Low-Molecular-Weight Thiols and Thioredoxins Are Important Players in Hg(II) Resistance in Thermus thermophilus HB27

ABSTRACT

Mercury (Hg), one of the most toxic and widely distributed heavy metals, has a high affinity for thiol groups. Thiol groups reduce and sequester Hg. Therefore, low-molecular-weight (LMW) and protein thiols may be important cell components used in Hg resistance. To date, the role of low-molecular-weight thiols in Hg detoxification remains understudied. The mercury resistance (mer) operon of Thermus thermophilus suggests an evolutionary link between Hg(II) resistance and low-molecular-weight thiol metabolism. The mer operon encodes an enzyme involved in methionine biosynthesis, Oah. Challenge with Hg(II) resulted in increased expression of genes involved in the biosynthesis of multiple low-molecular-weight thiols (cysteine, homocysteine, and bacillithiol), as well as the thioredoxin system. Phenotypic analysis of gene replacement mutants indicated that Oah contributes to Hg resistance under sulfur-limiting conditions, and strains lacking bacillithiol and/or thioredoxins are more sensitive to Hg(II) than the wild type. Growth in the presence of either a thiol-oxidizing agent or a thiol-alkylating agent increased sensitivity to Hg(II). Furthermore, exposure to 3 μM Hg(II) consumed all intracellular reduced bacillithiol and cysteine. Database searches indicate that oah2 is present in all Thermus sp. mer operons. The presence of a thiol-related gene was also detected in some alphaproteobacterial mer operons, in which a glutathione reductase gene was present, supporting the role of thiols in Hg(II) detoxification. These results have led to a working model in which LMW thiols act as Hg(II)-buffering agents while Hg is reduced by MerA.

IMPORTANCE The survival of microorganisms in the presence of toxic metals is central to life's sustainability. The affinity of thiol groups for toxic heavy metals drives microbe-metal interactions and modulates metal toxicity. Mercury detoxification (mer) genes likely originated early in microbial evolution in geothermal environments. Little is known about how mer systems interact with cellular thiol systems. Thermus spp. possess a simple mer operon in which a low-molecular-weight thiol biosynthesis gene is present, along with merR and merA. In this study, we present experimental evidence for the role of thiol systems in mercury resistance. Our data suggest that, in T. thermophilus, thiolated compounds may function side by side with mer genes to detoxify mercury. Thus, thiol systems function in consort with mer-mediated resistance to mercury, suggesting exciting new questions for future research.

INTRODUCTION

Mercury (Hg) is one of the most toxic and widely distributed heavy metals. Mercury toxicity is due in part to its high affinity for low-molecular-weight (LMW) thiols, including homocysteine, N-acetylcysteine, and cysteine, as well as thiol-based cellular redox buffers, such as glutathione (GSH) (1, 2). Not surprisingly, free thiol groups have a high binding constant for Hg(II) (3, 4). Mercury can also bind thiols present in proteins (1, 2, 5) and can oxidize thioredoxins (6, 7). Thus, the main biological effects of Hg are related to this high affinity for sulfhydryl groups (2).

Bacteria commonly possess two main thiol systems that maintain the redox state of the cell. The first is an LMW thiol system, typically utilizing GSH (8) or bacillithiol (BSH) (9), and the second is a thioredoxin system (10). LWM thiol systems vary among bacteria, are present at millimolar concentrations, and act as redox buffers in the cell (10, 11). Thioredoxins are present among all kingdoms, and the system consists of a thioredoxin protein(s) and an enzyme, thioredoxin reductase (10, 12). In Escherichia coli, there are two thioredoxins, Trx1 (encoded by trxA) and Trx2 (encoded by trxC), which differ in an extra N-terminal extension (10). Thioredoxins become oxidized when their targets are reduced, and thioredoxin reductase (encoded by trxB) uses electrons from NADPH to maintain thioredoxins in a reduced state (10, 12).

To overcome Hg toxicity, some bacteria and archaea employ the Hg resistance (mer) operon (13). The composition of the mer system varies among organisms, but they all have merA, which encodes a mercuric reductase that reduces inorganic Hg(II) to Hg(0); Hg(0) is volatile and is partitioned out of the cell. Proteobacterial mer operons are the most studied of these operons (13). These operons have several genes that can encode transcriptional repressors/activators (merR and merD) or specific transporters (merC, merE, merF, merG, merP, and merT), as well as merA (14). Some mer operons can detoxify organomercurial compounds, in addition to Hg(II), by the inclusion of an organomercurial lyase (encoded by merB). With the advance of whole-genome sequencing, new mer operons were discovered, and simpler operons were found in early microbial lineages (14). These simpler operons were found in some thermophilic microbes, such as the crenarchaeote Sulfolobus solfataricus (15), some bacteria belonging to the Aquificaceae (16), and Thermus thermophilus (17). These organisms were subsequently shown to have merA-dependent resistance to Hg(II).

The mer operon in T. thermophilus HB27 is unique because it consists of two classical mer operon genes (merA and merR) and the thiol biosynthesis-related gene (oah2) (Fig. 1A) (17). In the mer operon of T. thermophilus HB27, merA encodes a mercuric reductase, as suggested by protein homology to other MerA proteins (17, 18), a higher susceptibility of the ΔmerA mutant to Hg(II), and the lack of MerA activity in this mutant (17). Although the function of MerR in T. thermophilus has not been established, protein homology suggests that it is a transcriptional regulator encoded in the mer operon. The third member of the mer operon, oah2, encodes an O-acetyl-l-homoacetylserine sulfhydrylase (Oah) that synthesizes homocysteine (Fig. 1C) (19), an intermediate in methionine biosynthesis (20). Our previous work found that oah2, merR, and merA are expressed as a polycistronic unit in the presence of Hg(II) (17). The colocalization and coexpression of oah2 with the mer genes suggest a link between Hg(II) resistance and the biosynthesis of LMW thiol compounds.

FIG 1.

mer and met operons and methionine metabolism pathways in T. thermophilus. (A) Diagram of the mer operon in HB27, which is composed of oah2 (TT_C0792), merR (TT_C0791), and merA (TT_C0789). (B) Diagram of the met operon in HB27, composed of oah1 (TT_C0408) and met2 (TT_C0407). Arrows indicate the direction of transcription. (C) Methionine metabolism in T. thermophilus (20). The met2 gene encodes homoserine O-acetyltransferase, and the oah genes encode O-acetyl-homoserine sulfhydrylases.

