Methyl methanethiosulfonate as an active site probe of serine hydroxymethyltransferase.

Using methyl methanethiosulfonate and other sulfhydryl group modification reagents we have studied the structure and function of sulfhydryl groups in rabbit liver cytosolic serine hydroxymethyltransferase. From a tryptic digest of the enzyme, seven cysteine-containing peptides were isolated and sequenced. These peptides contained a total of 8 cysteine residues. There are no disulfide bonds in this enzyme. Of the eight sulfhydryl groups, four react with methyl methanethiosulfonate. Two sulfhydryl groups react rapidly with this reagent without altering enzyme catalytic activity. The remaining two sulfhydryl groups react more slowly and cause loss of greater than 90% of the catalytic activity of the enzyme. This nearly inactive enzyme contains pyridoxal-P and can form an enzyme-substrate complex. However, the complex dissociates from the active site suggesting that one possible role for a sulfhydryl group is to stabilize the enzyme-substrate complex. The sequence of the cysteine-containing peptide which is responsible for the mechanism-based inactivation of serine hydroxymethyltransferase by D-3-fluoroalanine was determined. This sulfhydryl group was shown not to be essential to the enzyme for catalytic activity. Also, the sequence of one of the cysteine peptides shows considerable homology to the active site cysteine peptide from tryptophan synthase.

Using methyl methanethiosulfonate and other sulfhydryl group modification reagents we have studied the structure and function of sulfhydryl groups in rabbit liver cytosolic serine hydroxymethyltransferase. From a tryptic digest of the enzyme, seven cysteine-containing peptides were isolated and sequenced. These peptides contained a total of 8 cysteine residues. There are no disulfide bonds in this enzyme. Of the eight sulfhydryl groups, four react with methyl methanethiosulfonate. T w o sulfhydryl groups react rapidly with this reagent without altering enzyme catalytic activity. The remaining two sulfhydryl groups react more slowly and cause loss of greater than 90% of the catalytic activity of the enzyme. This nearly inactive enzyme contains pyridoxal-P and can form an enzyme-substrate complex. However, the complex dissociates from the active site suggesting that one possible role for a sulfhydryl group is to stabilize the enzyme-substrate complex.
The sequence of the cysteine-containing peptide which is responsible for the mechanism-based inactivation of serine hydroxymethyltransferase by D-3-fluoroalanine was determined. This sulfhydryl group was shown not to be essential to the enzyme for catalytic activity. Also, the sequence of one of the cysteine peptides shows considerable homology to the active site cysteine peptide from tryptophan synthase.
Serine hydroxymethyltransferase (EC 2.1.2.1) catalyzes the conversion of serine to glycine and 5,lO-methylenetetrahydrofolate (1). This reaction is usually considered to be one of the primary sources of one-carbon groups required for a variety of biosynthetic pathways in the cell (1). Both cytosolic and mitochondrial forms of the enzyme have been purified to homogeneity from rabbit liver (2). The role of pyridoxal-P and tetrahydrofolate in the mechanism of the cytosolic enzyme has been extensively studied (3). However, little is known about what amino acid residues form the active site pocket and what role these residues play in the mechanism of this enzyme. We have previously published evidence that there is at least one sulfhydryl group at the active site The work reported here extends this previously published data on the number and role of cysteine residues at the active site of the rabbit liver cytosolic enzyme.
We fist observed that the reagent DTNB' reacted rapidly * This work was supported by Grant PCM-8110363 from the National Science Foundation and Grant GM28143 from the National Institutes of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
' The abbreviations used are: DTNB, 5,5'-dithiobis(nitrobenzoic with two sulfhydryl groups on the enzyme (5). This reaction leads to the release of the pyridoxal-P and complete loss of catalytic activity. We later showed that removal of pyridoxal-P from the active site uncovers a single sulfhydryl group which reacts rapidly with iodoacetate. A chymotryptic dodecapeptide containing the carboxymethylcysteine residue was isolated and sequenced (4). One problem with these studies was that both the DTNB and iodoacetate-inactivated enzymes do not bind pyridoxal-P. This makes it impossible to probe the possible function of the sulfhydryl group in the binding of the amino acid substrates or the mechanism of the enzyme. We were also unable to determine if the sulfhydryl group which reacted with DTNB in the holoenzyme was the same one which reacted with iodoacetate in the apoenzyme.
