Molecular Characterization of Rifampin-Resistant Isolates of Mycobacterium tuberculosis from Hungary by DNA Sequencing and the Line Probe Assay

Abstract

Two regions of rpoB associated with rifampin resistance were sequenced in 29 rifampin-resistant (determined by the proportion method) isolates of Mycobacterium tuberculosis obtained from patients from three counties in Hungary. Of the 29 resistant strains, 27 had a mutation in either the 81-bp region (26 strains) or the N-terminal region (1 strain), while the other 2 strains had no mutations in either region. The locations and frequencies of the mutations differed from those previously reported. The most common mutation in this study, D516V, was found in 38% of the Hungarian strains, a frequency 2 to 10 times higher than that found in studies from other countries. These same 29 isolates were also evaluated with the Inno-LiPA Rif. TB test (LiPA), a reverse hybridization assay for the rapid detection of rifampin resistance. Although LiPA detected the presence of an rpoB mutation in 26 of the resistant isolates, the type of mutation could not be determined in 4 isolates because the mutations present were not among those included on the LiPA strip. In addition, a silent mutation in one of the rifampin-susceptible control strains was interpreted as rifampin resistant by LiPA. These findings demonstrate the importance of validating this rapid molecular test by comparison with DNA sequence results in each geographic location before incorporating the test into routine diagnostic work.

The recent worldwide increase in the incidence of drug-resistant strains of Mycobacterium tuberculosis has highlighted the need for faster and more accurate detection of resistance to rifampin (RMP), one of the most important antituberculosis drugs (17). RMP is most effective in killing actively metabolizing M. tuberculosis, and resistance to RMP often results in high clinical relapse rates (5, 15). Because of the prolonged turnaround time for conventional susceptibility testing, patients infected with drug-resistant tuberculosis may be inadequately treated and thus remain infectious for longer periods than those infected with susceptible strains.

Based on collective observations that mutations resulting in an amino acid change within the 81-bp core region of the RNA polymerase β-subunit (rpoB) gene are found in more than 96% of RMP-resistant M. tuberculosis strains, several molecular methods have been developed for the rapid (24- to 48-h) detection of mutations in this region (3, 10, 16, 18, 24, 25, 28–30). In addition, other studies revealed that mutations associated with RMP resistance can also occur in other regions of the rpoB gene, although less frequently (6, 7, 23). It has also been shown elsewhere that the information provided by these molecular tests can serve as a molecular epidemiological marker since the relative frequency of the alleles associated with resistance can vary geographically (10, 20).

Therefore, the aim of the present study was to determine the drug resistance profile of 29 RMP-resistant M. tuberculosis isolates obtained in East Hungary and to detect and identify mutations present in the rpoB gene. Two molecular assays were used. In the first, two regions of rpoB that have been associated with RMP resistance were amplified by PCR and the DNA sequence was determined. The results of the DNA sequencing were then compared with results from a commercially available rapid test, the PCR-based reverse hybridization line probe assay (Inno-LiPA Rif. TB Test [LiPA]; Innogenetics N.V., Ghent, Belgium).

After 20 years of decline, the incidence of pulmonary tuberculosis in Hungary increased by 18.1% between 1990 and 1999 (rising from 31.0 to 36.6 per 100,000 inhabitants) (1). In addition, East Hungary (Borsod-Abaúj-Zemplén, Hajdú-Bihar, and Szabolcs-Szatmár-Bereg counties) is the part of the country with the highest incidence of tuberculosis generally (38.7, 51.5, and 56.7 per 100,000 inhabitants, respectively) and drug-resistant tuberculosis specifically (1). In 1999, these three counties collectively reported 888 of the 3,912 (22.7%) tuberculosis cases in Hungary. The 29 RMP-resistant isolates examined in this study were isolated from patients in these three counties during 1999 (1).

The M. tuberculosis H37Rv ATCC 27294 strain and six clinical M. tuberculosis isolates that were pansusceptible for all four first-line antituberculosis drugs were used as controls. All cultures were identified by means of the AccuProbe culture identification test (Gen-Probe Inc., San Diego, Calif.) and conventional biochemical tests (11, 14).