The genome of T. thermophilus carries two oah gene orthologs: oah2 and oah1. oah1 is located in an operon with met2, which encodes an enzyme that catalyzes the first step in methionine biosynthesis (Fig. 1B and C) (20). It was determined that Oah1 and Oah2 have sulfhydrylase activity in vitro (19, 21); however, Oah2 has a lower K_m for the substrates homoserine and sulfide (19). This information suggested that these enzymes may have similar, but possibly not identical, cellular functions. In the present study, the hypothesis that LMW thiols and the thioredoxin system play a role in Hg(II) resistance in T. thermophilus was tested by using a combination of gene expression, phenotypic, and metabolite analyses. We report that exposure to Hg(II) induced the expression of LMW thiol biosynthesis and thioredoxin genes and decreased the bioavailability of reduced bacillithiol and cysteine. Moreover, phenotypic analysis found that strains lacking Oah, BshA/C, or TrxA were more sensitive to Hg(II) than the wild-type (WT) strain. Database searches indicate that all Thermus sp. mer operons have an oah2 gene and that this phenomenon is not exclusive to Thermus spp.

RESULTS

Transcription of thiol-related genes is induced by Hg(II).

To begin to examine the role of thiols in Hg(II) resistance in T. thermophilus, the effects of Hg(II) on transcript levels of genes that encode thiol-related enzymes were evaluated. Bioinformatic analysis determined that the genome of strain HB27 lacks GSH biosynthesis genes, as has been reported for Deinococcus radiodurans (another member of the Deinococcus-Thermus phylum) (22). Three putative genes for the biosynthesis of bacillithiol (BSH) were detected (bshA, bshB, and bshC), as were the cysE (23) and cysK/oas (24) cysteine biosynthesis genes and the met2, oah2, oah1, and metH methionine biosynthesis genes (Fig. 1C) (20). Also, gene homologs encoding two thioredoxins (trxA1 and trxA2) and one thioredoxin reductase (trxB) were present in the genome (for locus tag identifiers of these genes, see Materials and Methods).

mRNA transcripts in WT cells exposed to 0 or 1 μM Hg(II) for 7.5, 15, or 60 min were quantified, and transcript abundances were normalized to that of gyrA (see Table S1 in the supplemental material) (25). Mercury treatment increased transcript levels of all thiol-related genes tested, from over 10-fold for oah1 and trxA1 to about 2-fold for bshC (Fig. 2). For LMW thiol genes, the highest induction was achieved after 7.5 min of Hg(II) exposure, followed by a decrease at 15 min. With regard to methionine biosynthesis genes (Fig. 2A), met2 showed fold inductions similar to those for oah1 at all time points tested, suggesting that these two genes are likely expressed from an operon. As expected, oah2 was induced (17), showing ∼5-fold induction, which was the same fold induction reached by metH. These data suggest that methionine and homocysteine biosynthesis genes are induced by short exposure to Hg(II). Likewise, oas was induced ∼7-fold by Hg(II). Transcript levels of bshA and bshB increased approximately 4-fold (Fig. 2B) upon Hg(II) exposure. Furthermore, transcript levels of thioredoxin-related genes were increased at 7.5 min (Table S2), but the highest fold induction was seen at 15 min, reaching ∼10-fold for trxA1 and trxB and ∼7-fold for trxA2 (Fig. 2C).

FIG 2.

Thiol systems are induced by Hg(II). Abundances of mRNA transcripts involved in methionine and cysteine biosynthesis (A), BSH biosynthesis (B), and the thioredoxin system (C) were measured in the WT strain after 7.5, 15, or 60 min of exposure to 1 μM Hg(II). Gene expression was normalized to the abundance of the gyrA transcript and graphed as fold induction relative to the control level (ΔΔC_T method). Averages and standard deviations represent triplicate samples from three independent trials. Statistical analysis was conducted using the two-tailed t test. *, P < 0.05; **, P < 0.03; ***, P < 0.001.

GSH and thioredoxins are oxidized or consumed when they reduce and sequester Hg(II) (5, 6, 26, 27), and this interaction with Hg can lead to disulfide stress. Mercury exposure can also produce reactive oxygen species (ROS), such as superoxide and hydrogen peroxide (6, 26, 27). Therefore, we evaluated the possibility that, in T. thermophilus, the 7.5-min induction of LMW thiol genes was due to these indirect effects triggered by Hg(II). For this purpose, known stressor agents that produce disulfide stress (diamide) and ROS were used. Exposure to 4 mM diamide (a thiol-oxidizing agent), 160 μM paraquat (a superoxide generator), or 1 mM hydrogen peroxide for 7.5 min did not induce oah1, oah2, oas, or bshA transcript levels in WT cells (Fig. S1). However, exposures induced expression of control genes for each stressor agent. Addition of diamide increased transcript levels of trxA1 (Fig. S1, control gene), suggesting that an increase in thioredoxin transcripts by Hg(II) could be an effect of disulfide stress. On the other hand, paraquat and hydrogen peroxide did increase SOD transcript levels (Fig. S1, control gene). These results suggest that induction of LMW thiol biosynthesis genes was likely triggered directly by Hg(II) rather than indirectly by disulfide stress (diamide) or accumulation of ROS (superoxide or hydrogen peroxide). Furthermore, the fast induction of LMW thiol genes caused by Hg(II) (7.5 min postexposure) suggests that this effect is not a consequence of unrelated Hg(II)-triggered responses that would likely require longer incubation times to be manifested. However, induction of protein-thiol systems, such as thioredoxins (Fig. 2C), might be related to Hg(II)-induced disulfide stress.

To evaluate if Hg(II) induction of thiol-related genes was unique to Thermus, the induction of homologous systems (thioredoxin, glutathione, cysteine, and methionine biosynthesis) was tested in Escherichia coli strain K-12, which lacks a mer system. When cells were exposed to 2 μM Hg(II) for 7.5 or 30 min, no increase in transcript level of any of the tested thiol genes was observed when the ssrA gene (Fig. S2) or rrsB (not shown) was used as a reference gene (28). There was no increase in transcript levels of genes involved in the biosynthesis of glutathione (gshA and gshB) or cysteine/homocysteine (cysK, cysM, or malY) or in protein thiol systems, such as glutathione reductase (GR) (gor) or thioredoxin reductase (trxB) (Fig. S2). It has previously been shown that zntA is induced by Hg(II) in E. coli (29), and as expected, Hg(II) exposure resulted in a 12-fold induction of the transcript level of this gene (Fig. S2).

In summary, exposure to Hg(II) elicited two distinct but complementary responses in T. thermophilus, but not in E. coli. First, LMW thiol biosynthesis genes showed a rapid increase in transcript levels followed by a rapid decline. Second, the thioredoxin system appeared to be induced more slowly than LMW biosynthesis genes, possibly as a consequence of disulfide stress that results from Hg exposure.

Thiol genes are involved in Hg(II) resistance.

A role for Oah and BSH in Hg(II) resistance has not been described. Thioredoxins are known to be oxidized by Hg(II) (6, 7), but little is known about their physiological role in modulating Hg(II) toxicity in bacteria. To begin to examine their roles in Hg(II) resistance, thiol-related genes were individually replaced with a thermostable kanamycin resistance gene (HTK gene) (30) in the WT HB27 background. These included two genes for the biosynthesis of methionine/homocysteine (oah1 and oah2), two genes involved in the biosynthesis of BSH (bshA and bshC), and two thioredoxin genes (trxA1 and trxA2).

