Abstract
We have developed a high-resolution melting (HRM) assay to scan for mutations in the rpoB, inhA, ahpC, and katG genes and/or promoter regions for the detection of rifampin and isoniazid resistance in Mycobacterium tuberculosis. For assay development, 23 drug-resistant isolates of M. tuberculosis having 29 different mutations, together with 40 drug-susceptible isolates, were utilized. All 29 mutations were accurately detected by our assay. We further validated the assay with a series of 59 samples tested in a blind manner. All sequence alterations that were within the regions targeted by the HRM assay were correctly identified. Compared against results of DNA sequencing, the sensitivity and specificity of our HRM assay were 100%. For the blinded samples, the specificities and sensitivities were 89.3% and 100%, respectively, for detecting rifampin resistance and 98.1% and 83.3%, respectively, for detecting isoniazid resistance, as isolates with mutations in regions not encompassed by our assay were not detected. A C-to-T sequence alteration at position −15 of the ahpC regulatory region, which was previously reported to be associated with isoniazid resistance, may possibly be a polymorphism, as it was detected in an isoniazid-susceptible M. tuberculosis isolate. HRM is a rapid, accurate, simple, closed-tube, and low-cost method. It is thus an ideal assay to be used in countries with a high prevalence of drug-resistant M. tuberculosis and where cost-effectiveness is essential. As a mutation-scanning assay for detecting drug-resistant M. tuberculosis, it can potentially lead to better treatment outcomes resulting from earlier treatment with the appropriate antibiotics.
The emergence of multidrug-resistant tuberculosis (MDR-TB) and extensively drug-resistant TB (XDR-TB) has hampered the control and treatment of TB (45). MDR-TB is defined as TB that is resistant to at least isoniazid (INH) and rifampin (RIF), two main first-line antitubercular drugs, while XDR-TB is MDR-TB that is additionally resistant to three or more second-line drugs. MDR-TB accounts for an estimated 5% of all TB cases (44); however, patients are often not expeditiously diagnosed, resulting in the delay of appropriate treatment as well as poorer treatment outcomes for patients and the propagation and spread of MDR-TB. Conventional methods for drug susceptibility testing of MDR-TB require an additional culture period, typically between 2 and 5 weeks. An easy-to-implement, cost-effective, and rapid method for drug susceptibility testing is thus of paramount importance to limit the spread of drug-resistant tuberculosis.
Drug resistance in Mycobacterium tuberculosis is due to mutations in genes or promoters of genes activating the drug or encoding the drug targets, which are detectable in the majority of drug-resistant isolates (41). Mutations associated with RIF resistance occur mainly in an 81-bp RIF resistance-determining region (RRDR) of the rpoB gene (codons 507 to 533; numbering according to the Escherichia coli rpoB sequence), with >95% of RIF-resistant isolates containing at least one mutation in this region (12, 13, 22, 28, 31, 36, 43). Mutations associated with INH resistance occur mainly in the katG gene (codon 315), the inhA gene and regulatory region, and the ahpC regulatory region (11, 20, 27, 29, 34, 40, 42).
While several molecular methods have been previously described for drug susceptibility testing of M. tuberculosis (2, 6, 7, 9, 26, 38), the cost and technical demands of the assays restrict their usage, especially in countries where funds are scarce. Another problem with the majority of PCR-based molecular methods is the requirement of downstream processing of PCR products, which exposes the PCR products to the environment, increasing the risk of cross-contamination of subsequent assays.
The high-resolution melting (HRM) analysis is a simple, cost-effective, closed-tube method with sensitivity and specificity reported to be higher than those of denaturing high-performance liquid chromatography (dHPLC) (3). HRM does not require the use of costly fluorescent probes and requires no post-PCR handling, making it an attractive alternative method for genotypic drug susceptibility testing of M. tuberculosis. The method involves performing a PCR with a saturating double-stranded DNA-binding dye such as Syto9, followed by a high-resolution melt analysis, whereby the amplicons are slowly heated to denaturation with real-time monitoring of the decrease in fluorescence during denaturation. By comparing the melting profile of the sample with a reference, any sequence variance can be detected. Homoduplexes are usually detected by a change in melting temperature (Tm), while heteroduplexes are usually detected by a change in the melt curve shape (24). As it is easier to identify a change in melt curve shape (10), the sample and reference DNA can be mixed together and amplified together by PCR to produce heteroduplexes, as in the method we have developed.
However, as HRM analysis detects all mutations within the PCR amplicon, known polymorphisms that lie within the amplicon can be excluded by the use of an unlabeled oligonucleotide probe as described by Zhou et al. (48). Briefly, a 3′-blocked unlabeled probe, designed to encompass the polymorphism, is included in the PCR.
In this study, we evaluated the efficacy of our assay in comparison with that of standard drug susceptibility testing for the detection of RIF- and/or INH-resistant M. tuberculosis strains from clinical specimens.
MATERIALS AND METHODS
M. tuberculosis clinical isolates and drug susceptibility testing.
A total of 23 drug-resistant and 40 drug-susceptible clinical isolates of M. tuberculosis used for the initial assay development were collected from the Central Tuberculosis Laboratory, Department of Pathology, Singapore General Hospital, and have been described previously (21). Phenotypic drug susceptibility testing was done using the BACTEC 460 system (Becton Dickinson, Towson, MD) with 2 μg/ml of RIF and 0.1 μg/ml of INH. The 59 clinical isolates of M. tuberculosis used for the screening in a blind manner were obtained from Hong Kong, and the agar proportion method was used for susceptibility testing for RIF and INH, as described previously (30).
DNA extraction.
DNA from the clinical isolates of M. tuberculosis used for the initial assay development was extracted from bacterial colonies grown on Löwenstein-Jensen slants as described previously (8). DNA from the clinical isolates of M. tuberculosis used for the blinded screening was extracted as described previously (46) and purified using phenol-chloroform-isoamyl alcohol (25:24:1) (Invitrogen). All DNA samples were quantified using the Nanodrop 1000 (Thermo Scientific, Waltham, MA).
Primers and probes.
