|Home | About | Journals | Submit | Contact Us | Français|
Mutations within codon 306 of the Mycobacterium tuberculosis embB gene modestly increase ethambutol (EMB) MICs. To identify other causes of EMB resistance and to identify causes of high-level resistance, we generated EMB-resistant M. tuberculosis isolates in vitro and performed allelic exchange studies of embB codon 406 (embB406) and embB497 mutations. In vitro selection produced mutations already identified clinically in embB306, embB397, embB497, embB1024, and embC13, which result in EMB MICs of 8 or 14 μg/ml, 5 μg/ml, 12 μg/ml, 3 μg/ml, and 4 μg/ml, respectively, and mutations at embB320, embB324, and embB445, which have not been identified in clinical M. tuberculosis isolates and which result in EMB MICs of 8 μg/ml, 8 μg/ml, and 2 to 8 μg/ml, respectively. To definitively identify the effect of the common clinical embB497 and embB406 mutations on EMB susceptibility, we created a series of isogenic mutants, exchanging the wild-type embB497 CAG codon in EMB-susceptible M. tuberculosis strain 210 for the embB497 CGG codon and the wild-type embB406 GGC codon for either the embB406 GCC, embB406 TGC, embB406 TCC, or embB406 GAC codon. These new mutants showed 6-fold and 3- to 3.5-fold increases in the EMB MICs, respectively. In contrast to the embB306 mutants, the isogenic embB497 and embB406 mutants did not have preferential growth in the presence of isoniazid or rifampin (rifampicin) at their MICs. These results demonstrate that individual embCAB mutations confer low to moderate increases in EMB MICs. Discrepancies between the EMB MICs of laboratory mutants and clinical M. tuberculosis strains with identical mutations suggest that clinical EMB resistance is multigenic and that high-level EMB resistance requires mutations in currently unknown loci.
Ethambutol (EMB) is a first-line antituberculosis drug that is often used in combination with other drugs to treat tuberculosis and to prevent the emergence of drug resistance. EMB also has a place in the treatment of drug-resistant and multidrug-resistant tuberculosis (2). The recent global increase in the incidence of drug-resistant tuberculosis has produced many strains that are resistant to EMB. Therefore, it is prudent to test isolates from all tuberculosis patients for their EMB susceptibility, especially when EMB is used to treat multidrug-resistant tuberculosis. Unfortunately, conventional culture-based EMB susceptibility test methods have poor intertest and interlaboratory reproducibilities (8, 21). This has made it difficult to firmly rule out the presence of EMB resistance by the use of conventional assays. Culture-based Mycobacterium tuberculosis drug susceptibility tests are also quite slow (12, 20).
Genetic tests for EMB resistance are potentially more rapid and more accurate than conventional culture-based resistance testing. Genetic assays identify resistance by detecting mutations that encode EMB resistance on the M. tuberculosis chromosome, principally within the embB gene (5, 17, 25). The results of genetic assays can be available within hours; they have high interassay reproducibilities and have the potential to have high sensitivities (5, 25). However, genetic testing for EMB resistance has been hindered by a persistent uncertainty concerning the role of specific mutations in EMB resistance. Initially, the role of mutations within codon 306 of the embB gene (embB306) was questioned. Although embB306 mutations were present in 30 to 68% of EMB-resistant clinical isolates (1, 13, 22), some studies had noted a widespread presence of embB306 mutations in EMB-susceptible isolates (1, 7, 9). The role of embB306 mutations was firmly established to be a cause of low- and moderate-level (two to seven times the MIC for the wild type) EMB resistance in a recent allelic exchange study (19). However, that study also demonstrated that embB306 mutations do not in themselves cause high-level (MICs > 20 μg/ml) EMB resistance. Furthermore, the cause of EMB resistance in the 32 to 70% of clinical EMB-resistant M. tuberculosis isolates that did not have embB306 mutations remained an open question.
