Mycobacterium tuberculosis embB Codon 306 Mutations Confer Moderately Increased Resistance to Ethambutol In Vitro and In Vivo

Abstract

Ethambutol (EMB) is a major component of the first-line therapy of tuberculosis. Mutations in codon 306 of embB (embB306) were suggested as a major resistance mechanism in clinical isolates. To directly analyze the impact of individual embB306 mutations on EMB resistance, we used allelic exchange experiments to generate embB306 mutants of M. tuberculosis H37Rv. The level of EMB resistance conferred by particular mutations was measured in vitro and in vivo after EMB therapy by daily gavage in a mouse model of aerogenic tuberculosis. The wild-type embB306 ATG codon was replaced by embB306 ATC, ATA, or GTG, respectively. All of the obtained embB306 mutants exhibited a 2- to 4-fold increase in EMB MIC compared to the wild-type H37Rv. In vivo, the one selected embB306 GTG mutant required a higher dose of ethambutol to restrict its growth in the lung compared to wild-type H37Rv. These experiments demonstrate that embB306 point mutations enhance the EMB MIC in vitro to a moderate, but significant extent, and reduce the efficacy of EMB treatment in the animal model. We propose that conventional EMB susceptibility testing, in combination with embB306 genotyping, may guide dose adjustment to avoid clinical treatment failure in these low-level resistant strains.

INTRODUCTION

Ethambutol (EMB) is used worldwide as one of the first-line drugs for the treatment of tuberculosis (TB). The antimycobacterial activity of EMB was initially described in 1961, and since then a lot of investigations have been carried out to identify the mode of action (6, 27). Since EMB shows various effects on mycobacterial cells, several hypotheses have been advanced, ranging from inhibition of nucleic acid metabolism (3), phospholipid metabolism (8), spermidine biosynthesis (14), glucose metabolism (21), and inhibition of mycolic acids transfer into the cell wall (24) to the induction of trehalose dimycolate accumulation (7). However, the current hypothesis suggests that the primary mode of action of EMB is the inhibition of the biosynthesis of cell wall arabinan (24).

Belanger et al. (4) first presented experimental evidence that the targets of EMB are arabinosyltransferases mediating the polymerization of arabinose into arabinan. These membrane-associated proteins are ubiquitous in mycobacteria and have no sequence similarity to any known protein family in other bacteria (25). The enzymes are encoded by the emb operon and the corresponding genes exhibit ∼65% similarity to each other. Gene transfer experiments in M. smegmatis and the presence of mutations in clinical EMB-resistant Mycobacterium tuberculosis strains suggested that mutations in the emb operon may play a key role in EMB resistance (1, 22, 25). Accordingly, the most frequent detected mutations that occurred in codon 306 of embB (embB306) in clinical EMB-resistant isolates were proposed as a molecular marker for the rapid determination of EMB resistance (13). Recently, it was demonstrated that the transfer of embB306 mutations into an EMB-susceptible M. tuberculosis strain enhance the EMB MIC in vitro (18, 23).

However, in several recent publications EMB-susceptible M. tuberculosis strains carrying a mutation in embB306 were described. These discordant results between phenotypic and genotypic resistance testing raised doubts about the significance of embB306 mutations for the development of EMB resistance. Shi et al. (20) and others (4, 19) even argued that alterations in embB306 are associated with the development of broad-spectrum drug resistance but are not the cause of clinically significant EMB resistance. Furthermore, no direct experimental evidence for the role of embB306 mutations for EMB resistance in vivo has been provided thus far.

Therefore, we applied allelic-exchange techniques to transfer embB306 point mutations of EMB-resistant clinical isolates into M. tuberculosis H37Rv in order to clearly elucidate the role of mutations in embB306 for the development of EMB resistance. For all generated mutants, the levels of phenotypic EMB resistance were first determined in vitro, and then the clinical relevance was investigated in vivo in an aerosol mouse model of anti-TB drug therapy.

