S-Nitrosomycothiol Reductase and Mycothiol Are Required for Survival Under Aldehyde Stress and Biofilm Formation in Mycobacterium smegmatis

Abstract

We show that Mycobacterium smegmatis mutants disrupted in mscR, coding for a dual function S-nitrosomycothiol reductase and formaldehyde dehydrogenase, and mshC, coding for a mycothiol ligase and lacking mycothiol (MSH), are more susceptible to S-nitrosoglutathione (GSNO) and aldehydes than wild type. MSH is a cofactor for MscR, and both mshC and mscR are induced by GSNO and aldehydes. We also show that a mutant disrupted in egtA, coding for a γ-glutamyl cysteine synthetase and lacking in ergothioneine, is sensitive to nitrosative stress but not to aldehydes. In addition, we find that MSH and S-nitrosomycothiol reductase are required for normal biofilm formation in M. smegmatis, suggesting potential new therapeutic pathways to target to inhibit or disrupt biofilm formation.

Introduction

Nitric oxide (NO) is produced by macrophages during the innate and adaptive immune responses to infection from a wide range of microbial pathogens (1,2). NO has diverse physiological functions, causing damage to both the pathogen and host cells by binding to metal and nitrosylating tyrosine and cysteine residues in proteins. It can also react with hydrogen peroxide to form peroxynitrite, causing oxidative damage. Paradoxically, nitrosylation is an important process in cell signaling as nitrosylation can activate or inhibit protein function (3). The functions of NO are not limited to its site of production, as NO is highly diffusible and low-molecular-weight (LMW) S-nitrosothiols can act as long-distance NO delivery vehicles. The LMW thiol, glutathione (GSH), reacts with NO to form nitrosoglutathione (GSNO), and GSNO can be catabolized to GSH and NO by S-nitrosoglutathione reductase (GSNOR), an enzyme common to most eukaryotes and prokaryotes (4,5). GSNOR is thus able to modulate effects of reactive oxygen and nitrogen species through catabolism of GSNO. Although its preferred substrates are GSNO and S-hydroxymethyl glutathione, an intermediate in formaldehyde metabolism, GSNOR has a dual NAD/H-dependent oxidoreductase activity toward a broad spectrum of aliphatic compounds including alcohols such as cinnamylalcohol, geraniol, ω-hydroxy fatty acids (6), and the aldehydes, glycolaldehyde and glyceraldehyde, which are byproducts of sugar metabolism. Indeed, a mutant of Haemophilus influenzae, disrupted in GSNOR, is sensitive to these toxic aldehydes (7).

Although GSH is the primary LMW thiol found in eukaryotes and Gram-negative bacteria, mycothiol (MSH) is the functional analog of GSH in actinobacteria, a phylum that includes a wide variety of bacteria, including the pathogen Mycobacterium tuberculosis (8). MSH contributes to cellular stress resistance against oxidants, antibiotics, toxins, and metals (9, 10). Additionally, a Mycobacterium smegmatis MSH-deficient mutant disrupted in mshA is sensitive to NO (11), and MSH acts as a cofactor for the MSH-dependent formaldehyde dehydrogenase/nitrosothiol reductase, MscR (12). Ergothioneine (ESH), another LMW thiol found in actinobacteria, is able to scavenge NO and oxidants in vitro (13) and in vivo (14).

Recently, a Nesseria meningitidis mutant lacking GSNOR demonstrated loss of viability in a biofilm (15). Whether MscR, MSH, or ESH play roles in biofilm formation in mycobacteria and other actinobacteria was not previously known. In this study, M. smegmatis mutants disrupted in mscR (SHP1), an MSH-deficient strain disrupted in mshC (S24) and an ESH-deficient strain disrupted in egtA (R119), were examined for their roles in detoxification of aldehydes and reactive nitrogen species. As biofilms produced during infection are highly resistant to broad-spectrum antibiotics and the host immune system and as the genus Mycobacterium includes pathogens such as M. tuberculosis and Mycobacterium avium, which are capable of forming biofilms (16), we also examined these mutants for biofilm formation in this study.

