- Email: [email protected]
Reversible inactivation of recombinant rat liver guanidinoacetate methyltransferase by glutathione disulfide
ARCHIVES
OF BlOCHEMISTRY
AND
BIOPHYSICS
Vol. 289, No. 1, August 15, pp. 90-96, 1991
Reversible Inactivation of Recombinant Guanidinoacetate Methyltransferase by Glutathione Disulfide Kiyoshi
Konishi
and Motoji
Fujiokal
Department of Biochemistry, Toyama Medical and Pharmaceutical Faculty of Medicine, Sugitani, Toyama 930-01, Japan
Received February
Rat Liver
University
25, 1991, and in revised form May 3, 1991
Recombinant rat liver guanidinoacetate methyltransferase is inactivated by glutathione disulfide (GSSG) following pseudo-first-order kinetics. A second-order rate constant of 20.8 Mm’ min-’ is obtained at pH 7.5 and 30°C. The inactivation is fully reversed by glutathione (GSH) in a pseudo-first-order fashion with a second-order rate constant of 11.1 Mm1 min-‘. The rate of inactivation is not affected by S-adenosylmethionine or guanidinoacetate, but complete protection against inactivation is observed in the presence of sinefungin plus guanidinoacetate. At equilibrium in the buffers containing various concentrations of GSH and GSSG, the enzyme shows activities that are dependent on the ratio but not on the total concentration of GSH and GSSG. A hyperbolic relationship is obtained between enzyme activity and [GSH]/[GSSG] ratio. The inactivation by GSSG is associated with the disappearance of -1 mol of sulfhydryl group per mole of enzyme. These results indicate that inactivation of guanidinoacetate methyltransferase by GSSG is the consequence of the formation of a mixed disulfide between a protein thiol and glutathione. The equilibrium constant for the redox reaction, E-SH + GSSG c E-SSG + GSH, obtained from the equilibrium data (1.69) is in good agreement with the value determined as the ratio of second-order rate constants for reactivation and inactivation (1.87). The cysteine residue engaged in the mixed disulfide with glutathione is identified as Cys-15 by peptide analysis after consecutive treatment of the GSSG-inactivated enzyme with Nethylmaleimide, 2-mercaptoethanol, and [‘4C]iodoacetate. The GSSG-inactivated enzyme binds S-adenosylmethionine but not guanidinoacetate in the presence and absence of sinefungin. Native guanidinoacetate methyltransferase binds guanidinoacetate in the presence of ’ To whom correspondence should be addressed at Department of Biochemistry, Toyama Medical and Pharmaceutical University Faculty of Medicine, 2630 Sugitani, Toyama 930-01, Japan. Fax: 81-764-34-4656.
sinefungin. The low overall redox equilibrium constant of 1.7-1.9 found for the reaction between guanidinoacetate methyltransferase and GSSG suggests that the activity of the enzyme is not amenable to modulation by the change in intracellular [GSH]/[GSSG] ratio. o 1991 Academic
Press,
Inc.
Guanidinoacetate methyltransferase (EC 2.1.1.2; GAMT)’ catalyzes the AdoMet-dependent methylation of guanidinoacetate to form creatine. The reaction catalyzed by GAMT is thought to constitute a major pathway by which AdoMet is converted to Ado&y in viva (1, 2). AdoHcy is hydrolyzed to homocysteine, which in turn is utilized for the synthesis of cysteine via cystathionine or remethylated to methionine by N5-methyltetrahydropteroylglutamates. Since the folate coenzymes exist mainly as N5-methyl derivatives in the cell, the conversion of homocysteine to methionine would be the key reaction for regeneration of tetrahydropteroylglutamates. Thus, in addition to the role of providing creatine, GAMT appears to play an important part in the metabolism of sulfur amino acids and one-carbon compounds. GAMT was obtained in pure form from the livers of rat (3) and pig (4), and the enzymes from both sources were shown to be monomeric proteins of similar size. The primary structure of rat liver GAMT has been determined from the cDNA nucleotide sequence (5). The enzyme, consisting of 235 amino acid residues, has 5 cysteine residues and no disulfide (6). One of the features of rat and ’ Abbreviations used: GAMT, guanidinoacetate methyltransferase; rGAMT, recombinant rat liver guanidinoacetate methyltransferase; AdoMet, S-adenosyl-L-methionine; AdoHcy, S-adenosyl-L-homocysteine; GSH, reduced glutathione; GSSG, oxidized glutathione; DTNB, 5,5’-dithiobis(Z-nitrobenzoic acid); DTT, dithiothreitol; NEM, N-ethylmaleimide.
90 All
0003.9861/91 $3.00 Copyright 0 1991 by Academic Press, Inc. rights of reproduction in any form reserved.
