Methylcobamide:Coenzyme M Methyltransferase Isozymes from ...

4 downloads 0 Views 232KB Size Report
From the Department of Biochemistry, Uniformed Services University of the Health Sciences,. Bethesda, Maryland 20814-4799. A comparative study was made ...
THE JOURNAL

OF

BIOLOGICAL CHEMISTRY

Vol. 271, No. 31, Issue of August 2, pp. 18725–18731, 1996 Printed in U.S.A.

Methylcobamide:Coenzyme M Methyltransferase Isozymes from Methanosarcina barkeri PHYSICOCHEMICAL CHARACTERIZATION, CLONING, SEQUENCE ANALYSIS, AND HETEROLOGOUS GENE EXPRESSION* (Received for publication, January 24, 1996, and in revised form April 3, 1996)

Gilles M. LeClerc and David A. Grahame‡ From the Department of Biochemistry, Uniformed Services University of the Health Sciences, Bethesda, Maryland 20814-4799

A comparative study was made on the physicochemical characteristics of two isozymes of methylcobamide:coenzyme M methyltransferase (MT2). Both isozymes catalyzed S-methylation of 2-thioethanesulfonate (coenzyme M) and exhibited similar apparent Km values for coenzyme M of 35 mM (MT2-A) and 20 mM (MT2-M). Weak binding to methylcobalamin was indicated by the apparent Km of 14 mM for both isozymes. Cob(I)alamin was established as the major product of the reaction, demonstrating heterolytic cleavage of the methylcobamide carbon-cobalt bond. The isozymes were shown to be zinc-containing metalloproteins. Metal ion chelators strongly inhibited both isozymes. A variety of coenzyme M analogs were tested for activity and/or inhibition. One alternative substrate 3-mercaptopropionate was discovered, with apparent Km 9 mM (MT2-A) and 10 mM (MT2M). The results suggested an active site geometry in which coenzyme M is bound both by S-coordination to zinc, and electrostatic interaction of the sulfonate with a cationic group on the enzyme. Methanosarcina barkeri genes cmtA and cmtM encoding both isozymes were cloned and sequenced. Both genes encoded proteins with 339 amino acids and predicted molecular masses of 36 –37 kDa. Active forms of both isozymes were expressed in Escherichia coli. A conserved segment with the potential for metal binding was found. The possibility of zinc involvement in catalysis of coenzyme M methylation is considered.

The methanogenic archaea are able to use for growth and methanogenesis only a relatively limited number of compounds with rather simple structures (1, 2). These compounds include substrates containing one-carbon groups such as carbon dioxide (plus H2), formate, methanol, carbon monoxide, dimethylsulfide, and various methylated amines, as well as the twocarbon compound acetate. Growth on a three-carbon compound (pyruvate) has been observed (3). In all methanogens the coenzyme 2-thioethanesulfonate (coenzyme M or HSCoM)1 acts as

the methyl group carrier in the final step of methane formation. Methane is produced by reductive demethylation of 2-methyl2-thioethanesulfonate (methylcoenzyme M, or CH3-SCoM) (4). Depending upon the growth substrate, different pathways lead to the production of CH3-SCoM. The pathways of methanogenesis from carbon dioxide, methanol, acetate, and pyruvate have been described in detail (3, 5–7). Burke and Krzycki have purified and characterized a 29-kDa corrinoid protein and shown that it functions in CH3-SCoM formation from monomethylamine (8). However, the pathways of CH3-CoM formation from this and other substrates are still not fully defined. In the conversion of methanol, synthesis of CH3-SCoM proceeds by two sequential reactions (9), as shown in Reactions 1 and 2. CH3OH1cob(I)amide-MT11H1 º Co-CH3cob(III)amide-MT11H2O REACTION 1 HSCoM1Co-CH3cob(III)amide-MT1 º cob(I)amide-MT11CH3-SCoM1H1 REACTION 2

The overall coupled reaction is given by Reaction 3. CH3OH1HSCoM º CH3-SCoM1H2O REACTION 3

Reaction 1 is catalyzed by the oxygen-labile enzyme methanol: 5-hydroxybenzimidazolylcobamide methyltransferase (MT1) (10). Activation of MT1 was found to require a reducing system (H2, hydrogenase, ferredoxin), ATP, and a separate methyltransferase activator protein (10, 11). In Reaction 2, the enzyme methylcobamide:CoM methyltransferase (MT2) catalyzes the transfer of the methyl group from the MT1-bound methylcobamide prosthetic group to coenzyme M. The ability to catalyze Reaction 4 is used as a means for routine assay of MT2 (12). HSCoM1methylcobalamin º CH3-SCoM1cob(I)alamin1H1

* This work was supported by Grant DMB 9304637 from the National Science Foundation and Grant DE-FG05-94ER20159 from the United States Department of Energy. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. The nucleotide sequence(s) reported in this paper has been submitted to the GenBankTM/EBI Data Bank with accession number(s) U38918 and U38919. ‡ To whom correspondence should be addressed. Tel.: 301-295-3555; Fax: 301-295-3512, E-mail: [email protected]. 1 The abbreviations used are: coenzyme M or HSCoM, 2-thioethanesulfonate; CH3-H4SPt, N 5-methyl-tetrahydrosarcinapterin; H4SPt, tet-

REACTION 4

In contrast to MT1, MT2 is not inactivated by exposure to air. Hitherto, no organic or inorganic cofactors have been reported

rahydrosarcinapterin; methylcoenzyme M or CH3-SCoM, 2-methyl-2thioethanesulfonate; MOPS, 3-(N-morpholino)propanesulfonic acid; MT2, methylcobamide:coenzyme M methyltransferase; MT1, methanol: 5-hydroxybenzimidazolylcobamide methyltransferase; PCR, polymerase chain reaction; kb, kilobase pair(s); bp, base pair(s); 3-MPA, 3-mercaptopropanoic acid.