Seven pairs of primers were designed to flank regions that were previously reported to be associated with RIF and INH resistance in M. tuberculosis. For rpoB, primers were designed to flank the RRDR region. For the mabA promoter, primers were designed to include the promoter, starting from 121 bp upstream of mabA. For inhA, three overlapping sets of primers were designed to detect the majority of mutations, including the 5′ end of the gene (including codon 110). For the ahpC promoter, the promoter was included starting from 144 bp upstream from the ahpC transcription start site. However, as there is a known polymorphism at position −46 of ahpC (11), we designed an additional unlabeled oligonucleotide probe (3′ blocked by an inverted deoxyribosylthymine [dT]) to specifically detect this polymorphism (48). All primer and probe sequences are listed in Table 1. Amplicon lengths were kept below 200 bp, as shorter amplicons produce melt curves of altered shapes that are easier to identify when there is a sequence variant (10).
TABLE 1.
Primer and probe sequences used for PCR
Primer namea | Primer/probe sequence | Amplicon size (bp) | Annealing temp (°C) | HRM temp range (°C) | Nucleotide positionsb |
---|---|---|---|---|---|
katG_F | 5′-CTGGAGCAGATGGGCTTG-3′ | 195 | 66 | 84 to 91 | 877 to 894 |
katG_R | 5′-GTCCTTGGCGGTGTATTGC-3′ | 1071 to 1053 | |||
mabA-Prom_F | 5′-ACACCGACAAACGTCACGAG-3′ | 143 | 67 | 85 to 92 | −121 to −102 |
mabA-Prom_R | 5′-CTTCAGTGGCTGTGGCAGTC-3′ | 22 to 3 | |||
inhA-Frag1_F | 5′-GCCACTGACACAACACAAGG-3′ | 153 | 63 | 83 to 92 | −26 to −7 |
inhA-Frag1_R | 5′-GGTCGAACCCGGTGAGC-3′ | 127 to 111 | |||
inhA-Frag2_F | 5′-CTGGTGCTCACCGGGTTC-3′ | 167 | 63 | 88 to 95 | 106 to 123 |
inhA-Frag2_R | 5′-ACCCCGTCGAGCTTGTTG-3′ | 272 to 255 | |||
inhA-Frag3_F | 5′-CGAACTCGACGTGCAAAAC-3′ | 148 | 62 | 85 to 94 | 183 to 201 |
inhA-Frag3_R | 5′-GTCGAAGAACGGGTTGATG-3′ | 330 to 312 | |||
ahpC-Prom_F | 5′-ACTGCTGAACCACTGCTTTG-3′ | 177 | 60 | 81 to 94 | −144 to −125 |
ahpC-Prom_R | 5′-CATCATCAAAGCGGACAATG-3′ | 33 to 14 | |||
ahpC_Probe-InvdT | 5′-GATGCCGATAAATATGGTGTGATATAT-3′ | 53 to 65 | −66 to −40 | ||
rpoB_F | 5′-CCGCGATCAAGGAGTTCTTC-3′ | 129 | 67 | 84 to 93 | 1256 to 1275 |
rpoB_R | 5′-CACGCTCACGTGACAGACC-3′ | 1384 to 1366 |
F, forward; R, reverse.
Nucleotide position is relative to the transcriptional start site of each gene.
Real-time PCR and high-resolution melting (HRM).
Prior to performing the HRM assay, a preliminary real-time PCR was done to ensure that the DNA concentration of each sample was correctly quantified. Real-time PCR was performed with the katG primer set, the assay that has the highest stringency for detecting degraded samples as it has the largest amplicon size (degraded nucleic acids, being shorter in length, may be amplifiable by a PCR assay which amplifies a short region but not by one which amplifies a longer region). No reference DNA was added, and 0.4 ng of sample DNA was used. Samples with threshold cycle (CT) values varying more than 2 cycles compared to a DNA sample of known high integrity were removed.
For the HRM assay, a PCR was performed in 10-μl reaction mixtures containing 0.2 ng sample DNA, 0.2 ng reference DNA (from Mycobacterium tuberculosis H37Rv), and 1× PCR buffer containing 1.5 mM MgCl2, 200 nM deoxynucleoside triphosphates (dNTPs), 200 nM each primer, 1.5 μM Syto9 (Molecular Probes, Eugene, OR), 0.5 U of HotStarTaq polymerase (Qiagen, Hilden, Germany), and 4 μl of mineral oil (Sigma Aldrich, St. Louis, MO). Mineral oil is essential for preventing any evaporation or condensation and to increase the accuracy of the assay.
The PCR for the probe-based assay for ahpC (used to detect a known polymorphism) was performed in 10-μl reaction mixtures containing 0.2 ng sample DNA, 0.2 ng reference DNA, and 1× PCR buffer containing 1.5 mM MgCl2, 200 nM dNTPs, 500 nM forward primer, 50 nM reverse primer, 500 nM probe, 300 nM 1.5 μM Syto9, 0.5 U of HotStarTaq polymerase, and 4 μl of mineral oil.
All PCRs were performed in duplicate. The PCR cycling and HRM analysis were performed on the Rotor-Gene 6000 (Corbett Research, Sydney, Australia). PCR cycling parameters were as follows: 95°C for 15 min, 40 cycles at 95°C for 20 s, and at the appropriate annealing temperature (Table 1) for 30 s. The melt curve was generated by heating, using the temperature ranges indicated in Table 1, at increments of 0.1°C/s, except for the ahpC probe, which had an additional melt at increments of 0.5°C/s in order to detect polymorphisms within the region encompassed by the ahpC probe. These melt curve parameters were optimized experimentally, and the parameters used are the ones that worked best to discriminate the mutants. The HRM curve was analyzed using the Rotor-Gene 1.7.87 software and the HRM algorithm included in the software, and the melt curves were normalized using the software and following the software instructions.
DNA sequencing.
The PCR products were sequenced as described previously (16, 21-23) or using the Applied Biosystems 3130XL genetic analyzer (Foster City, CA). The clinical isolates of M. tuberculosis used for the blinded screening were sequenced for katG, mabA-inhA, and rpoB genes and promoter regions using an Applied Biosystems 3130 genetic analyzer using the primer sequences listed in Table 2.
TABLE 2.