INACTIVATION
OF GUANIDINOACETATE
pig GAMTs is their slow inactivation upon storage. The lost activity is readily restored by the addition of thiol compounds (3,7), indicating the involvement of sulfhydryl groups in the inactivation process. Indeed, recent chemical modification and site-directed mutagenesis studies have shown that rat GAMT has three reactive cysteine residues, Cys-15, Cys-90, and Cys-219, apparently in close proximity in the three-di:mensional structure; upon incubation of the enzyme with 1 eq of DTNB a disulfide is readily formed between Cys-15 and Cys-90 (6), and the same treatment of mutant GAMT in which Cys-90 is replaced with alanine effects the disulfide cross-linking between Cys-15 and Cys-219 (8). In each case, the disulfide bond formation is accompanied by a large but not complete loss of activity. An early study in this laboratory has shown that rat GAMT is subjected to complete inactivation by the naturally occurring disulfide GSSG and the inactivation is fully reversed by sulfhydryl compounds including GSH (3). GAMT has four cysteine residues (Cys-15, Cys-90, Cys-207, and Cys-219) that react with DTNB, iodoacetate (6), and 2-nitro-5-thiocyanobenzoate (8). None of these cysteines is essential for activity (8), and simultaneous modification of more than two residues is apparently required to eliminate activity by DTNB or iodoacetate (6). Since it seems difficult to imagine that a charged, bulky molecule like GSSG is sterically accessible to multiple thiol residues of the enzyme, the complete inactivation observed with GSSG would be of mechanistic interest. In addition, the intracellular concentrations of GSH and GSSG are known to vary under certain physiological and pathological conditions, and a number of enzymes are reported to undergo regulation by the in viva redox status of GSH/GSSG (9). The present study was undertaken to elucidate the mechanism by which GSSG causes complete inactivation of GAMT, and to determine if the intracellular activity of GAMT could be modulated by the change in the in uiuo ratio of [GSH]/[GSSG]. MATERIALS
AND
METH.ODS
Materials. Reagents were obtained from the sources cited in parentheses: AdoMet (chloride salt), DTNB, sinefungin, and adenosine deaminase (type VI) (Sigma); a-chymotrypsin (Worthington); GSSG (PL Biochemicals); GSH, NEM, and the amino acid calibration mixture (Wako Pure Chemical Industries, Osaka, Japan); iodo[2-‘4C]acetic acid (55 mCi/mmol) and S-adenosyl-L-[methyl-‘4C]methionine ([14C]AdoMet) (42 mCi/mmol) (Amersham). [2-3H]Guanidinoacetic acid was synthesized from S-ethylthiourea hydrobromide and [2-3H]glycine (Du Pont-New England Nuclear) according to Brand and Brand (10). AdoMet was purified before use by passage through a Cls cartridge (SepPak, Waters Associates) as described (ll), and its concentration was determined spectrophotometrically with a value of tZ6,,= 15.4 X lo3 M-’ cm-‘. Iodoacetic acid (Merck) was recrystallized from hot chloroform. Other chemicals were of the highest purity available from local commercial sources. AdoHcy hydrolase was obtained from rat liver as described (12). GAMT used in the present study was a recombinant enzyme produced in Escherichia coli JM109 transformed with the plasmid pUCGATS-1 that contained the coding region of the cDNA for rat liver GAMT (5).
METHYLTRANSFERASE
BY GSSG
91
The recombinant rat GAMT (rGAMT) lacks the N-terminal acyl group present in the liver enzyme. Except for this difference, both enzymes show the same catalytic and physicochemical properties so far examined (5). The enzyme was purified to homogeneity as described previously using buffers containing DTT (5). Prior to experiments, the enzyme was dialyzed extensively against 50 mM Tris-HCl, pH 7.5/l mM EDTA to remove the DTT that had been added during purification. The purified enzyme was in the fully reduced state as determined by titration with Storage of the DTNB in the presence of 6 M guanidine hydrochloride. dialyzed enzyme at 0°C with no added thiols did not result in significant oxidation of sulfhydryl groups nor a decrease in enzyme activity at least for 1 week. Molar concentrations of the enzyme were calculated on the basis of M, 26,140 (5). Protein concentration was determined by the method of Lowry et al. (13) with purified rGAMT as the standard. Activity assay. The GAMT activity was determined spectrophotometrically by a coupled assay with AdoHcy hydrolase and adenosine deaminase (6). rGAMT was incubated with 20.0 pM AdoMet and 0.5 mM guanidinoacetate in 2.0 ml of 50 mM potassium phosphate, pH 8.0, in the presence of sufficient amounts of AdoHcy hydrolase and adenosine deaminase, and the decrease in absorbance at 265 nm due to the conversion of adenosine to inosine was followed continuously in a spectrophotometer. Reaction of guanidinoacetate methyltransferase with GSSG. rGAMT pH 7.5/l mM EDTA at was incubated with GSSG in 