When the role of the methionine/homocysteine biosynthesis pathway in Hg(II) resistance was evaluated in complex medium, the Δoah2 and Δoah1 strains were as resistant to Hg(II) as the WT strain (Fig. S3A). For reasons that are not currently understood, these mutant strains were unable to grow with sulfate as the sole sulfur source (data not shown), and therefore growth and Hg(II) resistance were analyzed in a chemically defined medium supplemented with 10 μM homocysteine. As illustrated in Fig. 3A, the Hg(II) 50% inhibitory concentrations (IC₅₀s) for the Δoah1 and Δoah2 strains were ∼2 μM, and the IC₅₀ for the WT strain was ∼4.5 μM. These results suggest that the cell's ability to synthesize homocysteine/methionine affects its Hg(II) resistance in defined medium. Thermus is found in hot springs, where the sulfur source is mostly sulfide or sulfate (31–33), indicating that these genes may have environmental relevance to Hg(II) resistance.

FIG 3.

Sensitivity to Hg(II) increases in mutant strains lacking LMW thiol synthesis genes or thiol homeostasis systems. The graphs show the effects of Hg(II) on growth in defined medium supplemented with 10 μM homocysteine after 24 h (A) or in defined medium supplemented with 5.2 mM sulfate after 20 h (B and C). Each point represents the average for three independent experiments, and standard deviations are shown. Statistical analysis was conducted using one-way ANOVA followed by the Tukey test. *, P < 0.001.

oah2 is part of the mer operon in T. thermophilus (17) (Fig. 1A); therefore, the difference in Hg(II) resistance between the Δoah2 strain and the WT strain may be due to a polar effect on merA expression. To test this possibility, we measured MerA activities in crude cell extracts of the WT, Δoah, and ΔmerA strains. MerA activities were statistically indistinguishable between the WT and Δoah2 strains in the presence of Hg(II), while the ΔmerA strain displayed negligible activity (Fig. S4).

The BSH mutant strains were slightly more sensitive to Hg(II) than the WT strain when they were grown in complex medium (Fig. S3B), and this difference was enhanced when cells were grown in defined medium containing sulfate as the sulfur source. The Hg(II) IC₅₀s for the ΔbshA and ΔbshC mutants were ∼0.2 μM, whereas the WT had an IC₅₀ of ∼1 μM (Fig. 3B). The lower Hg(II) resistance of the strains in defined medium than that in complex medium was likely due to a decrease in exogenous Hg(II) ligands, which affect Hg(II) bioavailability (34). Furthermore, the higher sensitivity of the WT to Hg(II) in the defined medium supplemented with sulfate (Fig. 3B and C) than that in medium supplemented with homocysteine (Fig. 3A) (1.6 times more sensitive) may have been due to the fact that homocysteine can bind Hg(II) (1, 2).

For the study of the thioredoxin system, ΔtrxA1 and ΔtrxA2 strains were constructed. Similar to the Δbsh strains, the ΔtrxA strains were slightly more susceptible to Hg(II) than the WT strain in complex medium (Fig. S3C). In defined medium, the ΔtrxA mutants were significantly more sensitive to Hg(II) than the WT. The Hg(II) IC₅₀ was <0.2 μM for the ΔtrxA1 and ΔtrxA2 mutants and ∼1 μM for the WT (Fig. 3C).

Taken together, the results in Fig. 3 and Fig. S3 clearly show that LMW thiols and thioredoxins enhance Hg(II) resistance in T. thermophilus HB27. It is environmentally significant that these genes have a larger impact on Hg(II) toxicity when sulfate is used as a sulfur source. In natural springs, Thermus spp. do not have access to complex sulfur sources to use as substrates (31–33).

Oxidation state of BSH and alternate LMW thiols is affected by Hg(II).

Due to the higher sensitivity of LMW thiol biosynthesis mutants to Hg(II), we examined how Hg(II) affected intracellular LMW thiol pools. The intracellular concentrations of LMW thiols were measured by utilizing the fluorescent probe monobromobimane (mBrB) (35, 36), and products were identified and quantified by high-pressure liquid chromatography (HPLC) analysis coupled with a fluorescence detector. mBrB reacts with free thiols (reduced thiols) to produce a fluorescent derivative, which is detected in the elute. Thus, this assay detects only reduced thiols.

The main redox buffer detected in T. thermophilus was BSH, which was present at 27.1 ± 8.5 nmol/g cellular dry weight (Fig. 4A; Table S3 and Fig. S5A and B). Two other LMW thiols were detected: cysteine (6.1 ± 3.1 nmol/g) and a large pool of sulfide (324.1 ± 88.4 nmol/g) (Fig. 4A; Fig. S5 and Table S3). The sulfide:BSH:Cys molar ratio in the WT strain was ∼ 12:1:0.2 (Fig. S5A and Table S3). As expected, BSH was not detected in the LMW thiol pools from the ΔbshA and ΔbshC strains (Fig. 4A; Table S3), confirming the role for both genes in the BSH biosynthetic pathway.

FIG 4.

BSH is the primary LMW thiol in T. thermophilus, and LMW thiols are responsive to Hg(II). (A) Cultures were grown in complex medium to an OD₆₀₀ of ∼0.4, and LMW thiols were quantified for the WT, Δoah1, Δoah2, and ΔbshC strains. (B) The WT strain was grown in complex medium to an OD₆₀₀ of ∼0.4, cells were exposed to different Hg(II) concentrations for 60 min, and small thiols were quantified. The thiol concentrations in cultures not exposed to Hg(II) were considered 100%. Averages and standard deviations represent triplicate samples from at least three independent trials. ND, not detected.

We next examined whether the LMW thiol pool is affected by exposure to Hg(II). Cells were incubated with Hg(II) for 1 h prior to determining the concentrations of free (reduced) LWM thiols. The addition of Hg(II) decreased free thiol pools in the WT strain (Fig. 4B; Table S3 and Fig. S5B) and the ΔmerA strain (Table S3). Upon exposure to 1 μM Hg(II), cysteine was undetected, and only 2% of reduced BSH remained in the WT strain (Fig. 4B; Table S3). BSH was undetected upon exposure of cells to 3 μM Hg(II) (Fig. 4B; Table S3 and Fig. S5B). On the other hand, BSH and Cys were undetected when the ΔmerA strain was exposed to 1 μM Hg(II) (Table S3), suggesting that Hg(II) removal by MerA in the WT strain alleviates Hg(II)-induced disulfide stress.

These results suggest that LMW thiols scavenge Hg(II). In T. thermophilus, BSH is the primary Hg(II) buffer, although cysteine, which is present at a much lower concentration, also acts as a Hg(II) buffer. The sequestration of Hg(II) by intracellular thiols is consistent with the hypothesis that induction of LMW thiol biosynthesis genes upon Hg(II) exposure (Fig. 2A and B) may serve to increase the concentration of thiol ligands as a primary response to Hg(II) toxicity. Expression of thioredoxins (Fig. 2C) may help to regenerate LMW thiol pools (37) that are sequestering Hg(II).

