Enzymatic Properties of S-Adenosylmethionine Synthetase from the Archaeon Methanococcus jannaschii *

S-Adenosylmethionine synthetase (ATP:l-methionine S-adenosyltransferase, MAT) catalyzes a unique enzymatic reaction that leads to formation of the primary biological alkylating agent. MAT from the hyperthermophilic archaeon Methanococcus jannaschii (MjMAT) is a prototype of the newly discovered archaeal class of MAT proteins that are nearly unrecognizable in sequence when compared with the class that encompasses both the eucaryal and bacterial enzymes. In this study the functional properties of purified recombinant MjMAT have been evaluated. The products of the reaction are AdoMet, PPi, and Pi; >90% of the Pioriginates from the γ-phosphoryl group of ATP. The circular dichroism spectrum of the dimeric MjMAT indicates that the secondary structure is more helical than the Escherichia coli counterpart (EcMAT), suggesting a different protein topology. The steady state kinetic mechanism is sequential, with random addition of ATP and methionine; AdoMet is the first product released, followed by release of PPi and Pi. The substrate specificity differs remarkably from the previously characterized MATs; the nucleotide binding site has a very broad tolerance of alterations in the adenosine moiety. MjMAT has activity at 70 °C comparable with that of EcMAT at 37 °C, consistent with the higher temperature habitat of M. jannaschii. The activation energy for AdoMet formation is larger than that for the E. coli MAT-catalyzed reaction, in accord with the notion that enzymes from thermophilic organisms are often more rigid than their mesophilic counterparts. The broad substrate tolerance of this enzyme proffers routes to preparation of novel AdoMet analogs.

thionine S-adenosyltransferase; also known as methionine adenosyltransferase (MAT), EC 2.5.1.6) catalyzes the only known route of AdoMet biosynthesis (6 -9). The synthetic process occurs in a unique reaction in which the complete triphosphate chain is displaced from ATP and a sulfonium ion formed (6,9). MATs from various organisms contain ϳ400-amino acid polypeptide chains. However we have recently found that the protein sequences comprise two categories, the extensively studied eucaryal-bacterial type (encoded by a catalytic subunit denoted ␣ (Ref. 10)) and the newly recently discovered archaeal type (encoded by a subunit that we denote ␥ (Ref. 11)); ␤ chains encode a regulatory subunit for the ␣ class (10). Within each category the sequences are highly conserved, e.g. there is 59% identity between the human and Escherichia coli MAT ␣ subunits. In contrast, the sequences of the two classes are widely diverged, e.g. there is only 22% identity between the Methanococcus jannaschii ␥-type MAT and either the E. coli or human ␣ subunits. The variation in sequence follows evolutionary lines and is not simply related to the habitat of the organism because MATs from hyperthermophilic bacteria and archaea fall in the same classes as MATs from mesophiles of the same kingdom (11). None of the MAT sequences is discernibly related to that of any other known protein.
The AdoMet synthetic reaction catalyzed by the ␣ class MATs is composed of two sequential steps, AdoMet formation and the subsequent hydrolysis of tripolyphosphate (PPP i ) as depicted below. L-methionine ϩ ATP ϩ H 2 O 3 ͓AdoMet ϩ PPP i ͔ ϩ H 2 O 3 AdoMet ϩ PP i ϩ P i Enzyme-bound REACTION 1 The P i formed during the MAT-␣-catalyzed reaction originates predominantly from the ␥-phosphoryl group of ATP (6). Both divalent cations (e.g. Mg 2ϩ ) and monovalent cations (e.g. K ϩ ) are required for maximal activity. Whether the same catalytic mechanism operates with the archaeal enzyme, particularly at elevated temperature, is of interest in view of the unclear role of PPP i hydrolysis in the thermodynamics of AdoMet formation (12).