50 mM Tris-HCl, 30°C. The buffer was prepared with nitrogen-degassed water that had been boiled for 20 min and then bubbled with nitrogen while cooling, and incubation was carried out under nitrogen atmosphere. The extent of inactivation was determined by measuring the residual enzyme activity. An aliquot (~20 ~1) removed from the incubation mixture was directly added to the assay mixture and the decrease in absorbance at 265 nm was followed. The concentrations of GSSG present in the assay mixture did not interfere with the assay; linear decreases in absorbance were observed during the assay period (<5 min). Reactivation of the GSSG-inactiuated enzyme with GSH. Aqueous solutions of GSH were prepared using nitrogen-degassed water before each experiment. rGAMT was treated with 2 mM GSSG as described above until no enzyme activity was detected. The reaction mixture was then subjected to gel filtration through a column of TSK G2000SW (0.75 X 30 cm) (Tosoh; Tokyo, Japan) equilibrated and eluted with 5 mM Tris-HCl, pH 6.8/l mM EDTA. The inactivated enzyme thus prepared was incubated with various concentrations of GSH in 50 mM Tris-HCl, pH 7.5/l mM EDTA under nitrogen. Reactivation was monitored by activity measurements on aliquots withdrawn from the reaction mixture. Labeling of cysteine residues modified by treatment with GSSG. rGAMT (0.5 mg/ml) was inactivated by GSSG to <5% residual activity. The inactivated enzyme was freed from excess reagent by gel filtration through TSK G20OOSW equilibrated and eluted with 50 mM potassium phosphate, pH 6.8/l mM EDTA, and concentrated by ultrafiltration using a collodion bag (Sartorius SM-13200) to -1 mg/ml. The enzyme solution was made 4 M in guanidine hydrochloride, and incubated with 6 mM NEM for 30 min at 30°C. Following dialysis against 20 mM potassium phosphate, pH 6.8/l mM EDTA/3 M guanidine hydrochloride, and then against water, the modified enzyme was lyophilized. The lyophilized sample was dissolved in 0.5 ml of 0.5 M Tris-HCl, pH 8.0/4 M guanidine hydrochloride/l mM EDTA, and treated with 14 mM 2mercaptoethanol at 50°C. After 4 h, [‘4C]iodoacetate (420 cpm/nmol) was added to a final concentration of 14 mM, and the mixture was further incubated for 1.5 h at 37°C. The reaction mixture was dialyzed as above, and the 14C-labeled enzyme was lyophilized. Proteolytic digestion and isolation ofpeptides. The GSSG-inactivated, NEM- and [‘4C]iodoacetate-treated rGAMT was incubated with a-chymotrypsin (100~1, w/w) in 0.1 M NH4HCOB for 12 h at 37°C. The resulting peptides were separated by HPLC on a Hitachi 638-30 liquid chromatograph with a TSK ODS-12OT reverse-phase column (0.46 X 25 cm) (Tosoh), using a linear gradient of acetonitrile in 0.05% trifluoroacetic acid. The effluent was monitored by absorbance at 220 nm and
92
KONISHI
AND
GSSG (m-M)
0 ? a 0.1 0.05
0.0 I
I
20 Time
40 (min)
FIG. 1. Time course of inactivation of rGAMT by GSSG. rGAMT (0.5 mg/ml) was incubated with GSSG at pH 7.5 and 30°C as described under Materials and Methods. The GSSG concentrations from top to bottom are: 0, 1, 2, and 4 mM. A0 represents the activity at Time 0, and A is the activity at a given time. The zero-time activity was 139 U/mg protein. One unit of enzyme activity is defined as the amount of enzyme catalyzing the disappearance of 1 nmol of AdoMet per minute.
collected in 1.6-ml fractions. Aliquots from each fraction were monitored for radioactivity. The radioactive fraction was rechromatographed on the same column with a gradient of acetonitrile in 10 mM ammonium acetate, pH 6.8. Isolated peptides Amino acid analysis and sequence determination. were hydrolyzed in 5.7 M HCl, containing 0.1% (v/v) phenol and 10 mM DTT for 24 h at 108°C. Amino acid compositions were determined based on reverse-phase separation of phenylthiocarbamoyl derivatives as described previously (14). Sequence determination was carried out by automated Edman degradation on an Applied Biosystems 470A gas-phase sequenator, equipped with a 120A high-performance liquid chromatograph system. Determination of total sulfhydryl groups. The contents of sulfhydryl groups of the native and GSSG-inactivated enzymes were determined by the reaction with DTNB in 6 M guanidine hydrochloride as described by Riddles et al. (15). Binding assay. The binding of [“C]AdoMet and [3H]guanidinoacetate to the GSSG-modified GAMT was determined by the equilibrium dialysis method. The enzyme (lo-20 pM) was allowed to equilibrate with 0.01-0.2 mM [W]AdoMet or with 0.01-0.5 mM [3H]guanidinoacetate in the presence and absence of 0.2 mM sinefungin in 50 mM potassium phosphate, pH 7.7/l mM EDTA at 4°C for 16 h. Other analytical procedures. Spectrophotometric and absorbance measurements were made with a Hitachi 320 spectrophotometer, and fluorescence measurements with a Farrand MK-2 spectrofluorometer. CD spectra were recorded with a Jasco J-500 spectropolarimeter.
