Geographic Differences in the Contribution of ubiA Mutations to High-Level Ethambutol Resistance in Mycobacterium tuberculosis

Abstract

Ethambutol (EMB) resistance can evolve through a multistep process, and mutations in the ubiA (Rv3806c) gene appear to be responsible for high-level EMB resistance in Mycobacterium tuberculosis. We evaluated the prevalence of ubiA and embB (Rv3795) mutations in EMB-resistant strains originating from Africa and South Korea. No differences in embB mutation frequencies were observed between strains from both origins. However, ubiA mutations were present in 45.5% ± 6.5% of the African EMB-resistant isolates but in only 9.5% ± 1.5% of the South Korean EMB-resistant isolates. The ubiA mutations associated with EMB resistance were localized to regions encoding the transmembrane domains of the protein, whereas the embB mutations were localized to regions encoding the extramembrane domains. Larger studies are needed to investigate the causes of increased ubiA mutations as a pathway to high-level EMB resistance in African countries, such as extended EMB usage during tuberculosis treatment.

INTRODUCTION

Ethambutol (EMB), a first-line antituberculosis drug, is often used in combination with other drugs to treat tuberculosis and prevent the emergence of drug resistance (1). Numerous studies have shown that mutations in the embCAB operon, particularly the embB gene, are a major cause of EMB resistance in Mycobacterium tuberculosis (2–8). A second set of mutations, which occur in the ubiA gene, has been associated with EMB resistance in clinical M. tuberculosis isolates (9–11). Mutations in ubiA almost always occur in EMB-resistant strains that also contain embB mutations, and ubiA appears to have multiplicative effects with embB mutations on MICs (9). The evolutionary path leading from low- to high-level EMB resistance has been studied in the laboratory (9). In these studies, high-level EMB resistance appears to develop through the stepwise acquisition of mutations in embB, ubiA, and embC.

In the study described here, we examined the prevalence of ubiA mutations in isolates from two different geographic regions. Our results confirm the association between ubiA mutations and the presence of embB mutations and EMB resistance. We also demonstrate that the prevalence of these mutations varies by geographic location, suggesting that local factors may play a role in the type of mutations which develop as M. tuberculosis strains evolve to become EMB resistant.

MATERIALS AND METHODS

Bacterial strains and culture conditions.

Fifty-four clinical M. tuberculosis strains from different geographic regions were selected from a collection of highly characterized M. tuberculosis isolates established by the World Health Organization Special Programme for Research and Training in Tropical Disease (TDR strains; see Table S1 in the supplemental material) (12). A second set of 39 isolates from the National Masan Hospital in Changwon, South Korea and 41 isolates from the National Reference Laboratory in Rwanda was also tested (see Table 3). The last two sets of isolates were part of a drug resistance survey and were cultured from patient sputum after obtaining informed consent from their institutional review boards of the respective institutions (13, 14). In this study, M. tuberculosis strains were cultured at 37°C either in Middlebrook 7H9 broth containing 0.05% (wt/vol) Tween 80 or on Middlebrook 7H10 agar supplemented with 0.5% (vol/vol) glycerol, both of which were enriched with 10% oleic acid-albumin-dextrose-catalase (OADC; Becton Dickinson).

TABLE 3.

Distribution of ubiA mutations in EMB-resistant M. tuberculosis strains isolated in laboratories of hospitals in South Korea and in Rwanda^a

Origin	No. of strains with the following UbiA codon change:												% MT
	Total	WT	Val148Ala (GTG → GCG) mix	Ser173Ala (TCC → GCC)	Trp175Gly (TGG → GGG)	Ile179Thr (ATC → ACC)	Val229Gly (GTG → GGG)	Ala237Val (GCT → GTT)	Arg240Cys (CGC → TGC)	Ala237Val (GCT → GTT) mix	Arg240Cys (CGC → TGC) mix	Ser173Ala (TCC→GCC), Ala278Val (GCG→GTG) DMT
South Korea	39	36		1			1					1	8
Rwanda	41	25	1		1	1		7	4	1	1		39

^aWT, wild type; mix, mixture of wild type and mutant; DMT, double mutant; MT, mutants.

DNA isolation, PCR, and DNA sequencing.