Our interest in the role of sulfhydryl groups was heightened by the report that D-3-fluoroalanine is a mechanism-based inactivator of this enzyme ( 6 ) . Wang et al. (6) have shown that the inactivation is due to a sulfhydryl group reacting with an enzyme-bound aminoacrylate intermediate. This suggests that a sulfhydryl group is located at the subsite for binding the amino acid substrate and implicates a mechanistic role for this group. Our continued interest in the chemistry of sulfhydryl groups on serine hydroxymethyltransferase is also due to the success of other investigators in being able to attach reported molecules such as 13C, "F, and spin labels, on sulfhydryl groups at or near the active site of enzymes. These reporter groups have served as useful tools for determining the pK values of adjacent groups at the active site (7-11).
The work reported in this paper determines the number and structure of all the cysteine-containing tryptic peptides of cytosolic serine hydroxymethyltransferase. The primary sulfhydryl-modifying reagent we have used is methyl methanethiosulfonate. This reagent reacts rapidly with sulfhydryl groups to form methyl disulfides which introduces a relatively small noncharged group a t each exposed cysteine residue (11-13). With the use of this and other reagents, we have demonstrated that there are at least two sulfhydryl groups at the active site but only one may be critical for catalytic activity. This critical sulfhydryl group is not the one involved in the D-3-fluoroalanine inactivation. Our understanding of the chemical reactivity of the sulfhydryl groups on this enzyme with methyl methanethiosulfonate will permit additional studies on the active site of this enzyme.
Cytosolic serine hydroxymethyltransferase was purified and crystallized by the method of Schirch and Peterson (2). Aposerine hydroxymethyltransferase was prepared by transamination of the pyridoxal-P with D-alanine to form pyridoxamine-P and pyruvate (15). The pyridoxamine-P was removed from the apoenzyme by gel fiitration.
Two methods were used to measure serine hydroxymethyltransferase activity in this study. The principal assay utilizes allothreonine as substrate. The product acetaldehyde was continuously measured by observing the decrease in absorbance at 340 nm upon reduction with NADH and alcohol dehydrogenase (16). This assay does not require H,folate as a co-substrate. When we wanted to monitor the effect of sulfhydryl reagents on Hlfolate function we utilized Lserine and tetrahydrofolate as substrates.
The production of the product 5,lO-methylenetetrahydrofolate was continuously monitored at 340 nm upon oxidation to 5,lO-methenyltetrahydrofolate with NADP' and formylmethenylmethylenetetrahydrofolate synthetase (14).
Methyl methanethiosulfonate was dissolved in absolute ethanol before each experiment. The concentration of this stock solution was determined by observing the decrease in absorbance at 412 nm upon the addition of an excess of p-nitro-n-carboxythiophenol. Stock solutions were normally made in the range of 50 mM.
The number of sulfhydryl groups on serine hydroxymethyltransferase was determined by incubating the enzyme with 0.02% sodium dodecyl sulfate and 0.5 mM DTNB at 90 "C for 15 min. The concentration of sulfhydryl groups was determined by dividing the absorbance at 412 nm by 13,600 (17). The subunit concentration of serine hydroxymethyltransferase was determined by dividing the absorbance at 280 nm by the subunit molecular weight of 54,000 and 0.72 which is the absorbance at 280 nm of a 1 mg/ml solution of holoenzyme. When apoenzyme was used, a value 0.62 was used for the 280 nm absorbance of a 1 mg/ml solution. The buffer used for all studies involving the measurement of sulfhydryl groups was 50 nm sodium N,N-bis[2-hydroxyethyl]-2-aminoethanesulfonate, pH 7.0.
