| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Technical Briefs |
1 Bernhard-Nocht Institute of Tropical Medicine, National Reference Centre for Tropical Infections, 20359 Hamburg, Germany
aaddress correspondence to this author at: Bernhard Nocht Institute of Tropical Medicine, Bernhard-Nocht-Strasse 74, 20359 Hamburg, Germany; fax 49-40-42818378, e-mail drosten{at}bni-hamburg.de
PCR is widely used in clinical laboratories to diagnose pulmonary, extrapulmonary, and disseminated tuberculosis. A multitude of primer pairs have been successfully applied, one of which has been studied most extensively because it is included in the Roche Amplicor MTB assay, the only Food and Drug Administration-cleared PCR-based test for clinical detection of Mycobacterium tuberculosis (1)(2)(3). The test amplifies a 584-bp fragment of the 16s rDNA of all mycobacteria and identifies members of the M. tuberculosis complex (MTC) by hybridization of a specific DNA probe. However, the hybridization step extends the turnaround time of this test, and obligatory license fees render it unaffordable for application in experimental studies or resource-limited settings.
Real-time detection technology has made it possible to establish noncommercial, probe-based PCR systems that provide stable operation, low contamination risk, and semiautomated interpretation of results (4)(5)(6)(7)(8)(9)(10)(11). We therefore aimed at adapting the Roche Amplicor assay to a real-time PCR protocol.
We did not modify the primers of the Amplicor test because of their demonstrated performance; thus, only the detection probe was adapted to the requirements of real-time PCR. Classic real-time probes [5'-nuclease, "TaqMan" (12)] require cleavage by Taq polymerase (13), which could not be accomplished efficiently in our assay because the amplicon was too long and the only probe binding site specific for MTC was too distant from any of the primers (9)(12)(14). As an alternative approach, we chose a pair of fluorescence resonance energy transfer (FRET) probes that do not have to be cleaved, making them less dependent on the above-mentioned factors (15). On neighboring hybridization to the target DNA, an excited FRET probe system generates long-wavelength fluorescent emission that can be read by a Roche LightCycler instrument.
A suitable pair of probes was identified based on empirical design guidelines (16) and use of Primer Express software (Applied Biosystems) for calculation of melting points. The binding site for the upstream probe was the same as for the probe used in the Roche Amplicor assay, whereas the adjacent downstream probe hybridized to a region conserved within the rDNA genes of all mycobacteria. After optimization of PCR, the probes gave specific fluorescence for MTC when combined with the Amplicor primers in the LightCycler. However, the sensitivity of the LightCycler protocol was inferior to that of a conventional thermal cycler protocol using the same reagent formulation (detectability of DNA in a limiting-dilution series reduced by a factor of 100).
In earlier experiments we had observed inefficient amplification of DNA fragments longer than 500 bp in LightCycler glass reaction capillaries (our unpublished data). We therefore adjusted the test to an Applied Biosystems 7700 SDS instrument that uses polypropylene reaction tubes similar to a conventional cycler. Because the long (640 nm) wavelength of FRET probes usually cannot be processed in this instrument, a spectral calibration run was performed according to the instrument manual with four replicate reaction vials containing 1 µM probe MTBP3 (carrying LCRed640, a dye emitting fluorescence at 640 nm) in 50 µL of 1x PCR buffer.
After reoptimization of the amplification protocol, a 50-µL reaction volume contained 20 µL of nucleic acids extract, 5 µL of 10x PCR buffer II, 200 µM each of the deoxynucleotide triphosphates, 4 mM MgCl2, 200 nM each of primers KY18 [5'-cacatgcaagtcgaacggaaagg-3' (1)] and KY75 [5'-gcccgtatcgcccgcacgctcaca-3' (1)], 100 nM probe MTBP5 (5'-accggataggaccacgggatgcatgtctt-3'; 3'-labeled with 6-carboxyfluorescein), 80 nM probe MTBP3 (5'-ggtggaaagcgctttagcggtgt-3', 5'-labeled with LCred640 and 3'-phosphorylated), and 1.25 U of AmpliTaq Gold DNA polymerase (all reagents from Applied Biosystems). Thermal cycling was as follows: 15 min at 95 °C, 50 cycles of 15 s at 95 °C, 20 s at 60 °C, and 40 s at 72 °C. The signal of LCRed640 was read at 60 °C and divided by the signal for 6-carboxyfluorescein (495 nm) for normalization.
The efficiency of this protocol in detecting M. tuberculosis DNA was assessed by amplifying various defined amounts of a photometrically quantified plasmid containing the cloned target region of the assay (for details, refer to the Data Supplement that accompanies the online version of this Technical Brief at http://www.clinchem.org/content/vol49/issue10/). The results were subjected to probit regression analysis using the Statgraphics 5.0 software package (Statistical Graphics) with default settings. A mean of 2.32 plasmids per PCR were calculated to yield a positive result with 95% probability. According to the Poisson distribution formula [P(a) = e-m(ma/a!), with P being the probability of a occurrences per test at a mean of m objects per volume unit; set P = 0.05, a = 0], 
2.99 plasmids would theoretically be required to achieve 95% positivity in a 100% efficient test. We therefore assumed a 100% efficiency of DNA amplification in our optimized PCR protocol for further calculations.