Primer sequences used for sequencing of a blinded series of 59 M. tuberculosis clinical isolates
Gene (Rv no.) | GeneID | Primer name | Lengtha | Tm | %GCb | Directionc | Sequence | Nucleotide positions |
---|---|---|---|---|---|---|---|---|
katG (Rv1908c) | 885638 | katG-pcrF | 23 | 58.2 | 42 | F | GATTGTTCGATATCCGACACTTC | −76 to −54 |
katG-1F | 21 | 59.78 | 52 | F | AGCAACACCCACCCATTACAG | 8 to 28 | ||
katG-2F | 18 | 58.96 | 67 | F | GGCACCTACCGCATCCAC | 331 to 348 | ||
katG-3R | 19 | 59.63 | 53 | R | AAAACAATCAGGTCCGCCC | 500 to 482 | ||
katG-4F | 18 | 59.74 | 61 | F | CGTCGGCGGTCACACTTT | 798 to 815 | ||
katG-5R | 18 | 59.47 | 61 | R | GGTTCCGGTGCCATACGA | 924 to 907 | ||
katG-6F | 20 | 59.11 | 55 | F | GTCCCGTTGCGAGATACCTT | 1262 to 1281 | ||
katG-7R | 22 | 60.52 | 50 | R | TGTGAGACAGTCAATCCCGATG | 1412 to 1391 | ||
katG-8F | 18 | 60.42 | 61 | F | CGCAGGAACAAACCGACG | 1724 to 1741 | ||
katG-9R | 20 | 59.02 | 50 | R | TGCCCTTTCCGAGGTAGTTT | 1805 to 1785 | ||
katG-10R | 18 | 59.48 | 56 | R | GCCGATCAACCCGAATCA | 15 to −2221 | ||
mabA-inhA (Rv1483- Rv1484) | 886551-886523 | inhA-pcrF | 21 | 60.15 | 52 | F | TTCGTAGGGCGTCAATACACC | mabA, −200 to −180 |
inhA-2F | 19 | 59.87 | 58 | F | TCACGGCGGTAGAAGAGCA | mabA, 215 to 233 | ||
inhA-3R | 18 | 59.83 | 67 | R | GCGTTGGACACCAGCACC | mabA, 266 to 249 | ||
inhA-4F | 18 | 59.49 | 56 | F | TTTATCCCAGCGAAGCGG | mabA, 613 to 630 | ||
inhA-5R | 18 | 60.32 | 67 | R | CCTCGGAAGCCAGGAAGC | mabA, 682 to 655 | ||
inhA-6F | 17 | 58.23 | 65 | F | GTTCATGCCGCAGACCG | inhA, 288 to 303 | ||
inhA-7R | 21 | 60.25 | 48 | R | AAGCATACGAATACGCCGAGA | inhA, 385 to 365 | ||
inhA-pcrR | 17 | 59.91 | 71 | R | GGTCCGCCGAACGACAG | inhA, 46 to 30 | ||
rpoB (Rv0667) | 888164 | rpoB-pcrF | 22 | 63.09 | 55 | F | CTTTGTTCGTGGTGAGCGTGAG | −565 to −544 |
rpoB-2F | 18 | 60.53 | 61 | F | TGCGGCTCAGCGGTTTAG | −147 to −130 | ||
rpoB-3R | 19 | 61.5 | 53 | R | GCCGCGATAATTTTGTCGG | −70 to −88 | ||
rpoB-4F | 22 | 62.22 | 55 | F | GTCGACGAGTGCAAAGACAAGG | 337 to 358 | ||
rpoB-5R | 22 | 61.45 | 55 | R | GAAGTCACCCATGAACACCGTC | 453 to 432 | ||
rpoB-6F | 16 | 60.29 | 69 | F | CGAGCCCCCGACCAAA | 834 to 849 | ||
rpoB-7R | 18 | 60.35 | 61 | R | TGCAGCCCGAGCTTCTTG | 950 to 933 | ||
rpoB-8F | 18 | 61.06 | 61 | F | CCCGATCGAAACCCCTGA | 1434 to 1451 | ||
rpoB-9R | 18 | 62.42 | 67 | R | GCGAGCCGATCAGACCGA | 1480 to 1463 | ||
rpoB-10F | 19 | 61.51 | 58 | F | GTGATGCACGACAACGGCA | 1960 to 1978 | ||
rpoB-11R | 18 | 59.6 | 67 | R | CGCATCCGGTAGGTACGC | 2000 to 1983 | ||
rpoB-12F | 19 | 61.04 | 68 | F | GGTGAGACCGAGCTGACGC | 2410 to 2428 | ||
rpoB-13R | 19 | 61.04 | 68 | R | GCGTCAGCTCGGTCTCACC | 2428 to 2410 | ||
rpoB-14F | 20 | 64.72 | 60 | F | ACGGCAAGGCCATGCTCTTC | 2996 to 3015 | ||
rpoB-15R | 19 | 59.41 | 63 | R | TGTAGGCAGCACCGTAGGC | 3232 to 3214 | ||
rpoB-pcrR | 22 | 63.62 | 50 | R | GTTGATCGTCTCCGGCTTTTTG | 149 to 128 |
Length, number of nucleotides.
%GC, number of G's and C's in the primer as a percentage of the total number of nucleotides.
F, forward; R, reverse.
RESULTS
Validation of the HRM assay with reference strains.
For the initial assay development, we tested the assay for 29 different mutations within the mabA promoter (n = 3), ahpC promoter (n = 6), katG (n = 6), inhA (n = 2), and rpoB (n = 12) from 23 drug-resistant M. tuberculosis isolates (Table 3). All 29 mutations were detected by our HRM assay. We also analyzed 40 drug-susceptible isolates to check for false-positive results. All 40 isolates were correctly identified as wild type. The normalized melt curves of PCR products of isolates with different mutations are shown in Fig. 1. Visually, samples with mutations (colored lines) are easily differentiated from the wild type (black lines) by the distinct differences in the shape of the melt curves.
TABLE 3.