MATERIALS AND METHODS

Bacterial strains and growth conditions.

M. tuberculosis H37Rv was grown at 37°C in Middlebrook 7H9 (Difco) supplemented with 10% oleic acid-albumin-dextrose-catalase (OADC), 0.5% glycerol, and 0.05% Tween 80. Transformants were grown on Middlebrook 7H10 or in 7H9 both supplemented with 4% OADC.

Escherichia coli XL10-Gold Kan (Stratagene, Amsterdam, Netherlands) or HB101 and their transformants were grown at 37°C in Luria-Bertani (LB) broth or on LB agar. Hygromycin (50 μg/ml) or ampicillin (100 μg/ml) were added when necessary.

Determination of the MIC.

For MIC determinations, suspensions were made from fresh cultures grown on Löwenstein-Jensen (LJ) medium. The suspensions were adjusted to the optical density of McFarland 1. An inoculum of 0.1 ml of a 10⁻² dilution was used for the drug-containing medium. The following drug concentrations were investigated on LJ slants: 0.125, 0.250, 0.500, 1, 2, and 4 μg/ml for EMB and 0.008, 0.031, 0.063, 0.250, 1, and 2 μg/ml for isoniazid (INH). Duplicates were inoculated for each concentration of both drugs. Two drug-free controls were included, one with 0.1 ml of a 10⁻² dilution (representing 100% growth) and the other one with 0.1 ml of a 10⁻⁴ dilution (representing 1% growth). Incubation was at 37°C for 4 weeks. The lowest concentration of EMB or INH, respectively, that inhibited visible growth was considered the MIC.

Cloning experiments.

Chromosomal DNA of EMB-resistant M. tuberculosis clinical isolates was isolated as described elsewhere (26). Fragments of the embB gene, including a point mutation at codon 306, were amplified by using Platinum Pfx DNA polymerase (Invitrogen, Karlsruhe, Germany) according to the manufacturer's instructions. The PCR was carried out with the primers embCP1 (5′-ATATTAATTAATGGGGATCGGTGGAGCAGT-3′, positions −122 to −104) and embCP2 (5′-ATATTAATTAACCGCGGTGAACAGGCATAG-3′, positions 1660 to 1642), which both contain a PacI site at the 5′ end (underlined). The cycling conditions were as follows: an initial denaturation step of 94°C for 2 min; followed by 35 cycles of 94°C for 15 s, 64°C for 30 s, and 68°C for 2 min; and finally 68°C for 10 min. The PCR products were cloned into the pPCR-Script Amp SK(+) by using the appropriate cloning kit (PCR-Script Amp cloning kit; Stratagene). Bacterial colonies containing plasmids with an insert were selected by blue-white color screening and confirmed by PCR and DNA sequencing.

Plasmid DNA was isolated with the NucleoSpin plasmid kit (Macherey Nagel, Düren, Germany) and digested with PacI. The purified embB insert was subcloned into the sacB suicide vector pYUB657 (15) and electroporated into E. coli HB101. Hygromycin-resistant clones were analyzed again by PCR and DNA sequencing of the embB fragment.

Plasmid DNA of clones with the correct fragments was prepared with NucleoSpin plasmid kit and examined for purity by measuring the optical density at 260 nm (OD₂₆₀)/OD₂₈₀ ratio in a photometer before used for transformation of H37Rv.

DNA sequencing.

To confirm that E. coli clones contain the correct plasmid, the whole 1,782-kb embB fragment was amplified by using the primers embCP1 and embCP2 and HotStar Taq-DNA polymerase (Qiagen, Hilden, Germany) according to the manufacturer's instructions. For sequencing, the primer embF545 was used (Table 1).

Table 1.

