INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Mycobacterium bovis, a member of the Mycobacterium tuberculosis complex, is the causative agent of bovine tuberculosis. This zoonotic disease continues to have considerable economic and public health implications (18, 19). National eradication programs employ tuberculin testing and slaughter strategies (17). Culture is used commonly to confirm infection in postmortem specimens from cattle slaughtered following a positive skin test reaction. Although culture is considered to be the "gold standard" for confirming tuberculosis, this procedure may take several weeks (15).

Consequently, rapid nucleic acid amplification techniques, including the PCR- and transcription-mediated amplification, have been applied to detect M. bovis directly in clinical specimens (1, 2, 15, 28; S. Roring et al., unpublished data). While PCR amplifications have enabled detection of nonviable mycobacteria (15, 16, 28), they are not as sensitive as culture. In comparison to M. tuberculosis in human sputa, which often contain large numbers of bacilli (28), bovine tuberculous tissues are associated with few bacilli. Extraction of mycobacterial DNA from sputa was considered to be less difficult than extraction from tissues, and consequently, improvements in nucleic acid extraction were recommended to increase sensitivity of PCR detection of M. bovis in tissue (6).

Mycobacterial DNA extraction efficiency has been improved with the development of a nucleic acid sequence capture procedure, which has enabled detection of mycobacteria in paucibacillary forms of tuberculosis (3, 16). Sequence capture PCR has been used successfully for simultaneous detection and strain typing of M. bovis from BACTEC cultures (22) and more recently from bovine lymph node tissue (23). These studies were performed in conventional heating block thermocyclers (HBTC) with end point detection of PCR products by agarose gel electrophoresis. Automation of the sequence capture PCR procedure should facilitate its use in routine diagnosis of M. tuberculosis in clinical specimens (3, 16).

Rapid-cycle PCR amplifications, using an air thermocycler (ATC), have increased the rapidity of M. tuberculosis detection by decreasing the amplification time (6, 13) and thus may assist automation. Additionally, alternative PCR product detection systems have the potential to improve automation, as traditional methods of detection may be laborious and time-consuming (12, 20).

Fluorimeter-based closed-tube PCR assays permit the continual monitoring of accumulating fluorescently labeled PCR products, termed real-time fluorescence (11). Rapid-cycle PCR in conjunction with fluorimeter-based closed-tube PCR assays have provided a rapid and sensitive method for identification and quantification of PCR products (21, 29). Real-time fluorescence has been used to detect M. tuberculosis in sputum using the TaqMan system (8). In this study, sequence capture is combined with rapid-cycle PCR and real-time fluorescence involving the use of an M. tuberculosis complex-specific fluorescence resonance energy transfer (FRET) probe in a LightCycler LC32 (Biogene Ltd., Kimbolton, United Kingdom) to detect M. bovis in bovine tissues.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Biological material. A bovine field isolate of M. bovis, characterized by restriction fragment length polymorphism analysis (24), was cultured in Middlebrook 7H9 broth (Difco Laboratories). Serial dilutions of culture broth (10³ to 10⁸) were made in sterile saline and subcultured on 7H9 agar plates. Aliquots of the culture dilution series (400 µl) were stored at 20°C and used for the determination of sequence capture PCR efficiency at a later date.