Several clinical studies have suggested that other mutations in the embCAB operon are responsible for at least some of the remaining EMB-resistant tuberculosis cases. The most commonly occurring embCAB mutations other than embB306 have been found in embB406 and embB497. Importantly, these two mutations have been detected in clinical isolates with high-level EMB resistance (11, 14). However, other studies identified embB406 mutations in EMB-susceptible clinical isolates (7, 15, 23). Other mutations in the embB and embC genes have also been identified in EMB-resistant clinical M. tuberculosis isolates (6, 22, 23), but at low frequencies, making it difficult to firmly establish associations with EMB resistance. Thus, the actual role of non-embB306 mutations in EMB resistance has not been proven.
In the study described here, we examined the role of embB mutations outside of the embB306 codon in EMB resistance. Using in vitro-selected mutants and allelic exchange techniques, our results demonstrate that non-embB306 mutations in the embCAB operon play an important role in EMB resistance, but like mutations in embB306, these mutations confer only a low to a moderate increase in EMB MICs. Our study strongly suggests that unrecognized mycobacterial gene targets for EMB resistance and high-level resistance remain to be discovered.
M. tuberculosis strains were cultured as described previously (18). The MICs for EMB, isoniazid (INH), and rifampin (RIF; rifampicin) were determined by the standard radiometric Bactec 460TB method (Becton Dickinson and Company, Sparks, MD) and by the 7H10 agar proportion method, as described previously (19). The Top10 Escherichia coli strain (Invitrogen) was used as the host for all plasmid constructions. Top10 was grown in Luria-Bertani broth or agar medium (Sigma), which was supplemented with 50 μg/ml kanamycin where appropriate.
Cultures of M. tuberculosis strain 210 were grown in liquid culture to an optical density at 600 nm of approximately 1.5, spun down, and resuspended in 2 ml of 7H9 medium. Approximately 109 M. tuberculosis cells were then plated on 7H10 medium containing either no EMB or EMB (Sigma) at a final concentration of 8, 16, or 32 μg/ml. The cells were then incubated at 37°C for 4 weeks. The resulting colonies were cultured in 7H9 medium without EMB prior to MIC testing. Genomic DNA was extracted from each colony after subculture in the absence of EMB. Mutations in embB406 and embB497 were detected by a real-time PCR assay consisting of the primers and molecular beacons listed in Table Table1,1, following the recommendations described previously (4). The entire embB gene of all EMB-resistant colonies that were found not to contain mutations in embB406 or embB497 were sequenced. Subsequently, the entire embCAB operon of colonies that were not found to have mutations in the embB gene was sequenced.
The PCR products of the embCAB locus were generated from genomic DNAs by using a TaKaRa LA PCR kit (version 2; Takara Bio) and standard PCR conditions. Five overlapping PCR products (amplicon 1, primers 1 [5′-CGCACATAACAGCTACACCC-3′] and 2 [5′-CGAAGGTCTGATCACGAAAG-3′]; amplicon 2, primers 3 [5′-CTTTCGTGATCAGACCTTCG-3′]) and 4 [5′-ACCAGCCAGTCCAGGAACAC-3′]; amplicon 3 primers 5 [5′-GTGTTCCTGGACTGGCTGGT-3′] and 6 [5′-GATCGACCGTTCCACCAACA-3′]; amplicon 4, primers 7 [5′-TGTTGGTGGAACGGTCGATC-3′] and 8 [5′-CCACCGACAACACAAAGCCA-3′]; amplicon 5, primers 9 [5′-GGTCCGTTCCTGTTCACC-3′] and 10 [5′-CGCTATGGACCAATTCGGATC-3′] were amplified from genomic DNAs to span the embCAB locus. Overlapping embCAB amplicons were purified with a QIAquick PCR cleanup column (Qiagen) and were used as templates for Sanger sequencing (Applied Biosystems). Sequence reads were assembled by using TIGR_Assembler (J. Craig Venter Institute) software. In order to identify sequence polymorphisms, the consensus sequence from each contig for each strain was compared to the sequences of M. tuberculosis strains 210 and NJT210GTG by using NUCMER sequence analysis software (3). Electropherograms were inspected by using Cloe sequence analysis software (http://cloe.sourceforge.net) for all polymorphisms identified to assess the quality of the read. Areas of low coverage or poor quality were resequenced to increase sequence coverage and quality. The genome coordinates of confirmed polymorphisms were determined by BLAST analysis by using M. tuberculosis strain H37Rv as a reference.
