for ap-plication in clinical diagnostics.

ACKNOWLEDGMENTS

We thank Kazunori Hochido for assistance with culture. We thankNoboru Fujinami and Yumiko Saito for stimulating discussions. Wealso thank Tadashi Matsunaga of the Tokyo University of Agricultureand Technology for critical advice regarding this study.

This study was supported by The New Energy and Industrial Tech-nology Development Organization of Japan.

REFERENCES

1. Bergmann, J. S., and G. L. Woods. 1996. Clinical evaluation of the RocheAMPLICOR PCR Mycobacterium tuberculosis test for detection of M. tuber-culosis in respiratory specimens. J. Clin. Microbiol. 34:1083–1085.

2. Brisson-Noel, A., B. Gicquel, D. Lecossier, V. Levy-Frebault, X. Nassiv, andA. J. Hance. 1989. Rapid diagnosis of tuberculosis by amplification of my-cobacterial DNA in clinical samples. Lancet ii:1069–1071.

3. Brisson-Noel, A., C. Aznar, C. Chureau, S. Nguyen, C. Pierre, M. Bartoli, R.Bonett, G. Pialoux, B. Gicquel, and G. Garrique. 1991. Diagnosis of tuber-culosis by DNA amplification in clinical practice evaluation. Lancet 338:364–366.

4. Cartuyvels, R., C. de Ridder, S. Jonckheere, L. Verbist, and J. van Eldere.1996. Prospective clinical evaluation of Amplicor Mycobacterium tuberculosisPCR test as a screening method in a low-prevalence population. J. Clin.Microbiol. 34:2001–2003.

5. Cloud, J. L., H. Neal, R. Rosenberry, C. Y. Turenne, M. Jama, D. R. Hillyard,and K. C. Carroll. 2002. Identification of Mycobacterium spp. by using acommercial 16S ribosomal DNA sequencing kit and additional sequencinglibraries. J. Clin. Microbiol. 40:400–406.

6. Cousins, D. V., S. D. Wilton, B. R. Francis, and B. L. Gow. 1992. Use ofpolymerase chain reaction for rapid diagnosis of tuberculosis. J. Clin. Mi-crobiol. 30:255–258.

7. Dawson, D. J. 1990. Tuberculosis in Australia: an unfinished fight. Med. J.Aust. 154:75–76.

8. De Wit, D., L. Steyn, S. Shoemaker, and M. Sogin. 1990. Direct detection ofMycobacterium tuberculosis in clinical specimens by DNA amplification.J. Clin. Microbiol. 28:2437–2441.

9. Eisenach, K. D., M. D. Cave, J. H. Bates, and J. T. Crawford. 1990. Poly-merase chain reaction amplification of a repetitive DNA sequence specificfor Mycobacterium tuberculosis. J. Infect. Dis. 161:977–981.

10. Eisenach, K. D., M. D. Sifford, M. D. Cave, J. H. Bates, and J. T. Crawford.1991. Detection of Mycobacterium tuberculosis in sputum samples using apolymerase chain reaction. Am. Rev. Respir. Dis. 144:1160–1163.

11. Fries, J. W. U., R. J. Patel, W. F. Piessens, and D. F. Wirth. 1990. Genus- andspecies-specific DNA probes to identify mycobacteria using the polymerasechain reaction. Mol. Cell. Probes 4:87–105.

12. Fukushima, M., K. Kakinuma, and R. Kawaguchi. 2002. Phylogenetic anal-ysis of Salmonella, Shigella, and Escherichia coli on the basis of the gyrB genesequence. J. Clin. Microbiol. 40:2779–2785.

13. Gaydos, C. A., K. A. Crotechfelt, N. Shah, M. Tennant, T. C. Quinn, J. C.Gaydos, K. T. McKee, Jr., and A. M. Rompalo. 2002. Evaluation of dry andwet transported intravaginal swabs in detection of Chlamydia trachomatisand Neisseria gonorrhoeae infections in female soldiers by PCR. J. Clin.Microbiol. 40:758–761.

14. Haas, W. H., W. R. Butler, C. L. Woodley, and J. T. Crawford. 1993. Mixed-linker polymerase chain reaction: a new method for rapid fingerprinting ofisolates of the Mycobacterium tuberculosis complex. J. Clin. Microbiol. 31:1293–1298.

15. Hance, A. J., B. Grandchamp, V. Levy-Frebault, D. Lecossier, J. Rauzier, D.Bocart, and B. Gicquel. 1989. Detection and identification of mycobacteriaby amplification of mycobacterial DNA. Mol. Microbiol. 3:843–849.