18725

18726

Methylcobamide:Coenzyme M Methyltransferase Isozymes

to be tightly bound to MT2. Direct evidence for the formation of cob(I)alamin as a product of the reaction has not been presented. Nevertheless, regeneration of the Co(I) form of the cobamide by heterolytic cleavage of the carbon-cobalt bond is portrayed in Reactions 2 and 4, based on results from studies on the reactivity of CH3-B12 with thiols (13). Two different isoenzyme forms of MT2 (MT2-A and MT2-M) have been identified in Methanosarcina barkeri (14). Both isozymes have similar molecular masses (' 34 kDa, as determined by SDS-polyacrylamide gel electrophoresis), but differ in overall charge (14). The two isozymes also exhibit different chromatographic and immunological properties (14). The isoenzymes are differentially expressed depending upon the substrate available for growth (14, 15). The specific metabolic functions of both isozymes were recently delineated (8, 16). Conversions of monomethylamine and dimethylamine to CH3SCoM are dependent upon MT2-A, and are not supported by MT2-M (8, 16). In contrast, MT2-M acts specifically in metabolism of methanol, but does not substitute for MT2-A in conversion of monomethylamine or dimethylamine (16). Nevertheless, both isozymes are capable of supporting the conversion of trimethylamine (16). It is proposed that functional specificity arises as a result of protein-protein interactions between the methylated corrinoid proteins acting as methyl donor substrates and the MT2 isozyme proteins catalyzing methyl transfer to HSCoM (16). Although cob(II)alamin was the predominant product found in nonenzymatic methyl transfer from methylcobalamin to thiols (13), herein we show that cob(I)alamin is by far the major product of the reaction catalyzed by MT2. A comparative study of the physicochemical properties of the two isozymes is presented that includes determination of kinetic parameters and characterization of methyl acceptor substrate specificities. The isozymes are shown to be zinc-containing metalloproteins that are strongly inhibited by metal ion chelators. The genes encoding MT2-A and MT2-M, which we designate as cmtA and cmtM, respectively (cmt 5 methylcobamide:CoM methyltransferase) were cloned in Escherichia coli. Furthermore, expression in E. coli is shown to produce both isozymes in an active state. Analyses of the deduced amino acid sequences of both isozymes are presented along with the tentative identification of a consensus zinc binding domain. The implications for geometric constraints on substrate binding at the active site, and the proposed function of zinc in methyltransferase catalysis are discussed. A preliminary account of this work has been presented (17). MATERIALS AND METHODS

General Procedures—Methylcobalamin was obtained from either Aldrich or Sigma. Carbonic anhydrase was a product of Worthington Biochemicals. Unless otherwise specified, all commercially available reagents were of analytical grade. Synthesis of 2-thioethanephosphonate was carried out from 2-bromoethylphosphonic acid using thiourea as sulfur donor. Base hydrolysis of the isothiauronium adduct followed by purification by anion exchange chromatography afforded material with free sulfhydryl/total phosphorous ratio of 1.00. Protein concentrations of the purified MT2 isozymes were determined based on absorbance at 280 nm by use of molar absorptivity coefficients calculated from the deduced amino acid compositions by the method of Perkins (18). The values for MT2-A and MT2-M were 0.884 and 0.558 mg/ml21 cm21, respectively. Direct determination of absorbance at 280 nm on a solution of MT2-M with concentration determined by quantitative amino acid analysis yielded a value of 0.617 mg/ml21 cm21. Analysis of MT-2 Activity—Methyltransferase activity was measured under anaerobic conditions in dim light. The standard assay was carried out at 37 °C in a mixture that contained 1 mM methylcobalamin and 2 mM 2-thioethanesulfonate with 50 mM MOPS buffer, pH 7.2. Demethylation of methylcobalamin was measured spectrophotometrically by the increase in absorbance at 684 nm using a value of 1.2 mM21 cm21 as the absorptivity coefficient of the cob(I)alamin product. In