Genomic DNA was extracted as described previously, with minor modifications (15, 16). To amplify DNA fragments for DNA sequencing, the PCR was performed using a mix containing 1 ng of genomic DNA, 5 pmol of each primer, 200 μM deoxynucleoside triphosphates, 1× PCR buffer, and 1 U of high-fidelity Pfx Taq polymerase (Invitrogen). All PCR products were purified using a gel extraction kit (Qiagen). Direct bidirectional Sanger sequencing of the embB and ubiA genes was performed with a BigDye Terminator kit and analyzed with an ABI 3100 genetic analyzer (Applied Biosystems). All primer sequences used in this study are described in Table 1.

TABLE 1.

Primers used for amplification and sequencing

Gene (Rv designation)	Size (bp)	Primer	Sequence	Product location
ubiA (Rv3806c)	909	F-ubiAa	ACGTTGAGCTTGAGGCTAGC	−116 to +15
		R-ubiAa	CGCTGTCGCGAATACTGCT
		Fin-ubiAseqb	GGTAGACGACCATTACCAGC	NAc
		Rin-ubiAseqb	GGTACTGGGAGTGACATCGC	NA
embB (Rv3795)	3297	F1-embBa	ATCGGTGGAGCAGTACCA	−117 to 877
		R1-embBa	ATGACATGCCAGAGCAGG
		F2-embBa	CACCGTCGTCGCACTGAT	683 to 1675
		R2-embBa	CGCAACATGATGAACACCG
		F3-embBa	CGTGGTATACCGAGAACCTG	1549 to +82
		R3-embBa	CATACCGAGCAGCATAGGAG
		Fin3-embBseqb	CGCATTCACACTGGGTGTGC	NA
		Rin3-embBseqb	GGCAGGATGAGGTAGTAG	NA

^aPrimers used for PCR and sequencing.

^bPrimers used only for sequencing.

^cNA, not applicable.

MIC testing and DST.

The EMB MICs of the TDR strains were determined in this study using the standard radiometric Bactec 460TB method (Becton Dickinson and Company, Sparks, MD) following the manufacturer's instructions with minor modifications (4). Each strain was tested against serial 2-fold increases in the antibiotic concentration. To confirm the results obtained by the Bactec method, 5 × 10³ CFU from the same inoculum was also spotted onto plates of 7H10 medium containing serial increases in the concentrations of EMB. The plates were incubated at 37°C for 2 to 3 weeks, and the MICs were determined to be the antibiotic concentration that inhibited growth compared to the growth of the 1/100 dilution on antibiotic-free 7H10 medium. The MICs determined by the Bactec method were not significantly different from those determined by the agar method. Drug susceptibility testing (DST) for determination of the EMB susceptibilities of the clinical strains from Rwanda and South Korea was performed previously by local clinical laboratories using the mycobacterial growth indicator tube (MGIT) method with a cutoff of 5 μg/ml, as indicated in the manufacturer's package inserts, and/or by the agar proportion method on Lowenstein-Jensen medium with a cutoff of 2 μg/ml, as described previously (17).

Statistical analysis.

For each gene, embB and ubiA, the associations between mutation status (presence, absence) and both the EMB MIC category and the geographic origin of the strain (Africa, South Korea) were assessed. Either the Pearson chi-square test or Fisher's exact test, if the expected sample counts were small, was used. The association between the frequency of embB mutations and ubiA mutations among all strains was assessed by McNemar's test for paired samples. A linear trend between mutations in each of the embB and ubiA genes and the EMB MIC was assessed using the exact Cochran-Armitage test for trend. All statistical tests were performed at a significance level of alpha equal to 0.05. All statistical analyses were performed using SAS (version 9.4) software (SAS Institute, Cary, NC).

Nucleotide sequence accession numbers.

In this study, we identified six new mutations in the ubiA gene, and the sequences were deposited in the GenBank database under accession numbers KX021867 for UbiA A38V, KX021868 for UbiA V55G, KX021869 for UbiA V148A, KX021870 for UbiA S173A, KX021871 for UbiA V229G, and KX021872 for UbiA S173A/A278V.

RESULTS

Examination of ubiA and embB genotypes in M. tuberculosis TDR strains isolated in South Korea and in African countries.