DNA isolation and PCRs were performed as described previously (18). To create the embB497 isogenic strains, a 3,500-bp DNA fragment spanning the embB497 codon was created in two separate PCR steps by using genomic DNA extracted either from M. tuberculosis strain 16703, which contained a single CAG-to-CGG point mutation at embB497, or from wild-type strain 210. The left and right fragments were amplified with primer pair F2-embBCL and R1-embBOEx and primers pair F2-embBOEx and R2-embBOEx (Table (Table1),1), respectively. The resulting left fragment was digested with Acc65I and BamHI, and the right fragment was digested with BamHI and HindIII (Promega). To create the embB406 isogenic strains, a 2,211-bp fragment carrying either the embB406 GCC (Ala), embB406 GAC (Asp), embB406 TGC (Cys), or embB406 TCC (Ser) codon was obtained in two separate PCR steps by use of a site-directed PCR mutagenesis method (16, 19), as follows. A first DNA fragment was amplified from strain 210 genomic DNA by using primers F-embB306CL and R1-embB406CL (Table (Table1).1). The amplicon was then purified and digested with HindIII and BamHI. To introduce different single point mutations at embB406, a second DNA fragment was obtained by five sequential PCRs with primer R-embB306CL and mutagenic primers F2-embB406, F3-embB406, F4-embB406, F5-embB406, and F6-embB406 (Table (Table1)1) in succession, as described previously (19). The final PCR products were purified and digested with BamHI and Acc65I. To obtain the embB497 or embB406 point mutations, the left and right amplicons for each mutation (or the wild-type control) were ligated and simultaneously cloned into the p2NIL vector, followed by the insertion of a PacI cassette containing the sacB and lacZ genes (10) (Table (Table1).1). The recombinant plasmids were transformed into M. tuberculosis strain 210, and the mutant colonies obtained by double-crossover events were screened by standardized molecular beacon assays (Table (Table1).1). We confirmed that only the intended single point mutations were introduced into the embB gene by direct DNA sequencing of the isogenic strain DNA, as described previously (19).
Mutations at embB406 and embB497 have frequently been described in clinical studies of EMB-resistant tuberculosis. We searched for embB406 and embB497 mutations in M. tuberculosis colonies that were picked after they were cultured on plates with different EMB concentrations (Table (Table2).2). Three of these 23 colonies had previously been identified as embB306 mutants (19). No embB406 mutants were detected in the 23 colonies screened; however, 4/23 colonies had embB497 mutations. Two of the 5 colonies selected with 32 μg/ml of EMB and 2 of 11 colonies selected with 16 μg/ml of EMB contained embB497 CAG-to-CGG (Gln-to-Arg) mutations (Table (Table2).2). All of these embB497 mutants had EMB MICs of 12 μg/ml (Table (Table3).3). This is in contrast to the findings of clinical studies, which have reported on embB497 mutants with much higher EMB MICs (14).
We performed additional sequencing of the in vitro-selected EMB-resistant colonies to search for mutations other than embB306, embB406, and embB497. We first sequenced the entire embB gene of each colony, and then we sequenced the embCA genes of all colonies that had wild-type sequence at embB. We detected mutations previously described in EMB-resistant clinical isolates at embB397, embB1024, and embC13. These mutants had very low levels of EMB resistance, with EMB MICs being between 3 and 5 μg/ml with the Bactec 460TB system and 8 μg/ml by the agar proportion method (Table (Table3).3). Three mutations that resulted in slightly higher levels of EMB resistance were identified at embB320, embB324, and embB445. The embB320 and embB324 mutants had EMB MICs of 8 μg/ml by the use of both the Bactec 460TB and the agar proportion methods (Table (Table3).3). Previous molecular studies also detected spontaneous mutants at emB320 or embB324 by in vitro selection methods (24); however, the embB445 mutation has not been described previously. The five embB445 mutants had identical embB445 CAG-to-CGG (Gln-to-Arg) changes; yet, surprisingly these colonies had a range of MICs by use of the Bactec 460TB system that varied from 2 to 8 μg/ml. These results were confirmed by repeat MIC testing two times. Complete sequencing of the entire embCAB operon confirmed the absence of any other mutation besides embB445 in each case, suggesting that additional mutations outside of the embCAB operon were responsible for the variability in the MICs. Interestingly, three EMB-resistant colonies (MICs, 4 to 6 μg/ml by the agar proportion method) had wild-type embCAB gene sequences. These colonies did not appear to be resistant by use of the Bactec 460TB method. It is possible that these clones contained mutations which were situated outside of the embCAB operon. These as-yet-unknown mutations may be at least partially responsible for the discrepancies in the results obtained between the liquid and solid EMB resistance testing methods sometimes observed in clinical studies (8, 12, 21). An additional three colonies had an intermediate EMB resistance phenotype, with the EMB MICs being 8 μg/ml by the agar proportion method and 3 to 4 μg/ml by use of the Bactec 460TB system. The significance of these MIC increases is unclear, even though they were reproducible.