FUJIOKA
tivation obeyed pseudo-first-order kinetics, the rate of which was a linear function of the GSSG concentration. A value of 20.8 MP1 min-’ was obtained for the secondorder rate constant. The presence of AdoMet (0.2 mM; >55 K,,,) or guanidinoacetate (1.0 mM; >30 K,) in the reaction mixture did not affect the rate of inactivation significantly. Sinefungin, a competitive inhibitor with respect to AdoMet, was also without effect at a concentration of 0.2 mM. However, in the presence of both sinefungin (0.2 mM) and guanidinoacetate (1.0 mM), complete protection against inactivation was observed (not shown). The GSSG-inactivated rGAMT exhibited the same behavior as the untreated enzyme on a gel filtration column TSK G2000SW (data not shown), indicating that the inactivation does not involve intermolecular disulfide formation. Reactivation of the GSSG-inactivated enzyme by GSH. Whereas no enzyme activity was recovered from the GSSG-inactivated enzyme by dilution or dialysis, reactivation was readily achieved by incubation with thiol compounds. With 20 mM DTT, recovery of activity to the original value occurred within 5 min of incubation at pH 7.5 and 30°C. Figure 2 shows the time course of regeneration of activity when the GSSG-inactivated enzyme was incubated with various concentrations of GSH. Plots of log [A,/(A, - A,)] against time yielded straight lines. (A, represents the maximum enzyme activity obtained by incubation with 20 mM DTT, and At is the activity at a given time of incubation with GSH.) A linear relationship was obtained between the slope and the GSH concentra-
I
Kinetics of inactivation of guanidinoacetate methyltransferase by GSSG. Incubation of rGAMT with excess GSSG at pH 7.5 resulted in a time- and concentrationdependent loss of enzyme activity (Fig. 1). No detectable activity remained after prolonged incubation. The inac-
1
1
GSH
(mM
/
20
IO RESULTS
1
20 30 Time (min)
40
FIG. 2. Reactivation of the GSSG-inactivated rGAMT by GSH. The GSSG-inactivated enzyme was prepared and incubated with GSH as described under Materials and Methods. The GSH concentrations from top to bottom are: 20, 10, 5, and 2.5 mM. A0 is the maximum activity attainable by incubation with 20 mM DTT (135 U/mg protein), and A, is the activity at a given time of incubation with GSH.
INACTIVATION
OF GUANIDINOACETATE
Time (min)
FIG. 3. Time course of change in GAMT activity in the buffer containing GSH and GSSG. The active (m) and the GSSG-inactivated (0) pH rGAMT (0.4 mg/ml) were incubated at 30°C in 50 mM Tris-HCl, 7.5/l mM EDTA, containing 5 mM GSH and 2.5 mM GSSG. The fractional activity recovered from the inactivated enzyme is the value obtained by comparison of the enzyme activity at a given time with the maximum activity attained by treatment with 20 mM DTT (131 U/mg protein).
tion. Thus, the reactivation is first order in GSH concentration, and the second-order rate constant for reactivation is calculated to be 11.1 M-l min-‘. The reactivation was not influenced by AdoMet or by guanidinoacetate in the presence and absence of sinefungin. Equilibrium study. When rGAMT was incubated in the buffer containing both GSH and GSSG, it exhibited the same final activity reigardless of whether the reaction was started with the active enzyme or the GSSG-inactivated enzyme (Fig. 3), indicating an equilibrium between the active and the inactive enzymes. If the loss of enzyme activity is due to the oxidation of a cysteine residue to mixed disulfide, the reaction may be written as E-SH + GSSG 2 E-SSG + GSH K eq Equation
= p-SSG][GSH] [ES-SH][GSSG]
’
PI
METHYLTRANSFERASE
93
BY GSSG
the plot is not a hyperbola and the [GSH]/[GSSG] required for half maximal activity shifts with [GSH]. Figure 4 shows the experiment performed to distinguish between these alternatives. The active rGAMT was incubated in the buffers containing 5, 10, and 20 mM GSH and various amounts of GSSG to give the indicated ratios, and the enzyme activity was measured on aliquots removed from the reaction mixtures. Three or more determinations of activity were carried out for each point to ensure that equilibrium had been established. In all cases, equilibrium was reached within 60 min. As shown in the figure, the points obtained at each fixed level of GSH fell on the same curve, indicating that the inactivation results from the formation of a protein-SSG mixed disulfide. A Keg value of 1.69 t 0.09 was obtained by fit of the data to Eq. [3]. A similar Keq value was obtained in the experiment started with the GSSG-inactivated enzyme (data not shown). Relationship between