MjMAT Purification-Recombinant M. jannaschii AdoMet synthetase was prepared from the strain BL21(DE3)(codon plus)/pMJ1208-1. This protein has a decahistidine tag on the N terminus (11). Cells were grown at 37°C in LB medium containing 50 g/ml carbenecillin. Induction of MjMAT expression was obtained by addition of 1 mM isopropyl-1-thio-␤-D-galactopyranoside to a 20-fold dilution of an overnight culture and shaking for 3 h. Cells were harvested by centrifugation and stored at Ϫ80°C until use.
The cell pellet was suspended in 50 mM Tris⅐HCl, 1 mM DTT, 30 M phenylmethylsulfonyl fluoride, pH 8, using 10 ml of buffer/g (wet weight) of cells. The cells were lysed by one pass through a French press at 10,000 p.s.i. Debris was removed by centrifugation at 13,000 ϫ g for 30 min. Ammonium sulfate was added to the supernatant to 20% saturation, and the insoluble material was removed by centrifugation. Ammonium sulfate was then added to the supernatant to 80% saturation, and the pellet was collected by centrifugation.
The scale of following steps is described for a preparation from 25 g (wet weight) of cells. The ammonium sulfate pellet was dissolved in ϳ20 ml of Buffer A (containing 50 mM Tris⅐HCl, 1 mM DTT, pH 8) and dialyzed overnight at 4°C against 2 liters of the same buffer. The protein was divided into 5-ml portions and heated in a 85°C water bath for 30 min. Precipitated materials were removed by centrifugation.
Chromatographic steps were performed at room temperature. The protein was dialyzed overnight into 50 mM Tris⅐HCl, pH 8, and then NaCl and imidazole were added to final concentrations of 300 and 5 mM, respectively. The sample was loaded onto a 2 ϫ 15-cm column of Ni 2ϩ -His⅐Bind resin (Novagen) that was equilibrated with 50 mM Tris⅐HCl, 300 mM NaCl, 5 mM imidazole, pH 8. The column was washed with 80 ml of starting buffer (at a flow rate of 1.0 ml/min) and eluted with a 200-ml gradient from the starting buffer to a final buffer of 50 mM Tris⅐HCl, 300 mM NaCl, 500 mM imidazole, pH 8. The fractions containing activity were dialyzed overnight against 50 mM Tris⅐HCl, 1 mM DTT, pH 8, and stored at Ϫ80°C.
Minor contaminants were subsequently removed by ion exchange and gel filtration chromatography. Before ion exchange chromatography, the protein was equilibrated in 25 mM MES, 1 mM DTT, pH 6, by dialysis. The protein was loaded onto a HiLoad Q-Sepharose (Amersham Biosciences) column (2.6 ϫ 12 cm) equilibrated and washed with 100 ml of dialysis buffer (at a flow rate of 2.5 ml/min), and the protein was eluted with a 500-ml gradient of this buffer also containing 0.7 M KCl. The active fractions were pooled and dialyzed against 50 mM Tris⅐HCl, 50 mM KCl, 1 mM DTT, pH 8. Final purification was obtained by gel filtration chromatography of 2-ml aliquots on a 2 ϫ 60-cm Superdex-200 column (Amersham Biosciences) equilibrated and eluted at 0.5 ml/min with 50 mM Tris⅐HCl, 50 mM KCl, 1 mM DTT, pH 8.0. The final protein displayed a single band on Coomassie Blue-stained 8 -25% gradient polyacrylamide Phast gels (Amersham Biosciences) in the presence and absence of SDS. The concentration of purified MjMAT was estimated from the absorbance at 280 nm using an extinction coefficient of 1.0 (mg/ml) Ϫ1 cm Ϫ1 as calculated from the amino acid composition (19). Approximately 50 mg of enzyme were obtained from 15 g (wet weight) of cells.
AdoMet synthetase activity was determined by a [ 14 C]AdoMet cation exchange filter binding method (20). Assays were performed at 55°C in 25 mM Hepes⅐(CH 3 ) 4 N ϩ at pH 8.0 with 50 mM KCl, and 10 mM MgCl 2 . Routine assays contained 9.5 mM ATP and 0.5 mM [methyl-14 C]methionine. Substrate saturation data were evaluated using the kinetic equations of Cleland (21) or the kinetics module of Sigmaplot 2000 (SPSS Science), which utilizes the Segel formalism (22). All fits were consistent with linear dependence of activity on substrate and/or inhibitor concentration. Inhibition studies were conducted with substrate concentrations near their K m values.
