Escherichia coli TehB Requires S-Adenosylmethionine as a Cofactor To Mediate Tellurite Resistance

Abstract

The Escherichia coli chromosomal determinant for tellurite resistance consists of two genes (tehA and tehB) which, when expressed on a multicopy plasmid, confer resistance to K₂TeO₃ at 128 μg/ml, compared to the MIC of 2 μg/ml for the wild type. TehB is a cytoplasmic protein which possesses three conserved motifs (I, II, and III) found in S-adenosyl-l-methionine (SAM)-dependent non-nucleic acid methyltransferases. Replacement of the conserved aspartate residue in motif I by asparagine or alanine, or of the conserved phenylalanine in motif II by tyrosine or alanine, decreased resistance to background levels. Our results are consistent with motifs I and II in TehB being involved in SAM binding. Additionally, conformational changes in TehB are observed upon binding of both tellurite and SAM. The hydrodynamic radius of TehB measured by dynamic light scattering showed a ∼20% decrease upon binding of both tellurite and SAM. These data suggest that TehB utilizes a methyltransferase activity in the detoxification of tellurite.

Tellurite (TeO₃²⁻) resistance, which is found in many microorganisms (37), is carried on plasmids of IncHI, IncHII, and IncP incompatibility groups (3, 36, 39) and/or on chromosomes (9, 10, 29, 33). Tellurite resistance (Te^r) determinants are very diverse, and at least five different determinants have been identified (37). Te^r determinants are unrelated, and probably different resistance mechanisms are involved. Except for that encoded by the Pseudomonas syringae tpm gene, Te^r mechanisms are not clearly understood. tpm encodes a methyltransferase that catalyzes S-adenosyl-l-methionine (SAM) methylation of 6-mercaptopurine, a substrate for human thiopurine methyltransferase. It was assumed that tellurite is also a substrate for this enzyme (11).

The Te^r determinant tehAB, an operon located at 32.3 min on the Escherichia coli K-12 chromosome (38), encodes proteins of 330 (TehA) and 197 (TehB) amino acids. TehA is a membrane protein with 10 transmembrane segments, whereas TehB is a soluble protein. The closest homologues of these two genes are found in Haemophilus influenzae (15).

A sequence search performed with the BLAST program (National Center for Biotechnology Information), followed by alignment analysis with the Genetics Computer Group software package (University of Wisconsin) (12), demonstrated that TehB displays some amino acid sequence similarity to many SAM-dependent non-nucleic acid methyltransferases. These proteins have three shared motifs, and the TehB proteins show homologies within all regions with comparable sequence interval distances (Fig. 1). Motif I is a glycine-rich region agreeing well with the consensus sequence h(D/E)hGxGxG, where h represents a hydrophobic residue and x is any residue (27, 46). Motif I is generally followed by an acidic residue on the C-terminal side, about 17 to 19 residues from the motif (27). In TehB, there are 16 residues before this aspartate (Fig. 1).

FIG. 1.

Alignment of sequences of E. coli TehB, H. influenzae TehB, and some non-nucleic acid methyltransferases to demonstrate the three motifs. The three homologous regions (motifs) of the methyltransferases are shown in the one-letter amino acid code. The numbers represent the residue number spacings between the motifs. The sequences are from the National Center for Biotechnology Information database. The names are abbreviated, and their accession numbers are in parentheses. EctehB, protein encoded by E. coli tellurite resistance gene tehB (M38696); HitehB, protein encoded by an H. influenzae tellurite resistance gene tehB homologue (U32807); RPIMT, rat protein l-isoaspartyl carboxyl methyltransferase (D11475); HPIMT, human protein l-isoaspartyl carboxyl methyltransferase (P22061); EcPIMT, E. coli protein l-isoaspartyl carboxyl methyltransferase (P24206); RDHPBMT, rat dihydroxypolyprenylbenzoate methyltransferase (L20427); EcUbiG, E. coli ubiquinone biosynthesis-related protein (M87509); YCoq3, Saccharomyces cerevisiae 3,4-dihydroxy-5-hexaprenylbenzoate methyltransferase (M73270); EcCFA, E. coli cyclopropane fatty acid synthetase (M98330); SaNMT, Streptomyces anulatus N-methyltransferase (X92429); EcBioC, E. coli biotin synthesis protein (P12999); EhBioC, Erwinia herbicola biotin synthesis protein; RbGNMT, rabbit glycine N-methyltransferase (D13307); RtGNMT, rat glycine N-methyltransferase (X06150); SgStsG, Streptomyces griseus methyltransferase involved in the N-methyl-l-glucosamine pathway (Y08763); RGAMT, rat guanidinoacetate methyltransferase (J03588).