loss of enzyme activity and number of sulfhydryl groups disappearing. To confirm that the GSSG inactivation is the result of oxidation of a single cysteine residue, we determined the extent of inactivation as a function of sulfhydryl group oxidation. rGAMT was incubated with GSH and GSSG at various ratios. After measuring the equilibrium enzyme activities, the modified enzymes were isolated by gel filtration and their sulfhydryl contents were determined by the reaction with DTNB under denaturing conditions (15). Complete loss of enzyme activity was associated with the disappearance of 1.17 mol of sulfhydryl group per mole of enzyme (Fig. 5). Identification of cysteine residue modified by treatment with GSSG. To identify the residue modified by GSSG,
‘*or----l
[21
[2] is rearranged to [E-SH] -=-
[&I
[GSH]/[GSSG]
KS + W=W[GSSGl) ’
[31
where [E,] is the total concentration of the enzyme. Therefore, plot of frac:tional enzyme activity versus [GSH]/[GSSG] should give a hyperbola which is dependent on the ratio but not on the total concentration of GSH and GSSG. If a protein disulfide rather than a mixed disulfide is formed, two molecules of GSH are produced for one molecule of GSSG consumed, and the [GSH] terms in Eqs. [2] and [3] must be squared. In this case
OoY
I
2I
31
4
5
1
CGSHl/ CGSSGI FIG. 4. GAMT activity as a function of [GSH]/[GSSG] ratio. rGAMT (0.5 mg/ml; 19.1 pM) was incubated at 30°C in 50 mM Tris-HCl, pH 7.5/l mM EDTA, containing 5 (O), 10 (A), and 20 mM (m) GSH and various amounts of GSSG to give the indicated ratios. The activities obtained after 60-min incubation were taken as equilibrium activities. Fractional activity represents the value obtained by comparison with the maximum activity attained by incubation with 20 mM DTT (136 U/mg protein). The curve was drawn according to Eq. [3] with K, = 1.69.
94
KONISHI
AND
the inactivated enzyme was first treated with NEM under denaturing conditions to block the unreacted, free sulfhydryl groups, and the cysteine residue engaged in the mixed disulfide was labeled with [14C]iodoacetate after reduction with 2-mercaptoethanol. Figure 6 shows the HPLC elution profile of chymotryptic peptides derived from the labeled enzyme. About 80% of the radioactivity applied to the column was found in only one fraction. Rechromatography of the radioactive fraction on the same column using a different solvent system revealed the presence of a number of peptides, but radioactivity was associated with only one peptide (Fig. 7). Amino acid analysis and sequence determination (Table I) showed that the peptide is the cysteine-containing peptide comprising residues 9-19. Thus, the results described above indicate that the inactivation of rGAMT by GSSG is the consequence of formation of a mixed disulfide between Cys-15 and glutathione. Binding of substrates to the GSSG-inactivated enzyme. The ability of the GSSG-inactivated enzyme to bind substrates was examined by equilibrium dialysis as described under Materials and Methods. The inactivated enzyme bound [14C]AdoMet. A Scatchard plot of the binding data indicated a stoichiometric binding of AdoMet with a dissociation constant of 17.0 + 3.9 PM at pH 7.7 and 4°C (plot not shown). A similar value of 10 PM is
0-v
lhmber df -SH groups disappeared
1.5
FIG. 5. Relationship between residual enzyme activity and the number of sulfhydryl groups disappeared. rGAMT (2 mg/ml) was incubated with GSH and GSSG at ratios of 4, 2.5, 1, 0.5, and 0.2 in 50 mM Tris-HCl, pH 7.5/l mM EDTA at 30°C until equilibrium was reached. After measurement of enzyme activity, the reaction mixture was subjected to gel filtration through a column of TSK G2000SW equilibrated and eluted with 5 mM Tris-HCl, pH 6.8/l mM EDTA, and the sultbydryl content of the modified enzyme was determined as described under Materials and Methods. The number of sulfhydryl groups that had disappeared was obtained by subtracting the experimental value from a value of 5.03 mol/mol of enzyme found for the untreated enzyme. The line is drawn by a least-squares linear regression.
FUJIOKA
FIG. 6. HPLC profile of a chymotryptic digest of the GSSG-inactivated rGAMT treated with NEM and with iodo[2-“Clacetate after reduction with 2-mercaptoethanol. About 5 nmol of chymotryptic peptides were fractionated on a TSK ODS-12OT column using a linear gradient of acetonitrile in 0.05% trifluoroacetic acid. The concentration of acetonitrile was varied from 0 to 50% over a period of 60 min starting at 10 min. Flow rate, 0.8 ml/min. Shadowed bars represent total radioactivity contained in each fraction (1.6 ml).