1

DNA extractions. DNA sequence capture PCR, as previously described (16) and modified (23), was applied to homogenates of the 38 decontaminated bovine lymph nodes with the following additional modifications. In brief, suspensions of decontaminated tissue homogenate (500 µl) were transferred to screw-cap microcentrifuge tubes containing 500 µl of 0.1-mm-diameter zirconium beads (Biospec Products Inc., Bartlesville, Okla.). Samples were centrifuged and washed as described previously (23). The pellet and beads were resuspended in 500 µl of TES [N-tris(hydroxymethyl)methyl-2-aminoethanesulfonic acid]-100 mM Tris-HCl (pH 7.4)-50 mM EDTA-150 mM NaCl and incubated in a sonicating water bath for 15 min. Samples were agitated in a FastPrep Bio 101 bead shaker (Savant Instruments Inc., Holbrook, N.Y.) at 6 m/s for 45 s. Proteinase K (Sigma, Poole, Dorset, United Kingdom) was added to a final concentration of 3 mg/ml, and the mixture was incubated at 50°C for 18 h. Samples were shaken in the FastPrep Bio 101 as before. Aliquots of proteinase K-treated homogenates (500 µl) were denatured at 100°C for 15 min and immediately transferred to ice for a further 5 min. The biotinylated capture oligonucleotides CapDRa (5'biotin-AAAAAGGTTTTGGGTCTGACGAC) and CapDRa (5'biotin-AAAAACCGAGAGGGGACGGAAAC) [Genosys Biotechnologies (Europe), Pampisford, Cambs., United Kingdom] were used to capture the DR region of the M. tuberculosis complex (16). Capture oligonucleotides (2.5 pmol of each) in 3.75 M NaCl solution were added to homogenates to a final concentration of 1 M NaCl, mixed well, and hybridized at 42°C with gentle agitation for 3 h. Following hybridization, 50 µg of streptavidin M-280 Dynal beads (Dynal, Oslo, Norway) was added, mixed well, and incubated at 36°C with gentle agitation for 2 h. The streptavidin M-280 Dynal beads were separated from the supernatant and washed using a magnetic bead separator (6 by 1.5 ml) (Stratagene, La Jolla, Calif.) in 750 µl of wash buffer (10 mM Tris-HCl [pH 7.4], 1 mM EDTA) and then in sterile deionized H₂O. Magnetic beads were resuspended in 25 µl of sterile deionized H₂O and stored at 20°C prior to PCR amplification. Immediately prior to amplification, captured mycobacterial DNA was released from magnetic beads by heat treatment at 100°C for 5 min. Following a brief centrifugation, the resulting supernatant was subjected to the appropriate amplification.

PCR amplification. Sequence-captured mycobacterial nucleic acid was subjected to rapid thermal cycling and continuous monitoring of PCR products in a LightCycler LC32 (Biogene Ltd.) and to conventional PCR amplification using an HBTC (model 480; Perkin-Elmer, Warrington, United Kingdom) (HBTC-PCR). Oligonucleotides specific for the M. tuberculosis complex IS6110 sequence (9) (Genosys) were used for the amplifications. A standard PCR protocol (LC-PCR) utilizing the LC32 instrumentation was adopted only after extensive modification and optimization of previously reported PCR conditions (9). Optimal magnesium ion and oligonucleotide concentrations were determined along with cycling parameters such as denaturation, annealing and elongation temperatures, incubation periods, and temperature transition rates. LC32 instrumentation variables were optimized as directed by the manufacturers.

Standard LightCycler protocol. After optimization, the following standard LC-PCR protocol was applied to all specimens. A commercial PCR master mixture (Bio/Gene Ltd.), containing 2.5 U of Taq polymerase, 250 µM deoxynucleoside triphosphates, and 3 mM MgCl₂, was pretreated with Taq-Start antibody (7 µl) (Sigma). The PCR mixture consisted of pretreated PCR master mixture supplemented with IS6110-specific oligonucleotides (500 nM), Syb Green 1 (1/60,000 dilution) (Biogene Ltd.), lambda DNA (5 pg/µl), and Cy5 3'-labeled FRET probe LCP (100 nM) together with 2.5 µl of target DNA in a final reaction volume of 10 µl. The FRET probe LCP (5'GCCCAGGTCGACACATAGG3'-Cy5), specific for the M. tuberculosis complex, was designed using OLIGO 5 primer design software (Biogene Ltd.) and to the specifications described by Biogene Ltd.