reactions containing low concentrations of methylcobalamin (#0.1 mM), demethylation was also monitored based on the decrease in absorbance at a more sensitive wavelength (524 nm). At high levels of methylcobalamin (.1 mM), measurements were made by a discontinuous method employing cyanide reactivity to quantify the extent of demethylation of methylcobalamin as a function of time, as described previously (14). Metal Analysis—The concentrations of 40 different elements were determined in protein samples submitted to the University of Georgia, Research Services, Chemical Analysis Laboratory for analysis by plasma atomic emission spectroscopy using a Jarrel Ash 965 Atom Comp plasma emission spectrometer. Zinc concentration was also determined independently using the metallochromic indicator 4-(2-pyridylazo)resorcinol (PAR) (19). A value of De 5 6.6 3 104 M21 cm21 at 500 nm was used as the molar absorptivity of the 2:1 PAR:Zn21 complex. A sample of the protein to be analyzed (0.5–2 mg) was applied to a column of Sephadex G-25 (1.0 3 8.5 cm) prewashed with a solution containing 10 mM EDTA and 50 mM MOPS buffer, pH 7.2. Absorbance at 280 nm was measured on fractions (1 ml) eluted with 50 mM MOPS, pH 7.2, in the absence of EDTA. In the final reaction mixtures (1.0 ml) samples from each fraction (200 –500 ml) were boiled for 10 min in the presence of 100 mM PAR, 150 mM methylmethanethiosulfonate, and 10% SDS, and the absorbance at 500 nm was recorded. When applied to carbonic anhydrase, a protein of known zinc content, the method yielded a value of 1.08 g-atom of zinc/mol of protein. DNA Manipulations—E. coli strains XL1-Blue MRF9 and XLOLR (Stratagene, Inc.) were grown routinely on LB broth at 37 °C with antibiotics, ampicillin, kanamycin, or tetracycline being added as required to final concentrations of 50, 75, and 15 mg/ml, respectively. Genomic DNA from M. barkeri was purified by phenol-chloroform extraction (20). Extraction and purification of plasmid DNA was performed using the Qiagen plasmid purifications kit (QIAGEN). For the production of single-stranded DNA, the helper bacteriophage VCS-M13 was used, and single-stranded DNA was subsequently extracted by standard methods (20). PCR Amplification—Approximately 20 ng of chromosomal DNA or plasmid DNA was added to a mixture (50 ml) containing 60 mM TrisHCl, pH 9.0; 2 mM MgCl2; the indicated oligonucleotide primers (0.5 mM each); dATP, dGTP, dTTP, dCTP (200 mM each); and 1.25 units of Ampli-Taq™ DNA polymerase (Perkin Elmer Corp.). A modified procedure was used to amplify DNA from plaques, as follows. Plaque cores were eluted with 200 ml of 10 mM Tris-HCl, 0.1 mM disodium EDTA, pH 8.0, and a 5-ml sample was added to the 50-ml PCR mixture supplemented with 0.1% Tween 20. Initial denaturation for 6 min at 94 °C was followed by 34 amplification cycles (denaturation, 1 min at 94 °C; annealing, 1 min at 46 °C; extension, 45 s at 72 °C) and a final incubation for 10 min at 72 °C. Oligonucleotide Probes—In order to generate probes for cloning the cmtA gene, we synthesized degenerate forward P2 and reverse P3 PCR primers based on the amino acid sequence of the NH2 terminus and an internal peptide fragment generated by cyanogen bromide cleavage. In a similar fashion, forward primer P4 and reverse primer P5 were prepared for cmtM (Figs. 3 and 5). These primers have the following sequences: P2 (59-AA(A/G)CA(A/G)CA(A/G)GT(A/T/G/C)GA(C/T)(C/A)G(A/T/G/C)ATGCC-39), P3 (59-TC(A/T/G/C)GC(A/T/G/C)CC(A/G)TT(C/ T)TC(A/T/G/C)ACCAT-39), P4 (59-GA(A/G)GG(A/T/G/C)AA(A/G)CC(A/ T/G/C)GT(A/T/G/C)GA(T/C)AA-39), and P5 (59-TG(C/T)C(T/G)(G/ A)TT(C/T)TT(A/T/G/C)GT(A/T/G/C)CCCAT-39). When used in PCR with M. barkeri genomic DNA, the cmtA gene specific primers P2 and P3 amplified a fragment of 550 bp designated dsPCR-P23. In a similar manner, the primers P4 and P5, specific for the cmtM gene, amplified a fragment of 220 bp (dsPCR-P45). DNA nucleotide sequence analysis of the dsPCR-P23 and dsPCR-P45 amplification products showed correspondence to the NH2-terminal amino acid sequences inclusive, and outside of the regions used for construction of the primers. This indicated that both PCR products contained DNA sequences that authentically encode portions of MT2-A and MT2-M. In Southern blots of M. barkeri DNA probed with dsPCR-P23 and dsPCR-P45, fragments of 4.5 kb (EcoRI) and 3.6 kb (HindIII) were revealed for cmtA and cmtM genes, respectively. No cross-hybridization was detected. The results suggested that each protein was encoded by a single locus. Construction and Screening of a Genomic Library—Total genomic DNA from M. barkeri was digested to apparent completion with HindIII and EcoRI and separated by electrophoresis on a 1% agarose gel. Fragments ranging from 2 to 5 kb (HindIII) and 4 to 6 kb (EcoRI) were extracted and purified with the GENECLEAN II kit (Bio 101, Inc.). A primary library was constructed with the HindIII genomic fragments using the pBluescript II SK(2) cloning vector (Stratagene, Inc.). Approximately 50 ng of extracted DNA was used for ligation into the