The presence of disulfides in the protein was determined by the procedure described in the previous paragraph except that the enzyme was incubated at 90 "C for 10 min in the presence of 0.02% sodium dodecyl sulfate and 50 n u dithiothreitol. One hundred pl of this enzyme were then passed through a Sephadex G-75 column in a Pasteur pipette to remove the excess dithiothreitol.
The isolation of tryptic peptides containing cysteine residues was accomplished by the following procedure. Solutions containing 5 to 10 mg of enzyme were made 5.5 M in guanidine hydrochloride and adjusted to pH 8.0. Twenty pCi of [ l-'4C]iodoacetate were added and the solution incubated at 37 "C for 30 min. After 30 min, nonradioactive iodoacetate was added to achieve a final concentration of 50 mM. This solution was incubated for an additional 60 min and then dialyzed against several changes of water. The precipitated protein was collected by centrifugation and resuspended in 0.1 M ammonium bicarbonate, pH 8.0. A freshly prepared solution of trypsin was added to a final enzyme to substrate ratio of 1:100 by weight. After incubation at 37 "C for 3 h, a second addition of trypsin was added for an additional 3 h. The digested protein was dried under vacuum and redissolved in 3 ml of pH 2.85 phosphoric acid solution. For identification of the number of carboxymethylated peptides, an aliquot containing 200 pg of peptides was chromatographed on a (2-18 reverse phase HPLC column and eluted with a phosphoric acid-acetonitrile gradient as previously described (18). The eluent from the HPLC column was monitored at 215 nm and collected in 1.3-ml fractions. Aliquots of 100 p1 of each fraction were counted in a Tri-Carb liquid scintillation counter using Bray's solution as the solvent (19).
For purification of the carboxymethylated peptides, the tryptic digest was placed on a phosphocellulose column (2 X 18 cm) equilibrated with a pH 2.85 phosphoric acid solution. The peptides were eluted with a linear gradient consisting of 400 ml of the equilibrating buffer in the mixing chamber and 400 ml of 0.4 M KCl, which was adjusted to pH 2.85 with phosphoric acid, in the reservoir. Fractions (4 n l ) were collected and analyzed for radioactivity. Five radioactive peaks were found. The contents of the tubes containing radioactivity were pooled and evaporated to dryness. The partially purified peptides were dissolved in 0.1% trifluoroacetic acid and subjected to chromatography on a C-18 reverse phase HPLC column (25 cm X 4 mm). Peptides were eluted with a linear gradient between buffer A (0.1% trifluoroacetic acid) and buffer B (0.1% trifluoroacetic acid in 70% acetonitrile). The gradient was from 0 to 100% in 90 min. The flow rate was 1.0 ml/min. Fractions from the HPLC column were collected and analyzed for radioactivity. Five of the seven cysteine peptides were pure after this column. The remaining two peptides were rechromatographed on the same HPLC system and by taking smalter fractions these peptides were also obtained free of contaminants.
Aliquots of the purified peptides containing carboxymethylcysteine residues (0.5-1.0 nmol) were hydrolyzed in sealed, evacuated tubes with constant boiling HCl for 24 h at 110 "C. Amino acid analyses were then determined using a Durmm MBF amino acid analyzer. The sequences of the peptides were determined by the manual Edman degradation method of Tarr (20). Usually 2 nmol of peptide were used. The released phenylthiohydantoin amino acids were identified by HPLC on a Beckman phenylthiohydantoin amino acid column or by amino acid analysis after hydrolysis with HI. The presence of phenylthiohydantoin Cys(Cm) was verified by counting an aliquot of those fractions which gave alanine after HI hydrolysis.