Susceptibility testing of all 36 isolates for isoniazid (INH), RMP, ethambutol (EMB), and streptomycin (SM) was carried out by the proportion method on Löwenstein-Jensen medium as described by Canetti et al. (2). The critical concentrations for INH, RMP, EMB, and SM were 0.2, 40, 1.0, and 10 μg/ml, respectively. Of the 29 RMP-resistant isolates, only 2 (6.9%) were resistant to RMP alone. Twenty-six (89.7%) were also resistant to INH (and thus classified as multidrug resistant), 18 (62.1%) were also resistant to EMB, and 9 (31.0%) were also resistant to SM (Table 1). In all, 20 of the 29 (70.0%) were resistant to at least three of the four first-line drugs.

TABLE 1.

Resistance patterns of RMP-resistant M. tuberculosis isolates from Hungary by the proportion method, the line probe assay, and DNA sequencing of the 81-bp region of the rpoB gene^a

Strain	Result by method:														Mutation
	Proportion method				LiPA
	SM	INH	RMP	EMB	TB-P	S1	S2	S3	S4	S5	R2	R4a	R4b	R5	LiPA	Seq.
1	S	R	R	R	+	+	−	+	+	+	+	−	−	−	D516V	D516V
2	S	R	R	S	+	+	+	+	+	−	−	−	−	+	S531L	S531L
3	S	R	R	R	+	+	−	+	+	+	+	−	−	−	D516V	D516V
4	S	R	R	S	+	+	+	+	+	−	−	−	−	+	S531L	S531L
5	R	S	R	S	+	+	+	+	+	−	−	−	−	+	S531L	S531L
6	S	S	R	S	+	+	+	−	−	+/−	−	−	−	−	UDR	Deletion 522–525
7	S	R	R	R	+	+	−	+	+	+	+	−	−	−	D516V	D516V
8	S	R	R	R	+	+	−	+	+	+	+	−	−	−	D516V	D516V
9	R	R	R	S	+	+	+	+	−	+	−	−	+	−	H526D	H526D
10	S	R	R	R	+	+	−	+	+	+	+	−	−	−	D516V	D516V
11	R	R	R	R	+	+	−	+	+	+	+	−	−	−	D516V	D516V
12	R	R	R	R	+	+	−	+	+	+	+	−	−	−	D516V	D516V
13	S	R	R	R	+	+	+	+	+	−	−	−	−	+	S531L	S531L
14	R	R	R	R	+	+	+	+	+	−	−	−	−	+	S531L	S531L
15	S	R	R	R	+	+	−	+	+	+	+	−	−	−	D516V	D516V
16	S	R	R	R	+	+	−	+	+	+	+	−	−	−	D516V	S509T/D516V
17	S	R	R	S	+	−	+	+	+	+	−	−	−	−	UDR	Q513K
18	S	R	R	R	+	+	+	+	+	+	−	−	−	−	UD	WT
19	S	R	R	R	+	+	+	+	+	−	−	−	−	+	S531L	S531L
20	R	R	R	S	+	+/−	+	+	+	+	−	−	−	−	UDR	Q513P
21	S	R	R	S	+	+	+	+	+	−	−	−	−	+	S531L	S531L
22	S	R	R	S	+	+	+	+	−	+	−	−	+	−	H526D	H526D
23	S	R	R	R	+	+	−	+	+	+	+	−	−	−	D516V	D516V
24	S	R	R	R	+	+	−	+	+	+	+	−	−	−	D516V	D516V
25	S	R	R	S	+	+	−	+	+	+	+	−	−	−	D516V	D516V
26	S	S	R	S	+	+	+	+	+	−	−	−	−	+	S531L	S531L
27	R	R	R	R	+	+	+	+	+	−	−	−	−	+	S531L	S531L
28	R	R	R	R	+	+	+	+	+	+	−	−	−	−	WT	WTb
29	R	R	R	R	+	+	+	+	+	+	−	−	−	−	WT	WT
Co. 1	S	S	S	S	+	+	+	+	−	−	−	−	−	−	UDR	R529R
Co. 2	S	S	S	S	+	+	+	+	+	+	−	−	−	−	WT	WT
Co. 3	S	S	S	S	+	+	+	+	+	+	−	−	−	−	WT	WT
Co. 4	S	S	S	S	+	+	+	+	+	+	−	−	−	−	WT	WT
Co. 5	S	S	S	S	+	+	+	+	+	+	−	−	−	−	WT	WT
Co. 6	S	S	S	S	+	+	+	+	+	+	−	−	−	−	WT	WT
H37Rv	S	S	S	S	+	+	+	+	+	+	−	−	−	−	WT	WT

^aAbbreviations: Co., control; Seq., DNA sequencing of the hypervariable region; S, susceptible; R, resistant; +, positive; −, negative; +/−, indeterminate; UD, undetermined; UDR, undetermined resistant; WT, wild type. Probes R2, R4a, R4b, and R5 are specific for mutations D516V, H526Y, H526D, and S531L, respectively. 