Primers used for PCR and sequencing of the embB gene

Primer	Sequence (5′–3′)	Function	Position
Rv3795aF	CCT GCG TGG GGA CTG GTA T	PCR, sequencing	−147 to −128
Rv3795aR	TGA ATG CGG CGG TAA CGA C	PCR, sequencing	1321 to 1303
Rv3795bF	TGG ACG GGC GGG GCT CAA T	PCR, sequencing	725 to 743
Rv3795bR	CAC AAC GCC AGC AGG AAG A	PCR, sequencing	1916 to 1898
Rv3795cF	TCC TGG CGG CGT TAT TCT T	PCR, sequencing	1883 to 1901
Rv3795cR	CAA CCG GGG TGA TGA TGG C	PCR, sequencing	+235 to +214
Rv3795aS	CAA CTT CGT CGG GCT CAA G	Sequencing	477 to 495
Rv3795bS	CCA GCA AAC CCG CCT ACT G	Sequencing	1136 to 1154
Rv3795cS1	TGG ACG GCG ATT CGG GTT CT	Sequencing	2300 to 2319
Rv3795cS2	GGA CTG GGC GGT CGG TTT G	Sequencing	2958 to 2976
embF545	CGT TCC GGC CTG CAT	Sequencing	764 to 778

For confirmation of M. tuberculosis mutants, the whole embB gene was sequenced. A total of three PCR products were generated by utilizing the primer pairs listed in Table 1. The parameters for all three PCRs were as follows: an initial denaturation step of 95°C for 15 min; 35 cycles of 94°C for 1 min, 65°C for 1 min, and 72°C for 1 min; and a final extension step of 72°C for 10 min.

Sequencing reaction was performed by using the suitable primers (see Table 1) and the ABI Prism BigDye Terminator kit v.1.1 according to the manufacturer's instructions. Sequence data provided by an ABI Prism 3100 capillary sequencer (Applied Biosystems, Darmstadt, Germany) were analyzed with the ABI software SeqScape v2.0.

Southern blot analysis.

Southern blot analysis was performed to screen for H37Rv transformants that have undergone a single-crossover event, as well as to analyze mutants emanated from a second crossover event. A 334-bp DNA fragment was generated by PCR amplification with the primers embFnew (5′-TGGACGGGCGGGGCTCAAT-3′) and embRnew (5′-GGCAGGCGCATCCACAGACT-3′) and used as a probe. Labeling was carried out with DIG High Prime according to the manufacturer's instructions (Roche, Mannheim, Germany). PvuII-digested genomic DNA was separated by agarose gel electrophoresis and blotted on Hybond N+ nylon membranes (Amersham Biosciences, Buckinghamshire, United Kingdom). Southern blot hybridization was performed by using the DIG system according to the instructions of the manufacturer (DIG Easy Hyb and DIG wash and block buffer set; Roche).

Transformation of mycobacteria.

Transformation of M. tuberculosis H37Rv was performed accordingly to the method of Pavelka and Jacobs (15).

In brief, for generation of electrocompetent cells, mycobacterial cultures were grown to an OD₅₉₅ of 0.8. Cells were collected by centrifugation (10 min, 3,000 × g, room temperature) and washed three times with an equal volume of 10% glycerol. The cell pellet was resuspended in a one-tenth volume of the culture with 10% glycerol. For electroporation, 400-μl portions of freshly prepared cells were mixed with 1 μg of DNA and put into prewarmed 0.2-cm cuvettes (Bio-Rad, Munich, Germany). After 1 min of incubation at 50°C, the cells were immediately electroporated. The cells were then directly transferred in 3 ml of 7H9/OADC, followed by incubation overnight at 37°C.

Transformants that integrated the plasmid into the chromosome by a single-crossover event were selected upon 7H10/OADC medium supplemented with Hygromycin (50 μg/ml). Hygromycin-resistant clones were cultured in 7H9/OADC without any other supplement. To select clones that underwent a second recombination event, transformants were grown to stationary phase and plated on 7H10/OADC containing 2% sucrose.

Experiments in mice.