We measured the EMB MICs of 54 M. tuberculosis clinical strains isolated in South Korea and in Africa selected from the TDR collection (see Table S1 in the supplemental material) and sequenced the embB and ubiA genes of each of these strains. We then classified each strain into one of three EMB MIC categories: susceptible (MIC, 2 μg/ml), low-level resistant (MIC, 4 to 16 μg/ml), and high-level resistant (MIC, 32 μg/ml) (Table 2). None of the susceptible strains contained a mutation in either embB or ubiA. However, embB mutations were identified in many of the low-level and high-level EMB-resistant strains. No differences in embB mutation frequency were observed between strains of African origin (10/10 [100%] high-level resistant strains and 9/11 [82%] low-level resistant strains) and strains of South Korean origin (5/5 [100%] high-level resistant strains and 17/23 [74%] low-level resistant strains) (P = 1.00). Mutations in ubiA were also identified in low-level and high-level EMB-resistant strains. Contrary to the findings for the embB mutations, ubiA mutations occurred more frequently in the strains of African origin (8/10 [80%] high-level resistant strains and 3/11 [27%] low-level resistant strains) than in strains of South Korean origin (2/5 [40%] high-level resistant strains and 1/23 [4%] low-level resistant strains) (Table 2). Combining all EMB-resistant strains (MICs, >2 μg/ml), ubiA mutations were significantly more prevalent in strains of African origin (11/21, 52%) than in strains of South Korean origin (3/28, 11%) (P = 0.001). Our results also confirmed that mutations in ubiA are associated with high-level EMB resistance (10/15 [67%] high-level resistant strains versus 4/34 [12%] low-level resistant strains; P > 0.001), supporting the findings of a previous study (9). All the ubiA mutations occurred in strains that also had embB mutations.

TABLE 2.

Distribution of embB and ubiA mutants among TDR strains classified by EMB MIC and geographic origin

EMB MIC (μg/ml)	No. of MTa/no. of strains tested (%b)
	embB			ubiA
	Africa	South Korea	Total	Africa	South Korea	Total
32	10/10 (100)	5/5 (100)	15/15 (100)	8/10 (80)	2/5 (40)	10/15 (67)
4–16	9/11 (82)	17/23 (74)	26/34 (76)	3/11 (27)	1/23 (4)	4/34 (12)
2	0/3 (0)	0/2 (0)	0/5 (0)	0/3 (0)	0/2 (0)	0/5 (0)

^aMT, mutant.

^bPercentage of mutants.

ubiA genotypes of M. tuberculosis EMB-resistant strains isolated in clinical laboratories of hospitals in Rwanda and in South Korea.

The ubiA mutation frequencies in African versus South Korean strains were then investigated by use of a second independent set of EMB-resistant strains obtained from Rwanda and South Korea. Because of the similar embB mutation frequencies described in African and South Korean EMB-resistant TDR strains, only the ubiA gene from this second set of strains was sequenced. Significant differences in ubiA mutation frequencies similar to those seen in the first set of strains tested were observed. Mutations in ubiA were detected in 16/41 (39%) EMB-resistant Rwandese isolates and 3/39 (8%) EMB-resistant South Korean isolates (P = 0.001) (Table 3). Three DNA samples from Rwanda had a mixture of wild-type and mutant ubiA sequences, and one DNA sample from South Korea had a double mutation in ubiA (Table 3).

Mutations in ubiA that are associated with EMB resistance are localized in the region encoding the transmembrane domains of the protein.

The ubiA mutations from this study and our previously published study provide the largest collection of ubiA mutations so far reported in M. tuberculosis (9). We previously identified two mutations, R76R and E149D, to be phylogenetic markers associated with ancestral M. tuberculosis strains (lineage 1) (9, 18). Indeed, both the R76R and E149D mutations are present in the published genome sequences of M. bovis species, M. africanum, and M. canettii strains in GenBank (ftp://ftp.ncbi.nlm.nih.gov/genomes/all) (see Table S2 in the supplemental material). We examined the location of the ubiA mutations in relationship to the predicted secondary structure of the UbiA protein. Computer models have predicted that UbiA and Emb are integral membrane proteins consisting of cytoplasmic loops, transmembrane domains, and extracytoplasmic loops (2, 3, 19). UbiA is predicted to be an α-helical protein with nine transmembrane domains and no large carboxy-terminal region (19). After eliminating the R76R and E149D mutations and other synonymous mutations and mutations found in susceptible clinical isolates, we found that the amino acid changes in the UbiA protein associated with EMB resistance were almost exclusively located in the predicted transmembrane domains of the protein (Fig. 1). Interestingly, the location of the UbiA resistance-associated mutations contrasted markedly with the predicted location of mutations associated with EMB resistance in the EmbB protein. EmbB is predicted to be an α-helical protein with 11 transmembrane domains and a carboxy-terminal region of approximately 375 amino acids (3). In contrast to UbiA, the amino acids of EmbB associated with EMB resistance are predicted to be located in the extracytoplasmic loops of the membrane (3). While the contrasting locations of the UbiA and EmbB mutations associated with EMB resistance are striking, the functional significance of this difference is unclear.