Reference strains with known mutations used for assay development
Sample | Gene | Nucleotide change(s) | Amino acid changea |
---|---|---|---|
IE1 | katG | TGG→TAG | Trp300Stop |
IS28 | katG | AGC→AGG | Ser302Arg |
I86 | katG | AGC→ACC | Ser315Thr |
I92 | katG | AGC→AAC | Ser315Asn |
IR4 | katG | AGC→ACC, ACG→ACA | Ser315Thr, Thr344Thr |
IE5 | katG | CTG→CGG | Leu336Arg |
IRS6 | mabA | −17 G→T | NA |
ISD3 | mabA | −8 T→A | NA |
IRS7 | mabA | −15 C→T | NA |
IR27 | inhA | GGA→GGC | Gly3Gly |
IRS7 | inhA | ATC→ACC | Ile21Thr |
IR27 | ahpC | −6 G→A | NA |
IR5 | ahpC | −12 C→T | NA |
IRS12 | ahpC | −30 C→T | NA |
IS3 | ahpC | −46 G→A | NA |
IRS6 | ahpC | −55 A→G | NA |
I86 | ahpC | −81 G→A | NA |
R3 | rpoB | CAA→GAA | Gln513Glu |
IRE6 | rpoB | TTC→CTC, GAC→GTC | Phe514Leu, Asp516Val |
IR12 | rpoB | GAC→GTC | Asp516Val |
MDR8 | rpoB | AAC deletion | Asn518 deletion |
IR25 | rpoB | TCG→TTG | Ser522Leu |
IR27 | rpoB | CAC→TAC | His526Tyr |
IR7 | rpoB | CAC→GAC | His526Asp |
R1 | rpoB | CAC→CGC | His526Arg |
R2 | rpoB | CAC→CTC | His526Leu |
MDR6 | rpoB | TCG→TTG | Ser531Leu |
IRS12 | rpoB | TCG→TGG | Ser531Trp |
IRE3 | rpoB | TCG→ATG | Ser531Met |
NA, not applicable.
FIG. 1.
High-resolution melt curves of katG (A), the mabA promoter (B), inhA (C), the ahpC promoter (D), the ahpC promoter (probe) (E), and rpoB (F), demonstrating the change in melt curve shape caused by mutations. Wild-type samples are shown in black, and samples with mutations are shown in color. Experiments were performed in duplicate.
Validation of the HRM assay with a blinded series of strains.
To further validate the assay, a series of 59 blinded samples was used to assess the sensitivity and specificity of the assay (Tables 4 and 5). Of the 28 RIF-resistant isolates in our blinded samples, 25 (89.3%) were detected to have mutations in rpoB. The three RIF-resistant isolates that were detected as wild type had mutations at I572F (two isolates) and at V169F (one isolate). These two mutations have been shown to associate with rifampin resistance upon transformation into Mycobacterium smegmatis (unpublished data). Both of these mutations are not within the RRDR region and were thus not detected by our HRM assay. All 31 RIF-susceptible isolates were correctly typed.
TABLE 4.
HRM screening of a blinded series of 59 Mycobacterium tuberculosis clinical isolates for rifampin and isoniazid resistance
Sample | Mutation statusa |
HRM interpretation of drug resistanceb |
|||||||
---|---|---|---|---|---|---|---|---|---|
rpoB | katG | mabA promoter | inhA fragment 1 | inhA fragment 2 | inhA fragment 3 | ahpC promoter | Rifampin | Isoniazid | |
1 | NM | NM | M | NM | NM | NM | NM | S | R |
2 | NM | M | NM | NM | NM | NM | NM | S | R |
3 | NM | NM (Q434STOP) | NM | NM | NM | NM | M | S | R |
4 | NM | NM (N655D) | NM | NM | NM | NM | M | S | R |
5 | NM | NM (Y98C) | NM | NM | NM | NM | NM | S | S (R) |
6 | NM | NM | M | NM | NM | NM | NM | S | R |
7 | NM | NM | NM | M | NM | NM | M | S | R |
8 | NM | M | NM | NM | NM | NM | NM | S | R |
9 | NM | M | NM | NM | NM | NM | NM | S | R |
10 | NM | NM | M | NM | NM | NM | NM | S | R |
11 | M | M | NM | NM | NM | NM | NM | R | R |
12 | M | NM | M | M | NM | NM | NM | R | R |
13 | M | NM | NM | NM | NM | NM | M | R | R |
14 | M | NM | M | NM | NM | NM | NM | R | R |
15 | M | M | NM | NM | NM | NM | NM | R | R |
16 | M | M | NM | NM | NM | NM | NM | R | R |
17 | M | NM | M | NM | NM | NM | NM | R | R |
18 | M | NM | M | NM | NM | NM | NM | R | R |
19 | M | M | NM | NM | NM | NM | NM | R | R |
20 | M | M | NM | NM | NM | NM | NM | R | R |
21 | M | M | NM | NM | NM | NM | NM | R | R |
22 | NM (I572F) | NM | M | NM | NM | NM | NM | S (R) | R |
23 | NM (V169F) | M | NM | NM | NM | NM | NM | S (R) | R |
24 | NM | NM | NM | NM | NM | NM | NM | S | S |
25 | NM | NM | NM | NM | NM | NM | M | S | R |
26 | NM | M | NM | NM | NM | NM | M | S | R |
27 | NM | NM | NM | NM | NM | NM | NM | S | S |
28 | NM | M | NM | NM | NM | NM | NM | S | R |
29 | NM | NM | M | NM | NM | NM | NM | S | R |
30 | NM | M | NM | NM | NM | NM | NM | S | R |
31 | NM | M | NM | NM | NM | NM | NM | S | R |
32 | NM | NM | M | NM | NM | NM | M | S | R |
33 | NM | M | NM | NM | NM | NM | NM | S | R |
34 | NM | NM | M | NM | NM | NM | NM | S | R |
35 | NM | NM | M | NM | NM | NM | NM | S | R |
36 | NM | NM | M | NM | NM | NM | NM | S | R |
37 | NM | NM | M | NM | NM | NM | M | S | R |
38 | NM | NM | M | NM | NM | NM | NM | S | R |
39 | M | NM | M | NM | NM | NM | NM | R | R |
40 | M | NM | M | NM | NM | NM | NM | R | R |
41 | M | NM | M | NM | NM | NM | NM | R | R |
42 | M | M | NM | NM | NM | NM | NM | R | R |
43 | M | M | NM | NM | NM | NM | NM | R | R |
44 | M | M | NM | NM | NM | NM | M | R | R |
45 | M | M | NM | NM | NM | NM | NM | R | R |
46 | M | NM (A478DEL) | NM | NM | NM | NM | M | R | R |
47 | M | NM | M | NM | NM | NM | NM | R | R |
48 | M | NM | NM | NM | NM | NM | M (P) | R | R (S) |
49 | NM | NM | NM | NM | NM | NM | NM | S | S |
50 | NM | NM | NM | NM | NM | NM | NM | S | S |
51 | NM | M | NM | NM | NM | NM | NM | S | R |
52 | NM | NM | M | NM | NM | NM | NM | S | R |
53 | M | M | NM | NM | NM | NM | NM | R | R |
54 | M | NM | M | M | NM | NM | NM | R | R |
55 | NM | NM | NM | M | NM | NM | M | S | R |
56 | M | M | NM | NM | NM | NM | M | R | R |
57 | M | M | NM | NM | NM | NM | NM | R | R |
58 | NM (I572F) | NM | M | NM | NM | NM | NM | S (R) | R |
59 | NM | NM | NM | NM | NM | NM | NM | S | S |
NM, no mutation; M, mutation; P, polymorphism. Discrepant results between HRM screening and DNA sequencing are shown in bold. The mutations detected by sequencing are in parentheses.