The activation energy for AdoMet synthesis, E a , was calculated from the slope of linear ranges of an Arrhenius plot of ln(k cat ) versus 1/T (slope ϭ ϪE a /R). The free energy of activation, ⌬G # , was calculated as ⌬G # ϭ RT(ln(k B T/h) Ϫ ln(k cat )) where R ϭ 1.9872 cal/mol-K, k B is the Boltzmann constant, h is Planck's constant, and k cat is the value measured at temperature T. The activation enthalpy, ⌬H # , was calculated as (E a Ϫ RT). The activation entropy ⌬S # was then obtained from (⌬H # Ϫ ⌬G # )/T (23). A more complete analysis of the temperature dependence of reaction rates considers that nonlinearity of an Arrhenius plot may arise from heat capacity change during the reaction, ⌬C p # (24). Thus, with respect to a reference temperature (T 0 , defined as 37°C), at a given temperature T, ). These relationships were used in fitting the observed data for MjMAT via nonlinear least squares using the program Scientist (MicroMath Inc.).
The phosphorus-containing reaction products were identified by evaluating PPP i , PP i , and P i formation from [␥-33 P]ATP in the presence and absence of inorganic pyrophosphatase. Compounds were separated by thin layer chromatography on polyethyleneimine-cellulose anion exchange thin layer sheets (EM Science) developed in 0.9 M LiCl, 50 mM EDTA, pH 7. The radiation was quantified using a Fuji phosphorimager. A 33 PPP i standard was prepared from the [␥-33 P]ATP by periodate oxidation followed by aniline cleavage (12, 25); 33 PP i was prepared from 33 PPP i using E. coli MAT. Solutions contained 5 mM ATP (4.4 ϫ 10 4 cpm/nmol), 5 mM methionine, 1 mg/ml MjMAT, 100 mM Hepes, pH 8, 50 mM KCl, 10 mM MgCl 2 ; when present, 0.025 units/l inorganic pyrophosphatase were added. Time points were taken from 1 to 60 min. ATP was removed from the reactions by adsorption to Norit before TLC (12). ATP, P i , PP i , and PPP i standards had R F values of 0.27, 0.8, 0.13, and 0.04, respectively. After 4 h, ϳ77% of the 33 P was present as P i .
The reverse reaction was examined in two ways: by looking for either the formation of [carboxy-14 C]methionine from [carboxy-14 C]AdoMet or the formation of [ 33 P]ATP or 33 PPP i from 33 P i . In the first case, 0.2 mM [carboxy-14 C]AdoMet (56 mCi/mmol), 5 mM PP i , and 5 mM P i were incubated with 1 mg/ml enzyme in 0.1 M Tris⅐Cl, 50 mM KCl, 10 mM MgCl 2 for up to 60 min at 55°C. Compounds were separated by thin layer chromatography on cellulose sheets developed in n-butyl alcohol: acetic acid:water (25:4:10) in which AdoMet and methionine have R F values of 0.26 and 0.56, respectively. Radioactivity was quantified by phosphorimaging. In the latter case, 5 mM AdoMet, 5 mM PP i , and 5 mM 33 P i (11 mCi/mmol) were incubated with 1 mg/ml enzyme for up to 60 min at 55°C. Compounds were separated by thin layer chromatography on polyethyleneimine-cellulose sheets and radioactivity quantified as described above.
Circular Dichroism (CD) Spectra-Protein secondary structure was assessed from circular dichroism spectra obtained on an Aviv model 62A spectropolarimeter. Samples (0.3 mg/ml protein in 25 mM Tris⅐HCl, 25 mM KCl, pH 8.0) were placed in 1-mm path length cells; spectra were recorded from 200 to 260 nm and were corrected for buffer contributions.