Inactivation of serine hydroxymethyltransferase by D-3-fluoroalanine was performed as follows. To several milligrams of enzyme in 0.5 ml of sodium 2-[bis(2-hydroyethyl)amino]ethanesulfonic acid buffer, pH 7.0, were added 5 mg of D-3-fluorodanine. A 10-pl aliquot was removed, and catalytic activity determined in the allothreonine assay containing 0.1 nm pyridoxal-P. The enzyme-~-3-fluoroalanine solution showed no loss of activity during a 10-min incubation at 30 "C. After 10 min, 20 pl of a 6 mM solution of tetrahydrofolate were added, and the enzymic activity determined at several intervals. The enzyme lost all activity during the next 20 min. The H4folate stock solution was prepared by dissolving 25 mg of tetrahydrofolate in 2.5 ml of 200 nm potassium phosphate, 50 mM ascorbate, pH 7.5. This solution showed the presence of some thiol compound as determined by reaction with DTNB. To remove this thiol from the stock tetrahydrofolate solution, DTNB was added to a final concentration of 1 mM.

Inhibition by Methyl
Methanethiosulfonate-When a 10fold excess of methyl methanethiosulfonate is added to serine hydroxymethyltransferase the enzyme loses activity over a 1h period as shown in Fig. 1. Increasing concentrations of the reagent increase the rate at which activity is lost although the loss of activity is never described by a simple fwst order reaction. The rate of inhibition by methyl methanethiosulfonate is decreased in the presence of the amino acid substrates serine and glycine (Fig. 1). However, removal of pyridoxal-P increases the rate of inactivation (Fig. 1). The assay used to measure enzyme activity in this study was allothreonine which does not require &folate as a co-substrate. When the study was repeated with L-serine and &folate as substrates, essentially identical curves with those recorded in Fig.  1 were obtained. The data in Fig. 1 suggest that critical sulfhydryl groups at the amino acid and pyridoxal-P binding sites are reacting with methyl methanethiosulfonate. There do not appear to be critical sulfhydryl groups at the Hlfolate binding site which react uniquely with methyl methanethiosulfonate.
Methyl methanethiosulfonate not only inactivates serine hydroxymethyltransferase but it also produces an enzyme which gives nonlinear kinetics with time (Fig. 2). Inactivation studies were performed by removing lo-$ aliquots of an enzyme-methyl methanethiosulfonate solution and adding it to an allothreonine assay solution. The formation of the product acetaldehyde was monitored by measuring the decrease in absorbance at 340 nm due to the alcohol dehydrogenase-catalyzed oxidation of NADH. Normally this assay is linear with time until the NADH is depleted. As shown in Fig.  2, for aliquots of enzyme from a methyl methanethiosulfonate inactivation study, the rate of product formation decreases with time. This nonlinearity appears to be the result of the loss of pyridoxal-P from the enzyme during the assay. Addition of 0.1 mM pyridoxal-P to the assay restores the reaction to near linearity with time. The pyridoxal-P in these partially The reaction conditions were the same as recorded for the holoenzyme in Fig. 1. The decrease in absorbance at 340 nm is due to the alcohol dehydrogenase-catalyzed reduction of the product acetaldehyde by NADH. The numbers on the curues are the minutes the enzyme had been preincubated with methyl methanethiosulfonate. inactivated enzyme studies is lost from the active site only in the presence of the amino acid substrates. This suggests that modifying sulfhydryl groups with methyl methanethiosulfonate results in a n enzyme in which the intermediates involving pyridoxal-P-substrate complexes are dissociating from the active site.