^bA V146F mutation was found in this isolate. 

On the day of detection of growth index 999, 200-μl aliquots from Bactec 12B subcultures of the susceptible control and RMP-resistant isolates were incubated at 80°C for 1 h to heat kill the mycobacterial cells. Using primers rpo95 (5′-CCACCCAGGACGTGGAGGCGATCACACCG-3′) and rpo397 (5′-GTCAACCCGTTCGGGTTCATCGAAACG-3′) flanking the 81-bp region of rpoB (9), a 329-bp product was generated from all 36 isolates. The same primers were used for DNA sequencing of both strands using the automated Applied Biosystems 377 DNA sequencer (Applied Biosystems, Foster City, Calif.). A recent study demonstrated that, in some RMP-resistant strains with the wild-type sequence in the 81-bp region, a V146F mutation was found in the N-terminal region (7). In order to detect the presence of this mutation, amplification and sequencing were performed using primers Tb176-f (5′-CTTCTCCGGGTCGATGTCGTTG-3′) and Tb176-r (5′-CGCGCTTGTCGACGTCAAACTC-3′) as described previously (7). A 365-bp product was generated and sequenced using the same primers.

The heat-killed samples were also used for production of a biotinylated 256-bp fragment of the rpoB gene using the LiPA kit according to the instructions of the manufacturer (Innogenetics). The biotin-labeled PCR product was denatured and hybridized to a strip with 10 specific oligonucleotide probes (19 to 23 bases long). One probe is specific for the M. tuberculosis complex (TB-P), while five partially overlapping wild-type probes (S1 to S5) encompass the region of the rpoB gene encoding amino acids 509 to 534. Four other probes are specific for the most common mutations, D516V, H526Y, H526D, and S531L (probes R2, R4a, R4b, and R5, respectively) (Innogenetics). Hybridized PCR product was detected, and the LiPA results were evaluated as described elsewhere (4).

In contrast with previous reports (Table 2) (8, 13, 19, 20, 22, 26, 27, 30, 31), the frequency of occurrence of particular mutations was different in the isolates from East Hungary, with 11 (37.9%) isolates carrying the less common D516V mutation. Nine (31.0%) isolates had an S531L mutation, and two (6.9%) isolates had an H526D mutation. These mutations were also correctly detected in the LiPA. In addition, DNA sequencing identified a double mutation (S509T and D516V) and a deletion (deletion 522–525) that have not been reported in the literature before (32). In these two cases, the LiPA was unable to detect the correct type of the mutation. However, it indicated the presence of the genetic alteration (Table 1). The LiPA also did not reveal the type of mutation in two additional strains with rare mutation patterns (Q513K and Q513P) (Table 1). Moreover, the test provided a false-resistant result with a pansusceptible control strain with a silent mutation (R529R) (Table 1).

TABLE 2.

Frequency of mutations in RMP-resistant M. tuberculosis isolates from different geographic regions

Mutated codon(s)	% Mutations in geographic region:
	Africaa (reference 22; n = 105)	Asiab (reference 8; n = 77)	Australia (reference 31; n = 33)	Brazilc (reference 27; n = 82)	Germanyd (reference 20; n = 36)	Greecee (reference 13; n = 17)	Italy (reference 18; n = 37)	United States (reference 30; n = 61)	Hungaryf(n = 29)
572			3.0
533	3.0			1.2				3.3
531	53.5	46.8	48.5	53.7	36.1	47.1	56.7	34.4	31.0
526	21.2	18.2	30.3	20.6	22.2	17.7	29.7	39.3	6.8
522	10g	1.3	6.1	2.4				1.6
521								1.6
518	1.0							1.6
516	13.1	16.9	9.1	7.4	2.8	5.9	2.7	6.6	37.9
514		2.6						1.6
513	3.0	6.5		1.2					6.8
511	2.0	1.3		1.2
508						17.7
505						5.9
504	2.0
531 and 526				1.2	2.8	5.9	2.7
No mutation	14.1	6.5	3.0	3.7	33.3			8.2	10.3h

^aSouth Africa (n = 87), Sierra Leone and Uganda (n = 16), and Namibia (n = 2). 