Recombinant
MjMAT was purified to electrophoretic homogeneity from E. coli as described under "Experimental Procedures." The protein was allowed to retain an N-terminal decahistidine tag, because we had found that the properties of the recombinant protein were comparable with those of the protein isolated from M. jannaschii (11). Recombinant MjMAT chromatographed with a M r of 86,000 on gel filtration, consistent with its being a dimer of 45-kDa subunits (11). The purified enzyme readily crystallizes under a variety of common conditions, 2 and attempts to obtain diffraction quality crystals are in progress.
CD spectra were compared for MjMAT and the E. coli enzyme to assess potential similarities in secondary structure (Fig. 1). MjMAT displays substantially larger ellipticity at 220 nm, a wavelength characteristic of the proportion of ␣-helix. The crystal structures of the E. coli and rat liver enzymes show that they are composed of ϳ25% ␣-helix and ϳ20% ␤-sheet (26,27). The differences in CD spectra suggest that the overall secondary structure of MjMAT is more helical than the eucaryal or bacterial MAT-␣.
Origin of the P i Formed in the Reaction-A notable feature of MAT-␣-catalyzed reaction is that the products released from the enzyme are PP i and P i , rather than PPP i initially formed in conjunction with AdoMet synthesis. The PPP i formed as an intermediate is primarily hydrolyzed even before it can reorient, with the result that Ͼ95% P i originates from the ␥-phosphoryl group of ATP (6). Our previous studies of MjMAT dem-onstrated that PP i and P i were formed as products (11). When the products formed in the MjMAT-catalyzed reaction from [␥-33 P]ATP were analyzed, 33 P i constituted 93% of the product and 7% was present in 33 PP i ; 33 PPP i was not detected. Thus, even at the 55°C temperature used in this experiment, the PPP i initially created upon AdoMet synthesis neither dissociates from the enzyme nor readily reorients before hydrolysis, and P i primarily originates from the ␥-phosphoryl group of ATP (11).
Irreversibility of the Reaction-Attempts were made to measure the reverse reaction by conversion of [carboxy-14 C]AdoMet or 33 P i to radiolabeled methionine or ATP, respectively (see "Experimental Procedures"). In neither case was any reaction observed, even using levels of enzyme and reaction times where 0.1% conversion could be measured; less than 0.1 eq of ATP, PPP i , or AdoMet, per enzyme subunit, was formed in a 2500fold longer time than that required for a single turnover in the forward direction. Thus, reversal of both the AdoMet forming and the PPP i hydrolytic steps appears to be unfavorable. This result is concordant with the established kinetic as well as thermodynamic irreversibility of AdoMet synthesis in reactions catalyzed by the ␣-type enzyme (6,12).
Cation Activation-No reaction was detected in the absence of Mg 2ϩ , consistent with the divalent cation requirement of other MATs. K ϩ both enhanced the k cat by 5-fold and decreased the K m values for both substrates, with half-maximal effect at 5 mM (Fig. 2). Saturating KCl decreased the K m for ATP from 1.4 to 0.22 mM and the K m for methionine from 1.1 to 0.3 mM. Plots of 1/v versus 1/K ϩ at different substrate concentrations (and the converse) did not intersect on the 1/v axis, showing that the cation and substrate binding are not related by an equilibrium ordered binding process. The data indicate that K ϩ is a stimulator rather than an essential activator, in common with the ␣-type MATs.
Kinetic Mechanism-Substrate saturation kinetic studies of the AdoMet synthetic reaction show that MjMAT catalyzes a sequential reaction in which both ATP and methionine bind before products are released (Fig. 3, A and B). The K m values do not vary substantially with the concentrations of co-substrate, indicating little synergism in binding affinity.