Cellular redox state modulates Hg resistance.

The depletion of LMW thiols by Hg(II) and the importance of thioredoxins in Hg(II) resistance led us to hypothesize that a disturbance in the balance of reduced/oxidized thiols would affect Hg(II) resistance. Diamide is a thiol-oxidizing agent that produces disulfide stress (38) by interfering with thiol-dependent processes in the cell. In the presence of diamide at either 1 mM (Fig. S6) or 2 mM (Fig. 5A), the WT, Δoah2, and ΔmerA strains were more sensitive to Hg(II). Resistance to Hg(II), as measured by the IC₅₀, decreased with increasing diamide concentrations in both the WT and ΔmerA strains. As expected, the WT strain was more resistant than the ΔmerA strain at all diamide concentrations (Fig. 5A; Fig. S6). The addition of diamide had the same effect as growth under sulfur-limiting conditions (Fig. 3A) in the Δoah2 strain: the Δoah2 strain was more sensitive to Hg(II) than the WT strain (Fig. 5A; Fig. S6).

FIG 5.

Decreased thiol availability enhances Hg(II) toxicity. (A) The effect of diamide on Hg(II) toxicity was evaluated in the presence (2 mM) (empty symbols) or absence (filled symbols) of diamide after 16 h of growth of the WT (circle), Δoah2 (triangle), and ΔmerA (square) strains. (B) The effect of the thiol-alkylating agent NEM on Hg(II) toxicity was evaluated after 24 h of growth of the WT strain. The culture optical density (A₆₀₀) with 0 μM Hg(II) in complex medium was considered 100% growth. Averages and standard deviations for at least 5 independent trials are shown. Statistical analysis was conducted using the two-tailed t test. *, P < 0.035; **, P < 0.001.

N-Ethylmaleimide (NEM) is a thiol-blocking agent that irreversibly binds to free thiols. When cells were exposed to 0.2 or 0.4 mM NEM, no difference in growth was detected compared to that of unexposed cells (not shown), indicating that NEM did not affect growth in the absence of Hg(II). When cells were coexposed to Hg(II) and NEM, they were more susceptible to Hg(II) than NEM-unexposed cells (Fig. 5B). The Hg(II) IC₅₀s for 0.2 mM and 0.4 mM NEM-exposed cells were ∼3 and ∼2.5 μM Hg(II), respectively, but the Hg(II) IC₅₀ was more than 4 μM for NEM-unexposed cells. Taken together, these findings clearly indicate that the intracellular redox state and thiol availability greatly affect Hg(II) toxicity.

Evolutionary aspects of thiol genes and mercury.

The presence of a gene related to the biosynthesis of thiol compounds, as we reported for oah2 in T. thermophilus's mer operon, has not been reported for other mer operons. To examine how general this phenomenon is, genomes from organisms belonging to the Deinococcus-Thermus phylum were searched for the presence of oah2 homologs or other thiol-related genes in the mer operons. Of the 75 Deinococcus-Thermus genomes available in June 2017, 32 belonged to Deinococcales and 43 to Thermales. Of the Thermales genomes, 30 genomes were from Thermus spp., and nine of them had mer operons (Table S4). All the Thermus sp. mer operons had the oah2 gene or, as annotated, O-acetylhomoserine (thiol)-lyase. Only four other Thermales genomes, all belonging to the Thermaceae family, had mer operons, but none of them included the oah2 gene. Only one Deinococcales genome had a mer operon lacking oah2. Thus, oah2 is exclusive and always present in mer operons of Thermus spp. In all of these operons, oah2 is located upstream of merR and merA. As in HB27, the genomes of all mer-containing Thermus spp. also had oah1. A phylogenetic analysis performed with Oah1 and Oah2 sequences showed a clear separation of the two proteins to two unique clades (Fig. 6; Fig. S7), suggesting that these two genes are evolving independently of each other.

FIG 6.

Molecular phylogenetic analysis of Oah proteins in Thermus spp. by the maximum likelihood method. The Oah proteins encoded in the Thermus sp. genomes were associated with mer operons (Oah2) or met operons (Oah1). The figure shows an enlargement of the original tree (see Fig. S7 in the supplemental material), presented to highlight the diversification of Oah homologs. Oas (CysK) protein sequences were used as an outgroup (for the full phylogenetic tree, refer to Fig. S7). The phylogeny with the highest log likelihood (−3,752.97) is shown. The tree is drawn to scale, with branch lengths measured in numbers of substitutions per site. Numbers at bifurcation points indicate bootstrap values. Protein IDs or locus tags are provided in Table S6.

In a survey of 272 mer operons available in databases in December 2011 (14), the only taxon containing a thiol-related gene, aside from T. thermophilus, was the Alphaproteobacteria (T. Barkay, unpublished data), where 8 of 25 operons had an open reading frame (ORF) annotated gor, which encodes a glutathione reductase (GR). The main LMW redox buffer in Alphaproteobacteria is GSH, and GR keeps GSH reduced by using NADPH as an electron donor (10). In April 2017, there were 505 finished and assembled alphaproteobacterial genomes, and when they were searched for the presence of MerA, 40 additional mer operons were found, 5 of which included GR. Among the 65 alphaproteobacterial genomes that had a mer operon, only 13 genomes had a GR gene present in the mer operon (Table S5). With one exception, GR was always located downstream from mer transport genes. Another commonality between Alphaproteobacteria and Thermus spp. is the presence of a GR-paralogous gene elsewhere in the genome. As designated for Thermus, GRs encoded in the mer operons were named GR2, and the non-mer-operon GRs were designated GR1. Similar to the Oah proteins, the GR2 proteins appear to be evolving independently from GR1, as suggested by the clustering pattern of the alphaproteobacterial GR phylogeny (Fig. S8).

These data indicate that the presence of a thiol gene in a mer operon is not unique to Thermus spp. The role of the GR system in Hg(II) resistance in Alphaproteobacteria remains to be studied.

DISCUSSION

The prior observation that in T. thermophilus HB27 oah2 is cotranscribed with the mer genes in response to Hg(II) (17) led us to discover the following about this bacterium: (i) expression of LMW and protein thiol genes are induced by Hg(II), (ii) BSH is the primary redox buffer, and (iii) BSH, along with other LMW thiols and thioredoxins, increases cellular resistance to Hg(II). Together these results highlight a role of LMW and protein thiols in mitigating Hg(II) toxicity, supporting their integration into the paradigm of cellular defense against this highly toxic heavy metal.