S, susceptible; R, resistant. Discrepant results between HRM screening and phenotypic drug susceptibility testing are shown in bold. The phenotypes detected by drug susceptibility testing are shown in parentheses.
TABLE 5.
Sensitivity and specificity of the high-resolution melting (HRM) assay in comparison with drug susceptibility testing
Drug susceptibility | No. of isolates (n = 59): |
Sensitivitya (95% CIb) | Specificityc (95% CI) | |
---|---|---|---|---|
Mutation positive by HRM | Mutation negative by HRM | |||
Rifampin | 25/28, 89.3% (72.0-97.1) | 31/31, 100% (90.4-100) | ||
Resistant | 25 | 3 | ||
Susceptible | 0 | 31 | ||
Isoniazid | 52/53, 98.1% (89.1-100) | 5/6, 83.3% (41.8-98.9) | ||
Resistant | 52 | 1 | ||
Susceptible | 1 | 5 |
[Number of drug-resistant isolates with mutations]/[number of drug-resistant isolates with mutations + number of drug-resistant isolates without mutation].
Statistical calculations were performed with the free software available from http://www.measuringusability.com/wald.htm using the Adjusted Wald method.
[Number of drug-susceptible isolates without mutations]/[number of drug-susceptible isolates with mutations + number of drug-susceptible isolates without mutations].
Of the 53 INH-resistant isolates in our blinded samples, 52 (98.1%) were detected to have mutations in katG, the mabA promoter, inhA, and/or ahpC. The isolate that was detected as wild type had a Y98C mutation in katG, which is not within the region covered by our assay. Five of six (83.3%) INH-susceptible isolates were correctly typed. Sample 48, although phenotypically INH susceptible, was found to contain a C-to-T sequence alteration at position −15 of the ahpC regulatory region. Thus, our results suggest that this sequence alteration, which was previously reported to be associated with INH resistance (11, 39), may possibly be a polymorphism.
Three INH-resistant isolates (sample 3, 4, and 46), although containing mutations in katG in regions that are not covered by our assay, were still correctly interpreted as INH resistant due to a concurrent mutation in the ahpC promoter (Table 4).
DISCUSSION
We have developed and evaluated a HRM-based assay for RIF and INH susceptibility testing of M. tuberculosis. This assay allows the detection of mutations that are commonly associated with RIF and INH resistance in five regions: RRDR of rpoB for RIF resistance and the mabA and ahpC promoters, katG, and inhA for INH resistance.
All 29 (100%) known mutations within the RIF- and INH-resistant strains used for assay development had clearly distinguishable melt curves using the HRM assay described in this study. When our assay was validated with a blinded series of samples, RIF resistance was detected with a sensitivity of 89.3% and a specificity of 100%. However, if the three RIF-resistant isolates with mutations not within the RRDR region were to be excluded, the sensitivity and specificity of our assay would be 100%. Since more than 95% of RIF resistance-associated mutations lie within the RRDR region (19, 35), it may not be necessary to screen outside the RRDR region, as this may increase the number of primer sets required for the assay and also the possibility of detecting polymorphisms, giving rise to false positives.
When evaluated for the detection of INH resistance in a blinded series of isolates, our assay performed with a sensitivity of 98.1% and a specificity of 83.3% (detecting five of six susceptible isolates). Four isolates had mutations in katG in regions not encompassed by our assay. The phenotypes of three of these isolates were, however, still interpreted correctly, as another concurrent mutation in the ahpC promoter was detected. Since the majority of INH resistance-associated mutations in katG are within codon 315 (11, 27, 40), codon 315 has been commonly used as the only target in katG for genotypic INH susceptibility assays, in research studies and commercially available kits (2, 15, 18, 33, 47), and thus it may not be necessary to screen other regions of katG.
A key advantage of genotypic drug susceptibility assays over phenotypic assays is the shorter time required for the assay, with genotypic assays requiring just hours to complete in contrast to phenotypic tests that can take weeks. Due to the need to grow the organism, phenotypic methods require weeks of culture, during which the patient may be treated with the wrong antibiotics, resulting in poorer treatment outcome or the transmission of resistant strains. As genotypic tests do not require additional culture beyond that required for the initial isolation, there will also be fewer biohazard-related risks. In addition, genotypic tests may also be able to provide information on drug susceptibility in cases in which the phenotypic assay is indeterminate due to primary culture contamination.