Thirty-four residues downstream from motif I is motif II, which comprises 8 residues. Motif II is unusually rich in aromatic amino acids (Tyr, Trp, and Phe), which are located around a central invariant aspartate residue (27). Aromatic rings are considered to be involved in binding SAM through cation-quadrapole interactions with the positively charged sulfonium moiety of SAM (13, 31). The importance of tyrosine in motif II of rat guanidinoacetate methyltransferase (RGAMT) has already been demonstrated (22). In TehB, four aromatic residues were found in this region; they are Phe (−5 position with respect to Asp), Tyr (−1), Phe (+1), and Phe (+10). Motif III is 22 residues away from motif II and consists of a 9-residue block. In the middle of the block are two highly conserved Gly residues. Motif III is consistent with the consensus sequence L(R/K)PGG(R/I/J)(L/I)(L/F/I)(I/L) (25, 27). However, the significance of motif III in methyltransferases has been questioned, because site-directed mutagenesis of several residues of RGAMT motif III minimally affected enzymatic activity (19).

To date, crystal structures of six methyltransferases have been elucidated; four are DNA methyltransferases: HhaI DNA methyltransferase (M.HhaI) (7), TaqI methyltransferase (M.TaqI) (30), HaeIII methyltransferase (M.HaeIII) (34), and PvuII methyltransferase (M.PvuII) (20). Two are small-molecule methyltransferases: catechol O-methyltransferase (45) and glycine N-methyltransferase (17). Generally DNA and RNA methyltransferases lack motifs II and III of the non-nucleic acid methyltransferases but possess a glycine-rich sequence which shows a weak resemblance to motif I (25, 27). Despite the low degree of sequence similarity among these six enzymes, the tertiary structures of their SAM-binding domains are strikingly similar, suggesting that many (if not all) methyltransferases may have a common structure for SAM binding (8, 17, 35). Escherichia coli TehB exhibits sequence similarities to all three motifs of glycine N-methyltransferase (Fig. 1); thus, it is likely that TehB has a similar SAM-binding site.

To determine if the SAM-binding motifs in TehB are involved in tellurite resistance, we performed site-directed mutagenesis on two aspartate residues in motif I and on three residues in motif II (Table 1). Mutagenesis was performed by the two-stage PCR-based overlapping extension method (23). MICs were determined as previously described (42) with the tehAB operon expressed from pTWT101, a pTZ19R (47) clone, in E. coli JM109. The MICs of K₂TeO₃ for all three motif I mutants and two motif II mutants (Phe98Tyr and Phe98Ala) were reduced to 2 μg/ml from 128 μg/ml. The Tyr96Ala and Asp97Asn mutations had only a partial effect on tellurite resistance, whereas Asp97Glu had no effect. The drop in MIC was not due to decreased expression of the mutants since sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) showed that the level of expression of mutant TehB proteins was similar to that of the wild type (data not shown). These results suggest that motifs I and II in E. coli TehB, like those that play a direct role in SAM binding in other non-nucleic acid methyltransferases, are directly involved in TehB's role in tellurite resistance.

TABLE 1.

Tellurite resistance encoded by plasmids with mutations in TehB methyltransferase

Mutation or plasmid	MICa (μg of K2TeO3/ml)	Motif containing mutationb
pTWT101c	128
pTZ19Rd	2
Asp36Ala	2	I
Asp36Asn	2	I
Asp59Ala	2	I
Tyr96Ala	16	II
Asp97Glu	128	II
Asp97Asn	64	II
Phe98Tyr	2	II
Phe98Ala	2	II

^aMICs were measured several times with no variation in the resistance end points. 

^bSee Fig. 1 for motifs in methyltransferase proteins. 

^cPositive control with wild-type tehAB expressed on a multicopy plasmid. 