Number of Groups Modified by Methyl Methanethiosulfo-
nate-In this experiment less than stoichiometric amounts of methyl methanethiosulfonate were added to 1 ml of a 9 mg/ ml solution of enzyme. After incubation for 15 min at 23 "C a 100-pl aliquot of the reaction mixture was passed through a small Sephadex G-75 column to remove any unreacted modifying reagent. The enzyme was analyzed for both catalytic activity and the total number of free sulfhydryl groups. The procedure was repeated by the addition of more methyl methanethiosulfonate to the stock enzyme solution until the enzyme had lost all of its activity. With no modifying reagent, the enzyme was found to have 7.9 to 8.0 sulfhydryl groups/ subunit. Fig. 3 shows that as one titrates the enzyme with methyl methanethiosulfonate, a total of four sulfhydryl groups are blocked in the holoenzyme. The data in Fig. 3 also show that the fiist two sulfhydryl groups that are blocked do not cause a loss of activity in the allothreonine assay. However, modifying the third and fourth sulfhydryl groups leads to a loss of 94% of the catalytic activity. The addition of more methyl methanethiosulfonate did not lead to any additional loss of free sulfhydryl groups.
This study was repeated using serine and H4folate as substrates to measure catalytic activity. As shown in Fig. 3, the same results were found as when allothreonine was used as substrate. This further c o n f i s that the four sulfhydryl groups which are blocked by methyl methanesulfonate probably do not play an important role in &folate binding or function.
Titration of sulfhydryl groups by methyl methanethiosulfonate was also performed on aposerine hydroxymethyltransferase. In the absence of pyridoxal-P, a total of five sulfhydryl groups are blocked with a complete loss of catalytic activity. The catalytically important sulfhydryl groups appear to react rapidly with methyl methanethiosulfonate since even the lowest concentration of reagent results in a loss of activity. This is consistent with the kinetics of inhibition for the apoenzyme recorded in Fig. 1.
When the modified and inactive enzyme is incubated with  Holoenzyme was titrated with methyl methanethiosulfonate as described in Fig. 3. In this experiment, the absorption properties of the active site pyridoxal-P was determined as a function of the number of sulfhydryl groups modified (indicated by numbers on each spectrum). The inset shows the percentage of change in absorbance at 430 nm between the enzyme with no blocked sulfhydryl groups and the enzyme with 4.0 blocked sulfhydryl groups.
10 m~ dithioerythritol for 30 min at 23 "C, 95% of the catalytic activity is recovered, demonstrating that modification by methyl methanethiosulfonate can be reversed.
These experiments suggest that in the holoenzyme two sulfhydryl groups are exposed but that neither of these is critical for catalytic activity. The data also suggest that at least one catalytically important sulfhydryl group is partially protected from modification by pyridoxal-P. The data do not rule out that as many as three sulfhydryl groups are protected by pyridoxal-P and substrates.
Properties of the Bound Pyridoxal-P during Titration with Methyl methanethiosulfonate-In a n experiment, similar to the one recorded in Fig. 3, we monitored the spectral properties of the active site pyridoxal-P during titration of the sulfhydryl groups with methyl methanethiosulfonate. These data are recorded in Fig. 4 and show that blocking the third and fourth sulfhydryl group causes a spectral shift in the bound pyridoxal-P from a 428-nm absorption maximum to one at 415 nm. The inset demonstrates that the spectral shift follows the same pattern as loss of catalytic activity as recorded in Fig. 3. The enzyme with sulfhydryl groups blocked can be dialyzed extensively without additional changes in the spectrum or loss of pyridoxal-P. However, if an amino acid substrate is added to the enzyme prior to dialysis the enzyme loses its bound pyridoxal-P. This further confirms the data reported in Fig. 2 that the nonlinear kinetics are due to the dissociation of the pyridoxal-P-amino acid substrate complex from the active site.

Properties of the Enzyme with Two Blocked Sulfhydryl
Groups-Holoserine hydroxymethyltransferase was titrated with methyl methanethiosulfonate until 2.2 to 2.4 sulfhydryl groups were blocked as methyl disulfides. The enzyme retained 80 to 90% of its catalytic activity. We define this as the enzyme with two blocked sulfhydryl groups.