Methylcobamide:Coenzyme M Methyltransferase Isozymes

18727

HindIII site of pBSK(2) vector using T4 DNA ligase (New England Biolabs, Inc.) at 4 °C for 16 h. The ligation mixture was transformed into the E. coli strain XL1-blue MRF9, and recombinants were selected on LB plates containing ampicillin, 40 mg/ml 5-bromo-4-chloro-3-indolyl b-D-galactoside, and 0.25 mM isopropyl-1-thio-b-D-galactoside. The white colonies were transferred onto positively charged nylon membranes (Life Technologies, Inc.) and screened for cmtM by colony hybridization (21) using digoxigenin-dUTP-labeled dsPCR-P45. In order to clone cmtA, a second library was constructed using the EcoRI fragments, because a HindIII restriction site was present in the dsPCR-P23 amplification product. Approximately 50 ng of extracted DNA was ligated into EcoRI-digested arms of the lZAP Express vector (Stratagene, Inc.) using T4 DNA ligase. DNA from the recombinants was then packaged according to the manufacturer’s recommendations (Gigapack II Gold packaging extract, Stratagene, Inc.) and assayed on the bacterial host XL1-Blue MRF9. A portion of the library was plated out at low density (104 plaque-forming units/plate), and the phage plaques were transferred to nylon membranes and screened for recombinants containing the cmtA gene using digoxigenin-dUTP-labeled dsPCR-P23. DNA Sequencing—Nucleotide sequences were determined by the dideoxy chain termination method of Sanger et al. (22). with a Sequenase kit (U.S. Biochemical Corp.) and [a-35S]dATP (1000 Ci mmol21; DuPont NEN) according to the instructions contained therein. The nucleotide sequence was determined from both strands using specific synthetic oligonucleotide primers. Nucleotide sequence data were analyzed with the Genetics Computer Group (University of Wisconsin, Madison) sequence analysis software package, version 8.0 (23). In order to prepare single stranded templates for both strands, we subcloned the 3.6-kb HindIII fragment (cmtM) in the pBluescript KS(1) vector generating the plasmid pMET84. The 4.5-kb EcoRI fragment (cmtA) was subcloned in both orientations (relative to the f1 replication origin) by use of pBluescript vectors SK(2) and KS(1), producing the plasmids pMA83 and pMA85, respectively. RESULTS AND DISCUSSION

Identification of Cob(I)alamin as the Major Product of Methyl Group Transfer—The finding of trace levels of cob(I)alamin in non-enzymic reactions of methylcobalamin with thiols was taken to indicate that cob(I)alamin, and not cob(II)alamin, is the immediate product of methyl group transfer (13). In order to identify the products formed by the methyltransferase isozymes, reaction mixtures were monitored by UV-visible spectroscopy. As shown in Fig. 1, large amounts of cob(I)alamin were produced with relatively little generation of cob(II)alamin. However, substantial amounts of cob(II)alamin were found in separate reactions (not shown) that contained either aerobic enzyme samples or high levels of thiols. These results are consistent with oxidation of cob(I)alamin by traces of disulfide present in thiol preparations, as mentioned in studies by Hogenkamp et al. (13). To our knowledge these are the first data presented that show the major product of the MT2-catalyzed reaction to be the Co11 form of the cobamide (as designated in Reactions 2 and 4). The findings strongly support the conclusion of heterolytic cleavage of the carbon-cobalt bond (13), thus indicating displacement of the methyl group by nucleophilic attack of thiolate. Inhibition of MT2 by Metal Ion Chelators—Both MT2 isozymes were markedly inhibited in the presence of metal ion chelators. The dependence of methyltransferase activity on chelator concentration was assessed in standard reaction mixtures (initiated by addition of purified MT2-A) that contained varying levels of EGTA, EDTA, and 1,10-phenanthroline. A high degree of inhibition resulted at relatively low chelator concentrations. Fifty percent inhibition was produced by EGTA at a level of 50 mM and by 1,10-phenanthroline at 70 mM. The strongest inhibitor was EDTA, which caused 50% reduction of activity at a concentration of about 5 mM. Inhibition by EDTA was relieved by excess addition of any of several different divalent metal ions, including Ca21, Mn21, and Ni21. Chromatography of a 1-ml sample of enzyme containing 5 mM EDTA on a column (1.5 3 17.5 cm) of Sephadex G-25 resulted in 90%

FIG. 1. Cob(I)alamin product of MT2-A and MT2-M catalysis. Reactions were carried out in dim light under strictly anaerobic conditions at 23 °C with 100 mM HSCoM in 25 mM MOPS buffer, pH 7.2. Total reaction volumes were 1.00 ml containing the purified isozymes MT2-A (0.275 unit) or MT2-M (0.348 unit). Methylcobalamin concentration was 56 mM in the reaction with MT2-A and 65 mM with MT2-M. Reactions were initiated by addition of HSCoM. Spectra plotted for the reaction with MT2-A were recorded at the following times (min) after addition of HSCoM: 0, 1.75, 3.75, 6.08, 9.08, 13.75, 19.75, and 44.75. The spectra from the MT2-M reaction were plotted at 0, 1.42, 2.75, 4.08, 5.42, 7.08, 9.75, and 23.75 min after start of the reaction. The peak at 388 nm is characteristic of cob(I)alamin, and absorbance at 388 nm is shown as a function of reaction time in the insets.