HBTC protocol. HBTC-PCR included either 2.5 µl of sequence-captured magnetic beads or a 5-µl inoculum of either genomic DNA or heat-inactivated M. bovis. A PCR master mix (50-µl final volume) contained 2.5 U of Taq polymerase (Sigma), a 500 nM concentration of each of the IS6110 primers, 250 µM deoxynucleoside triphosphates, and 1.75 mM MgCl₂, with a mineral oil overlay (Sigma). Conditions for cycling were 94°C for 5 min, followed by 30 cycles of 94°C for 1 min, 68°C for 2 min, and 74°C for 1 min. Amplification mixtures were incubated for a further 7 min at 74°C, with a final 4°C soak.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Optimization of light cycler reactions. (i) Preliminary experiments. PCR amplifications and monitoring of fluorescence emission during PCR were performed using a LightCycler LC32. An amplification program, recommended by the LC32 manufacturer, was modified so that annealing temperatures similar to those previously published for this primer set were used (9). Preliminary experimentation using positive controls of heat-killed mycobacterial suspension together with negative controls of water established suboptimal reaction conditions of 3 mM magnesium (range of 2 to 5 mM tested in 1 mM increments), annealing temperature of 60°C (range of 60 to 66^oC tested in 1°C increments), 500 µM primer (100 and 500 nM tested), and 100 nM LCP (10, 50, and 100 nM tested).

View larger version (43K): [in this window] [in a new window]   FIG. 1.   Screen capture image (CorelDRAW; Corel Corporation, Ottawa, Ontario, Canada) of run profile analysis demonstrating the accumulation of F1 fluorescence during PCR of 10-fold serial dilutions of M. bovis genomic DNA representing 2.5 × 104 to 2.5 × 101 IS6110 gene copies using Syb Green 1 only.

View larger version (44K): [in this window] [in a new window]   FIG. 2.   Screen capture (CorelDRAW) of F1 melting-peak analysis of 10-fold serial dilutions of M. bovis genomic DNA representing 2.5 × 104 to 2.5 × 101 IS6110 gene copies using Syb Green 1. Each trace graphically displays the rate of change of F1 fluorescence emissions versus temperature; a peak represents the Tm of the PCR product. The IS6110 123-bp species displays a Tm of 89°C, while a primer artifact displays a peak at 86°C.

View larger version (48K): [in this window] [in a new window]   FIG. 3.   Screen capture (CorelDRAW) of F2 melting-peak analysis of an M. bovis genomic DNA dilution series representing 2.5 × 104 to 2.5 × 101 IS6110 gene copies using FRET probe LCP and Syb Green 1. The Tm of each PCR product is identified as a peak; each trace represents the rate of change of F2/F1 fluorescence emissions versus temperature. The IS6110 123-bp species displays a Tm of 60 to 65°C, while the primer artifact displays none. The F1 emissions are responsible for the F1 peaks at the right side of the graph.

(ii) Improved reproducibility and sensitivity. In preliminary experiments, amplifications performed on consecutive days were found to lack reproducibility (data not shown). Positive control run profiles demonstrated inconsistencies in signal acquisition. This lack of reproducibility of the LC-PCR was attributed to Syb Green 1 instability. Revised protocols from the manufacturer (Biogene Ltd.) required that Syb Green 1 be stored at 20°C, in an undiluted form, for storage of up to 1 year without loss of activity. Once diluted 1/1,000 in the dilution buffer supplied by the manufacturer, Syb Green 1 must be stored in the dark at 4°C for no more than 1 month and shaken vigorously for 1 min immediately prior to use.

Sensitivity. LightCycler PCR amplifications using the standard protocol reproducibly detected 25 copies of the M. bovis IS6110 gene DNA with melt peak analysis (dF/dT versus temperature) of both F1 and F2 emissions (Fig. 2 and 3). LC-PCR amplification of sequence-captured M. bovis culture dilution series detected 36 bacilli using the F1 channel and 360 bacilli using the F2 channel. A single round of HBTC-PCR amplification of the same dilution series of M. bovis IS6110 gene DNA as used for the LC-PCR amplifications reproducibly detected 625 gene copies. A second round of HBTC amplification, however, was required in order to detect 62 gene copies. Two rounds of HBTC amplifications were also required to detect 36 sequence-captured M. bovis bacilli.

Clinical specimens. Lymph node tissues were cultured and examined histopathologically. Of the 38 lymph node specimens from animals with lesions, 13 were from cattle which were skin test positive (Table 1), and 11 of these 13 specimens were culture positive. Twenty-seven of the 38 specimens were culture positive using both L-J slope and BACTEC media (Table 1). All culture isolates were confirmed as M. tuberculosis complex by Accuprobe (Gen-Probe) (data not shown). Three were BACTEC cultured positive after 7 days, and the remaining 24 were positive after 14 days. Histopathological examination identified 26 bovine tuberculosis-positive specimens. Five of the 28 culture-positive specimens were negative by histopathology, and 3 of the 10 culture-negative specimens (Table 1) were positive by histopathology. One specimen, identified with morphology typical of actinomycetes by histopathology, was BACTEC culture negative; however, this specimen was L-J culture positive after 28 days.

