Identification of Novel Mutations in LprG (rv1411c), rv0521, rv3630, rv0010c, ppsC, and cyp128 Associated with Pyrazinoic Acid/Pyrazinamide Resistance in Mycobacterium tuberculosis

ABSTRACT

Despite progress, the mechanisms of action and resistance of pyrazinamide (PZA) are not well understood. We characterized 109 mutants resistant to pyrazinoic acid (POA), the active form of PZA, and found that while most (n = 101 [93%]) mutants had panD mutations and 4 (4%) had clpC1 mutations (S91G), new mutations in lprG (rv1411c) and rv0521 (n = 4 [4%]), rv3630, rv0010c, ppsC, and cyp128 (cytochrome P450 128) were identified, shedding new light on the mechanisms of action and resistance of PZA in M. tuberculosis.

TEXT

There is currently considerable interest in understanding the mechanisms of action of pyrazinamide (PZA), a critical frontline tuberculosis (TB) drug that shortens TB therapy due to its unique activity against Mycobacterium tuberculosis persisters that are not killed by other TB drugs (1, 2). Despite the importance of PZA in the treatment of drug-susceptible and drug-resistant TB and its simple structure, its mechanisms of action are complex and not well understood (1, 2). PZA is a prodrug that is converted to the active form pyrazinoic acid (POA) by nicotinamidase/pyrazinamidase (PZase) encoded by the pncA gene (3), whose mutation is the most common mechanism of PZA resistance in M. tuberculosis (3–5). However, some low-level PZA-resistant strains (MIC, 200 to 300 μg/ml, pH 6.0) do not have mutations in the pncA gene (5, 6). rpsA, which encodes the ribosomal protein S1 involved in translation and trans-translation processes, has been proposed as a target of PZA (7), but a recent study claims RpsA is not a target of PZA (8). rpsA mutations seem to play a minor role in PZA resistance without pncA mutations in clinical isolates (6, 9), and rpsA polymorphisms seem to be found in some PZA-susceptible clinical isolates closely related to certain geographic regions. In addition, mutations in panD encoding aspartate decarboxylase were identified as a new mechanism of PZA resistance from in vitro mutants resistant to PZA, and the PanD protein was found to be another target of PZA (10–12). panD mutations were initially found in mutants resistant to PZA (10), then in mutants resistant to POA (11, 13), and later in clinical isolates (9). It is worth noting that POA-resistant mutants were not previously isolated at acid pH (4), which is required for higher activity of PZA against M. tuberculosis. However, we were able to successfully isolate POA-resistant mutants with high POA concentrations at close to neutral pH (pH 6.8) (11), which led to discovery of new genes involved in POA and PZA resistance. For example, clpC1, which was also isolated from mutants resistant to PZA (14), was identified in mutants resistant to POA (14, 15).