Our inability to identify embB406 mutations in any of the in vitro-selected EMB-resistant colonies caused us to question the significance of this mutation, despite previously observed associations between embB406 mutations and EMB resistance in clinical studies. We performed allelic exchange experiments to definitively explore this issue (Table (Table4).4). We exchanged the wild-type clinical M. tuberculosis 210 strain embB406 GGC (Gly) codon for either the embB406 GCC (Ala), embB406 GAC (Asp), embB406 TGC (Cys), or embB406 TCC (Ser) codon, creating isogenic mutants NJT210GCC, NJT210GAC, NJT210TGC, and NJT210TCC, respectively. These mutations were selected because each of them has been reported in clinical studies of EMB-resistant tuberculosis (7, 11, 14). The unique transfer of the intended point mutation into embB was confirmed by sequencing the entire embB gene. Our results show that embB406 mutations are responsible for small increases in EMB MICs. Strains NJT210GCC, NJT210GAC, and NJT210TGC each showed an EMB MIC of 7 μg/ml and strain NJT210TCC had an MIC of 6 μg/ml (Table (Table3),3), whereas the EMB MIC was 2 μg/ml for parental strain 210. The small increase in the EMB MICs that we observed may explain our inability to detect embB406 mutants by in vitro selection.
Mutations at embB497 have been reported to be associated with high-level EMB resistance in clinical strains. We performed allelic exchange studies at embB497 to study the impact of mutation of this allele on the EMB MIC. We replaced the wild-type embB497 CAG (Gln) codon in strain 210 with the embB497 CGG (Arg) codon, creating isogenic mutant strain NJT210CGG (Table (Table4).4). This mutant showed a moderately increased EMB MIC (12 μg/ml). This increase is comparable to that seen when the more commonly occurring mutant embB306 GTG allele is moved into wild-type strain 210. However, neither mutation produces high-level EMB resistance. We also reintroduced the wild-type embB497 CAG sequence back into NJT210CGG, creating strain NJT210CGG-CAG. This was done to confirm that the increased EMB MIC that we observed was not caused by an accidental introduction of mutations at other locations within the M. tuberculosis genome. As expected, the reintroduction of the wild-type embB497 sequence into NJT210CGG-CAG caused its EMB MIC to revert to the wild-type level.
We had previously shown that embB306 mutants show less growth inhibition in the presence of subinhibitory concentrations of INH or RIF. An embB306 CTG isogenic mutant was also able to outcompete its wild-type parental strain during coculture in subinhibitory concentrations of INH (19). We performed similar growth studies with INH and RIF using the mutants created in this study. Upon exposure to either drug, none of the embB406 or embB497 mutants showed a difference in growth compared to the growth of the parental control.