In recent years, numerous genotypic assays (2, 6, 7, 9, 26, 38), including several commercial kits (1, 14, 15, 25), have been described for the detection of drug resistance in M. tuberculosis. Although rapid, most of them are too costly, labor-intensive, or technically demanding. Another major problem faced with the use of these assays, including the commercial kits, is the risk of cross-contamination with PCR products (44). This is because almost all genotypic drug susceptibility tests involve PCR, followed by the subsequent processing of the PCR products for the detection of mutations, increasing the potential risk of cross-contamination of subsequent reactions. This is often not a problem for a well-set-up molecular biology laboratory, which will have separate rooms or areas for pre-PCR and post-PCR work as well as skilled molecular biologists with stringent work practices. However, most TB labs in high-burden countries are unfortunately not as well equipped in molecular biology and may have space constraints, which may lead to poor-quality results, thus making implementation of these assays problematic.
In comparison to a previously reported HRM RIF susceptibility testing method (17) as well as other genotypic assays which are not closed tube, as they require the reopening of tubes after PCR to mix the samples and reference DNA together prior to HRM, our method has the advantage of being closed tube, thus lowering the risk of cross-contamination. This in itself makes the HRM assay easier to implement, as there will not be a need to make major changes to the lab. Furthermore, the simplicity of the assay, its low cost (estimated at approximately $0.30 per HRM reaction), and its low level of technical requirements makes it even more appealing as a rapid diagnostic test. In comparison to probe-based assays such as the Genotype MTBDRplus assay (Hain Lifescience, Nehren, Germany), our assay is more comprehensive, as it targets the inhA gene and ahpC promoter in addition to the katG gene and mabA promoter. This increases the sensitivity of our assay by about 4 to 10% in comparison to results from assays targeting solely the katG and the mabA promoter (4, 5, 11, 32, 37). HRM is also able to detect mutations over a much broader region (including novel mutations), making it more feasible for detecting drug resistance-associated mutations that are scattered throughout large regions. With increased coverage, the sensitivity of the assay is correspondingly increased. Each HRM reaction is also highly sensitive, requiring only 0.2 ng of sample DNA.
As all genotypic methods for detecting drug susceptibility rely on the detection of mutations associated with drug resistance, the reliability of genotypic methods ultimately depends on our knowledge of molecular mechanisms for drug resistance. Our current knowledge of molecular mechanisms for drug resistance remains incomplete, especially for second-line antituberculosis drugs, and future research into determining these molecular mechanisms will result in improved genotype-based drug susceptibility assays.
Acknowledgments
We acknowledge the Central Tuberculosis Laboratory, Department of Pathology, Singapore General Hospital, for providing isolates.
This work was supported by a grant from the Biomedical Research Council (BMRC) of Singapore (BMRC 07/1/31/19/511).
Footnotes
Published ahead of print on 17 February 2010.
REFERENCES
- 1.Bang, D., Å. Bengard Andersen, and V. Ø. Thomsen. 2006. Rapid genotypic detection of rifampin- and isoniazid-resistant Mycobacterium tuberculosis directly in clinical specimens. J. Clin. Microbiol. 44:2605-2608. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Bergval, I. L., R. N. Vijzelaar, E. R. Dalla Costa, A. R. Schuitema, L. Oskam, A. L. Kritski, P. R. Klatser, and R. M. Anthony. 2008. Development of multiplex assay for rapid characterization of Mycobacterium tuberculosis. J. Clin. Microbiol. 46:689-699. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Chou, L. S., E. Lyon, and C. T. Wittwer. 2005. A comparison of high-resolution melting analysis with denaturing high-performance liquid chromatography for mutation scanning: cystic fibrosis transmembrane conductance regulator gene as a model. Am. J. Clin. Pathol. 124:330-338. [DOI] [PubMed] [Google Scholar]
- 4.Dalla Costa, E. R., M. O. Ribeiro, M. S. Silva, L. S. Arnold, D. C. Rostirolla, P. I. Cafrune, R. C. Espinoza, M. Palaci, M. A. Telles, V. Ritacco, P. N. Suffys, M. L. Lopes, C. L. Campelo, S. S. Miranda, K. Kremer, P. E. da Silva, S. Fonseca Lde, J. L. Ho, A. L. Kritski, and M. L. Rossetti. 2009. Correlations of mutations in katG, oxyR-ahpC and inhA genes and in vitro susceptibility in Mycobacterium tuberculosis clinical strains segregated by spoligotype families from tuberculosis prevalent countries in South America. BMC Microbiol. 9:39. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Doustdar, F., A. D. Khosravi, P. Farnia, M. R. Masjedi, and A. A. Velayati. 2008. Molecular analysis of isoniazid resistance in different genotypes of Mycobacterium tuberculosis isolates from Iran. Microb. Drug Resist. 14:273-279. [DOI] [PubMed] [Google Scholar]
- 6.Edwards, K. J., L. A. Metherell, M. Yates, and N. A. Saunders. 2001. Detection of rpoB mutations in Mycobacterium tuberculosis by biprobe analysis. J. Clin. Microbiol. 39:3350-3352. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Gegia, M., N. Mdivani, R. E. Mendes, H. Li, M. Akhalaia, J. Han, G. Khechinashvili, and Y. W. Tang. 2008. Prevalence of and molecular basis for tuberculosis drug resistance in the Republic of Georgia: validation of a QIAplex system for detection of drug resistance-related mutations. Antimicrob. Agents Chemother. 52:725-729. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Goyal, M., N. A. Saunders, J. D. van Embden, D. B. Young, and R. J. Shaw. 1997. Differentiation of Mycobacterium tuberculosis isolates by spoligotyping and IS6110 restriction fragment length polymorphism. J. Clin. Microbiol. 35:647-651. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Gryadunov, D., V. Mikhailovich, S. Lapa, N. Roudinskii, M. Donnikov, S. Pan'kov, O. Markova, A. Kuz'min, L. Chernousova, O. Skotnikova, A. Moroz, A. Zasedatelev, and A. Mirzabekov. 2005. Evaluation of hybridisation on oligonucleotide microarrays for analysis of drug-resistant Mycobacterium tuberculosis. Clin. Microbiol. Infect. 11:531-539. [DOI] [PubMed] [Google Scholar]