Unlike the unblocked native enzyme, the enzyme with two blocked sulfhydryl groups is not inhibited by DTNB (Fig. 5) and does not show any rapidly reacting sulfhydryl groups with this reagent. This suggests that the two sulfhydryl groups which react rapidly with DTNB in the native enzyme and result in loss of activity are now blocked as methyl disulfides. Since the enzyme is still active, these sulfhydryl groups do not appear to be critical for catalytic activity. We repeated this experiment with the enzyme containing two blocked sulfhydryl groups and found that the inactivation by this substrate analog is greatly reduced (Fig. 5 ) . Since there were trace amounts of a sulfhydryl-containing compound in the Hlfolate solution, the slow inactivation could be due to some unblocking of the methyl disulfides in the enzyme. However, this experiment suggests that the sulfhydryl group which attacks the D-3-fluoroalanine intermediate is blocked as a methyl disulfide. This would place at least one of the two blocked sulfhydryl groups near the active site.
Native rabbit liver enzyme forms complexes with glycine which absorb at 343, 425, and 495 nm. The structure and sequence of formation of these complexes have been previously described. The interconversion of the complexes absorbing at 343 and 425 nm is perturbed by a group on the enzyme with a pK of 7.0 (21). This pK could possibly be a reflection of the sulfhydryl group which reacts with D-3-fluoroalanine. We repeated these pH spectral titration studies of the enzymeglycine complexes with the enzyme containing two blocked sulfhydryl groups. The spectral changes were the same as the native enzyme and suggest that the sulfhydryl group near the active site is not the group with a pK of 7.0.

Total Number and Structure of Sulfhydryl Groups in Serine
Hydroxymethyltranserase-Denaturation of the enzyme with sodium dodecyl sulfate and reaction of the sulfhydryl groups with DTNB shows that there are eight reactive groups/subunit. If the enzyme is denatured in sodium dodecyl sulfate in the presence of dithiothreitol for 30 min, we still f i d only eight DTNB reactive sulfhydryl groups after rapid removal by gel fitration of the excess dithiothreitol. This shows that serine hydroxymethyltransferase has no disulfide bonds. We next determined the number and structure of the tryptic peptides containing cysteine residues. This was done by labeling the enzyme in 5.5 M guanidine HC1 with [l-'4C]iodoacetate. After removal of the guanidine HC1 the carboxymethylated enzyme was digested with trypsin. About 200 pg of the trypsin digest were eluted from an HPLC C-18 reverse phase column with a phosphoric acid-acetonitrile gradient. The peptide eIution profie monitored by absorbance a t 215 nm is shown in Fig. 6B. Fractions, containing 1 ml of eluate, were collected and analyzed for radioactivity. Seven radioactive peaks containing carboxymethylcysteine were found as shown in Fig. 6A. FIG. 6. Elution profile of tryptic peptides of cytosol serine hydroxymethyltransferase from a C-18 reverse phase column on high performance liquid chromatography. A, profde of radioactivity of enzyme which had been reacted with [l-'4-C]iodoacetic acid before digestion with trypsin. B, profde of peptides monitored at 215 nm of tryptic digest.  In order to purify the cysteine-containing peptides, the trypsin hydrolysate was first fractionated on a phosphocellulose column. Five fractions containing radioactivity from this column were placed on an HPLC C-18 reverse phase column for isolation. The purified peptides corresponding to the radioactive peaks in Fig. 6A were sequenced by the Edman degradation procedure (20). The sequences of the peptides are recorded in Table I. Peptide 4 contains 2 cysteine residues which accounts for all eight sulfhydryl groups observed on titration with DTNB. Peptide 5 is the active site peptide previously sequenced by us. The cysteine in this peptide is the one which reacts with iodoacetate in the apoenzyme (4). This purified active site tryptic peptide does not contain Lys or Arg. This is probably a chymotryptic peptide which was formed due to the presence of a Lys or Arg-Pro bond near the COOH-terminal end of this peptide.