The interaction of thiols with metals is one of life's foundations; this interaction controls functional and structural attributes of molecules and cells. In fact, sequestration of soft metals by thiol compounds, such as metallothioneins and phytochelatins, is a well-established mechanism of metal resistance in which the biosynthesis of the sequestering molecules is induced by exposure to the metals (39). The central paradigm for microbial resistance to Hg(II) is transformation to Hg(0) (13), but although the role of GSH in Hg(II) resistance is known (27, 40), not much information is available about other thiol agents. The Thermus system combines MerA-dependent reduction (17) with thiol-based sequestration (this study). Upon Hg exposure, the induction of genes encoding LMW and protein thiol systems along with merA suggests a role for both in Hg(II) detoxification. Note that mercury-dependent induction of thiol biosynthesis genes did not occur in E. coli (see Fig. S2 in the supplemental material), consistent with a previous study in which Hg(II) exposure did not induce expression of the glutathione transferase gene (28).

This study showed an induction of various LMW thiol biosynthesis genes by Hg(II). Moreover, knockout mutants of all thiol biosynthesis genes were more susceptible to Hg(II) than the WT was, connecting these genes and their Hg(II)-induced expression with a role in resistance to Hg(II). We propose the following model (Fig. 7) to explain how Hg(II) sequestration by thiols and reduction work together. In Hg(II)-exposed cells, the expression of thiol biosynthesis and mer genes is induced. LMW thiols bind Hg(II) and prevent Hg(II) from damaging sensitive targets in the cell. Oah synthesizes homocysteine that can be used as a precursor in cysteine biosynthesis (Fig. 7, dashed arrow); homocysteine may also bind Hg(II) with its free thiol (not shown). Cysteine availability is ensured by Hg(II)-induced overexpression of oas, and this cysteine can be used by BshC to finalize the biosynthesis of BSH. In addition, thioredoxins might directly reduce or sequester Hg(II), and the resulting thioredoxin-Hg complex can be resolved by NADPH-dependent thioredoxin reductase leading to Hg detoxification. The expression of merA results in the reduction of Hg(II) to Hg(0), leading to removal of the metal. This model accounts for the two cellular lines of defense against Hg(II) toxicity, i.e., sequestration by thiol-based redox buffers and a mer-based detoxification system. The role of LMW thiol agents, including BSH, in Hg(II) resistance among prokaryotes warrants additional study.

FIG 7.

Proposed model for a two-tiered response to Hg(II) toxicity in T. thermophilus. Hg(II) exposure (black arrows) induces expression of mer, LMW thiols, and thioredoxin systems. The latter two aid in sequestering Hg(II) until reduction of Hg(II) to the less toxic form Hg(0) by MerA. Abbreviations: BSH, bacillithiol; Hcys, homocysteine; Cys, cysteine; hp, hypothetical protein; TR, thioredoxin reductase; Trx, thioredoxin.

The important question raised by this model is how thiol agents interact with MerA in Hg detoxification. One possibility is suggested by the fact that in Thermus MerA lacks NmerA, the N-terminal extension of 70 amino acids (41). In proteobacterial MerA, NmerA delivers Hg(II) to the catalytic core of the enzyme; when NmerA is absent, Hg(II) can be transferred, less efficiently, by thioredoxins and GSH (41). Generally, taxa carrying the core MerA exhibit millimolar concentrations of LMW thiols, but those carrying full-length MerA (i.e., including NmerA) exhibit lower concentrations of LMW thiols (18). The lack of the NmerA domain in Thermus spp. might suggest that in addition to their role in intracellular sequestration of Hg(II), thioredoxins and/or BSH may play a role in tolerance by delivering Hg(II) to the enzyme.

Our results demonstrate a role for the thioredoxin/thioredoxin reductase system in Hg(II) resistance. Most studies on thioredoxin systems and Hg(II) have been performed in eukaryotic cells. These studies found that Hg(II) induces trx expression (42) but not enzymatic activities (7, 42), likely due to inhibition of the proteins by Hg(II) (7). Bacterial thioredoxin systems are different from those of mammalian cells. The little that is known about bacterial thioredoxin systems under Hg(II) stress indicates that bacterial thioredoxin, unlike the human enzyme, does not dimerize in the presence of Hg(II) (7). Moreover, Hg(II) oxidizes thioredoxin in Geobacter sulfurreducens but not in Shewanella oneidensis (6), and Hg(II) exposure increases the amount of thioredoxin reductase in Corynebacterium glutamicum (43). Whereas our study adds new information on the role of the thioredoxin system in response to Hg(II) in a thermophilic bacterium, additional studies are needed to better understand the role of thioredoxin interactions with toxic metals, considering the diversity in redox homeostasis systems among prokaryotes (11).

Identification of BSH as a major LMW thiol agent in Thermus is another new finding of our research. Whereas this was expected because BSH was reported to be the main LMW thiol agent in the related taxon Deinococcus (22), we showed that Δbsh mutants are more susceptible to Hg(II) than the WT. We also determined that similar to what has been reported for Bacillus subtilis (44), T. thermophilus BSH mutants were more sensitive to Zn(II) than the WT strain (data not shown). This information suggests that BSH has a general role in metal ion tolerance and/or buffering (45). Thermus mer operons are not the only ones that contain thiol-related genes; some alphaproteobacterial mer operons include gor, a glutathione reductase gene. Although studies have reported Hg resistance in environmental alphaproteobacterial isolates (46, 47) and the presence of alphaproteobacterial merA in soil metagenomes (48), the roles of gor and the mer operons in managing alphaproteobacterial Hg(II) stress have received little attention. Nevertheless, the presence of the gor gene (see Fig. S8 and Table S5 in the supplemental material) clearly suggests that the mer operon of T. thermophilus may not be the only example of the integration of genes involved in thiol systems. It is possible that in some alphaproteobacteria the supply of reduced glutathione is increased upon Hg(II) exposure, but this remains to be tested. Interestingly, Thermus sp. and alphaproteobacterial genomes carry a second copy of the thiol-related gene, and in both the two paralogs appear to be evolving independently of each other (Fig. 6; Fig. S8), suggesting the occurrence of evolution in response to different selective pressures. If this is so, the mer-associated thiol gene might have a unique function that is not shared with the chromosomally located gene.

In our prior research on the evolution of mer operons (14), we suggested a gradual evolution from simple, constitutively expressed operons (merA and a possible transporter gene) in early thermophilic lineages to a highly efficient and more complex detoxification system in the Proteobacteria. It is possible that LMW and protein thiols are another line of defense against Hg toxicity in at least one early bacterial lineage. We previously argued that the mer system originated among thermophilic bacteria from geothermal environments (14, 18) with high Hg levels of geological origin (32). The discovery of an LMW thiol-based defense against Hg toxicity in Thermus may suggest that the thiol-dependent cellular defense strategy originated early in the evolution of microbial life as well.

MATERIALS AND METHODS

Bacterial stains and growth conditions.

T. thermophilus HB27 (DSMZ 7039) and its mutants were cultured at 65°C in 461 Castenholz TYE medium (complex medium) as described by Wang et al. (17). Chemically defined medium was prepared as described by Tanaka et al. (49). For culture in liquid medium, cells were grown in 3 ml of medium in 13-ml test tubes shaken at 200 rpm. Solid culture medium was supplemented with 1.5% noble agar (Sigma). Where present, kanamycin (Kan) was used at 25 μg/ml (Sigma). E. coli strains were grown at 37°C in Luria-Bertani (LB) medium. Liquid cultures were shaken at 180 rpm, and solid medium was supplemented with 1.5% agar.