16. Hermans, P. W. M., A. R. J. Schuitema, D. Van Soolingen, C. P. H. J.Verstynen, E. M. Bik, J. E. R. Thole, A. H. J. Kolk, and J. D. A. van Embden.1990. Specific detection of Mycobacterium tuberculosis complex strains bypolymerase chain reaction. J. Clin. Microbiol. 28:1204–1213.

17. Higgins, D. G., A. J. Bleasby, and R. Fuchs. 1992. CLUSTAL V: improvedsoftware for multiple sequence alignment. Comput. Appl. Biosci. 8:189–191.

18. Holberg-Petersen, M., M. Steinbakk, K. J. Figenschau, E. Jantzen, J. Eng,and K. K. Melby. 1999. Identification of clinical isolates of Mycobacteriumspp. by sequence analysis of the 16S ribosomal RNA gene. Experience froma clinical laboratory. APMIS 107:231–239.

19. Kasai, H., T. Ezaki, and S. Harayama. 2000. Differentiation of phylogeneti-cally related slowly growing mycobacteria by their gyrB sequences. J. Clin.Microbiol. 38:301–308.

20. Kearns, A. M., A. J. Turner, G. J. Eltringham, and R. Freeman. 2002. Rapiddetection and quantification of CMV DNA in urine using LightCycler-basedreal-time PCR. J. Clin. Virol. 24:131–134.

21. Kirschner, P., J. Rosenau, B. Springer, K. Teschner, K. Feldmann, and E. C.Bottger. 1996. Diagnosis of mycobacterial infection by nucleic acid amplifi-cation: 18-month prospective study. J. Clin. Microbiol. 34:304–312.

22. Maidak, B. L., N. Larsen, M. J. McCaughey, R. Overbeek, G. J. Olsen, K.Fogel, J. Blandy, and C. R. Woese. 1994. The Ribosomal Database Project.Nucleic Acids Res. 22:3485–3487.

23. Maldonado, R. R., M. L. Espinosa, C. S. Suarez, W. G. Beattie, and K. L.Beattie. 1999. Hybridization of glass-tethered oligonucleotide probes to tar-get strands preannealed with labeled auxiliary oligonucleotides. Mol. Bio-technol. 11:1–12.

24. Pao, C. C., T. S. B. Yen, J.-B. You, J.-S. Maa, E. H. Fiss, and C.-H. Chang.1990. Detection and identification of Mycobacterium tuberculosis by DNAamplification. J. Clin. Microbiol. 28:1877–1880.

25. Patel, J. B., D. G. Leonard, X. Pan, J. M. Musser, R. E. Berman, and I.Nachamkin. 2000. Sequence-based identification of Mycobacterium speciesusing the MicroSeq 500 16S rDNA bacterial identification system. J. Clin.Microbiol. 38:246–251.

26. Patel, R. J., J. W. U. Fries, W. F. Piessens, and D. F. Wirth. 1990. Sequenceanalysis and amplification by polymerase chain reaction of a cloned DNAfragment for identification of Mycobacterium tuberculosis. J. Clin. Microbiol.28:513–518.

27. Ramsay, G. 1998. DNA chips: state-of-the art. Nat. Biotechnol. 16:40–44.28. Reid, S. M., N. P. Ferris, G. H. Hutchings, K. De Clercq, B. J. Newman, N. J.

Knowle, and A. R. Samuel. 2001. Diagnosis of foot-and-mouth disease byRT-PCR: use of phylogenetic data to evaluate primers for the typing of viralRNA in clinical samples. Arch. Virol. 146:2421–2434.

29. Roberts, G. D., E. W. Koneman, and Y. K. Kim. 1991. Mycobacterium, p.304–340. In A. Balows, W. J. Hausler, K. L. Hermann, H. D. Isenberg, andH. J. Shadomy (ed.), Manual of clinical microbiology, 5th ed. AmericanSociety for Microbiology, Washington, D.C.

30. Ross, B. C., and B. Dwyer. 1993. Rapid, simple method for typing isolates ofMycobacterium tuberculosis by using the polymerase chain reaction. J. Clin.Microbiol. 31:329–334.