In our previous study, only 30 POA-resistant mutants of M. tuberculosis were analyzed, and all the mutations were mapped to the panD gene (11). The fact that some PZA-resistant strains, such as strain 9739 (5), do not have any mutations in the known genes pncA, rpsA, panD, or clpC1 involved in PZA resistance (W. Shi and Y. Zhang, unpublished observations) suggests that new mechanisms of POA/PZA resistance still exist. In this study, to identify possible new mechanisms of PZA resistance, we characterized an additional 109 POA-resistant spontaneous mutants (MIC, 200 μg/ml POA, pH 6.8) of M. tuberculosis H37Ra (MIC, <100 μg/ml POA) derived from our previous study (11). The genomic DNA from the 109 POA-resistant mutants was extracted and subjected to PCR DNA sequencing of the known genes pncA, rpsA, panD, and clpC1 involved in PZA resistance. It is noteworthy that 101 of 109 POA-resistant mutants (93%) had panD mutations, of which 75 had the same dominant M117I mutation previously identified (11). However, 8 POA-resistant mutants did not have panD mutations, and 4 of the mutants had the same mutation in clpC1 where A271G nucleotide change caused amino acid substitution S91G. Note that the clpC1 mutation S91G is different from our previously identified mutation (G99D) (14) and those identified by Yee et al. (15). However, the remaining 4 POA-resistant mutants were also resistant to PZA and were 2- to 4-fold more resistant (50 to 100 μg/ml PZA, pH 5.8) than the parent strain M. tuberculosis H37Ra (25 μg/ml PZA, pH 5.8); however, these 4 POA-resistant mutants did not have any mutations in the known PZA resistance genes pncA, rpsA, panD, or clpC1. Whole-genome sequencing of the 4 POA-resistant mutants was performed as described previously (10). Briefly, genomic DNA from 4 POA-resistant mutants, and the parent strain M. tuberculosis H37Ra, was sequenced using MiSeq (Illumina, Inc.), except that paired-end sequencing libraries were constructed using Nextera XT DNA sample preparation kits (Illumina, USA) following the manufacturer's instructions. For each isolate, 800 M to 1.5 G base (180-fold to 350-fold genome coverage) sequences were generated after barcodes were trimmed. Single-nucleotide variants and insertions and deletions (indels) ranging from 1 to 5 bp were sorted and called at a minimum coverage of 8 reads using the M. tuberculosis H37Ra genome (NC_009525) as a reference. Mutations in PE/PPE family genes and regions having repetition sequences were excluded from the analysis. Mutations in the parent strain M. tuberculosis H37Ra comparing with the genome online (NC_009525) were also excluded from the analysis. Heterogeneous percentage was calculated as mutated reads. However, additional mutations mapping in such excluded regions during WGS analysis may have played a role in PZA resistance. The sequence analysis revealed that all 4 mutants had the same 2 mutations, rv1411c (lipoprotein LprG C672T change leading to W224STOP) and rv0521 (methyltransferase, C295T causing amino acid substitution R99W) (Table 1). Sanger sequencing of the PCR products confirmed the mutations identified by whole-genome sequencing. We looked at possible mutations in LprG and Rv0521 in some PZA-resistant clinical isolates and did not find a clear association of mutations in these candidate genes and PZA resistance. In 38 PZA-resistant clinical isolates analyzed, only 1 of each isolate had a mutation in lprG and rv3630, and both had pncA mutations. No mutations were found in rv0521. In addition, analysis of 20 PZA-susceptible clinical strains found no lprG or rv0521 mutations.

TABLE 1.

Mutations identified in POA-resistant mutants of M. tuberculosis that do not have known PZA resistance mutations

Gene name	Gene product	Nucleotide change	Amino acid change	POA-resistant straina
				3G1	1H2	2F9	3H4
rv1411c	Lipoprotein LprG	C672T	W224STOP	SNV	SNV	SNV	SNV
rv0521	Methyltransferase	C295T	R99W	SNV	SNV	SNV	SNV
rv3630	Hypothetical protein	G1021A	A341T		Heterogeneous (86/236 reads, 36%)	Heterogeneous (37/193 reads, 19%)
rv0010c	Hypothetical protein	T-69C			Low heterogeneity (10/99 reads, 10%)		Low heterogeneity (11/107 reads, 10%)
ppsC	Phenolpthiocerol synthesis type-I polyketide synthase C	C2284A	H762N				Low heterogeneity (12/118 reads, 10%)
cyp128	Cytochrome P450 128 (Cyp128)	T1018C	T340A				Low heterogeneity (29/228 reads, 12%)

^aSNV, single-nucleotide variant.