Mutant construction.

Construction of the ΔmerA mutant was described previously (17). To create gene replacements of the oah2, oah1, bshA, bshC, trxA1, and trxA2 genes with a kanamycin resistance gene (“Δgene::HTK”; denoted “Δgene” strains), the upstream and downstream flanking regions of the target gene were PCR amplified and the products fused with the thermostable Kan resistance gene, HTK (30) (see Fig. S9 in the supplemental material). Two strategies were used to fuse the PCR products. For the oah2 gene, different restriction sites were added to the 3′ ends of PCR fragments; for the other five genes, fusion PCR was performed (50). The final constructs were cloned into pUC19 and used to transform Max Efficiency DH5α competent cells (Invitrogen), with transformants selected on LB plates supplemented with 100 μg/ml ampicillin (Amp; Sigma). Transformants were grown in liquid LB medium supplemented with 100 μg/ml Amp, plasmids were extracted using a Wizard Plus SV miniprep DNA purification system (Promega), and purified plasmids were used to transform T. thermophilus as described by Koyama et al. (51). T. thermophilus was grown for 2 to 3 days on complex medium plates containing 25 μg/ml Kan until transformed colonies appeared. The in-frame replacement of the gene was confirmed by sequencing of the insert by use of primers 5 and 6 for each strain (Table 1). All transformants had the respective native promoter controlling the expression of the HTK gene cassette. Primers used for construction of the knockout strains are listed in Table 1.

TABLE 1.

PCR primers used for construction of T. thermophilus HB27 knockout mutants

Mutant strain	Primer name	Primer sequencea	Product size (bp)	Target
Δoah2 strain	oas 720 RI for1	GGATGCAGCAGCCCTGGAATTCGTAGTAGG	418	oah2 upstream sequence
	oas 1100 KpnI Rev1a	GTCTTCGGGAAGCCCGGCGAGGGTACCGAGGGTGGTGTA
	pUC-HTK 670 for	GCTTGCATGCGGGGTACCCTAGAATTCAT	800	HTK gene
	pUC-HTK 1450 Xba Rev	GAGGTCATCGTCTAGAATGGTATGC
	oas 2310 XbaI for2	CCTGGAGGCGGTCCATATGCTCTAGACCAT	470	oah2 downstream sequence
	oas 2750 PstI Rev2	GGTTCTTCTCCATGCCTGCAGGCTAGACC
	oas del for3	GGTCTTCACGGGCTTGCCCTGAAAGG	595	Sequencing
	HTK 5′ rev	GCTGAACTCTACTCCCTCTGTTGACAGAACAC
Δoah1 strain	A HTKoah1upfor	GACTTTGGAAAGGAGGCCAAGGATGAAAGGACCAATAATAATGA	761	HTK gene
	B HTKmet2rev	CGAGGGCGATCTCGCTCATTCAAAATGGTATGCGT
	E oah1uprev	TCATCCTTGGCCTCCTTTCCAAAGT	559	oah1 upstream sequence
	1 oah1 ecoRI	CCTCGAATTCCACGTCCCCG
	F met2for	ATGAGCGAGATCGCCCTCG	366	oah1 downstream sequence
	4 met2rev	GGCGGGGTCCAGGATCC
	htkoah1uprev	TCATTATTATTGGTCCTTTCATCCTTGGCCTCCTTTCCAAAGTC	1,772	Sequencing
	htkmet2for	ACGCATACCATTTTGAATGAGCGAGATCGCCCTCG
ΔbshC strain	B spo htk rev	CTAACCCTTCTTCAAAATGGTATGCGTTTTGAC	786	HTK gene
	A crp htk for	TAGGCTTTAAAGCATGAAAGGACCAATAATAATGAC
	E htk crp rev	GGTCCTTTCATGCTTTAAAGCCTAAAGTTCC	513	bshC upstream sequence
	1 crp bamHI for	CTCGTCGGGATCCACCGG
	F htk spo for	GCATACCATTTTGAAGAAGGGTTAGGATGTTCG	378	bshC downstream sequence
	4 spodown psti rev	AGGTAGTCCTGCAGGGGAAGG
	5 bshC for	GTCCTTTGCGCTTGAGGCG	1,774	Sequencing
	6 bshC rev	ACGGCCTGGGCCTTCA
ΔbshA strain	A perm HTK for	CTAGGCTTAGGGCATGAAAGGACCAATAATA	785	HTK gene
	B bshA htk rev	GGGCGTAGACTCAAAATGGTATGCG
	E htk perm rev	TTATTGGTCCTTTCATGCCCTAAGCCTAG	458	bshA upstream sequence
	1perm ecoR1 for	GCGCCGGGAATTCCGG
	F HTK bshA for	ACCATTTTGAGTCTACGCCCAGG	714	bshA downstream sequence
	4 prot PstI rev	GAGAAGGACCTGCAGGCCTAC
	5 perm upins for	AGACCCTCACCCTCAAGGACT	2,062	Sequencing
	6 pro downins rev	GCCCTCCTCCGCAAGG
ΔtrxA2 strain	A trx2upHTK	GCGGAGGGGGTACCTATGAAAGGACCAATAATAAT	796	HTK gene
	B hsp20HTK rev	TCCCTTCTAGGTAGGGCCTTCAAAATGGTATGCG
	E HTKuptrx2 rev	ATTATTGGTCCTTTCATAGGTACCCCCTCCG	539	trxA2 upstream sequence
	1 trx2 ecoR1 for	GAAAGGGAATTCCAGCTTGCGGG
	4 hsp20 bamH1 rev	CGGGGGATCCGCACCT	403	trxA2 downstream sequence
	F HTKtrx2for	GCATACCATTTTGAAGGCCCTACCTAGAAGG
	5 trx2upst for	CTTCTGGACCTGAGGGCGA	1,465	Sequencing
	6 hsp20downst rev	TTTCCTGAATGGGTATCCGCACG
ΔtrxA strain	A trx1 htk for	CGTAGAGTAGGGGGGTATGAAAGGACCAATAATAATG	791	HTK gene
	B2 trx1	GTCCAGGGCGATGCGCATCCTCAAAATGG
	E htktrx1 rev	ATTGGTCCTTTCATACCCCCCTACT	486	trxA1 upstream sequence
	1 trx1 ecoR1 for	CGCGGGTGAATTCCGCG
	F htkphos for	CATACCATTTTGAGGATGCGCATCGC	397	trxA1 downstream sequence
	4 phos bamH1 rev	TCAAGCTGGGGATCCCCAC
	5 trx1 upsisn for	GGACCTTCGCCGAGCTCC	1,752	Sequencing
	6 Phos downins rev	GGACGAGGAAGGAAGGGCG

^aUnderlined sequences indicate restriction enzyme cutting sites.

Mercury resistance in complex and defined media.