^dNegative control with vector lacking tehAB. 

To further assess whether TehB possesses SAM-binding ability, we performed in vitro experiments on purified TehB. The coding sequence for E. coli TehB was cloned by PCR which allowed for insertion of an N-terminal His₆ fusion construct into the vector pRSETC (Invitrogen), resulting in pTWT124. E. coli C41(DE3) (32), a derivative of BL21(DE3) (hsdS gal λcIts857 ind1 Sam7 nin5 lacUV5-T7 gene 1) was found to be the best expression host for the purification of His₆-TehB. Growth conditions, time of induction, and harvest were optimized to achieve the highest possible level of accumulation of soluble protein. As much as 90% of the total TehB was lost in the low-speed-centrifugation pellet as inclusion bodies depending on the method and host chosen. The best conditions involved addition of a 2% inoculant of overnight culture into 1 liter of Luria-Bertani (LB) medium and growth for 3 h at 37°C. After the 3-h time period (optical density at 600 nm [OD₆₀₀], ∼0.5), IPTG (isopropyl-β-d-thiogalactopyranoside) was added to 0.5 mM and the culture was incubated for another 3 h. Cells were harvested at 5,500 × g for 10 min, resuspended in 50 ml of buffer (50 mM morpholinepropanesulfonic acid [MOPS; pH 7.5], 2 mM EDTA, 1 mM dithiothreitol [DTT], 100 μM phenylmethylsulfonyl fluoride [PMSF]) per liter of culture, and repelleted at 5,500 × g for 10 min. The cell pellet was resuspended (1 ml of buffer per 0.5 g of paste), and the cells were lysed by two passages through a French pressure cell (16,000 lb/in²). Unbroken cells and cellular debris were removed by centrifugation at 10,000 × g for 10 min.

To purify the His₆-TehB, a 2.5-ml column bed of Superflow nickel-nitrilotriacetic acid resin (Qiagen) was poured and equilibrated with buffer (50 mM NaH₂PO₄ [pH 8.0], 300 mM NaCl, 10 mM imidazole), and 2 to 3 ml of the cell extract was loaded. An equimolar amount of MgCl₂ was added to the cell extract to coordinate the EDTA of the lysis buffer. The column was washed with 10 bed volumes of wash buffer (50 mM NaH₂PO₄ [pH 8.0], 300 mM NaCl, 20 mM imidazole); His₆-TehB was eluted with 50 mM NaH₂PO₄ (pH 8.0)–300 mM NaCl–250 mM imidazole, and 1-ml fractions were collected. All fractions were electrophoresed on an SDS–12% polyacrylamide gel, and purity was evidenced as a single band stained with Coomassie brilliant blue G250 and R250 (Fig. 2). Fractions with the highest TehB content and purity were pooled, and the A₂₈₀ of the pooled fractions was measured. The concentration was estimated by using an ɛ₂₈₀ of 12,000 M⁻¹ cm⁻¹ for TehB.

FIG. 2.

SDS-polyacrylamide gel electrophoresis analysis of the purification of E. coli TehB. Lane 1, protein molecular size standards with molecular masses of (from top to bottom) 97.4, 66.2, 45.0, 31.0, 21.5, and 14.4 kDa; lane 2, cell extract of induced E. coli C41(DE3)/pTWT124; lane 3, cell extract of uninduced E. coli C41(DE3)/pTWT124; lane 4, freshly purified TehB from the nickel-nitrilotriacetic acid affinity resin; lane 5, TehB after storage at 4°C for 2 weeks, showing the presence of a high-molecular-mass aggregate protein (∗).

The protocol described above affords the production of ∼3 mg of pure soluble protein per liter of culture. The pure TehB was found to be unstable under conditions of low ionic strength and formed insoluble aggregates (Fig. 2, lane 5) upon freezing, dialysis, extended storage (more than 1 week), or concentrating (to >10 mg/ml). Therefore, all enzymatic and biophysical studies were carried out within 1 day of purification.

The hydrodynamic radius (R_h) of TehB was studied by dynamic light scattering (DLS) in the presence and absence of SAM and TeO₃²⁻. Enzymes generally undergo conformational changes upon substrate binding, thereby trapping substrates and positioning key active-site residues and substrates for catalysis. Such conformational changes are often reflected by changes in the R_h, which can readily be detected by DLS. Therefore, we employed this technique to identify whether TehB does indeed bind SAM and/or tellurite.