recovery of activity, indicating that upon removal of the chelator inhibition was reversible. The findings suggested the presence of an essential metal ion bound to the enzyme. Identification of MT2 as a Zinc-containing Enzyme—In order to determine whether or not metals were present in the enzyme, samples of MT2-M were analyzed by plasma emission spectroscopy, as described under “Materials and Methods.” The only elements found at concentrations comparable to that of the protein were copper and zinc, accounting for 0.08 and 0.4 g-atom/mol protein, respectively. No significant levels were found of any of the other elements tested, including Mg, Ca, Mn, Fe, Co, Ni, Se, and Mo. Exposure of the enzyme to 5 mM EDTA followed by gel filtration in the presence of EDTA (2 mM) had no effect on the level of either copper or zinc. Metal content of MT2-M measured by a separate method using a metallochromic indicator, as described under “Materials and Methods,” yielded a value of 0.5 g-atom zinc/mol MT2-M. By this method MT2-A was found to contain 0.9 g-atom zinc/mol. The levels of contaminating zinc in buffer solutions used with the enzymes were not detectable by either of the methods as employed. The results provide evidence that MT2 isozymes are zinc-containing metalloproteins, and suggest that inhibition by chelators may be due to interference with Zn21 coordination by the

18728

Methylcobamide:Coenzyme M Methyltransferase Isozymes

TABLE I Substrate specificities and kinetic parameters of methyltransferase isozymes

a Compounds were tested at 20 mM final concentration added to the standard assay in place of HSCoM, with the exception that H4SPt and CH3H4SPt were both 140 mM. b Km measurements for all methyl acceptor substrates were performed at 50 mM CH3-B12 in the presence of 1 mM dithiothreitol. c Activity was lower than with either HSCoM or 3-mercaptopropionate. d The Km for methylcobalamin was determined under standard conditions in the presence of saturating HSCoM (2 mM).

enzyme. Methyl Acceptor Substrate Kinetics and Specificity—In previous studies, methyl acceptor substrate activity was tested with several different thiol compounds (14, 15). However, until now no compounds have been reported to be active other than the natural substrate HSCoM. Thus, a high degree of substrate specificity is indicated. In order to characterize more fully the substrate specificity, and to determine the kinetic properties of alternative substrates, we have extended the range of compounds tested as methyl acceptors (Table I). With the exception of weak inhibition by 2,3-dimercaptopropanesulfonate, none of the compounds listed caused a decrease in activity when present at 20 mM in reaction mixtures containing 2 mM HSCoM. As shown in Table I, activity was found with two compounds other than HSCoM, i.e. with 3-mercaptopropanoic acid (3-MPA), and 2-thioethanephosphonic acid. Neither the higher homolog of coenzyme M, 2-thiopropanesulfonate, nor the lower homolog of 3-MPA, thioglycolate, was active, indicating that carbon chain length may be critical for interaction of the substrate at the active site. The data further suggest that the substrate must contain minimally a free sulfhydryl group positioned at a specific distance from an anionic moiety (e.g. carboxylate, phosphonate, or sulfonate). Steady state kinetic experiments showed that both MT2 isozymes have similar values of Km(app) for coenzyme M (35 mM

FIG. 2. Strategy for cloning of cmtA. The steps employed to clone a 4.5-kb EcoRI fragment containing cmtA into pBK-CMV are shown schematically. Details of the procedures used are described under “Materials and Methods.” Results obtained at the various stages are presented under “Results and Discussion.”

for MT2-A, and 20 mM for MT2-M), as given in Table I. The apparent Km values for 3-mercaptopropanoic acid were 300 – 500 times higher than the values for coenzyme M, possibly indicating a weaker affinity for the non-natural substrate. Nevertheless, both isozymes also displayed similarly high values of Km(app) for 3-MPA (Table I). The methyl donor substrate kinetics were also similar in that both isozymes displayed rather high Km(app) values for methylcobalamin of 14 mM (Table I). This finding indicates the possibility of weak binding to methylcobalamin. Low affinity for methylcobalamin has been previously considered to be a consequence of the fact that the physiological methyl donor substrates are different corrinoid proteins exhibiting specific recognition of MT2 isozymes via protein-protein interactions (16). Cloning and Sequencing of MT2 Isozyme Genes cmtA and cmtM—A DNA library from M. barkeri consisting of genomic EcoRI fragments ranging from 4 to 6 kb was prepared in the lZAP Express vector, as described under “Materials and Methods.” Positive plaques containing the cmtA gene were screened initially by plaque hybridization using the 550-bp dsPCR-P23 amplification product as probe (Fig. 2). A second screening was performed by direct PCR amplification of the DNA eluted from 32 positive plaques using primers P2 and P3. From this, two positive clones Z-8 and Z-32 were purified for further characterization. The recombinant clone Z-8 was used for in vivo excision of the phagemid pMA81 (Fig. 2). Restriction mapping analysis indicated that the phagemid contained a 4.5-kb EcoRI insert (Fig. 2) that hybridized with the dsPCR-P23 probe. The presence of a region encoding MT2-A was further confirmed by PCR amplification with primers P2 and P3. The DNA sequence of cmtA was determined on both strands, as described under “Materials and Methods.” Analysis of the nucleotide sequence

Methylcobamide:Coenzyme M Methyltransferase Isozymes

FIG. 3. Nucleotide and deduced amino acid sequence of cmtA (MT2-A). The deduced amino acid sequence is shown directly below the second base of each codon. The stop codon is indicated by End. A potential zinc binding site TVLHICG is given in bold. The degenerate oligonucleotides P2 and P3 were designed from amino acid sequences within the shaded boxes. A sequence corresponding to the ribosome binding site is double underlined. The potential TATA box A and box B are marked above the sequences by asterisks and dots, respectively. The inverted repeats inv-1 and inv-2 are indicated by converging arrows below the sequence. A putative transcription terminator sequence is indicated by a dashed underline. Downstream direct tandem repeats are underlined.