We next determined which peptide contained the cysteine which reacts with D-3-fluoroalanine. This was achieved by denaturing the enzyme which had been inactivated with fluoroalanine and reacting all of the exposed sulfhydryl groups with [l-'4C]iodoacetate. The HPLC elution profde of the tryptic digest of this enzyme was missing peptide 3.
We also attempted to determine which 4 of the cysteine residues are not blocked by methyl methanethiosulfonate in the holoenzyme. This was done by denaturing the enzyme with four blocked sulfhydryl groups in guanidine HC1 and reacting it with [l-'4C]iodoacetate. Tryptic peptides were isolated as before and the elution profile on the HPLC System determined. Only peptides 1 and 7 (Table I) had greatly reduced radioactivity. The fact that only two peptides are missing in the enzyme with four blocked sulfhydryl groups suggests that during the denaturation of the enzyme with guanidine HC1, some disulfide interchange took place.
Inhibition of Serine Hydroxymethyltransferase by Periodate-Stoichiometric concentrations of periodate have been shown to be a mild procedure for oxidizing vicinal sulfhydryl groups in enzymes (22). We tested the effect of a 4-fold excess of sodium periodate on the apoenzyme and holoenzyme. In each experiment the enzyme lost all activity in less than 1 min of incubation. Denaturation of the oxidized enzyme in sodium dodecyl sulfate and reaction with DTNB showed the presence of only four sulfhydryl groups. This suggests that periodate has formed two disulfide bonds. Addition of substrates or phosphate did not protect the enzyme from reaction with periodate. Addition of methyl methanethiosulfonate or DTNB to the periodate-oxidized enzyme did not result in a further decrease in the number of DTNB-reactive sulfhydryl groups. This suggests that the four sulfhydryl groups which react with periodate are the same four which react with methyl methanethiosulfonate.
The periodate-oxidized apoenzyme can be reactivated to about 90% activity on incubation for 30 min at 25 "C with 10 mM dithiothreitol. The spectrum of the periodate-oxidized holoenzyme shows an absorption maximum at 420 nm and is essentially identical with the spectrum of the enzyme inhibited with methyl methanethiosulfonate as recorded in Fig. 4.

DISCUSSION
Sulfhydryl groups have been implicated as being at or near the active site of the following pyridoxal-P enzyme. Glutamate decarboxylase (23), isoleucine decarboxylase (24), lysine decarboxylase (25), ornithine decarboxylase (26), cysteine sulfinate decarboxylase (27), tryptophan synthase (28), aspartate aminotransferase (29), L-alanine aminotransferase (30), and D-amino acid transaminase (31). Since all pyridoxal-P enzymes probably have similar intermediates it seems likely that a sulfhydryl group at the active site may have a similar function in these enzymes. However, the function of a sulfhydryl group in the mechanism of any pyridoxal-P enzyme has yet to be elucidated. One approach to determining if sulfhydryl groups are playing similar roles in pyridoxal-P enzymes is to isolate cysteine-containing peptides from the various enzymes and look for similarities in amino acid sequence. This approach has been done with the active site pyridoxal-P peptide from a wide variety of B6-enzymes. As we pointed out earlier the pyridoxyl peptide from rabbit liver cytosolic serine hydroxymethyltransferase shows a great deal of homology to the sequences of all bacterial decarboxylases and tryptophan synthase (32). Unfortunately, the only enzyme of this group to have the active site cysteine peptide sequenced is tryptophan synthase. The amino acid sequence around the active site cysteine in this enzyme is . . . Ala-Leu-Thr-Lys-Cys-Gln-Asn . . . (33). In peptide 6 from serine hydroxymethyltransferase we find the sequence . . . Ala-Leu-Gly-Ser-Cys-Leu-Asn . . .
where four of the seven amino acids are in the same location. This sequence is also present in mitochondrial serine hydroxymethyltransferase.2 Although peptide 6 (Table I), is not one of the peptides known to be at the active site of the enzyme, we cannot eliminate it as being at that location. We feel that the methodology reported in this paper can be used on other pyridoxal-P enzymes in order to compare sequences around active site cysteine residues.