DLS was performed in a DynaPro-MSTC dynamic light-scattering instrument (Protein Solutions Inc.). Protein samples were analyzed for their translational diffusion coefficient (D_T), which, assuming Brownian motion, was translated, using the Stokes-Einstein equation, to the R_h. Measurements of complexes of TehB with TeO₃²⁻ and SAM were obtained for 1.5- to 4.0-mg/ml protein solutions incubated with 1 mM SAM and 3 mM K₂TeO₃, providing at least a 100-fold molar excess of each ligand. Experiments were performed at least three times for different protein expression-purification trials, and each measurement was performed 50 times, with the results being analyzed by using the DYNALS software package. The mean R_h ± standard deviation in phosphate buffer, with or without TeO₃²⁻ or SAM, was 3.21 ± 0.02 nm. Addition of both TeO₃²⁻ and SAM gave an R_h of 2.59 ± 0.02 nm.

The sample was not found to be perfectly monodispersed. Although a single, well-defined major species was observed in all experiments, a small amount of a heterogeneous high-molecular-weight aggregate was also observed in all of the samples analyzed. The contributions of the aggregate increased with increased storage time (over weeks), further demonstrating the instability and propensity of this protein to aggregate (Fig. 2). Various buffers and buffer additives (such as glycerol and detergents) helped to diminish this aggregation but did not completely resolve it. These problems likely reflect the localization of TehB as a peripheral membrane protein; as such, it probably interacts with the integral membrane protein TehA. Since freshly purified samples were practically devoid of aggregate, they were always used in the DLS studies.

The calculated formula mass of the His₆-tagged TehB is 26,256 Da, and it migrates on an SDS–12% polyacrylamide gel to a position corresponding to between 27 and 28 kDa. The observed R_h of 3.21 nm relates to a molecular mass of 49.0 kDa assuming that the protein is in the shape of a sphere. This suggests that TehB is a dimer. Although no significant change in R_h was observed when SAM or tellurite was added individually, when the two were added simultaneously the R_h was reduced to 2.6 nm, which correlates with a molecular mass of ∼29.1 kDa. This indicates that there is on the order of a 20% change in TehB molecular free volume upon cofactor-ligand binding. Alternatively, the protein may become a monomer in the presence of SAM and tellurite. These experiments provide evidence that TehB is a SAM- and TeO₃²⁻-binding protein and that a SAM-dependent activity forms part of the Te^r mechanism.

SAM binding mutants of TehB were also PCR subcloned and purified, using His₆ tagging in the same fashion as was done for the wild type. Although no difference in size or shape and no effect of SAM addition on conformation change were expected, the mutant TehB(Phe98Tyr) produced unexpected results. We observed R_h values ranging from 6.1 to 7.5 nm for TehB(Phe98Tyr), which suggests that this particular mutant either is more susceptible to aggregation or oligomerizes into a decamer. The other mutants showed no change with SAM and tellurite binding, which is in agreement with studies of motifs I and II of a rat non-nucleic acid methyltransferase examined by a variety of spectroscopic methods (22). The study observed only minor conformational changes with mutations to key residues and concluded that the SAM-binding site has strict residue requirements.

To determine if the shape change was a general effect of tellurite-SAM addition, a control protein that had been purified in a similar manner was chosen at random. The periplasmic siderophore-binding protein FepB, which has a molecular weight similar to that of TehB, was used. Although a slight change in the R_h was observed upon addition of tellurite alone, no change was observed when SAM and tellurite were added together.