An ATP analog that contains a C5Ј-S-P linkage was not a substrate for AdoMet formation, but was a competitive inhibitor with respect to ATP and noncompetitive with respect to methionine (Fig. 3, C and D), with a K i comparable with the K m for ATP (Table I). The alternate substrate GTP was also a competitive inhibitor with respect to ATP and noncompetitive with methionine (AdoMet formation was measured with [8-14 C]ATP in these inhibition experiments). A nonreactive analog of methionine that has high affinity was not identified. Cycloleucine, 1-amino-1-carboxy-cyclopentane, is a commonly used dead-end inhibitor of the ␣-type enzymes (28). However, 20 mM cycloleucine gave no detectable inhibition of MjMAT, even with both the substrates present at their K m values. Furthermore, MjMAT was not significantly inhibited (Ͻ10%) by 10 mM concentrations of L-homocysteine, L-norleucine, or L-cis-2-amino-4-methoxy-3-butenoic acid (the most potent methionine analog inhibitor of MAT-␣ (Ref. 18)). The alternate substrates L-ethionine and L-methionine methyl ester were competitive inhibitors with respect to methionine for AdoMet formation from [methyl-14 C]methionine, and noncompetitive inhibitors with respect to ATP (cf. Fig. 3, E and F). The observed pattern of inhibition by nonreactive compounds and alternate substrates indicates that substrate binding is random. Both ethionine and methionine methyl ester have much smaller K m values than their K i values for inhibition of the reaction with methionine (6-and 70-fold, respectively), sug-gesting the presence of substantial kinetic contributions to the K m . The noncompetitive inhibition toward methionine by nonreactive ATP analogs reflects formation of dead-end enzymesubstrate-inhibitor complexes (21,22).
Product inhibition studies showed that AdoMet is a noncompetitive inhibitor with respect to both methionine and ATP with K i values near 2 mM (Fig. 3, G and H). This high K i contrasts with that of most ␣-type MATs for which AdoMet is a potent inhibitor with K i values in the ϳ10 Ϫ5 M range. Pyrophosphate is a competitive inhibitor with respect to ATP (K i ϭ 0.83 mM) and noncompetitive with respect to methionine (Fig.  3, I and J). Phosphate is a competitive inhibitor toward both ATP and methionine, with K i values comparable with the K i for PP i (Fig. 3, K and L). The product inhibition results indicate that product release is ordered with AdoMet dissociating before PP i and P i , which subsequently dissociate randomly. Apparently, P i binding prevents access of both ATP and methionine to their binding sites. This kinetic mechanism is illustrated in Scheme 1. The finding that none of the inhibition patterns are uncompetitive, and that at least one product is a competitive inhibitor with each substrate, in conjunction with the dead-end and alternative substrate inhibition data, shows that the only compatible kinetic scheme for the forward reaction has random addition of ATP and methionine and partially random product release (22). The noncompetitive inhibition by AdoMet toward both substrates reflects formation of dead-end enzyme-substrate-AdoMet complexes with both ATP and methionine, whereas the noncompetitive inhibition by PP i with respect to methionine reflects formation of a dead-end enzyme-methionine-PP i complex. Because the kinetics of the reverse reaction could not be studied, preferential binding order in that direction could not be further evaluated.
The nonhydrolyzable PPP i analog diimidotriphosphate (O 3 P-NH-PO 2 -NH-PO 3 ) is a potent inhibitor of the ␣-type EcMAT with a K i of 2 nM (17). Approximately 50% inhibition of MjMAT was observed at 2 M PNPNP when enzyme, ATP, and methionine were present at 5 M, 0.5 mM, and 2.4 mM, respectively (at 55°C). The same result was obtained when the reaction was initiated by addition of enzyme or by addition of substrates to a mixture of enzyme and PNPNP. Clearly PNPNP has high affinity with a K i of less than 2 M. Because of the complexities of quantitative analysis of tight binding inhibition (29), the details of the inhibition have not yet been pursued.