reported for the unmodified enzyme under the same conditions (6). GAMT binds substrates in an obligatory order: AdoMet is bound first followed by guanidinoacetate (16). Since sinefungin, a structural analog of AdoMet, can bind to the free enzyme, the effectiveness of the compound to support binding of guanidinoacetate was tested. Incubation of the unmodified enzyme with sinefungin (0.2 mM) and [3H]guanidinoacetate (0.01-0.5 mM) under the same conditions as above revealed binding of guanidinoacetate to the rGAMT-sinefungin complex with a Kd value of 70.1 f 9.8 PM. However, no appreciable binding of [3H]guanidinoacetate to the GSSG-modified enzyme was observed in the presence or absence of sinefungin. That the modified enzyme could bind sinefungin was demonstrated by its competition with [14C]AdoMet binding. enSpectroscopic properties of the GSSG-inactivated zyme. The spectroscopic properties of rGAMT remained unchanged after GSSG inactivation. The UV absorption, CD, and fluorescence spectra of the native and modified enzymes were almost identical in shape and intensity (spectra not shown). Therefore, no gross change in the secondary and tertiary structure appears to accompany the formation of glutathione mixed disulfide at Cys-15. DISCUSSION
The complete inactivation of rat liver GAMT by GSSG observed previously (3) has now been confirmed and shown to be the consequence of mixed disulfide formation between glutathione and Cys-15 as follows. (a) In the
INAlCTIVATION
0
OF GUANIDINOACETATE
IO 20 30 Time (min 1
FIG. 7. Rechromatography of the radioactive fraction of Fig. 6. The radioactive peptides eluting at 38-40 min were rechromatographed on a TSK ODS-12OT column with an acetonitrile gradient in 10 mM ammonium acetate, pH 6.8. Acetonitrile gradient: O-16% between 10 and 15 min; 16-57% between 15 and 65 min. The arrow shows the radioactive peptide.
presence of both GSH and GSSG, GAMT activity reaches an equilibrium value, whilch is dependent on the [GSH]/ [GSSG] ratio but not on the absolute concentrations of the reagents. Plotting equilibrium activities against [GSH]/[GSSG] ratios yields a hyperbolic curve (Fig. 4). These results show that the inactivation is due to the formation of a protein-SSG mixed disulfide, not to an intramolecular disulfide cross-linking. The equilibrium constant of 1.69 found in this experiment is in good agreement with that obtained as the ratio of the bimolecular rate constants for inactivation and reactivation reactions (k,/lz, = 1.87, Reaction [l]). (b) A 1:l stoichiometry between loss of enzyme activity and the number of sulfhydryl groups that had disappeared (Fig. 5) and the identification of Cys-15 as the sole residue modified in the inactivated enzyme indicate that the formation of one glutathione mixed disulfilde is sufficient to eliminate all activity. It has been shown that treatment of rGAMT with an equimolar amount of DT’NB results in the formation of a disulfide bond between ~Cys-15 and Cys-90 (6), and the same treatment of mutant GAMT in which Cys-90 is mutated to alanine leads to ldisulfide cross-linking between Cys-15 and Cys-219 (8). Therefore, in view of numerous examples of GSSG-mediated formation of protein disulfides, it is rather surprisi:ng to find that the glutathione moiety fixed on Cys-15 is not easily displaced. It is possible that the redox equilibrium involving the protein aryl mixed disulfide and protein thiol lies so far toward aryl thiol, while the reverse is true for the protein-SSG mixed disulfide and protein thiol. Another possibility is that co-
METHYLTRANSFERASE
95
BY GSSG
valent bonding of the charged tripeptide group onto Cys15 causes local structural perturbation thereby preventing the attack of Cys-90 or Cys-219 on the mixed disulfide. A drastic conformational change, however, is unlikely because the modified enzyme exhibits spectroscopic properties almost identical to those of the unmodified enzyme. The complete loss of activity that occurs upon disulfide formation between Cys-15 and glutathione appears to reflect the inability of the modified enzyme to bind guanidinoacetate. A previous study has shown that the removal of N-terminal eicosapeptide from rGAMT by trypsinolysis results in an enzyme that is catalytically inactive but weakly binds AdoMet and guanidinoacetate (16). Also, the enzyme in which Cys-15 is cross-linked to Cys-90 or Cys-219 appears to have affinities for substrates not greatly different from those of the native enzyme (8). Therefore, it is highly improbable that the peptide segment around Cys-15 is part of the active site. GAMT binds substrates in a compulsory order: guanidinoacetate binds only to the enzyme-AdoMet binary complex (16). Thus, the binding of AdoMet must induce a conformation change to generate or expose the binding site for guanidinoacetate. It is possible that this process is disrupted in some way in the modified enzyme. The protection against GSSG inactivation afforded by sinefungin plus guanidinoacetate may be interpreted as the result of substantial movement of the portion around Cys-15 upon ternary complex formation. It has been suggested that this portion of the enzyme lies on the protein surface and is flexible (16). The structural change induced by binding of both substrates might be a prerequisite for normal catalysis since the loss of flexibility caused by cross-linking of Cys-15 with either Cys-90 or Cys-219 leads to a large loss of activity (6,8), and the removal of 20 amino acids from the N-terminus results in complete inactivation (16). Based on the results of the present and previous studies it seems likely that the N-terminal segment around Cys-
TABLE
I
Edman Degradation of Radioactive Peptide from the GSSGInactivated Enzyme Labeled with [ 14C]Iodoacetate Cycle 1 2 3 4 5 6 7 8 9 10 11 ’ S-Carboxymethylcysteine.