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

TABLE 3. Comparison of results of microarray analyseswith those of the AMPLICOR system

AMPLICOR test specificityand result

No. of specimens withthe following result by

microarray analysis:

Positive Negative

M. tuberculosisPositive 10 0Negative 0 112

M. aviumPositive 6 (2)a 0Negative 0 116

M. intracelluarePositive 5 (1)b 0Negative 0 117

a Two cases of dual infection with M. avium-M. intracelluare and M. avium-M. kansasii were detected by microarray analysis.

b One dual infection with M. intracellulare and M. kansasii was detected bymicroarray analysis.



http://jcm.asm

.org/D

ownloaded from

http://jcm.asm.org/

32. Saitou, N., and M. Nei. 1987. The neighbor-joining method: a new methodfor reconstructing phylogenetic trees. Mol. Biol. E. 4:406–425.

33. Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecular cloning: alaboratory manual, 2nd ed. Cold Spring Harbor Laboratory Press, ColdSpring Harbor, N.Y.

34. Scarparo, C., P. Piccoli, A. Rigon, G. Ruggiero, D. Nista, and C. Piersimoni.2001. Direct identification of mycobacteria from MB/BacT alert 3D bottles:comparative evaluation of two commercial probe assays. J. Clin. Microbiol.39:3222–3227.

35. Schena, M., D. Shalon, R. W. Davis, and P. O. Brown. 1995. Quantitativemonitoring of gene expression patterns with a complementary DNA mi-croarray. Science 270:467–470.

36. Shankar, P., N. Manjunath, K. K. Mohan, K. Prasad, M. Behari, G. K.Shriniwas, and K. Ahuja. 1991. Rapid diagnosis of tuberculous meningitis bypolymerase chain reaction. Lancet 337:5–7.

37. Shawar, R. M., F. A. K. El-Zaatari, A. Nataraj, and J. E. Clarridge. 1993.Detection of Mycobacterium tuberculosis in clinical samples by two-step poly-merase chain reaction and nonisotopic hybridization methods. J. Clin. Mi-crobiol. 31:61–65.

38. Siobring, U., M. Mecklenburg, A. B. Andersen, and H. Miorner. 1990. Poly-merase chain reaction for detection of Mycobacterium tuberculosis. J. Clin.Microbiol. 28:2200–2204.

39. Takewaki, S.-I., K. Okuzumi, H. Ishiko, K.-I. Nakahara, A. Ohkubo, and R.Nagai. 1993. Genus-specific polymerase chain reaction for the mycobacterialdnaJ gene and species-specific oligonucleotide probes. J. Clin. Microbiol.31:446–450.

40. Thierry, D., M. D. Cave, K. D. Eisenach, J. T. Crawford, J. H. Bates, B.Gicquel, and J. L. Guesdon. 1990. IS6110, an IS-like element of the Myco-bacterium tuberculosis complex. Nucleic Acids Res. 18:188.

41. Tortoli, E., A. Nanetti, C. Piersimoni, P. Cichero, C. Farina, G. Mucignat, C.Scarparo, L. Bartolini, R. Valentini, D. Nista, G. Gesu, C. P. Tosi, M.Crovatto, and G. Brusarosco. 2001. Performance assessment of new multi-plex probe assay for identification of mycobacteria. J. Clin. Microbiol. 39:1079–1084.

42. Troesch, A., H. Nguyen, C. G. Miyada, S. Desvrenne, T. R. Gingeras, P. M.Kaplan, P. Cros, and C. Mabilat. 1999. Mycobacterium species identificationand rifampin resistance testing with high-density DNA probe arrays. J. Clin.Microbiol. 37:49–55.

43. Wilson, S. M., R. McNerney, P. M. Nye, P. D. Godfrey-Faussett, N. G.Stoker, and A. Voller. 1993. Progress toward a simplified polymerase chainreaction and its application to diagnosis of tuberculosis. J. Clin. Microbiol.31:776–782.

44. Yamamoto, S., and S. Harayama. 1995. PCR amplification and direct se-quencing of gyrB genes with universal primers and their application to thedetection and taxonomic analysis of Pseudomonas putida strains. Appl. En-viron. Microbiol. 61:1104–1109.

45. Yamamoto, S., and S. Harayama. 1996. Phylogenetic analysis of Acineto-bacter strains based on the nucleotide sequences of gyrB genes and on theamino acid sequences of their products. Int. J. Syst. Bacteriol. 46:506–511.

46. Yamamoto, S., and S. Harayama. 1998. Phylogenetic relationships ofPseudomonas putida strains deduced from the nucleotide sequences of gyrB,rpoD, and 16S rRNA genes. Int. J. Syst. Bacteriol. 3:813–819.



http://jcm.asm

.org/D

ownloaded from

http://jcm.asm.org/

detection and identiﬁcation of mycobacterium species ...genomics research institute,1 center for...

Documents