Materials and Methods

Culture Conditions

M. smegmatis mc²155 wild-type cultures were routinely started from glycerol stocks onto Middlebrook 7H10 solid medium with 0.5% glycerol supplemented with 1% glucose. Colonies were then transferred to Middlebrook 7H9 broth (Difco) with 0.05% Tween-80 supplemented with 1% glucose or 10% OADC (oleic acid, albumin, dextrose, and catalase). Antibiotics were added when appropriate (25 μg/mL kanamycin for S24 and R119, 75 μg/mL hygromycin for SHP1, and 25 μg/mL kanamycin and 75 μg/mL hygromycin for complemented strains, S24C and SHP1C). Cultures were incubated at 37 °C with shaking at 160 rpm.

Creation and Screening of a Transposon Mutant Library and Identification of the Disrupted Gene in S24 and R119

An M. smegmatis transposon mutant library (EZ::TN <kan-2> Tnp transposase and Tn5 kanamycin resistance marker) was screened for diamide-sensitive mutants as described by Rawat et al. (17). To identify the site of insertion, genomic DNA was digested with restriction enzymes Sa1 I and PstI followed by self-ligation and PCR amplification with primers complementary to the transposon. The sequence of the amplified PCR product was compared with the M. smegmatis genome sequence (17).

Minimum Inhibitory Concentrations

Starter cultures were used to inoculate 7H9 medium containing 1% glucose supplemented with 0.05% Tween-80. These cultures were grown to mid-log phase and then diluted to an OD₆₀₀ of 0.2. Formaldehyde, methylglyoxal, glycolaldehyde, glyceraldehyde, or GSNO were serially diluted twofold in 100 μL and added to 100 μL of culture to obtain a final OD₆₀₀ of 0.1 in ELISA plates. These cultures were incubated for 48 h, and then the minimum inhibitory concentration (MIC) was determined. Assays were performed in quadruplicate and repeated at least three times.

Growth Curves

Cultures were grown to mid-log phase and then diluted to an OD₆₀₀ of 0.1. Ten milliliters of cultures in 50-mL tubes was treated with 0.75 mM formaldehyde, 0.75 mM methylglyoxal, 2.25 mM glycolaldehyde, 9 mM glyceraldehyde, 0.75 mM NaNO₂, or 0.75 mM GSNO, concentrations that were 75–90% of the determined MIC. Time points for growth curves were taken until stationary phase, and the growth curves were performed three times in triplicate or more. To analyze the effects of oxygen tension on growth, the headspace–volume ratio was adjusted to 4.0 (high oxygen tension), 1.5 (moderate oxygen tension), or 0.66 (low oxygen tension) in 50-mL tubes (18).

Creation of mscR Mutant, SH1, and Complementation of Mutant with M. Tuberculosis mscR Gene

The mscR (MSMEG_4340) gene and flanking regions were amplified using primers 5′-CTCGAGTCACAACATCACCAC-3′ and 5′-GAATTCATGAGTCAGACGGTG-3′. The resulting PCR amplicon was purified and subsequently ligated into the TA cloning vector, pCR2.1, containing a kanamycin resistance cassette to obtain pSH1. A hygromycin resistance cassette was excised from pHINT using the restriction enzymes PstI and SmaI, blunted with the Klenow fragment, and then inserted into the StuI site in the mscR gene. The resulting vector, pSH2, was then electroporated into competent M. smegmatis mc²155, and electroporated cells were plated on selective media with hygromycin, followed by replica plating of mutant colonies on media with kanamycin to rule out nonhomologous recombination events. Confirmation of mutation was performed using PCR with the above primers and Southern blot hybridization. One confirmed colony, SHP1, was further characterized.