Evidence for a SAM-dependent enzymatic reaction of TehB with tellurite was explored. No volatile tellurium compounds were observed in the head gas of cultures exposed to tellurite (R. J. Turner, V. Van Fleet-Stalder, and T. G. Chasteen, unpublished results). Therefore, the final product is likely more complex than a simple methylated telluride. Because the product of the reaction is unknown at this time, only the loss of substrate could be monitored. Two different assays were developed to examine the SAM-dependent loss of TeO₃²⁻. Sodium diethyl dithiocarbamate (DDTC) has been successfully utilized to measure TeO₃²⁻ levels in bacterial cultures (44). The assay was modified to measure tellurite in buffered solutions. A typical assay mixture consisted of 780 μl of 50 mM Tris (pH 7.5), 10 μl of a 200-mg/ml solution of SAM, 10 μl of a ∼1-mg/ml solution of TehB, and 100 μl of K₂TeO₃ (10 to 10,000 μmol; added at time zero). Samples were incubated for 30 min at 37°C, after which 100 μl of 10 mM DDTC was added to each. The mixtures were allowed to incubate a further 5 min at 37°C, and the A₃₄₀s were recorded (ɛ₃₄₀ = 3.04 × 10² M⁻¹ cm⁻¹), allowing for the determination of any TeO₃²⁻ remaining.

The second procedure involved the reaction of tellurite with DTT followed by assessment of the remaining DTT with dithionitrobenzoic acid (DTNB) as a coupled assay. DTT becomes oxidized by tellurite in a 1:1 molar ratio, and DTNB reacts with DTT in a 2:1 molar ratio. The reaction mixture consisted of 2.4 ml of 50 mM MOPS (pH 7.5), 30 μl of a 1-mM solution of SAM, and 30 μl of a 1-mg/ml TehB solution. At time zero, 300 μl of 0.03 mM K₂TeO₃ was added, and 400-μl aliquots were removed at regular time intervals for 30 min. After addition of 50 μl of 0.1 mM DTT to each of these aliquots, mixing, and a 1-min incubation, 50 μl of 1 mM DTNB was added to each sample. The A₄₁₂ of each sample was recorded, and the concentration of reduced DTT (ɛ₄₁₂ = 1.36 × 10⁴ M⁻¹ cm⁻¹), which is related to the number of moles of TeO₃²⁻ remaining in the reaction after incubation, was determined.

Although neither of these assays is ideal for examination of the enzymology of TehA and TehB, it was possible to observe a SAM-dependent loss of tellurite on the order of 350 nmol of TeO₃²⁻/mg of TehB/min.

Microbial methylation of metalloids to yield various organoderivatives is a well-known phenomenon. The methylated derivatives are typically quite volatile, which provides a means of removing the toxicant. Many bacteria and fungi, and even the protozoan Tetrahymena thermophila, have been found to methylate arsenic compounds (24, 41) and selenium compounds (4, 18, 40), as well as antimony compounds (21, 26) and sulfide (14). While much less work has been done on tellurium methylation, it has been observed that both fungi and bacteria are capable of methylating tellurium compounds. For example, several Penicillium strains (P. brevicaule, P. chrysogenum, and P. notatum) and the bacterium Pseudomonas fluorescens were found to produce dimethyltelluride [Te(CH₃)₂] (2, 6, 16). In addition, the fungi Acremonium falciforme and Penicillium citrinum produce both Te(CH₃)₂ and dimethylditelluride [Te₂(CH₃)₂] as well as another, unknown compound (6). A Penicillium sp. isolated from evaporation pond water was also capable of yielding Te(CH₃)₂ and Te₂(CH₃)₂ (24). After administration of tellurium compounds to rats and dogs (28), guinea pigs (1), and ducks (5), a garlic-like breath odor was noted. In these reports, Te₂(CH₃)₂ was considered to be the odorous ingredient, although other compounds of tellurium, selenium, and sulfur also have this property. A study of a nonhomologous tellurite resistance gene (tpm) from Pseudomonas syringae pathovar pisi provided evidence that methylation might be involved in tellurite resistance (11). The chemical nature of the methylated tellurium product was not determined in this study. It is possible that methylation of tellurium compounds is a widespread phenomenon in both prokaryotic and eukaryotic organisms.

Since TehB likely has a methyltransferase activity, a detoxification mechanism for tellurite resistance by tehAB is a logical inference. The chemical modification of TeO₃²⁻ with methyl groups would provide a means for detoxification. However, it is clear that in our case the final product of the reaction is not a volatile form of methylated tellurium. A detoxification mechanism is supported by tellurite accumulation experiments. The tehAB determinant was found to continuously remove tellurite from the growth medium, whereas other Te^r determinants would more rapidly become saturated (43). Such an observation is consistent with the occurrence of a modification to tellurite within the cell.