One of our current interests is to identify the sequences of peptides which define the active site of serine hydroxymethyltransferase. We had isolated and sequenced two peptides previously and in this paper we have added a third sequence. The inactivation studies with D-3-fluorodanine show that the cysteine in peptide 3 (Table I) lies near the a-carbon of the amino acid substrate-binding site. This sulfhydryl group reacts rapidly with methyl methanethiosulfonate and DTNB (Fig.  3) but only results in inactivation with DTNB modification. There is at least one more sulfhydryl group at the active site which reacts with methyl methanethiosulfonate. The reaction of this group results in an inactive enzyme and is protected by the addition of substrates (Fig. 1). This sulfhydryl group may be the same one we have shown previously to react with iodoacetate in the apoenzyme and is identified as peptide 5 (Table I) in the HPLC profile ( Fig. 6) (22). Even though this sulfhydryl group when blocked gives an inactive enzyme, we cannot assign it a functional role in the mechanism of the enzyme. Blocking this group appears to change the environment of the pyridoxal-P as evidenced by the spectral shift from 428 to 415 nm (Fig. 4). This change in environment may mean the pyridoxal-P is no longer in line with critical residues required for catalysis. The critical sulfhydryl group may play a role in stabilizing the enzyme-substrate complex, however, since when it is blocked the pyridoxal-P-substrate complex readily dissociates from the active site (Fig. 2).
There are several experiments which suggest that there may be even more than the two sulfhydryl groups in peptides 3 and 5 (Table I) at the active site. Methyl methanethiosulfonate reacts with four sulfhydryl groups in the holoenzyme but removal of pyridoxal-P uncovers another reactive sulfhydryl group. At this time we can only rule out cysteines in peptides 1 and 7 (Table I) as being buried and not accessible to methyl methanethiosulfonate.
Knowing the location and reactivity of sulfhydryl groups on an enzyme can provide useful tools for mechanistic studies. In this paper we use methyl methanethiosulfonate as a sulfhydryl group-modifying reagent to probe the structure and function of cysteine residues in cytosolic serine hydroxymethyltransferase. The advantage of this reagent is that it modifies sulfhydryl groups as methyl disulfides which should have a minimum steric and charge effect on the structure of the enzyme. The reagent also appears to be specific for sulfhydryl groups. It does react with lysine and histidine, but the reaction is slow and not reversed by thiol compounds (34). The rapid and reversible inactivation of serine hydroxymethyltransferase strongly suggests that in this enzyme we are looking only at sulfhydryl group modifications. In future studies, we hope to be able to use I3C-labeled methyl methanethiosulfonate as a method of introducing an NMR probe into specific locations in this enzyme.
In addition to furthering OUT understanding of the number and function of sulfhydryl s o u p s in this enzyme, this study has suggested two additional areas of research. The fist is the use of modifying reagents to probe for other active site residues. In the past, we have attempted to use such reagents as phenylglyoxal, ethoxyformic anhydride, and cyanate to test for arginyl, histidyl, and lysyl groups at the active site. In each case, we observed both inhibition and loss of at least one sulfhydryl group. We could not, however, correlate inhibition with only the loss of the sulfhydryl group. It became apparent that we could not probe for the other active site residues unless we could reversibly block the active site sulfhydryl groups during the modification reaction. The experiments in this paper show that both methyl methanethiosulfonate and periodate can serve this function.
A second area of future study is the effect of blocking the sulfhydryl groups as methyl disulfides on the stability of the enzyme. Reacting the enzyme with methyl methanethiosulfonate decreases the stability of the enzyme. With four blocked sulfhydryl groups, the denaturation temperature is decreased by 15 "C as determined by a differential scanning calorimetry study.* This suggests that the sulfhydryl groups may be important in maintaining the proper tertiary structure of cytosolic serine hydroxymethyltransferase.