Female, 6- to 8-week-old, specific-pathogen-free C57BL/6 mice were purchased from Charles River Laboratories (Sülzfeld, Germany) and maintained in individually ventilated cages (Ebeco, Castrop-Rauxel, Germany) under biosafety level III conditions. All animal experiments were approved by the Ministry of Environment, Nature Protection, and Agriculture (Kiel, Germany).

For infection experiments, M. tuberculosis H37Rv or embB306 mutant 14-9 cultures were grown to mid-log phase, harvested, divided into aliquots, and frozen at −80°C. After thawing, viable cell counts were determined by plating serial dilutions on Middlebrook 7H10 agar supplemented with 10% bovine serum (Biowest, Nuaillé, France). Mice were infected with 100 CFU/lung using an inhalation exposure system (Glas-Col, Terre-Haute, IN). Mice were exposed for 40 min to an aerosol generated by nebulizing approximately 5.5 ml of a suspension containing 10⁷ live bacteria. The inoculum size was confirmed at 24 h postinfection by determining the bacterial load in the lung.

EMB treatment began 7 days after infection and was given by oral gavage (200 μl) five times per week until the mice were sacrificed (11). Mice in the control group received unsupplemented water by gavage.

To determine the bacterial loads in lungs, livers, and spleens at 4 and 7 weeks after infection, groups of five mice were sacrificed, and the organs were removed aseptically, weighed, and homogenized in distilled water. Tenfold serial dilutions of organ homogenates were plated on Middlebrook 7H10 agar supplemented with 10% bovine serum, followed by incubation at 37°C for 21 days. The detection limit in the lung was 12 CFU/organ, and in the liver and spleen it was 16 CFU/organ. Quantifiable data were expressed as means ± the standard deviation (SD) and subjected to statistical analysis using a Kruskal-Wallis test with Dunn's multiple-comparison post test, comparing all individual treatment groups against untreated, infected controls. A P value of ≤0.05 was considered significant.

RESULTS

Construction of unmarked H37Rv mutants with altered embB306 alleles.

M. tuberculosis H37Rv mutants, containing a point mutation in embB306, were generated by two-step allelic exchange via homologous recombination as described in Materials and Methods. The embB306 wild-type codon ATG (Met) was replaced by inserting the mutated embB306 fragments with codons ATA (Ile), ATC (Ile), or GTG (Val) in the genome of H37Rv, respectively. Sucrose-resistant but hygromycin-susceptible clones were analyzed by PCR and Southern blotting. In addition, we sequenced the embB gene to confirm potential H37Rv mutants (data not shown). Overall, we obtained one mutant carrying the mutated codon ATC (strain 2-7), five mutants with codon ATA (strains 3-1, 3-2, 3-3, 3-4, and 3-5), and four clones with codon GTG (strains 14-5, 14-9, 14-10, and 14-13). Compared to the H37Rv embB wild-type sequence, there was no sequence variation in the mutant strains 2-7, 3-1, 3-2, 3-3, 3-4, 3-5, and 14-9 except for the inserted point mutation in codon embB306 (data not shown). For three of the four embB306 GTG mutants (14-5, 14-10, and 14-13) an additional mutation (C→T) was detected in the embB gene, which resulted in the amino acid replacement Arg464Trp.

In vitro effect of embB306 mutations on EMB resistance.

To investigate the relationship between the introduction of embB306 mutations and the level of EMB resistance, we determined the MIC of M. tuberculosis H37Rv wild-type and of all (n = 10) generated H37Rv embB306 mutants over a range of concentrations between 0.125 and 4 μg of EMB/ml in LJ medium (Table 2).

Table 2.

MICs for EMB and INH^a

Strain (embB306, amino acid)	MIC (μg/ml)
	EMB	INH
H37Rv (ATG, Met)	1	0.063
2-7 (ATC, Ile)	2	0.063
3-1 (ATA, Ile)	4	0.063
3-2 (ATA, Ile)	4	0.063
3-3 (ATA, Ile)	4	0.063
3-4 (ATA, Ile)	4	0.063
3-5 (ATA, Ile)	4	0.063
14-5 (GTG, Val)	4	0.063
14-9 (GTG, Val)	4	0.063
14-10 (GTG, Val)	4	0.063
14-13 (GTG, Val)	4	0.063

^aCultures were grown as duplicates for 3 weeks on LJ medium containing the appropriate drug concentrations.