(Fig. 3) revealed an open reading frame encoding a protein of 339 amino acids with a calculated molecular mass of 36.7 kDa. The deduced molecular mass agreed with the previous estimation by SDS-polyacrylamide gel electrophoresis (8, 9, 14, 15). A putative TATA-like sequence (box A), which matched the archaeal consensus sequence 59-(T/C)TTA(T/A)A-39 (24, 25) was identified 103 nucleotides upstream from the initiation codon ATG (position 202) (Fig. 3). A potential ribosome binding site was identified 7 bp upstream from the initiation codon. In the promoter region, two DNA segments were found with the potential to form stem and loop structures (Fig. 3). The first of these inverted repeats (inv-1, located at positions 129 –158) encompassed a putative transcription start signal (archaeal box B: 59-ATGC-39) located 27 nucleotides downstream from the box A. The second (inv-2, at positions 162–199) overlapped the ribosome binding site. In addition, two tandem direct repeats were also present in the promoter region. It has been suggested that such structures play a role in regulation of gene transcription (26). A potential methanogen transcription terminator was found at positions 1297–1327. Downstream from this (positions 1385–1443) there existed four direct tandem repeats composed of 15 bp (t/CATTTTTCAATCTCG) punctuated by a guanine residue (Fig. 3). Similar tandem repeats have been described in other methanogens, but their function is still unknown (26, 27).

18729

FIG. 4. Strategy for cloning of cmtM. The strategy used for cloning of the 3.6-kb HindIII fragment encoding cmtM into pBluescript II SK(2) is shown. Details of the methods employed and the results obtained are given in the text.

In order to isolate the gene encoding MT2-M, an M. barkeri genomic library was constructed in the pBSK(2) vector containing HindIII fragments of size varying from 2 to 5 kb. The digoxigenin-labeled 220-bp dsPCR-P45 amplification product specific to MT2-M was used as a probe (Fig. 4). From a single positive colony, the plasmid pMET8 was isolated and subjected to restriction endonuclease digestions. The results showed that two fragments (3.4 and 3.6 kb) had been jointly ligated into the same vector. Subcloning into pBSK(2) produced plasmid pMET82 (Fig. 4) that contained only the 3.6-kb HindIII fragment and was able to hybridize with the dsPCR-P45 probe and produce the 220-bp dsPCR-P45 product in PCR amplification. The amino acid sequence deduced from DNA sequence analysis of the 59 end matched perfectly the NH2-terminal region of MT2-M, confirming that the 3.6-kb HindIII fragment contained either a large portion or, as later demonstrated, the entire cmtM gene. The complete sequence of DNA containing cmtM is shown in Fig. 5. A single open reading frame of 1,017 bp was found encoding a 339-amino acid protein with a molecular mass of 35.9 kDa. The amino acid composition of MT2-M deduced from the DNA sequence was very close to the composition of the protein obtained by hydrolysis and amino acid analysis (data not shown). A putative TATA box A matching to the consensus archaeal box A sequence (24, 25) was identified 294 bp upstream from the initiation codon ATG (position 406). A consensus ribosome binding site was found preceding the initiation codon at positions 393–398. In addition, there existed at 19 nucleotides downstream from the TATA box A, a potential methanogenic transcription start site (archaeal box B), with

18730

Methylcobamide:Coenzyme M Methyltransferase Isozymes

FIG. 6. Comparison of the amino acid sequences of MT2-A and MT2-M. Deduced amino acid sequences of the M. barkeri MT2 isozymes were aligned using the GCG software (23). Dotted positions indicate gaps. Identical residues are shaded.

FIG. 5. Nucleotide and deduced amino acid sequence of cmtM (MT2-M). The sequence is annotated as described in the legend to Fig. 3. Degenerate oligonucleotides P4 and P5 were devised from the shaded amino acid sequences.

sequence identical to the consensus 59-ATGC-39 (26).2 As in the case of cmtA, sequences capable of secondary structure formation were found in the promotor region (Fig. 5). One of these (inv-1) was localized at positions 250 –295, and another (inv-2) was found at positions 191–219. A third region located at positions 98 –119 contained a transcription terminator-like sequence, overlapping the TATA box A. In addition, we have identified in the promotor region six short tandem direct repeats of 7– 8 bp in length. A potential transcription terminator was found immediately downstream from the stop codon at positions 1434 –1448 (Fig. 5). Analysis of the Deduced Amino Acid Sequences from cmtA and cmtM Genes—Comparison of the deduced amino acid sequences from cmtA and cmtM revealed 39% identity and 58% similarity, as shown in Fig. 6. Based on the content of charged amino acids, the calculated net charges at pH 9.0 were 220 and 217 for MT2-A and MT2-M, respectively. This finding is consistent with previous results, which showed that MT2-A has greater electrophoretic mobility at pH 9 than MT2-M (14). A large number of the identical amino acids were found grouped in clusters. These conserved regions might be involved with catalytic functions common to both isozymes. Conversely, variation in other parts of the protein might be related to their immunological and corrinoid protein substrate specificity differences (14, 16). The sequence TVLHICG located in the carboxyl-terminal region (residues 236 –242) is the longest stretch 2 Monocistronic mRNA transcripts of 1100 bp (cmtA) and 1300 bp (cmtM) were reported very recently by Harms and Thauer (39). The lengths of these transcripts correspond well to the distances between the positions of the proposed transcription start sites (box B) and 39termination sequences shown in Figs. 3 and 5.