Residue Phe Ala Pro GUY
Glu Asp Cys (cm)’ GUY Pro Ala Trp
pm01 41.0 96.4 11.7 32.3 22.8 9.5 15.4 12.1 21.5 7.5 3.1
96
KONISHI
AND
15 is functionally linked to the active site. Because of concurrent modification of Cys-15 and Cys-90 by DTNB and iodoacetate (6), it has not been possible to determine the consequence of modification of individual residues. The results obtained here indicate that modification of Cys-15 alone exerts a profound effect on enzyme catalysis. The predominant intracellular low-molecular-weight thiol and disulfide are GSH and GSSG. Recent measurements have shown that the levels of GSH and GSSG in normal liver, assuming their homogeneous distribution, are 8-10 mM and 0.02-0.5 mM, respectively (17, 18). The intracellular concentrations of GSH and GSSG, besides undergoing changes under conditions of induced oxidative stress (19,20), can vary depending on the metabolic state of an animal. For example, fasting causes a significant reduction in the concentration of GSH (to 3-4 mM) (17, 21). Thus, the value of [GSH]/[GSSG] in uiuo could vary between 100 and 400 under physiological conditions. For the activity of a given enzyme to be modulated by the change in the cellular oxidation-reduction state of glutathione, the thiol/disulfide redox equilibrium of the protein sulfhydryls should be near the ilz uiuo [GSH]/[GSSG] ratio, and the thiol/disulfide interchange reaction must be kinetically competent (22). Although it might be possible that the thiol/disulfide interchange is a catalyzed reaction in uiuo, the low overall redox equilibrium constant of GAMT (1.7-1.9) is consistent with the prediction that the enzyme exists predominantly in the reduced, active form under normal metabolic conditions. Thus, it appears unlikely that the GAMT activity is modulated in response to physiological alterations in the [GSH]/[GSSG] ratio. REFERENCES 1. Mudd, S. H., and Poole, J. R. (1975) Metab. Clin. Exp. 24, 721735. 2. Mudd, S. H., Ebert, M. H., and Striver, C. R. (1980) Metab. Clin. Exp. 29, 707-720.
FUJIOKA 3. Ogawa, H., Ishiguro, Y., and Fujioka, M. (1983) Arch. Biochem. Biophys. 226, 265-275. 4. Im, Y. S., Chiang, P. K., and Cantoni, G. L. (1979) J. Biol. Chem. 254, 11,047-11,050. 5. Ogawa, H., Date, T., Gomi, T., Konishi, K., Pitot, H. C., Cantoni, G. L., and Fujioka, M. (1988) Proc. N&l. Acad. Sci. USA 85, 694698. 6. Fujioka, M., Konishi, K., and Takata, Y. (1988) Biochemistry 27, 7658-7664. 7. Cantoni, G. L., and Vignos, P. J., Jr. (1954) J. Biol. Chem. 209, 647-659. 8. Takata, Y., Date, T., and Fujioka, M. (1991) &o&em. J., in press. 9. Gilbert, H. F. (1990) Ado. Enzymol. Relat. Areas Mol. Biol. 63,69172. 10. Brand, E., and Brand, F. C. (1955) in Organic Syntheses (Horning, E. C., Ed.), Collective Vol. 3, pp. 440-442, Wiley, New York. 11. Fujioka, M., and Ishiguro, Y. (1986) J. Biol. Chem. 261,6346-6351. 12. Fujioka, M., and Takata, Y. (1981) J. Biol. C/rem. 256, 1631-1635. 13. Lowry, 0. H., Rosebrough, N. J., Farr, A. L., and Randall, R. J. (1951) J. Biol. Chem. 193,265-275. 14. Gomi, T., Ogawa, H., and Fujioka, M. (1986) J. Biol. Chem. 261, 13,422-13,425. 15. Riddles, P. W., Blakeley, R. L., and Zerner, B. (1983) in Methods in Enzymology (Hire,, C. H. W., and Timasheff, S. N., Eds.), Vol. 91, pp. 49-60, Academic Press, New York. 16. Fujioka, M., Takata, Y., and Gomi, T. (1991) Arch. B&hem. Biophys. 285,181-186. 17. Alpert, A. J., and Gilbert, H. F. (1985) Anal. Biochzm. 144, 553562. 18. Akerboom, T. P. M., Bilzer, M., and Sies, H. (1982) J. Biol. Chem. 257, 4248-4252. 19. Collison, M. W., and Thomas, J. A. (1987) Biochim. Biophys. Acta 928,121-129. 20. Lauterburg, B. H., Smith, C. V., Hughes, H., and Mitchell, J. R. (1984) J. Clin. Znoest. 73, 124-133. 21. Tateishi, N., and Sakamoto, Y. (1983) in Glutathione: Storage, Transport, and Turnover in Mammals (Sakamoto, Y., Higashi, T., and Tateishi, N., Eds.), pp. 3-9, Japan Scientific Society Press, Tokyo. 22. Gilbert, H. F. (1984) in Methods in Enzymology (Weld, F., and Moldave, K., Eds.), Vol. 107, pp. 330-351, Academic Press, Orlando.