FIG 1.

Schematic representation of UbiA amino acid changes associated with EMB resistance. M.tb, M. tuberculosis. Numbers 1 to 9 indicate transmembrane domains; I, II, III, IV, and V indicate cytoplasmic loops.

DISCUSSION

This study confirms the previously documented association between ubiA mutations and EMB resistance, particularly with high-level EMB resistance (9). We found that ubiA mutations are always found together with embB mutations in EMB-resistant strains but that embB mutations can occur in the absence of ubiA mutations. This observation suggests that mutations occur in embB as the first step in the evolution of EMB resistance and that ubiA mutations occur as secondary mutations, usually in association with high-level EMB resistance.

We have previously identified the association of the R76R and E149D ubiA mutations with lineage 1 of M. tuberculosis and not with EMB resistance (9). It is well reported that the changes in resistance genes should be evaluated when developing molecular diagnostic assays and managing treatment decisions (20). The V188A, A237V, R240C, and A249G mutations in ubiA have been conclusively shown, using allelic exchange studies, to cause EMB resistance (9); and using overexpression studies, K174T, W175C, and F176L mutations have been strongly implicated as a cause of EMB resistance (10). Thus, similar studies are needed to evaluate the contribution of other ubiA mutations in EMB resistance, further refine the association of all common ubiA mutations and EMB resistance, and improve genotypic drug resistance testing (20).

We found that ubiA mutations occurred more frequently in a collection of EMB-resistant isolates of African origin than a collection of EMB-resistant isolates of South Korean origin. This observation was confirmed with a second set of representative EMB-resistant isolates from Rwanda and South Korea. The reasons behind these regional differences in ubiA mutations remain unclear. It could be due to regional variations in M. tuberculosis strain types. Indeed, virtually all M. tuberculosis strains isolated from South Korea are members of the Beijing family of lineage 2 (21, 22), while most strains in Africa are members of lineage 1 or 3, with the T2 family being the most predominant genotype in Rwanda (23–26). Alternatively, variations in treatment practices among geographic regions could induce different resistance mutation profiles among EMB-resistant isolates. Standardized antituberculosis regimens in Rwanda use EMB for both the initial treatment and retreatment of tuberculosis, and drug susceptibility tests (DSTs) were not used to tailor treatments during the sampling period. In the absence of routine DSTs, clinicians may continue to use EMB even in EMB-resistant cases, potentially leading to the development of secondary mutations in ubiA and high-level EMB resistance. In contrast, clinicians in South Korea typically adjust their treatment regimens according to the results of DSTs that are performed at the start of treatment. These adjustments may prevent EMB-resistant isolates from acquiring additional EMB resistance mutations, including ones in ubiA. Routine DST is thought to improve treatment outcomes and to prevent the acquisition of resistance to new drugs (27). Our results also suggest that the more widespread use of DST could prevent the further evolution of resistance to a drug to which an M. tuberculosis strain is already resistant.

Our large panel of mutations permitted us to map UbiA resistance mutations to several transmembrane domains. The active site of the UbiA protein has been shown to be in the N-terminal region located in cytoplasmic loops II and IV (19). Thus, it is not clear how transmembrane mutations could lead to increased decaprenylphosphoryl-β-d-arabinose production and EMB resistance, as previously shown (9). One possibility is that UbiA is involved in a large enzyme complex with Emb proteins and other arabinan biosynthetic pathway components and that mutations in the transmembrane domains of UbiA could affect the stability or efficacy of this complex.

This study suggests that EMB resistance continues to evolve in isolates that are already EMB resistant through the acquisition of additional mutations. Our results also suggest that molecular tests for drug resistance should be implemented in antituberculosis health programs throughout the world to reduce the misuse of antimicrobials and to control the emergence of resistant strains. More clinical studies are needed to determine the contribution of embCAB, ubiA, and aftA mutations to the evolution of EMB resistance in different geographic regions. These future studies should help to identify biomarkers for low-level and high-level EMB resistance, which will improve treatment decisions and prevent the emergence of multidrug-resistant organisms.