of identical amino acids. The primary structures of zinc-containing proteins frequently contain closely spaced sulfhydryl and imidazole groups at the site of metal ion binding (28). Thus, the conserved TVLHICG sequence is a potential candidate for binding of zinc. The GenBank/EMBL data base was searched for translated sequences homologous to MT2 isozymes. A FASTA homology search showed 21.7% identity with 8 gaps within a 335-amino acid region overlap between MT2-M and the uroporphyrinogen decarboxylase from Anacystis nidulans R2 (29). Significant homology was also revealed between MT2-A and the Rhodobacter capsulatus uroporphyrinogen decarboxylase (30). However, 12 gaps were required yielding 24% identity overall. Moreover, both MT2 isozymes showed high levels of homology ('30% identity and 53% similarity) to the 40-kDa subunit of the M. barkeri 480-kDa corrinoid enzyme.3 Functional Expression of cmtA and cmtM Genes in E. coli—In order to determine whether or not the cloned M. barkeri genes were expressed in E. coli, cell extracts were prepared from E. coli containing plasmids pMA81 (cmtA), pMET82 (cmtM), and pMET81 (control). The plasmid pMET81 was used as a negative control because it contained an anonymous 3.4-kb HindIII fragment, as shown in Fig. 4. Western blot analyses revealed that both cmt genes were expressed in E. coli, and that both protein products exhibited electrophoretic mobilities identical to those found with M. barkeri extracts. A band at the position corresponding to MT2 was absent in the extract from cells containing pMET81. The extracts were then assayed for MT2 activity, and the results are presented in Table II. Extracts from cells containing pMA81 and pMET82 were both capable of rapidly catalyzing methyl group transfer from methylcobalamin to HSCoM (Table II). However, activity was not detectable in extracts of E. coli cells harboring pMET81. Although neither cmtA nor cmtM was under the control of the lacZ promotor (see Figs. 2 and 4), the results show that both archaeal genes are recognized by the E. coli transcriptional apparatus, and that substantial levels of both isozymes are produced in an active form. These findings suggest that in the A1T-rich 59-untranslated region there may be a sequence that mimics the E. coli promotor consensus sequence, thereby providing a site for initiation of transcription. Such a sequence has been reported in Methanococcus vannielii at which E. coli RNA polymerase binds and initiates transcription of the hisA gene (31). Mechanism of Catalysis—Alkylation of metal-bound thiolate 3 The sequence of the 40-kDa component of the M. barkeri 480-kDa corrinoid protein was recently determined (L. Paul and J. A. Krzycki, submitted for publication).

Methylcobamide:Coenzyme M Methyltransferase Isozymes TABLE II Methylcobamide:CoM methyltransferase activity in E. coli expressing MT2-A and MT2-M Bacterial strain/pDNA (cell-free extract)

MT2 specific activitya

mmol CH3-SCoM/min/mg

M. barkeri (trimethylamine-grown) M. barkeri (methanol-grown) E. coli XLOLR/pMA81 (cmtA) E. coli XL1-Blue/pMET82 (cmtM) E. coli XL1-Blue/pMET81 (control) a

2.2 7.5 1.21 0.154 0.0000

Activity measured under standard assay conditions.

ligands has been described in transition metal complex systems (32, 33). Furthermore, the involvement of zinc in the mechanism of methylthioether formation at a specific cysteine residue in the repair of DNA methylphosphotriesters by the E. coli Ada protein has received considerable attention (34 –38). In the Ada system, spectroscopic evidence shows that zinc binds to, and activates the sulfhydryl of cysteine 69 for nucleophilic attack on the methyl group of a DNA methylphosphotriester, forming the methylthioether (35). Further studies may now be considered in order to determine whether or not the MT2 isozymes operate by an analogous mechanism. Acknowledgments—We thank Guy LeClerc and Rowena Matthews for helpful discussions. We appreciate technical assistance provided by Kenneth Gable and Gary Herring. We are very thankful to L. Paul and J. A. Krzycki for generously communicating their unpublished sequence of the 40-kDa protein. We are grateful to Eric Eccleston at the University of Minnesota, Department of Biochemistry, Microchemical Facility for quantitative amino acid analysis of MT2-M. We also thank Rebecca Auxier at the University of Georgia, Research Services, Chemical Analysis Laboratory for carrying out plasma emission spectroscopic analyses. REFERENCES 1. Whitman, W. B. (1985) in The Bacteria (Woese, C. R., and Wolfe, R. S., eds) Vol. 8, pp. 3– 84, Academic Press, Inc., New York 2. Ferry, J. G. (1992) CRC Crit. Rev. Biochem. Mol. Biol. 27, 473–503 3. Bock, A.-K., Prieger-Kraft, A., and Scho¨nheit, P. (1994) Arch. Microbiol. 161, 33– 46 4. Wolfe, R. S. (1991) Annu. Rev. Microbiol. 45, 1–35 5. Ferry, J. G. (1992) J. Bacteriol. 174, 5489 –5495 6. DiMarco, A. A., Bobik, T. A., and Wolfe, R. S. (1990) Annu. Rev. Biochem. 59,