This study demonstrates that mutations at embB497 and embB406 are definitive causes of EMB resistance in M. tuberculosis. Other than mutations at embB306, these mutations are the ones that are the most frequently detected in clinical EMB-resistant strains. This study demonstrates that the clinically observed mutations at embB397, embB1024, and embC13 also increase the EMB MICs, although their effects are much smaller (resulting in 1.5- to 2-fold increases in the MICs). We also confirm that mutations in embB320 and embB324 result in EMB resistance, supporting the findings of a previous in vitro selection study (24); and we identified a new mutation at embB445 to be a cause of EMB resistance. The level of proof supplied by our in vitro selection studies is somewhat lower than that supplied by the allelic exchange experiments performed to investigate embB406 and embB497 mutations. We cannot rule out a contribution of mutations outside of the embCAB operon that could have been coselected during growth in the presence of EMB. In fact, our observation that the five separate clones harboring the same embB445 Gln-to-Arg mutation had EMB MICs ranging from 2 to 8 μg/ml strongly suggests the coselection of additional mutations, along with the mutation of the embB445 allele.
Our study provides further support for the hypothesis that several mutations, one within the embCAB operon and one or more outside of the embCAB operon, may be required for high-level EMB resistance. Safi et al. previously demonstrated that embB306 mutations were necessary but not sufficient for high-level EMB resistance by performing allelic exchange studies with several clinical M. tuberculosis strains (19). In the current study, we show that embB497 and embB406 mutations also confer a relatively low level of EMB resistance when they are introduced as single mutations. This finding is in contrast to those of studies that have observed the same mutations in clinical strains with high-level EMB resistance (11, 14). Similarly, the embB397, embB1024, and embC13 mutations which were selected in vitro had very small elevations in EMB MICs, even though the same mutations are found in clinical M. tuberculosis isolates that have much higher EMB MICs. Together, these results demonstrate that EMB resistance is much more complex than was previously thought.
Our results suggest that phenotypic tests for EMB resistance may have trouble consistently detecting some embB mutations, despite their clear role in altering EMB susceptibility. The transfer of GCC, GAC, TGC, and TCC embB406 codons into wild-type M. tuberculosis increased the EMB MICs only from 2 to 6 or 7 μg/ml. Most clinical laboratories use breakpoints of 5.0 or 7.5 μg/ml to define EMB resistance. Thus, it is not surprising that many embB406 mutants have been identified as EMB susceptible in clinical studies. It is not possible for us to determine whether these small increases in the EMB MIC can adversely affect treatment outcomes by themselves. However, it is reasonable to assume, at the least, that these mutations will decrease the genetic barrier against high-level EMB resistance and, at the worst, that these mutations could render EMB ineffective at standard doses.
It has previously been shown that embB306 mutations confer on strains the ability to grow more vigorously than wild-type strains in the presence of sub-MICs of INH and RIF. Antibiotic synergy is also adversely affected. We did not observe the same phenomenon in the embB497 or embB406 mutants. The reasons for this difference are unclear. It is possible that embB306 codon changes affect cell wall permeability by altering the arabinogalactan or the lipoarabinomannan content or structure within the M. tuberculosis cell wall. This may not occur to the same extent in embB406 and embB497 mutants.
This study may have important implications for the design of genetic drug resistance tests and for the unraveling of the molecular mechanism of EMB resistance. More effective therapeutic and diagnostic measures for tuberculosis are desperately needed. An improved understanding of the basic mechanisms of M. tuberculosis drug resistance is central to the development of new molecular diagnostic assays and new drugs. Our study reveals the complexity of EMB resistance in M. tuberculosis. These results strongly suggest that embB mutations confer a low to moderate level of EMB resistance and that other mutations elsewhere in the chromosome are needed for full resistance. More clinical studies are needed to fully identify the range of EMB resistance-associated mutations. This should be complemented by whole-genome sequencing investigations to discover the unrecognized target genes for EMB resistance. In addition to providing improved diagnostic assays, this work is likely to also reveal new EMB drug targets, and it may identify new ways in which M. tuberculosis can be inhibited or killed by using novel classes of drugs.
This work was supported by National Institutes of Health grants AI065273 and AI080653 to D. Alland. This work was also partially funded through the NIAID Pathogen Functional Genomics Resource Center (contract N01-AI-15447), managed and funded by the Division of Microbiology and Infectious Diseases, NIAID, NIH, DHHS, and operated by the J. Craig Venter Institute.
Published ahead of print on 12 October 2009.