View larger version (54K): [in this window] [in a new window]   FIG. 4.   (a) Agarose gel electrophoresis of HBTC-PCR-amplified DNAs from lymph nodes from animals with bovine tuberculosis lesions. Amplification products from five specimens listed in Table 1 are displayed. The size of the amplified target DNA is shown. (b) Screen capture (CorelDRAW) of F2 melting peak analysis of LC-PCR amplification of DNAs from lymph nodes from five animals with bovine tuberculosis lesions (Table 1).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

This work is the first reported application of sequence capture PCR and real-time fluorescence detection of M. bovis in clinical tissue. The benefits of target enrichment and removal of inhibitory substances attributed to sequence capture were combined successfully with the alternative PCR product detection technology offered by real-time fluorescence. Several nucleic acid amplification-based techniques have become accessible to clinical mycobacteriological laboratories in recent years. However, poor performance in paucibacillary situations and high costs have limited their widespread application (20). Requirement of a high degree of operator skill to perform and interpret molecular analyses has also hindered adoption of molecular methods in diagnostic laboratories (10). The application of PCR to disease diagnosis is dependent on specificity, sensitivity, ease of use, and high sample throughput (3, 16, 20). In this study improved DNA extraction and fluorescent detection methodologies were used to examine these criteria and to facilitate PCR diagnosis of M. bovis in bovine tuberculosis specimens.

Culture of M. bovis from clinical specimens has been regarded as the standard against which other M. bovis diagnostic methodologies have been measured (15). However, culture of M. bovis is slow, and incubation periods of several weeks may be required (15, 28). Radiometric culture (BACTEC) permits more rapid detection of mycobacteria; however, minimum incubation periods of several days are still required (14). In this study, only 4 of the 28 culture-positive specimens were BACTEC culture positive within 7 days. A further 23 were culture positive within 14 days. The remaining specimen was BACTEC culture negative and exhibited only poor growth on L-J slopes. The inability to culture mycobacteria from tissue specimens has previously been associated with sample autolysis or microbial contamination (15, 28), which may account for the latter result.

Histopathological examination, used commonly for routine diagnosis of bovine tuberculosis, permits rapid identification of lesions but may not differentiate between those caused by M. bovis and by other mycobacterial or closely related agents (7). In addition to this lack of specificity, the sensitivity of microscopic examination is limited (31). The results of this study indicated the limitations of histopathological analysis, as only 82% (23 of 28) of the culture-positive specimens were positive by histopathological analysis. The five histopathologically negative specimens were both culture and PCR positive. Three histopathologically positive specimens were negative by both culture and PCR, suggesting that these histopathology results may represent false positives.

Limited sensitivity has also been associated with molecular technologies for mycobacterial diagnosis (16). PCR has been applied widely to the detection of M. tuberculosis in sputa and bronchiolar lavages from human tuberculosis patients with success but not so for bovine tuberculosis, where the bacterial load is substantially less (8, 28).

The application of sequence capture and conventional PCR amplification methodologies in this study demonstrated M. bovis detection sensitivities comparable to those of culture. This is in agreement with other sequence capture studies (3, 16, 23). Ninety-three percent of culture-positive specimens were detected using two rounds of PCR amplification in the HBTC followed by analysis of the PCR products by agarose gel electrophoresis. This was comparable to a sensitivity of 91% achieved in earlier PCR studies (1, 28). PCR amplification and product detection, using HBTC, took approximately 9 h, which was much shorter than the period required for detection by culture.

Further reductions in the PCR detection time for M. tuberculosis have been achieved using an ATC (6, 13). In these studies, air thermal cycling used small-volume glass capillaries and high-velocity heated air to generate temperature transition rates in excess of 10°C/s and culminated in amplification times of less than 30 min. In addition, the use of reduced reaction volumes associated with an ATC resulted in a substantial reduction in reagent costs (6). Until recently, PCR amplification and product analysis have been sequential procedures. The LC32 used in this study measures real-time fluorescence, combining the rapid-cycle capabilities of the ATC with continual fluorimetric detection of accumulating PCR products.