The AdoMet metabolites S-adenosylhomocysteine (20 mM) and 5Ј-methylthioadenosine (5 mM) produced less than 10% inhibition when ATP and methionine were present at 0.2 and 0.5 mM, respectively, suggesting that these compounds are unlikely to be significant physiological regulators of enzyme activity. The AdoMet analog sinefungin was also a poor inhibitor, with 12 mM sinefungin giving less than 25% inhibition under these conditions. Substrate Specificity-Because AdoMet is the substrate of a large number of enzymes, the ability to prepare a variety of AdoMet analogs would be a valuable tool. EcMAT has been extensively used to prepare isotopically labeled AdoMet and some analogs despite limitations from severe product inhibition and significant substrate selectivity (14,30,31). The potential of MjMAT as a tool for synthesis of AdoMet analogs was assessed. Tables II and III compare the results for the use of ATP and methionine analogs by MjMAT and the E. coli enzyme. The tolerance of the ATP site of MjMAT is substantially greater than the bacterial enzyme, and MjMAT readily accepts modifications in both in the adenine and ribose moieties. For example, GTP, UTP, and CTP are substrates for MjMAT but not for EcMAT. Both 2Ј-deoxy-and 3Ј-deoxy-ATP are substrates for MjMAT, whereas only the latter is accepted by EcMAT.
Neither MjMAT nor EcMAT catalyzes AdoMet formation from analogs in which the scissile C-O bond is replaced by a C-S SCHEME 1. Steady state kinetic mechanism for MjMAT. The inferred intermediate E-AdoMet-PPP i complex is not shown. Dead-end ternary complexes formed in the presence of products or nonreactive ATP analog inhibitors (xTP) are also shown.
NC, K is ϭ 0.52, K ii ϭ 2.6 P i C, K i ϭ 0.65 C, K i ϭ 0.38 PPP i C, K i ϭ 0.12 NC, K is ϭ 0.14, K ii ϭ 0.76 A(S)TP C, K i ϭ 0.32 NC, K is ϭ 0.36, K ii ϭ 1.4 GTP C, K i ϭ 1.5 NC, K is ϭ 1.9, K ii ϭ 4.6 L-Ethionine NC, K is ϭ 31, K ii ϭ 140 C, K i ϭ 50 L-Methioninemethyl ester NC, K is ϭ 46, K ii ϭ 25 C, K i ϭ 16 a C, competitive; NC, noncompetitive. b K is , K i value from slope; K ii , K i value from intercept (for noncompetitive inhibitors). Reactions were conducted in 25 mM Hepes⅐ (CH 3 ) 4 N ϩ at pH 8.0 with 50 mM KCl and 10 mM MgCl 2 at 55°C. When methionine was varied the ATP concentration was fixed at 0.5 mM, and when ATP was varied the methionine concentration was fixed at 0.4 mM.
c When the methionine concentration was fixed at 5 mM, the inhibition was noncompetitive with K is ϭ 2.9 mM, K ii ϭ 6.5 mM.
d When the ATP concentration was fixed at 5 mM, the inhibition was noncompetitive with K is ϭ 5.2 mM, K ii ϭ 8.1 mM. or C-N bond. The C5Ј-S and C5Ј-N containing compounds are good competitive inhibitors with respect to ATP, with K i values near the K m for ATP for both enzymes.
The methionine site is rather restrictive in both MjMAT and EcMAT, but it is quantitatively different for the two enzymes (Table III). For both enzymes, the ethyl analog of L-methionine (i.e. L-ethionine) is a substrate, as is L-methionine methyl ester, but which is the better substrate differs for MjMAT and Ec-MAT. D-Methionine is also a substrate for both MjMAT and EcMAT. No activity was seen with either enzyme with the aldehyde methional, the alcohol methioninol, or 3-methylthiopropylamine, all of which lack one of the polar attachments present in methionine.