Cells were grown overnight (O/N) in complex medium and diluted to an optical density at 600 nm (OD₆₀₀) of 0.1 in fresh complex medium or in defined medium supplemented with sulfate, and HgCl₂ was added to individual tubes at concentrations ranging from 0 to 10 μM. Growth was monitored by use of the OD₆₀₀ (Spectronic 20 Genesys spectrophotometer; Spectronic Instruments). Resistance was assessed as the percentage of growth observed 18 or 20 h after Hg(II) addition relative to the growth of no-Hg(II) control cultures (100% growth). All experiments were performed in triplicate unless stated otherwise.

Mercury resistance under sulfur-limiting conditions.

All strains were grown O/N in complex medium, washed twice in chemically defined medium without sulfate, and resuspended in this medium to an OD₆₀₀ of 0.1. Cell suspensions were divided into different tubes, and 10 μM homocysteine (a concentration determined in preliminary experiments to be growth limiting) was added to each tube. Finally, HgCl₂ was added from 0 to 6 μM. Resistance was assessed as the percentage of growth (OD₆₀₀) observed 24 h after the addition of Hg(II) relative to the growth of no-Hg(II) control cultures.

Diamide and NEM assays.

Cells were grown O/N in complex medium, diluted to an OD₆₀₀ of 0.1, and divided into different tubes containing 0 to 2 mM diamide (Sigma) or 0 to 0.4 mM N-ethylmaleimide (NEM). HgCl₂ was added (at concentrations ranging from 0 to 4 μM), and growth (OD₆₀₀) was measured after 16 (diamide) or 24 (NEM) h. For exposed cultures, including those exposed only to diamide or NEM, growth in the absence of Hg(II) was considered 100% growth.

RNA extraction and cDNA synthesis.

Cells from O/N cultures of T. thermophilus were diluted to an OD₆₀₀ of 0.1 in complex medium and incubated to an OD₆₀₀ of ∼0.4, and then 1 μM HgCl₂ was added. For E. coli K-12 strains, an O/N culture was diluted to an OD₆₀₀ of 0.1 in LB medium and incubated to an OD₆₀₀ of ∼0.8, and then 2 μM HgCl₂ was added. Three-milliliter aliquots of cell suspensions were removed 7.5, 15, 30, or 60 min after the addition of Hg(II); an unexposed control was included at each time point. The removed aliquots were mixed with 0.5 volume of RNAprotect (Qiagen) and incubated for 5 min at room temperature. Cells were washed once with TE buffer (10 mM Tris-HCl, 1 mM EDTA, pH 8.0), resuspended in 300 μl of lysis buffer (20 mM sodium acetate, 1 mM EDTA, 0.5% SDS, pH 5.5), and incubated for 5 min at 65°C. One milliliter of TRIzol (Invitrogen) was added, and RNA extraction was performed as instructed by the manufacturer. RNA integrity was checked in a 1.5% agarose gel, and the concentration was measured by NanoDrop spectrophotometry (ND-1000 spectrophotometer; NanoDrop Technologies Inc., Wilmington, DE). RNA was treated with a Turbo DNA-free kit (Ambion) to remove DNA, and the DNA-free RNA was used to synthesize cDNA by use of a high-capacity cDNA reverse transcription kit (Applied Biosystems), using 1 μg of RNA for T. thermophilus. For E. coli, 2 μg of RNA was used to synthesize cDNA by use of the SuperScript III first-strand synthesis system (Life Technologies). For both kits, cDNAs were synthesized following the manufacturers' instructions.

Quantitative PCR.

T. thermophilus transcripts of oah2 (WP_011173224), oah1 (WP_011172856), met2 (WP_041443334), metH (WP_041443334), oas (WP_011174005), bshA (WP_011173182), bshB (WP_011173270), bshC (WP_011173609), trxA1 (WP_011173768), trxA2 (WP_011173531), and trxB (WP_011173929) were quantified by qPCR, using cDNA as the template and primers specific for each gene (Table 2). For E. coli, transcripts of gor (Ga0175964_11222), trxB (Ga0175964_112966), gshA (Ga0175964_111051), gshB (Ga0175964_11784), cysK (Ga0175964_111321), malY (Ga0175964_112139), cysM (Ga0175964_111314), and zntA (Ga0175964_11254) were quantified as described for T. thermophilus. Reaction mixtures contained 12.5 μl of 2× SYBR Green JumpStart Taq ReadyMix (Sigma), 1 μl cDNA, and an optimized concentration of each primer set (Table 2) in a final volume of 25 μl. Thermal cycling (iCycler iQ; Bio-Rad Laboratories Inc., Hercules, CA) conditions were as follows: initial denaturation at 95°C for 5 min; 30 cycles of denaturation (95°C for 15 s), annealing (at the temperature shown in Table 2, for 15 s), and extension (72°C for 15 s); fluorescence measurement; and a final melt curve (60 to 99°C). Results for at least three biological replicates were averaged in all cases. A reaction mixture with DNA-free RNA was run as a control for detection of DNA contamination. Transcript abundance was normalized to that of gyrA (TT_C0990) for T. thermophilus (25) and to that of ssrA (Ga0028711) for E. coli (28), using the ΔΔC_T method (52).

TABLE 2.