All strains carrying a mutation in embB306 showed an increase in EMB MICs compared to the H37Rv wild-type (1 μg/ml). The GTG mutation (Val) resulted in a 4-fold increase in the MIC for all four investigated mutants (Table 2). There were no differences detectable between the three double mutants (14-5, 14-10, and 14-13) and the mutant with the single point mutation (14-9). The same results, i.e., a 4-fold increase in the MIC for EMB, were obtained for the five clones containing an embB306 mutation from ATG to ATA (Ile). The EMB MIC of the one strain harboring a mutation on the third base from G to C was only 2 μg/ml, although this mutation also encodes isoleucine.

Although there is a clearly increased MIC for EMB of all mutant strains, there was no comparable effect for INH since no difference between the INH MICs for the H37Rv wild-type and all 10 embB306 mutants was observed (Table 2).

In vivo effect of embB306 mutations on EMB resistance.

To investigate whether the embB306 mutations resulting in increased MIC in vitro also confer EMB resistance in vivo, we compared the replication of the H37Rv strain and the congenic embB306 mutant strain 14-9 in the mouse model of aerosol infection with M. tuberculosis.

Based on human studies (16), drug therapy was performed by oral gavage using different EMB dosages (6.25, 12.5, 25, or 50 mg/kg of body weight/day, respectively). In preliminary experiments, serum levels after treatment were measured by gas chromatography: 1 h after gavage of 50 mg of EMB/kg of body weight, a serum level of 5.78 mg/liter was determined; the 25-mg/kg dose resulted in 2.2 mg/liter; the 12.5-mg/kg dose resulted in 0.54 mg/liter; and the 6.25-mg/kg dose resulted in 0.196 mg/liter in the serum 1 h after gavage. Since MICs for the H37Rv strain and the 14-9 strain fell within this range, treatment of mice using these four dosing regimens was started 7 days after infection and was continued until day 49. Relatively early treatment was performed because EMB since monotherapeutic agent shows, at best, only moderate bacteriostatic activity. Mice were sacrificed to determine the bacterial burden in the lung, liver, and spleen at days 28 and 49 after infection.

EMB-treated mice infected with wild-type H37Rv showed a dose-dependent reduction in bacterial multiplication in all three investigated organs after 4 and 7 weeks of infection (Fig. 1). Lower doses (6.25 and 12.5 mg/kg/day) of EMB therapy had no significant effect on the growth of H37Rv in the lung, liver, and spleen compared to the untreated mice. In contrast, mice treated with the intermediate-dose (25 mg/kg/day) and high-dose (50 mg/kg/day) EMB therapy showed a significant reduction in CFU counts in the lung compared to the control groups (Fig. 1A). Treatment with 50 mg of EMB reduced bacterial growth by 2.1 log₁₀ CFU in the lung (P < 0.05), by 2.3, log₁₀ CFU in the liver (P < 0.05), and by 2.3, log₁₀ CFU in the spleen (P < 0.05) after 49 days of infection.

Fig. 1.

Bacterial growth of wild-type H37Rv and embB306 GTG mutant 14-9 in EMB-treated mice compared to untreated control groups. C57BL/6 female mice were inoculated via aerosol with 100 CFU of M. tuberculosis H37Rv (solid bars) or embB306 GTG mutant 14-9 (striped bars). The bacterial burden of both strains (d1, n = 3) was determined in the lung 24 h after infection resulting in 2.43 CFU (mean, log₁₀) for H37Rv and 2.40 CFU (mean, log₁₀) for mutant 14-9. The EMB treatment started 7 days after infection. Different EMB dosages (6.25, 12.5, 25, or 50 mg/kg/day, respectively; increasing gray shades in columns indicate increases in dose per treatment group) were administered by gavage. Control mice received nonsupplemented sterile water. The absolute bacterial load in the lung (A), in the liver (B) and in the spleen (C) was determined at day 28 and 49 after infection (n = 5, mean CFU + the SD). *, P < 0.05 for comparisons between individual treatments versus nontreated, infected controls.