Interestingly, the POA-resistant strain 3G1 had the 2 identical mutations in LprG (rv1411c) and rv0521, whereas the remaining 3 mutants, besides having the 2 mutations in LprG (rv1411c) and rv0521, had an additional low percentage of heterogeneous mutations in rv3630 (hypothetical protein), rv0010c (hypothetical protein), ppsC (phenolpthiocerol synthesis type-I polyketide synthase C), and cyp128 (cytochrome P450 128) (Table 1). Strain 3H4 had a point mutation T199C, causing amino acid change T67A in cytochrome P450 128 (cyp128, Rv2268c), which encodes a heme-thiolate monooxygenase. The role of these low-percentage mutations is unclear, but they may be involved in compensating for the lprG and rv0521 mutations or providing an additional level of resistance. It would be of interest to isolate single colonies characterized by single-nucleotide variants to determine the phenotype of the triple mutants and address the effect of these mutations on PZA resistance in the future.

The LprG (Rv1411 lipoprotein)-Rv1410 (efflux protein) locus has been shown to participate in triacylglyceride (TAG) transport in M. tuberculosis, as loss of its function leads to intracellular TAG accumulation, and is critical for regulating bacterial growth and metabolism during carbon starvation and infection in mice (16). The finding that all 4 POA-resistant mutants have the same W224STOP mutation in LprG suggests that the nonfunctional LprG will lead to higher TAG accumulation and thus cause higher metabolism, which is known to be a condition that antagonizes PZA/POA activity (2). This provides a plausible explanation for the loss-of-function mutation in LprG being a likely cause of PZA/POA resistance. However, the role of the mutation in rv0521, which encodes a putative methyltransferase, in causing POA resistance is less clear, and it is not known why the 4 POA-resistant mutants all had the same 2 mutations in LprG and Rv0521. Because 1 POA-resistant mutant, 3G1, has only the 2 identical mutations and no other mutations, it can be inferred that LprG, Rv0521, or both must be involved in POA/PZA resistance. Future studies are needed to address the contribution of each individual gene, lprG and rv0521, in causing POA/PZA resistance and its possible role in the mode of action of PZA.

However, the contribution of the low-abundance mutations in rv3630 (hypothetical protein), rv0010c (hypothetical protein), ppsC (phenolpthiocerol synthesis type-I polyketide synthase C), and cyp128 (cytochrome P450 128) is less clear (Table 1). Previous studies have found that mutations in ppsA through ppsE loci involved in synthesis of virulence factor phthiocerol dimycocerosate (PDIM) are associated with PZA resistance (13), but the role of the pps operon gene mutations, including ppsC mutations, in PZA resistance remains to be determined. In M. tuberculosis, cytochrome P450 proteins have many functions, such as virulence, persistence, cholesterol metabolism, lipid degradation, regulator, and resistance to azole drugs (17). cyp128 is one of three essential cytochrome P450s (cyp121, cyp125, cyp128) in M. tuberculosis and is part of an operon with rv2269c; and sft3, a sulfotransferase involved in biosynthesis of sulfomenaquinone (SMK) and disruption of SMK synthesis, affects M. tuberculosis respiration and fitness (18). cyp128 is upregulated during starvation and is essential for survival in mice (17). Although we found that a single mutation in cyp128 (T199C) caused amino acid change T67A in one POA-resistant mutant 3H4 (Table 1), future studies are needed to address the possible involvement of cytochrome P450 protein Cyp128 in resistance to PZA. While a causative role of the low-abundance gene mutations cannot be ruled out, it is likely that these mutations may serve to elevate resistance levels or allow the mutant to adapt to the 2 mutations in LprG and methyltransferase. These possibilities remain to be tested in future studies.

Despite the complexity of the mode of action of PZA, significant progress has been made in recent years in our understanding of this peculiar and yet critical drug, as PZA has been shown to interfere with multiple targets, including trans-translation (RpsA), energy production (PanD), and protein degradation (mediated by ClpC1), all of which are involved in persister survival in M. tuberculosis (2). This study identified novel genes associated with POA/PZA resistance, which may shed new light on the mechanisms of action of PZA. Future studies are needed to address the role of the identified mutations in PZA resistance and how they might be involved in the mode of action of PZA in M. tuberculosis.