qPCR primers used to measure gene expression of the listed genes^a

Primer	Sequence	Final concn (μM)	Annealing temp (°C)	Product size (bp)
gyrase-F	GGGCGAGGTCATGGGC	1	61	134
gyrase-R	CGCCGTCTATGGAGCCG	0.25
oah2-F3	GAGCTCTGGCGGAACTAC	0.25	56	104
oah2-R3	AAGGTGCGGACCCTTTC	1
merR4-F	AGCTTGAGGACATCGCCTGGAT	0.5	61	214
merR4-R	TCCAAATAGACGCAGCGGTCCC	1
merA for	GCCTTCAAGATCGTGGTGGACGAAGAG	0.5	62	186
merA rev	CCTGGGCCACGAGCCTTATCC	0.25
oah1-F2	GGAGCTCGCCTTCATCGTCA	1	59	137
oah1-R2	TGTTTTCCACGTGACGCTCG	0.25
met2-F	GAATCGCCATGATGAGCTACC	1.25	57	92
met2-R	CTGGTAGTCCAGGTAGGTTTCC	1.25
oas-F	CGCTACCTCAAGGAAAGGATC	1.25	56	145
oas-R	GGGAAAGGTCCAGGTTCTCG	1.25
metH-F	CCTCTTTGACCTCCTTACCTTCC	2.5	57	98
metH-R	TCTCCCGTAGCTCCTCTATGG	2.5
bshA-F	CTAGACCTGAGCGCAAGAGG	2.5	58	128
bshA-R	GAGGCGCTTCCGGATTT	2.5
bshB-F	CGCCGTTACTTTGGGAACTA	2.5	58	76
bshB-R	GCACGTAGAGCACGGGAAG	2.5
bshC-F2	GGAGGCGGAGACGCTTTC	1.25	55	81
bshC-R2	GGGTCAAAGGGCACAAGC	1.25
trxB-F	GGTGCGGCTTAAGAACCTAAAG	1.25	58	100
trxB-R	CTTGAGGAAAGCGGTGTTGG	1.25
trxA1-F	GACCAGAACTTTGACGAGACCC	1.25	60	132
trxA1-R	AAGCTTCCCCTCGTACTCTTTGG	1.25
trxA2-F	CCCACCCTGGTCCTCTTCC	1.25	60	105
trxA2-R	CCTCCCTTCTAGGTAGGGCCTG	1.25
gor-R e coli	GCCGTGAATACCGACAATCT	0.4	60	130
gor-F e coli	CGACGATCAGGTGAAAGTGTAT	0.4
trxB-F e coli	ACAGTCGGGTATTCATGGTAATG	0.4	60	110
trxB-R e coli	CTGTACCGGCCGAAGTAATG	0.4
gshB-F e coli	CCCATCGCAAACATCAACATC	0.4	60	101
gshB-R e coli	AGATCGCCCATCTCCATATAGT	0.4
gshA-F e coli	CGGATGTGGCCGTTAAGTAT	1	60	91
gshA-R e coli	TAAAGCGTCCGGTGTTAGAAG	1
zntA-F	ATCAGGTGCAGGTGTTGTT	1	60	113
zntA-R	TTCATCGCGCAGGGAATAG	1
cysM-F ecoli	ATACGCCTCTGGTGAAGTTG	0.5	60	89
cysM-R ecoli	CGAACCTGCCGGGTTATT	0.5
cysK-F ecoli	CGATCTCAAGCTGGTCGATAAA	0.4	60	116
cysK-R ecoli	CAGCTGCTCCAGAAGAGATAC	0.4
MalY-F ecoli	GATGAGTTTCTCGCGGCTATT	0.5	60	93
MalY-R ecoli	GATGACAGAAGGGCCATACAC	0.5
ssrA-F Ecoli	CGCCCGTCACGAATACTTTA	0.4	60	110
ssrA-R Ecoli	ACGTAGCTGTCGCTGATATTG	0.4

^aAll primers were designed for this study, except for the merA primers, which were from the work of Wang et al. (17).

Mercuric reductase assay.

Cells from an O/N culture were diluted in fresh complex medium to an OD₆₀₀ of 0.1 and grown to an OD of ∼0.4, 1 μM HgCl₂ was added, and cultures were incubated for 30 additional minutes. Each experiment included an unexposed control. Cultures (35 ml) were centrifuged and washed once in phosphate-buffered saline (PBS; 8.01 g/liter NaCl, 0.2 g/liter KCl, 1.78 g/liter Na₂HPO₄, 0.27 g/liter KH₂PO₄, pH 7.4), and cell pellets were frozen until further use. Crude cell extracts were prepared as previously described (53). MerA assays were performed at 70°C as described by Wang et al. (17), using 200 μM NADH and 20 μl of protein extract. Oxidation of NADH was monitored spectroscopically at 340 nm (Aviv model 14 UV-VI biomedical spectrophotometer) each second for 1 min. Control reaction mixtures were set up with NADH and cell extract but no HgCl₂, and the cell extract-only activities were subtracted from rates for complete assay mixtures to determine Hg(II)-dependent NADH oxidation. Specific activities were defined in units per milligram of protein, with 1 U corresponding to 1 μmol of NADH oxidized per min. The Bradford assay (Bio-Rad Laboratories, Inc., Hercules, CA) was used to determine protein concentrations.

Thiol content assay.

Cells from an O/N culture were diluted to an OD₆₀₀ of 0.1 and grown to an OD of ∼0.4, 0 to 3 μM HgCl₂ was added, and growth was continued for an additional 30 or 60 min. Cells from a 25-ml culture were harvested and washed twice with PBS. Cell pellets were resuspended in 500 μl of D-mix (9.4 mM monobromobimane [mBrB; Sigma], 50% acetonitrile, 50 mM HEPES, 5 mM EDTA, pH 8.0) and incubated for 15 min at 60°C in the dark. Derivatization was stopped by adding methanesulfonic acid to a final concentration of 25 mM. Samples were stored at −20°C until high-performance liquid chromatography (HPLC) analysis. Reversed-phase HPLC analysis and fluorescence detection of the bimane derivatization products were performed as previously described by Rethmeier et al. (35).

Bioinformatic analysis.

The presence of oah2 gene homologs in all finished, permanent draft, and draft Deinococcus-Thermus genomes available in the IMG/MER database (https://img.jgi.doe.gov/cgi-bin/mer/main.cgi) on 8 June 2017 was determined as follows. First, mer operons were identified by use of BLASTX (https://img.jgi.doe.gov/cgi-bin/mer/main.cgi?section=FindGenesBlast&page=geneSearchBlast), with MerA of HB27 (accession number TTC0789) as a query and a cutoff value of 1e−50. Hits were manually examined for the presence of amino acid residues characteristic of MerA (13), and neighboring genes were examined for the presence of Oah2 homologs in the JGI's gene detail page.

The presence of the GR gene in the finished and assembled alphaproteobacterial genomes was determined on 18 April 2017 using a blastp search (https://img.jgi.doe.gov/cgi-bin/mer/main.cgi?section=FindGenesBlast&page=geneSearchBlast), with MerA of Aurantimonas manganoxydans S185-9A1 (accession no. EAS49959.1) as a query and a cutoff value of 1e−50. Hits were manually examined for the presence of amino acid residues characteristic of MerA (13) and neighboring genes for the presence of GR homologs using the Joint Genome Institute's gene detail page.

Phylogenetic analysis.

Protein sequences of Oah and GR were aligned using ClustalX (ver. 2.0) (54, 55). The resulting alignments were used for phylogenetic analysis, which was inferred by using the maximum likelihood method based on the JTT matrix-based model (56). The initial tree(s) for the heuristic search was obtained automatically by applying neighbor-joining and BioNJ algorithms to a matrix of pairwise distances estimated using a JTT model and then selecting a topology with a superior log likelihood value. Evolutionary analyses were conducted in MEGA7 (57).

For both trees, all positions containing gaps and missing data were eliminated. The Oah tree was constructed of 21 amino acid sequences, with a total of 248 positions in the final data set. The outgroup used was O-acetylserine sulfhydrylase (Oas) enzymes because they carry out reactions similar to those of Oah but have substrate specificity different from that of Oah (21). Met17 was used as an internal outgroup for the Oah1 proteins due to its high homology to the latter (21). For the GR tree, the analysis included 30 amino acid sequences, with a total of 323 positions in the final data set. LpdA was used as the outgroup because it is considered to be ancestral to GR in the FAD-dependent pyridine nucleotide-disulfide oxidoreductase family (58, 59).

Statistical analysis.

One-way analysis of variance (ANOVA) followed by Tukey test analysis was performed for multiple-group comparisons. For two-group comparisons, the t test was run.