To validate the phenotype of the mscR mutant, the native gene from M. tuberculosis, Rv2259, was amplified with fdhco5, 5′-CTCGAGTTCGCTCTGCATCGTCAC-3′, and fdhco3, 5′-GGTACCGACCAGCCAGATGTTGTT-3′, primers with XhoI and KpnI restriction sites engineered, respectively, into the primer sequence. The 1.5-kb gene fragment containing the gene and 345 bp upstream of the start site was purified from a 1% agarose gel and TA cloned into pCR2.1, and the resulting plasmid, pSH3, was then subjected to digestion with XhoI and KpnI for cloning into the integrative vector, pAINT that had been similarly restriction digested. Selection of this new plasmid, pSH4, was performed on LB media supplemented with kanamycin followed by transformation into SHP1 competent cells and plating on 7H10 supplemented with 10% OADC and hygromycin and kanamycin. Confirmation of an integrated copy of the M. tuberculosis mscR was carried out by PCR using primers fdhco5 and fdhco3. The M. smegmatis strain SHP1C contained a mutated copy of mscR and a copy of the M. tuberculosis mscR.

RNA Extraction and qPCR

To analyze gene expression, M. smegmatis mc²155 was grown in Middlebrook 7H9 media with 1% glucose at 37 °C with shaking to an OD₆₀₀ of 0.5. Aliquots of 10 mL of culture were treated with 0.75 mM formaldehyde, 0.75 mM methylglyoxal, 2.25 mM glycolaldehyde, 9 mM glyceraldehyde, or 0.75 mM GSNO in quadruplicates. After 1-h incubation, the aliquots were centrifuged at 4,000 rpm at 4 °C for 10 min, and pelleted cells were flash frozen and stored at −80 °C. RNA was extracted using a Qiagen Rneasy Mini Kit with the optional on-column DNase treatment (Qiagen) as specified in the manufacturer’s protocol, except cells were lysed with glass beads using a ribolyzer (three 45-sec cycles at maximum speed). The RNA extraction was followed by an additional Turbo DNase treatment (Life Technologies). Samples were then checked for DNA contamination using PCR. One microgram of RNA and 0.5 μg of random primers were heated for 5 min at 70 °C, cooled on ice for 5 min, and then reverse transcribed using a Promega M-MLV Reverse Transcriptase Kit according to the manufacturer’s instructions. After confirmation of synthesis of cDNA, qPCR reactions were performed using Thermo Scientific ABsolute SYBR Green ROX Mix with 50 ng RNA and 200 nM primers in a reaction volume of 25 μL. The qPCR conditions were an initial 50 °C cycle for 15 min, followed by a 95 °C cycle for 15 min, 40 cycles of 95 °C for 30 sec, 55 °C for 30 sec, and 72 °C for 1 min, and a final elongation step of 72 °C for 4 min. Primer efficiency was calculated by generating a standard curve and by analysis of the melting curve. Real-time quantitative PCR (qPCR) was performed in an Eppendorf Realplex² Mastercycler with primers (Table 1). Relative gene expression was calculated using the −2^ΔΔCT method (ΔC_t – C_{t sample} − C_{t control}) and reported as fold change in gene expression of each sample normalized to the mysA relative to the untreated control. Each experiment was performed in triplicate.

TABLE 1.

Primers for quantitative real-time PCR

Name	Primer sequence
mscR-Fwd	CCT GGA AGC AGG CCT TCT A
mscR-Rev	GCA GGT ACA GCG AGA TCA G
mysA-Fwd	ATC GAC GAG CCG TCC GAG AA
mysA-Rev	CCT TGC CGA TCT GCT TGA GG
egtA-Fwd	GGT GAT GCT GGT GAA CTC T
egtA-Rev	CGT CGT CAG GTG ATA GTC C
mshC-Fwd	ATC GCC GAG GTC GTC GAG AT
mshC-Rev	GGC GAA CAA CGT GAG CAT GG

Mycobacterium Smegmatis Biofilm Assay

M. smegmatis wild type and mutant strains were incubated on Trypticase Soy Agar (TSA) plates containing antibiotics if necessary at 37 °C for 48 hours. A single colony from each strain was inoculated in 4 ml of Trypticase Soy Broth (TSB) with 0.05% Tween 80 with antibiotics if needed. The cultures were incubated at 37 °C at 225 rpm for 72 hours. When the OD₆₀₀ reached 2.0, the cells were diluted 1:1000 and incubated in TSB media supplemented with 1% Procalamine (B. Braun Medical Inc.), which enhances biofilm formation in M. smegmatis, and incubated at 30 °C for 72 hours with no disturbance. Four replicates for each strain were seeded for biofilm formation. After 72 hours, the biofilms, visible on the liquid-air interface were documented. The thickness of the biofilms was quantitated by measuring absorbance at OD₆₃₀ in a BioTek EON plate reader after removal of the media from each well.