Strikingly different results were observed for mice infected with the mutant 14-9. Overall, the mutant 14-9 showed a reduced replication in vivo compared to the wild-type strain H37Rv. At days 28 and 49 after infection, no significant reduction in the bacterial burden was detectable in the livers and spleens of mice infected with mutant 14-9 compared to untreated infected mice. In the lung, only the highest treatment dose (50 mg of EMB/kg of body weight) significantly (P < 0.05) inhibited the growth of the mutant (by ∼1.3 log₁₀ CFU at day 28 and by ∼0.8 log₁₀ CFU at day 49 after infection compared to untreated mice; Fig. 1A); all other doses were ineffective in vivo. In conclusion, mutant 14-9 was more resistant to EMB treatment than was the wild-type strain, requiring a higher dose to achieve growth inhibition in the lung.

DISCUSSION

The data presented in this investigation provide support for the contention that resistance to EMB conferred by mutations in the embB codon 306 is moderate, but significant, and translates into clinically relevant EMB resistance. Although there is experimental evidence that mutations in embB codon 306, such as are found in many clinical isolates, result in the development of resistance to EMB in vitro, as determined by a cutoff MIC of 2 μg/ml (on LJ medium), our data question whether they alone can confer the observed high level of resistance to EMB in many clinical isolates. Our experimental therapy data also suggest that the moderate levels of EMB resistance conferred by embB306 mutations may be overcome by moderately increased dosing of EMB.

All of the M. tuberculosis H37Rv embB306 mutants generated here by using allelic-exchange techniques showed up to 4-fold increased MICs for EMB in vitro. Furthermore, in a proof-of-principle approach, the clinical relevance of the conferred MIC increase was tested for one selected mutant strain in aerosol-infected mice treated with EMB. Treatment with a dose of 25 mg of EMB/kg (body weight)/day, which resulted in an EMB serum level of 2.2 mg/liter by 1 h after administration, significantly reduced the growth of wild-type H37Rv (MIC = 1 mg/liter) but not that of mutant 14-9 (GTG; MIC = 4 mg/liter) in the lungs of mice. In contrast, a higher dose of 50 mg of EMB/kg (body weight)/day, which resulted in an EMB serum level of 5.78 mg/liter by 1 h after administration, significantly reduced the growth of both wild-type H37Rv and mutant 14-9 in the lungs of infected mice. These data suggest that a moderate increase in EMB dose can overcome the moderate degree of drug resistance conferred by the embB306 mutation. However, in the spleen and liver, this higher dose inhibited growth only of the wild-type H37Rv and not of the mutant 14-9 strain. This is likely due to the fact that strain 14-9 was attenuated for growth in vivo, resulting in relatively low and variable CFU counts in extrapulmonary organs, and that the trend toward a small bacteriostatic effect of the highest dose of EMB did not reach statistical significance in the liver and spleen.

The in vitro results of our study are in accordance with gene transfer experiments carried out in M. smegmatis. Belanger et al. inserted the embAB genes of a resistant M. avium strain into the chromosome of M. smegmatis mc²155 and noticed a 2-fold increase in EMB resistance (2). Similar results were presented by Telenti et al. (25), who integrated a 40-kb fragment of a highly resistant M. smegmatis mutant into the genome of M. smegmatis mc²155 wild type. The transformants exhibited an increased EMB MIC of 1 μg/ml compared to 0.5 μg/ml for the wild type. Only recently, Alland and coworkers (18) demonstrated by a similar approach that embB306 mutations lead to an increased EMB MIC in M. tuberculosis Beijing genotype strains. Our own data show a similar increase of EMB MIC (2 to 4 μg/ml compared to 1 μg/ml on LJ medium) by introducing embB306 alleles from EMB-resistant clinical isolates in M. tuberculosis H37Rv, also resulting in a resistant test result in routine diagnostic EMB susceptibility testing with a critical concentration of 2 μg/ml for LJ medium (13). The critical concentration for EMB resistance testing as well as the observed MIC values depend strongly on the medium uses for testing.