Interpretation of fluorimetric emissions generated by Syb Green 1 was confirmed initially by agarose gel electrophoresis analysis of PCR products evacuated from the capillary reaction vessels and resolved on agarose gels, as was the case in earlier studies utilizing fluorescent detection systems (27). The presence of the expected 123-bp PCR product along with PCR artifacts established that the molecular enzymology of the LC32 had resulted in the synthesis of the predicted IS6110 PCR product. The sensitivity achieved in this report with the IS6110 target was 25 IS6110 gene copies, which was comparable to that of previous studies (16). Accumulation of fluorescence may not equate to specific PCR product synthesis, as Syb Green 1 detects all double-stranded DNA, including PCR artifacts (21, 27).

In this present study PCR product differentiation was achieved, in most cases, by analyzing the melting curve characteristics of the PCR products. Melting curves have previously differentiated multiple PCR products in concordance with those products identified by gel electrophoresis. PCR products with T_ms differing by only 2°C have been distinguished (21). Again in the present study, the 123-bp IS6110-specific product exhibited a T_m of 89°C, while the PCR primer artifact displayed a T_m of 86°C. Primer artifacts occasionally generated shoulder effects on F1 signal melt peaks of specific PCR products, making their interpretation difficult. Primer artifacts with T_ms similar to those of specific products have been reported in other studies (27, 30). These reports suggested that melt curve analysis alone is not definitive in identifying specific PCR products.

Sequence-specific detection using FRET hybridization probes (specific oligonucleotide probes labeled with fluorescent dyes) has been reported. Several FRET formats have been developed (12). One format, hybridization probes, which employs two single fluorescence-labeled oligonucleotides that hybridize to adjacent regions of target DNA, has been used in conjunction with LightCycler instrumentation (4). In the present study, because the target was small and GC rich, only a single-oligonucleotide FRET probe was used as a hybridization probe. Resonance energy, received from Syb Green 1 intercalated between the oligonucleotide and PCR product, was transferred to the Cy5 fluorophore attached to the oligonucleotide's 3' terminus. The broad emission spectrum of the fluorescence of Syb Green 1 overlaps that of Cy5, but the melt curve analysis compensates by dividing the F2 signal by the F1 signal. This FRET probe enabled detection of an M. tuberculosis complex target which exhibited a T_m of 60 to 65°C. A sensitivity of 25 IS6110 gene copies, similar to that achieved using the DNA dilution series and Syb Green 1 only, was demonstrated. The FRET probe circumvented the previously mentioned difficulties with interpretation of the F1 signal melt peak. Few studies have used this single-probe approach, but those that have report a greater sensitivity than with the double-probe format (4).

The Cy5-labeled FRET hybridization probe LCP described in this study detected M. bovis genomic DNA in decontaminated tissue homogenates of 20 of the 28 culture-positive specimens. This sensitivity of detection was lower than the 26 of 28 culture-positive specimens detected by PCR in the HBTC. The reduced sensitivity with the FRET probe was attributed in part to the perturbation of fluorescent signals by the red iron oxide present in the sequence capture magnetic beads. This perturbation was alleviated by detachment of captured mycobacterial DNA from magnetic beads by heat treatment at 100°C for 5 min and centrifugation prior to amplification. This treatment resulted in significant improvements in signal interpretation. PCR in the HBTC was probably more sensitive than that in the LC32 due to the second round of amplification employed in the HBTC. It is clear that the benefits of PCR reamplification need to be balanced against the associated increased risk of sample cross contamination and increased assay time. Further optimization of the LC32 system is expected to elevate the sensitivity to that of HBTC. This study demonstrated the rapidity of PCR diagnosis compared to culture. Rapid-cycle PCR and fluorimetric detection technologies have reduced detection time further to several minutes and would allow rapid screening of tissue specimens for the presence of M. bovis. The same fluorescent technology should facilitate development of a high-throughput molecular diagnostic assay and provide a more practical approach for confirmation of tuberculosis directly from clinical specimens.