Results and Discussion

Mutants Disrupted in mscR, mshC but not egtA Are Sensitive to Nitrosative Stress

We had previously reported that MSH was required for growth of M. smegmatis when cultures were exposed to gaseous NO; here we confirm that this is also true with GSNO (11). To determine if mutants deficient in mscR, MSH, or ESH are sensitive to aldehyde and nitrosative stress, MICs were determined and growth curves were performed using sub-lethal concentrations of stressors. The transposon mutant, S24, disrupted in mshC, which codes for MSH ligase, the fourth step of a five step MSH biosynthetic pathway, was most sensitive to killing by GSNO (Fig. 1j), followed by SHP1 (Fig. 1e), a deletion mutant of mscR, and then the transposon mutant, R119, disrupted in egtA, which codes for a γ-glutamate-cysteine ligase that provides sulfur for the synthesis of ESH (Fig. 1o), when compared with wild type. ESH thus appears to play a smaller role in nitrosative stress resistance than MSH in M. smegmatis despite the fact that ESH is an excellent scavenger of peroxynitrite (19) and GSNO (13).

FIG 1.

Growth of M. smegmatis in 7H9 media treated with 0.75 mM formaldehyde, 0.75 mM methylglyoxal, 2.25 mM glycolaldehyde, 9 mM glyceraldehyde, or 0.75 mM GSNO; wild type (□), SHP1 (mscR⁻) (◊), SHP1C (◆), S24 (mshC⁻) (○), S24C (•), R119 (egtA⁻) (∇), and R119C (▼).

Inside macrophages, NO produced by the inducible NO synthase (iNOS) system can combine with GSNO; both NO and GSNO are known to be inhibitory to bacteria such as M. tuberculosis (20). In bacteria with GSH as the major LMW thiol, such as Escherichia coli, or in bacteria that import GSH from the environment, such as Haemophilus influenzae (21) and Streptococcus pneumoniae (22), S-nitrosothiol reductases that catabolize GSO have been found to be protective against nitrosative stress. Indeed, in S. pneumoniae, nitrosothiol reductase is required for survival in blood after an intraperitoneal challenge experiment, but does not affect the adherence of pneumococci to the nasopharynx in vivo (22). The situation may be different in MSH-containing bacteria, where GSH and GSNO are known to control the growth of M. tuberculosis within macrophages (23). GSNO produced by macrophages could be catabolized to NO and GSH, either chemically or enzymatically through the actions of carbonyl reductase 1, the thioredoxin system, or Cu–Zn superoxide dismutase (24), and the NO released could be used to form MSNO, the substrate for MscR. Interestingly, a mutant disrupted in the M. tuberculosis mscR gene has not been obtained, implying that this gene is essential (25), whereas a Corynebacterium glutamicum mscR mutant is viable and is able to grow as well as the wild type. In addition, a C. glutamicum double mutant disrupted in a gene encoding an acetaldehyde dehydrogenase, ald, and mscR and another double mutant disrupted in disrupted in ald and mshC were severely impaired in their ability to oxidize formaldehyde (26).