Temperature Dependence of Kinetic Parameters- Fig. 4 illustrates an Arrhenius plot of the temperature dependence of k cat for MjMAT and EcMAT. The temperature dependence of k cat for MjMAT is nonlinear. Because the enzyme was stable at elevated temperatures and care was taken to measure initial reaction rates and verify substrate saturation, the change in slope of the Arrhenius plot apparently reflects a negative heat capacity (⌬C p ) of activation (24). Analysis of the temperature dependence of k cat using ⌬H 310 # , ⌬S 310 # , and ⌬C p310 # as parameters yielded values of 23 kcal/mol, ϩ11 entropy units, and Ϫ0.3 kcal/mol-K, respectively, and the fit shown in Fig. 4. The negative heat capacity change is consistent with a protein conformational alteration during catalysis that results in a net decrease in exposure of nonpolar surface area (32). Interestingly, the MjMAT and EcMAT have comparable k cat values at approximately 37 and 70°C, consistent with the optimal growth temperature for M. jannaschii of 87°C. The ⌬H 310 # for EcMAT is 16 2% of the V max for methionine when present at 10 mM. b Ͻ5% inhibition at 10 mM. kcal/mol, lower than that for MjMAT. The activation entropy for the EcMAT-catalyzed reaction, ⌬S 310 # , is Ϫ7 entropy units. The K m values vary less than 3-fold over the ranges studied for both enzymes, increasing with temperature in both cases. DISCUSSION Despite the widely different sequences of the ␣and ␥-types of MAT, substantial similarities are present. Both classes of MAT require Mg 2ϩ for activity and are activated by K ϩ , although the monovalent cation is not essential. The enzymes hydrolyze the PPP i , initially formed from ATP, to yield PP i and P i ; the P i largely originates as the ␥-phosphoryl group of the nucleotide, indicating that motion of PPP i is restricted within the active site. Consistent with the tightly bound nature of the PPP i intermediate, the nonhydrolyzable analog diimidotriphosphate is a potent inhibitor of both MAT classes.
The steady state kinetic mechanism of MjMAT shows random substrate binding and ordered product release, AdoMet dissociating before PP i or P i . Random substrate binding was previously found for MAT from E. coli, whereas some eucaryal MAT have ordered binding wherein ATP associates before methionine (8). The order of product release in the MjMAT reaction is AdoMet before PP i and P i ; with other MATs, both the same and different orders of product release have been observed. A remarkable property of MjMAT is the low affinity for AdoMet, with millimolar K i values, which contrasts with the ␣-type MATs for which AdoMet typically inhibits with a K i in the physiologically significant 10 Ϫ5 M range.
The substantial substrate activity of common naturally occurring nucleotides other than ATP, such as 2Ј-deoxy-ATP, GTP, CTP, and UTP, is surprising. Although the higher K m values compared with ATP and the probable lower intracellular concentrations of the other triphosphates may render the in vivo synthesis of large amounts of the corresponding sulfonium ions unlikely, whether physiologically significant quantities are made in vivo is as yet unclear. Thus, it is possible that there is an additional in vivo role of this enzyme in producing hitherto unknown sulfonium metabolite. Such an as yet unidentified enzymatic role might rationalize why the primitive bacterium Aquifex aeolicus, which has one of the smallest genomes known for a free living organism, harbors genes for both classes of MAT (11).
The comparable activity of the M. jannaschii and E. coli enzymes near their physiological temperatures reflects the "principle of corresponding temperature" (33) and suggests a similar extent of utilization of AdoMet in these organisms. The thermal lability of AdoMet, both in chirality at the sulfonium center and in covalent bonding (34), requires a substantial metabolic flux through the synthetic reaction to maintain a pool of substrate for further metabolic requirements. The larger apparent activation energy for the MjMAT in the 0 -37°C range reflects an increase in the apparent enthalpy of activation, whereas there is a more favorable entropy of activation. These observations are consistent with the notion that enzymes from organisms that inhabit warmer environments are "stiffer" than those from cooler habitats (23).
The tolerance of MjMAT for chemical alterations of both the methionine and ATP moieties, the low affinity for the product AdoMet, in conjunction with the stability of the protein, suggests that this enzyme may be a useful synthetic tool for preparation of AdoMet analogs. In view of the myriad of roles of AdoMet in metabolism, and the paucity of available analogs, the feasibility of these syntheses are being explored.