Mycobacterial DNA was extracted from sputum samples or tissue biopsies by proteinase K digestion and lysozyme treatment in combination with the Puregene Blood Kit (Gentra Systems). In brief, 2-mm3 pieces of skin biopsies or, alternatively, bacteria pelleted from 1 mL of sputum samples treated with of 25 mL/L NaOH–N-acetylcysteine (17), were incubated overnight at 55 °C in 300 µL of buffer CLS (included in the reagent set) containing a final proteinase K concentration of 300 mg/L (Sigma). After the samples were heated to 95 °C for 15 min and cooled to room temperature, egg-white lysozyme (Sigma) was added to a final concentration of 250 mg/L and incubated for 1 h at 37 °C. The further procedure was according to the manufacturer’s instructions. Pelleted DNA was resuspended in 200 µL of Tris-EDTA buffer.
Compared with nearly cell-free specimens such as cerebrospinal fluid, urine, or decontaminated sputum samples, MTC DNA extracted from biopsies would contain an interfering background of tissue DNA. The preparation procedure was therefore tested with human skin biopsies (3 mm3 in size) to which we added during the lysis step photometrically counted stock suspensions of cultured M. tuberculosis. Samples containing different amounts of bacteria were each extracted three times in parallel; the obtained DNA solutions were then analyzed in five parallel PCR reactions each (15 reactions per concentration). Probit analysis of these data (see the online Data Supplement) showed that a positive result could be expected in 95% of tests at a mean copy number of 7.36 copies/assay. Knowing that, theoretically, 2.99 bacteria per test would have achieved the same positivity rate if their DNA had been liberated with 100% efficiency in the purification procedure and that the recovery of liberated DNA is 100% efficient in our procedure (as determined above using pure plasmids), the liberation of DNA from mycobacteria would be 41% efficient by relating these numbers (2.99/7.36).
The analytical detection limit of 
7.36 mycobacteria/test was in the same range as that of the commercial Roche Amplicor assay [20 mycobacteria/test (3)]. The sensitivity was further assessed by use of a small collective of samples from 20 inpatients of the Bernhard-Nocht Institute treated for pulmonary or extrapulmonary tuberculosis. Twelve, 13, and 10 of 20 samples were positive by PCR, culture, and Ziel-Neelsen-stained smear, respectively. The resulting sensitivity of real-time PCR in culture positive samples was 92% (95% confidence interval, 66–99%; Table 1
). Similar values have been found in large studies comparing the Amplicor PCR with culture and microscopy (3). To directly compare the clinical sensitivity with that of the Roche Amplicor assay, we tested 29 stored clinical samples in both assays. Instructions for the Amplicor assay were strictly followed. The results were equivalent, as shown in Table 1
. To confirm the specificity of our test, we tested strains of M. africanum, M. bovis, M. leprae, M. ulcerans, M. chelonae, M. smegmatis, M. gordonae, and M. avium (one suspension of cultured bacteria each, pretested with a pan-Mycobacterium PCR). Fluorescent signals were obtained with M. africanum and M. bovis, but not with any other strain.
|
Overall expenses for reagents and consumables are US $5–6 per sample, as opposed to approximately US $30 in the Roche Amplicor assay. Hands-on time in our assay is the same as with the Roche test, but the time required for thermal cycling and product analysis is considerably reduced (2.5 h in our assay compared with 5 h in the Roche assay).
Real-time PCR has already been shown to be appropriate for testing for MTC (8)(9)(10)(11), but all published assays use novel primers that have not been studied as extensively as the primers used here. Because the analytical sensitivity of our test was in the same range as that of the Roche Amplicor assay and the clinical sensitivity was equivalent, we believe that it will also yield a comparable clinical sensitivity when applied to larger patient collectives. Reconfiguration for reading out FRET probes extends the capabilities of the ABI 7700, an instrument that is in operation in many laboratories. The novel approach of using these probes on such an instrument may also be useful in similar settings in which a compatible 5'-nuclease probe cannot be combined with established primers. We believe that this low-cost approach could enable experimental studies and diagnostics in resource-limited settings.
Acknowledgments
We are grateful to Gisela Bretzel and Elvira Richter for providing cultured and clinical mycobacteria samples. This work was funded by the German Ministry of Health.
References
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  S. Burggraf, U. Reischl, N. Malik, M. Bollwein, L. Naumann, and B. Olgemoller Comparison of an Internally Controlled, Large-Volume LightCycler Assay for Detection of Mycobacterium tuberculosis in Clinical Samples with the COBAS AMPLICOR Assay J. Clin. Microbiol., April 1, 2005; 43(4): 1564 - 1569. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