Mutants Disrupted in mscR and mshC but Not egtA Are Sensitive to Aldehyde Stress

Endogenous sources of reactive aldehydes include sugars, such as fructose whose metabolism results in glyceraldehyde and glycolaldehyde intermediates, the latter being further oxidized to glyoxal (27) and glucose, whose metabolism can result in the formation of methylglyoxal (28). Both SHP1 (Figs. 1b–1d) and S24 (Figs. 1g–1i) were sensitive to killing by various short-chain aliphatic aldehydes, formaldehyde, methylglyoxal, glycolaldehyde, and glyceraldehyde. SHP1 and S24 were also sensitive to formaldehyde, a reactive byproduct of methanol metabolism, as well as demethylation reactions in the cell (Figs. 1a and 1f, respectively). R119 showed no significant sensitivity to aldehydes, indicating that ESH is not involved in the protection against these stresses (Figs. 1k–1n). Taken together, these data indicate that both MscR and MSH also play roles in the detoxification of short-chain aldehydes, which are formed as metabolic byproducts and are toxic to cells.

Oxygen Tension Increases the Stress of Aldehydes

Growth curves were set up with tubes having varying head space–volume ratio to determine if oxygen tension has an effect on killing in wild type, SHP1, and SHP1C. There was no obvious difference in growth when the cultures were treated with NaNO₂ (data not shown), which is not surprising as most bacteria can generate NO either aerobically by NO synthases or anaerobically from nitrite. Under aldehyde stress, SHP1 was more sensitive than wild type to killing by aldehydes, and the complemented strain (SHP1C) had intermediate sensitivity under all headspace–volume ratios (Fig. 2). However, SHP1 was substantially more sensitive to formaldehyde or glycolaldehyde when grown under high oxygen tension (Figs. 2i and 2k) when compared with lower oxygen tension (Figs. 2a and 2c). Short-chain aldehydes can oxidize to form dicarbonyl species in the presence of oxygen, and these highly reactive dicarbonyl species may be one reason for the increased sensitivity of SHP1 to formaldehyde and glycolaldehyde under aerobic conditions (29).

FIG 2.

Growth of M. smegmatis in 7H9 media treated with 0.75 mM formaldehyde, 0.75 mM methylglyoxal, 2.25 mM glycolaldehyde, and 9 mM glyceraldehyde. Cultures were grown in tubes with a headspace–volume ratio of 4.0 (a–d), 1.5 (e–h), and 0.66 (i–l); wild type (□), SHP1 (mscR⁻) (◊), and SHP1C (◆).

mscR and mshC expression increases under aldehyde stress and. mscR, mshC, and egtA Expression Increases under Nitrosative Stress

Induction of bacterial S-nitrosthiol reductases has been studied in several organisms and seems to be triggered by distinct stimuli. The GSNOR gene of H. influenzae is induced by formaldehyde but not GSNO (21), whereas the S. pneumoniae gene is induced in response to GSNO (22). An increase in mscR expression in response to formaldehyde has been documented in Rhodococcus erythropolis, a methylotrophic bacterium (30), and the expression of both adhCs (nitrosothiol reductases) present in Acinetobacter baumannii is significantly enhanced by the presence of subinhibitory concentrations of formaldehyde in the culture media (31). To determine how the expression of mscR, mshC, and egtA changes in response to nitrosative stress and aldehyde stress in M. smegmatis, qPCR was performed to monitor gene expression. A fivefold to 10-fold increase in the expression of mscR and mshC was detected in response to aldehyde and nitrosative stress. In contrast, expression of egtA did not significantly change on exposure to aldehyde stress; however, there was a significant increase in the expression of egtA in response to GSNO, consistent with the finding that the R119 mutant is sensitive to GSNO but not aldehyde stress (Fig. 3).

FIG 3.

Expression of mscR (black), mshC (gray), and egtA (white) in M. smegmatis mc²155 after 1-h incubation in 0.75 mM formaldehyde, 0.75 methylglyoxal, 2.25 mM glycolaldehyde, 9 mM glyceraldehyde, or 0.75 mM GSNO.