- 10.Gundry, C. N., J. G. Vandersteen, G. H. Reed, R. J. Pryor, J. Chen, and C. T. Wittwer. 2003. Amplicon melting analysis with labeled primers: a closed-tube method for differentiating homozygotes and heterozygotes. Clin. Chem. 49:396-406. [DOI] [PubMed] [Google Scholar]
- 11.Hazbón, M. H., M. Brimacombe, M. Bobadilla del Valle, M. Cavatore, M. I. Guerrero, M. Varma-Basil, H. Billman-Jacobe, C. Lavender, J. Fyfe, L. García-García, C. I. León, M. Bose, F. Chaves, M. Murray, K. D. Eisenach, J. Sifuentes-Osornio, M. D. Cave, A. Ponce de León, and D. Alland. 2006. Population genetics study of isoniazid resistance mutations and evolution of multidrug-resistant Mycobacterium tuberculosis. Antimicrob. Agents Chemother. 50:2640-2649. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Heep, M., B. Brandstätter, U. Rieger, N. Lehn, E. Richter, S. Rüsch-Gerdes, and S. Niemann. 2001. Frequency of rpoB mutations inside and outside the cluster I region in rifampin-resistant clinical Mycobacterium tuberculosis isolates. J. Clin. Microbiol. 39:107-110. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Herrera, L., S. Jiménez, A. Valverde, M. A. García-Aranda, and J. A. Sáez-Nieto. 2003. Molecular analysis of rifampicin-resistant Mycobacterium tuberculosis isolated in Spain (1996-2001). Description of new mutations in the rpoB gene and review of the literature. Int. J. Antimicrob. Agents 21:403-408. [DOI] [PubMed] [Google Scholar]
- 14.Hillemann, D., S. Rüsch-Gerdes, and E. Richter. 2006. Application of the Genotype MTBDR assay directly on sputum specimens. Int. J. Tuber. Lung Dis. 10:1057-1059. [PubMed] [Google Scholar]
- 15.Hillemann, D., S. Rüsch-Gerdes, and E. Richter. 2007. Evaluation of the GenoType MTBDRplus assay for rifampin and isoniazid susceptibility testing of Mycobacterium tuberculosis strains and clinical specimens. J. Clin. Microbiol. 45:2635-2640. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Ho, Y. M., Y. J. Sun, S. Y. Wong, and A. S. Lee. 2009. Contribution of dfrA and inhA mutations to the detection of isoniazid-resistant Mycobacterium tuberculosis isolates. Antimicrob. Agents Chemother. 53:4010-4012. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Hoek, K. G., N. C. Gey van Pittius, H. Moolman-Smook, K. Carelse-Tofa, A. Jordaan, G. D. van der Spuy, E. Streicher, T. C. Victor, P. D. van Helden, and R. M. Warren. 2008. Fluorometric assay for testing rifampin susceptibility of Mycobacterium tuberculosis complex. J. Clin. Microbiol. 46:1369-1373. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Jalava, J., and H. Marttila. 2004. Application of molecular genetic methods in macrolide, lincosamide and streptogramin resistance diagnostics and in detection of drug-resistant Mycobacterium tuberculosis. APMIS 112:838-855. [DOI] [PubMed] [Google Scholar]
- 19.Jenkins, C. 2005. Rifampicin resistance in tuberculosis outbreak, London, England. Emerg. Infect. Dis. 11:931-934. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Kelley, C. L., D. A. Rouse, and S. L. Morris. 1997. Analysis of ahpC gene mutations in isoniazid-resistant clinical isolates of Mycobacterium tuberculosis. Antimicrob. Agents Chemother. 41:2057-2058. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Lee, A. S., I. H. Lim, L. L. Tang, A. Telenti, and S. Y. Wong. 1999. Contribution of kasA analysis to detection of isoniazid-resistant Mycobacterium tuberculosis in Singapore. Antimicrob. Agents Chemother. 43:2087-2089. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Lee, A. S., I. H. Lim, L. L. Tang, and S. Y. Wong. 2005. High frequency of mutations in the rpoB gene in rifampin-resistant clinical isolates of Mycobacterium tuberculosis from Singapore. J. Clin. Microbiol. 43:2026-2027. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Lee, A. S., L. L. Tang, I. H. Lim, and S. Y. Wong. 2002. Characterization of pyrazinamide and ofloxacin resistance among drug-resistant Mycobacterium tuberculosis isolates from Singapore. Int. J. Infect. Dis. 6:48-51. [DOI] [PubMed] [Google Scholar]
- 24.Liew, M., R. Pryor, R. Palais, C. Meadows, M. Erali, E. Lyon, and C. Wittwer. 2004. Genotyping of single-nucleotide polymorphisms by high-resolution melting of small amplicons. Clin. Chem. 50:1156-1164. [DOI] [PubMed] [Google Scholar]
- 25.Mäkinen, J., H. J. Marttila, M. Marjamaki, M. K. Viljanen, and H. Soini. 2006. Comparison of two commercially available DNA line probe assays for detection of multidrug-resistant Mycobacterium tuberculosis. J. Clin. Microbiol. 44:350-352. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.McCammon, M. T., J. S. Gillette, D. P. Thomas, S. V. Ramaswamy, E. A. Graviss, B. N. Kreiswirth, J. Vijg, and T. N. Quitugua. 2005. Detection of rpoB mutations associated with rifampin resistance in Mycobacterium tuberculosis using denaturing gradient gel electrophoresis. Antimicrob. Agents Chemother. 49:2200-2209. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Mokrousov, I., O. Narvskaya, T. Otten, E. Limeschenko, L. Steklova, and B. Vyshnevskiy. 2002. High prevalence of KatG Ser315Thr substitution among isoniazid-resistant Mycobacterium tuberculosis clinical isolates from northwestern Russia, 1996 to 2001. Antimicrob. Agents Chemother. 46:1417-1424. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Musser, J. M. 1995. Antimicrobial agent resistance in mycobacteria: molecular genetic insights. Clin. Microbiol. Rev. 8:496-514. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Musser, J. M., V. Kapur, D. L. Williams, B. N. Kreiswirth, D. van Soolingen, and J. D. van Embden. 1996. Characterization of the catalase-peroxidase gene (katG) and inhA locus in isoniazid-resistant and -susceptible strains of Mycobacterium tuberculosis by automated DNA sequencing: restricted array of mutations associated with drug resistance. J. Infect. Dis. 173:196-202. [DOI] [PubMed] [Google Scholar]