18731

355–394 7. Keltjens, J. T., and Vogels, G. D. (1993) in Methanogenesis (Ferry, J. G., ed) pp. 253–303, Chapman & Hall, New York 8. Burke, S. A., and Krzycki, J. A. (1995) J. Bacteriol. 177, 4410 – 4416 9. van der Meijden, P., Heythuysen, H. J., Pouwels, A., Houwen, F., van der Drift, C., and Vogels, G. D. (1983) Arch. Microbiol. 134, 238 –242 10. van der Meijden, P., te Bro¨mmelstroet, B. W., Poirot, C. M., van der Drift, C., and Vogels, G. D. (1984) J. Bacteriol. 160, 629 – 635 11. Daas, P. J. H., Gerrits, K. A. A., Keltjens, J. T., van der Drift, C., and Vogels, G. D. (1993) J. Bacteriol. 175, 1278 –1283 12. Taylor, C. D., and Wolfe, R. S. (1974) J. Biol. Chem. 249, 4886 – 4890 13. Hogenkamp, H. P. C., Bratt, G. T., and Sun, S. (1985) Biochemistry 24, 6428 – 6432 14. Grahame, D. A. (1989) J. Biol. Chem. 264, 12890 –12894 15. Yeliseev, A., Ga¨rtner, P., Harms, U., Linder, D., and Thauer, R. K. (1993) Arch. Microbiol. 159, 530 –536 16. Ferguson, D. J., Krzycki, J. A., and Grahame, D. A. (1996) J. Biol. Chem. 271, 5189 –5194 17. LeClerc, G. M., and Grahame, D. A. (1995) Ann. Meet. Am. Soc. Microbiol., May 21–25, 1995, Washington, D. C., p. 320 18. Perkins, S. J. (1986) Eur. J. Biochem. 157, 169 –180 19. Hunt, J. B., Neece, S. H., and Ginsberg, A. (1985) Anal. Biochem. 146, 150 –157 20. Sambrook, J., Fritsch, E. F., and Maniatis, T. (1989) Molecular Cloning: A Laboratory Manual, 2nd Ed., Cold Spring Harbor Laboratory, Cold Spring Harbor, NY 21. Grunstein, M., and Hogness, D. S. (1975) Proc. Natl. Acad. Sci. U. S. A. 72, 3961–3965 22. Sanger, F., Nicklen, S., and Coulson, A. R. (1977) Proc. Natl. Acad. Sci. U. S. A. 74, 5463–5467 23. Devereux, J., Haeberli, P., and Smithies, O. (1984) Nucleic Acids Res. 12, 387–395 24. Palmer J. R., and Daniels, C. J. (1995) J. Bacteriol. 177, 1844 –1849 25. Hain, J., Reiter, W.-D., Hu¨depohl, U., and Zillig, W. (1992) Nucleic Acids Res. 20, 5423–5428 26. Reeve, J. N. (1992) Annu. Rev. Microbiol. 46, 165–191 27. Latimer, M. T., and Ferry, J. G., (1993) J. Bacteriol. 175, 6822– 6829 28. Vallee, B. L., and Auld, D. S. (1990) Biochemistry 29, 5647–5659 29. Kiel, J. A., Ten Berge, A. M., and Venema, G. (1992) DNA Seq. 2, 415– 418 30. Ineichen, G., and Biel, A. J. (1995) Plant Physiol. 108, 423 31. Brown, J. W., Thomm, M., Beckler, G. S., Frey, G., Stetter, K. O., and Reeve, J. N. (1988) Nucleic Acids Res. 16, 135–150 32. Lindoy, L. F., and Busch, D. H. (1974) Inorg. Chem. 13, 2494 –2498 33. Elder, R. C., Kennard, G. J., Payne, M. D., and Deutsch, E. (1978) Inorg. Chem. 17, 1296 –1303 34. Myers, L. C., Terranova, M. P., Nash, H. M., Markus, M. A., and Verdine, G. L. (1992) Biochemistry 31, 4541– 4546 35. Myers, L. C., Terranova, M. P., Ferentz, A. E., Wagner, G., and Verdine, G. L. (1993) Science 261, 1164 –1167 36. Myers, L. C., Verdine, G. L., and Wagner, G. (1993) Biochemistry 32, 14089 –14094 37. Myers, L. C., Jackow, F., and Verdine, G. L. (1995) J. Biol. Chem. 270, 6664 – 6670 38. Wilker, J. J., and Lippard, S. J. (1995) J. Am. Chem. Soc. 117, 8682– 8683 39. Harms, U., and Thauer, R. K. (1996) Eur. J. Biochem. 235, 653– 659