Approximately 80% of ethambutol is absorbed after oral administration. Following intake of a dose of 15 mg/kg (body weight), a peak serum concentration of approximately 2 to 4 mg/liter is achieved in 2 to 4 h. This forms the basis of the laboratory definition for the determination of clinical resistance (5, 9). However, our in vivo results suggest that single point mutations in embB306 are not sufficient to confer EMB resistance upon M. tuberculosis to an extent, making it completely refractory to treatment. In our hands, moderate dose adjustment (50 mg of EMB/kg (body weight)/day in mice) led to significant growth inhibition in the lung of a mutant strain carrying an embB306 mutation (strain 14-9). High-level EMB MIC values of 20 to 40 μg/ml or even more have been reported in clinical isolates by some studies report, which obviously cannot be fully explained by the single point mutations in embB306 (22). This makes it likely that multiple consecutive genetic events, involving an accumulation of several different mutations, are necessary for the acquisition of high-level and clinically significant levels of EMB resistance in M. tuberculosis.

Previous experiments with M. smegmatis support the assumption that the attainment of a high-level resistant phenotype is a multistep process (25). The occurrence of polymorphisms in promoter regions or in regulatory genes can lead to overexpression of Emb proteins. An alternative way to overcome the inhibition by the drug is the appearance of mutations in the emb operon resulting in structural changes in the Emb proteins. Nonetheless, it should be stated that several EMB-resistant clinical isolates tested in our laboratory had MIC levels comparable to that of the embB306 mutant strains reported here (data not shown). The interpretation of current EMB susceptibility testing results in terms of their validity for predicting clinical EMB resistance should therefore be based on actually determined MIC values, since only moderately increased doses might be sufficient to overcome low-level resistance. To be fair, it is also not known how the different pharmacokinetics of mice and men may be adequately compared and how the doses and treatment modality used in our study correspond to attainable tissue levels in humans. As a consequence, controlled clinical studies are necessary to demonstrate whether EMB treatment in patients with moderately increased doses is truly an option when low-level resistance (e.g., conferred by single point mutations in embB306) is diagnosed.

Since most studies investigating the molecular basis of EMB resistance have analyzed only a short fragment of the embB gene, there is only limited information on additional genetic variations in clinical strains that, e.g., developed EMB resistance under therapy (10, 12, 28). In light of the data presented here, there is an urgent need to perform more extensive genome studies to decipher further mechanisms contributing to the higher levels of EMB resistance observed in some clinical isolates. Few publications that have analyzed the whole embCAB operon, as well as the putative regulatory genes embR and Rv3124 of EMB-resistant M. tuberculosis isolates, have identified several polymorphisms in all three genes of the emb operon and in the intergenic region (between embC and embA) which possibly play a role for the acquisition of EMB resistance (17).

In all likelihood, other, as-yet-unidentified, genes contribute to EMB resistance. It is already known for other drugs that resistance can be a multifunctional mechanism. For example, INH resistance can be a result of mutations in the genes katG, inhA, ahpC, kasA, and/or ndH, and a streptomycin-resistant strain can have acquired alterations in the gene rpsL and/or in rrs (29).

In conclusion, the data presented here suggest that it may be useful to distinguish between low levels of resistance to EMB, such as that conferred by embB306 mutations, and high levels of resistance which are observed in many clinical isolates and are possibly due to variations in additional loci. Our mouse experiments are compatible with the interpretation that low to moderately increased levels of resistance against EMB may still be treated in the clinic by a moderate adjustment of the dosing regimen.