Mutants Disrupted in mscR and mshC Have Decreased Biofilm Formation

M. smegmatis forms biofilms that appear as surface pellicles at the air–liquid interface due to the presence of large amounts of free mycolic acids in its extracellular matrix, instead of polysaccharides commonly found in the matrix of other biofilms (32). The biofilm appears as a complex network of ridges with a bulk fluid sequestered underneath the surface pellicle (32). Examination of biofilms produced by S24 and SHP1 revealed a striking difference in the appearance of these biofilms when compared with the wild type; both mutants displayed thin biofilms with less ridges, and this phenotype was partially reversed in the complemented strains (Fig. 4 and Table 2). MSH and ESH can chemically react with NO directly or in a transnitrosylation reaction with GSNO, and thus, the lack of either thiol should lead to increased levels of NO and GSNO, as would the disruption of mscR, resulting in increased protein nitrosylation and presumably inactivation and/or alteration of cell signaling involving NO. The role of NO signaling during the biofilm lifecycle has been examined in P. aeruginosa, where NO was discovered to be produced at the same time as cell death and dispersal (33), and thus, NO may also play important functional roles throughout the developmental stages of the mycobacteria biofilm lifecycle.

FIG 4.

M. smegmatis biofilms formed in TSB media supplemented with 1% procalamine and incubated at 30 °C for 72 h with no disturbance. Wild-type strain compared with SHP1 (mscR⁻), SHP1C (complemented SHP1), S24 (mshC⁻), S24C (complemented S24), R119 (egtA⁻), and R119C (complemented R119).

TABLE 2.

Biofilm formation in M. smegmatis strains

Strain	OD630	P-value*
Wild type	1.153 ± 0.012
A1 (mshA−)	1.057 ± 0.028	0.01514
A1C	1.116 ± 0.020	0.0180**
S24 (mshC)−	0.852 ± 0.132	0.04119
S24C	1.159 ± 0.034	0.88105
R119 (egtA)−	1.188 ± 0.003	0.02736
R119C	1.054 ± 0.030	0.01847
SHP1 (mscR)−	0.713 ± 0.101	0.00249
SHP1C	1.120 ± 0.009	0.03352

Biofilms were assayed at OD₆₃₀ for four replicates.

^*P-value when compared with wild type calculated using a paired Student’s two-tailed t-test.

^**Partial complementation due to inducible nature of the acetamidase promoter controlling the mshA expression in pALACE vector (22).

Mutants, disrupted in mshC, such as S24, are sensitive to oxidative stress (34), which may hinder biofilm formation due to the accumulation of reactive oxygen species. Interestingly, A1, disrupted in mshA, which codes for a glycosyl transferase catalyzing the first step in the MSH biosynthetic pathway (35) also had reduced biofilm formation albeit not as severely reduced as S24 and SHP1 (Table 2), and this was partially reversed in the complemented strain. R119 had slightly improved biofilm formation when compared with wild type, and this was reversed in the complemented strain R119C. The less attenuated phenotype of A1 may be due to the overproduction of Ohr, an organic hydroperoxide resistance protein, and increased levels of ESH, which compensate for the lack of MSH and reduce oxidative stress in this mutant (36). Recently, Trivedi et al. reported that M. tuberculosis exposed to DTT, which causes an increase in reductive stress, produced more biofilm. Genes involved in protection against oxidative stress, such as katG, ahpC, and ahpD, were upregulated on DTT exposure (37). It is also possible that in A1 and R119, an enzyme that neutralizes NO and/or GSNO maybe upregulated in a similar manner as Ohr. Mutants disrupted in other genes involved in oxidative stress protection, such as the superoxide dismutase mutant in Listeria monocytogenes (38) and a mutant disrupted in glutathione reductase in N. gonorrhoeae (39), were previously found to be impaired in biofilm formation. There was no difference in biofilm formation in Pseudomonas aeruginosa catalase mutants (40), and an alkyl hydroperoxide reductase mutant of Campylobacter jejuni (41) had increased biofilm formation.

As the formation of a biofilm often leads to persistent and chronic infections, which increase morbidity and mortality, our results suggest that novel strategies to inhibit and/or disrupt biofilm formation and possibly to alter biofilm dispersal (inducing or preventing it) may exist through targeting LMW thiols and S-nitrosothiol reductases in bacteria.