- 30.National Committee for Clinical Laboratory Standards. 2000. Susceptibility testing of mycobacteria, nocardia and other aerobic actinomycetes. Tentative standard MT24-T2. National Committee for Clinical Laboratory Standards, Wayne, PA. [PubMed]
- 31.Nikolayevsky, V., T. Brown, Y. Balabanova, M. Ruddy, I. Fedorin, and F. Drobniewski. 2004. Detection of mutations associated with isoniazid and rifampin resistance in Mycobacterium tuberculosis isolates from Samara Region, Russian Federation. J. Clin. Microbiol. 42:4498-4502. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Nusrath Unissa, A., N. Selvakumar, S. Narayanan, and P. R. Narayanan. 2008. Molecular analysis of isoniazid-resistant clinical isolates of Mycobacterium tuberculosis from India. Int. J. Antimicrob. Agents 31:71-75. [DOI] [PubMed] [Google Scholar]
- 33.Park, H., E. J. Song, E. S. Song, E. Y. Lee, C. M. Kim, S. H. Jeong, J. H. Shin, J. Jeong, S. Kim, Y. K. Park, G. H. Bai, and C. L. Chang. 2006. Comparison of a conventional antimicrobial susceptibility assay to an oligonucleotide chip system for detection of drug resistance in Mycobacterium tuberculosis isolates. J. Clin. Microbiol. 44:1619-1624. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Piatek, A. S., A. Telenti, M. R. Murray, H. El-Hajj, W. R. Jacobs, Jr., F. R. Kramer, and D. Alland. 2000. Genotypic analysis of Mycobacterium tuberculosis in two distinct populations using molecular beacons: implications for rapid susceptibility testing. Antimicrob. Agents Chemother. 44:103-110. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Ramaswamy, S., and J. M. Musser. 1998. Molecular genetic basis of antimicrobial agent resistance in Mycobacterium tuberculosis: 1998 update. Tuber. Lung Dis. 79:3-29. [DOI] [PubMed] [Google Scholar]
- 36.Sajduda, A., A. Brzostek, M. Poplawska, E. Augustynowicz-Kopeć, Z. Zwolska, S. Niemann, J. Dziadek, and D. Hillemann. 2004. Molecular characterization of rifampin- and isoniazid-resistant Mycobacterium tuberculosis strains isolated in Poland. J. Clin. Microbiol. 42:2425-2431. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Saribas, Z., P. Yurdakul, A. Alp, and A. Gunalp. 2005. Use of fluorescence resonance energy transfer for rapid detection of isoniazid resistance in Mycobacterium tuberculosis clinical isolates. Int. J. Tuberc. Lung Dis. 9:181-187. [PubMed] [Google Scholar]
- 38.Sekiguchi, J., T. Miyoshi-Akiyama, E. Augustynowicz-Kopeć, Z. Zwolska, F. Kirikae, E. Toyota, I. Kobayashi, K. Morita, K. Kudo, S. Kato, T. Kuratsuji, T. Mori, and T. Kirikae. 2007. Detection of multidrug resistance in Mycobacterium tuberculosis. J. Clin. Microbiol. 45:179-192. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Silva, M. S., S. G. Senna, M. O. Ribeiro, A. R. Valim, M. A. Telles, A. Kritski, G. P. Morlock, R. C. Cooksey, A. Zaha, and M. L. Rossetti. 2003. Mutations in katG, inhA, and ahpC genes of Brazilian isoniazid-resistant isolates of Mycobacterium tuberculosis. J. Clin. Microbiol. 41:4471-4474. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Slayden, R. A., and C. E. Barry III. 2000. The genetics and biochemistry of isoniazid resistance in Mycobacterium tuberculosis. Microbes Infect. 2:659-669. [DOI] [PubMed] [Google Scholar]
- 41.Soini, H., and J. M. Musser. 2001. Molecular diagnosis of mycobacteria. Clin. Chem. 47:809-814. [PubMed] [Google Scholar]
- 42.Telenti, A., N. Honoré, C. Bernasconi, J. March, A. Ortega, B. Heym, H. E. Takiff, and S. T. Cole. 1997. Genotypic assessment of isoniazid and rifampin resistance in Mycobacterium tuberculosis: a blind study at reference laboratory level. J. Clin. Microbiol. 35:719-723. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Tracevska, T., I. Jansone, L. Broka, O. Marga, and V. Baumanis. 2002. Mutations in the rpoB and katG genes leading to drug resistance in Mycobacterium tuberculosis in Latvia. J. Clin. Microbiol. 40:3789-3792. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.World Health Organization. 2008. Expert group report: molecular line probe assays for rapid screening of patients at risk of multidrug-resistant tuberculosis (MDR-TB). World Health Organization, Geneva, Switzerland.
- 45.World Health Organization. 2008. Global tuberculosis control: surveillance, planning, financing: WHO report 2008. WHO/HTM/TB/2008.393. World Health Organization, Geneva, Switzerland.
- 46.Yam, W. C., K. Y. Yuen, S. Y. Kam, L. S. Yiu, K. S. Chan, C. C. Leung, C. M. Tam, P. O. Ho, W. W. Yew, W. H. Seto, and P. L. Ho. 2006. Diagnostic application of genotypic identification of mycobacteria. J. Med. Microbiol. 55:529-536. [DOI] [PubMed] [Google Scholar]
- 47.Zhang, S. L., J. G. Shen, P. H. Xu, D. X. Li, Z. Q. Sun, L. Li, Z. R. Yang, and Q. Sun. 2007. A novel genotypic test for rapid detection of multidrug-resistant Mycobacterium tuberculosis isolates by a multiplex probe array. J. Appl. Microbiol. 103:1262-1271. [DOI] [PubMed] [Google Scholar]
- 48.Zhou, L., A. N. Myers, J. G. Vandersteen, L. Wang, and C. T. Wittwer. 2004. Closed-tube genotyping with unlabeled oligonucleotide probes and a saturating DNA dye. Clin. Chem. 50:1328-1335. [DOI] [PubMed] [Google Scholar]