^bBangladesh (n = 2), India (n = 4), Indonesia (n = 1), Malaysia (n = 14), Myanmar (n = 15), Nepal (n = 2), the Philippines (n = 4), Thailand (n = 14), and Yemen (n = 21). 

^cLess common mutations (1.2% each) were as follows: 531 and 514; 516 and 511; 515 and 511; 526 and 531; 524, 525, and 526 and deletion 527; deletion 515, 516, 517, and 518; 513 and deletion 514, 515, and 516. 

^dA rare (2.8%) double mutation was detected at codons 526 and 511. 

^eRare (2.7% each) triple mutations were detected at codons 553, 541, and 526; 525, 523, and 516; and 516, 512, and 511. 

^fA deletion (522–525) that has not been reported previously was detected in one strain (3.4%). 

^gSilent mutation. 

^hIn one of these three strains, an N-terminal V146F mutation was detected. 

The LiPA has been reported to be an easy-to-use test for the rapid detection of RMP resistance. The test is available in a kit format and, therefore, especially useful for routine work in clinical laboratories that are not capable of carrying out DNA sequencing (4, 21). In the present study, LiPA was able to detect a genetic alteration in 26 (89.7%) of the 29 RMP-resistant strains and to identify the particular mutation in 22 strains (75.9%) (Table 1).

The LiPA can provide the type of mutation for only the four most common mutations of the rpoB gene (S531L, H526Y, H526D, and D516V), while in cases of other mutations it indicates only the presence of a genetic alteration. Our findings suggest that in a geographic area such as East Hungary where less common or novel mutations of the rpoB gene occur more frequently, the interpretation of the LiPA results may be more difficult. In such an environment, the characterization of the type of the mutation (and the resultant amino acid change) by DNA sequencing is indispensable in order to avoid the report of any false RMP resistance results.

Although more than 96% of the RMP-resistant strains have a mutation within the 81-bp core region of the rpoB gene, a recent study revealed that a mutation associated with RMP resistance can also occur in other regions of the gene, although less frequently (7). The present study revealed three RMP-resistant isolates where neither the DNA sequencing of the 81-bp region nor the LiPA detected any mutations (Table 1). However, the DNA sequencing for the detection of the V146F mutation was positive in one of these three isolates (strain 28, Table 1). All the other study isolates had the wild-type sequence in this region. The two isolates with no mutations in either region indicate that at present the confirmation of molecular results by conventional tests is still warranted. In addition, since the routinely applied DNA sequencing methods usually examine only the 81-bp region of the rpoB gene (32), in cases where resistance is demonstrated in conventional susceptibility testing but no mutation is found we also suggest screening for the V146F mutation. If this assay also fails to detect a mutation, then other rare mutations of the rpoB gene, heteroresistance (a mixture of susceptible and resistant strains), or, less likely, another mechanism of resistance may be involved (6, 12, 23).

In conclusion, this study demonstrated that frequencies of particular mutations in RMP-resistant M. tuberculosis isolates from East Hungary are different from those that have been reported for isolates from other geographic areas (Table 2). DNA sequencing of the two regions of the rpoB gene identified mutations in 27 (93.1%) of the investigated 29 RMP-resistant isolates, and the LiPA identified mutations in 26 isolates (89.7%). However, the rapid LiPA was unable to determine the type of mutation in 4 of the 26 strains because these isolates contained unique mutations not included on the test strip. In addition, the one isolate that contained a V146F mutation in the N-terminal region of the rpoB gene was falsely interpreted as RMP susceptible in the LiPA. Finally, the LiPA gave a false-resistant result with one of the RMP-susceptible control strains that contained a silent mutation. These findings demonstrate the importance of validating molecular tests for the detection of RMP resistance using DNA sequence analysis and thus determining the frequencies of particular mutations in the test region before introducing the assay into routine clinical service.