OF BlOCHEMISTRY
AND
BIOPHYSICS
Vol. 289, No. 1, August 15, pp. 90-96, 1991
Reversible Inactivation of Recombinant Guanidinoacetate Methyltransferase by Glutathione Disulfide Kiyoshi
Konishi
and Motoji
Fujiokal
Department of Biochemistry, Toyama Medical and Pharmaceutical Faculty of Medicine, Sugitani, Toyama 930-01, Japan
Received February
Rat Liver
University
25, 1991, and in revised form May 3, 1991
Recombinant rat liver guanidinoacetate methyltransferase is inactivated by glutathione disulfide (GSSG) following pseudo-first-order kinetics. A second-order rate constant of 20.8 Mm’ min-’ is obtained at pH 7.5 and 30°C. The inactivation is fully reversed by glutathione (GSH) in a pseudo-first-order fashion with a second-order rate constant of 11.1 Mm1 min-‘. The rate of inactivation is not affected by S-adenosylmethionine or guanidinoacetate, but complete protection against inactivation is observed in the presence of sinefungin plus guanidinoacetate. At equilibrium in the buffers containing various concentrations of GSH and GSSG, the enzyme shows activities that are dependent on the ratio but not on the total concentration of GSH and GSSG. A hyperbolic relationship is obtained between enzyme activity and [GSH]/[GSSG] ratio. The inactivation by GSSG is associated with the disappearance of -1 mol of sulfhydryl group per mole of enzyme. These results indicate that inactivation of guanidinoacetate methyltransferase by GSSG is the consequence of the formation of a mixed disulfide between a protein thiol and glutathione. The equilibrium constant for the redox reaction, E-SH + GSSG c E-SSG + GSH, obtained from the equilibrium data (1.69) is in good agreement with the value determined as the ratio of second-order rate constants for reactivation and inactivation (1.87). The cysteine residue engaged in the mixed disulfide with glutathione is identified as Cys-15 by peptide analysis after consecutive treatment of the GSSG-inactivated enzyme with Nethylmaleimide, 2-mercaptoethanol, and [‘4C]iodoacetate. The GSSG-inactivated enzyme binds S-adenosylmethionine but not guanidinoacetate in the presence and absence of sinefungin. Native guanidinoacetate methyltransferase binds guanidinoacetate in the presence of ’ To whom correspondence should be addressed at Department of Biochemistry, Toyama Medical and Pharmaceutical University Faculty of Medicine, 2630 Sugitani, Toyama 930-01, Japan. Fax: 81-764-34-4656.
sinefungin. The low overall redox equilibrium constant of 1.7-1.9 found for the reaction between guanidinoacetate methyltransferase and GSSG suggests that the activity of the enzyme is not amenable to modulation by the change in intracellular [GSH]/[GSSG] ratio. o 1991 Academic
Press,
Inc.
Guanidinoacetate methyltransferase (EC 2.1.1.2; GAMT)’ catalyzes the AdoMet-dependent methylation of guanidinoacetate to form creatine. The reaction catalyzed by GAMT is thought to constitute a major pathway by which AdoMet is converted to Ado&y in viva (1, 2). AdoHcy is hydrolyzed to homocysteine, which in turn is utilized for the synthesis of cysteine via cystathionine or remethylated to methionine by N5-methyltetrahydropteroylglutamates. Since the folate coenzymes exist mainly as N5-methyl derivatives in the cell, the conversion of homocysteine to methionine would be the key reaction for regeneration of tetrahydropteroylglutamates. Thus, in addition to the role of providing creatine, GAMT appears to play an important part in the metabolism of sulfur amino acids and one-carbon compounds. GAMT was obtained in pure form from the livers of rat (3) and pig (4), and the enzymes from both sources were shown to be monomeric proteins of similar size. The primary structure of rat liver GAMT has been determined from the cDNA nucleotide sequence (5). The enzyme, consisting of 235 amino acid residues, has 5 cysteine residues and no disulfide (6). One of the features of rat and ’ Abbreviations used: GAMT, guanidinoacetate methyltransferase; rGAMT, recombinant rat liver guanidinoacetate methyltransferase; AdoMet, S-adenosyl-L-methionine; AdoHcy, S-adenosyl-L-homocysteine; GSH, reduced glutathione; GSSG, oxidized glutathione; DTNB, 5,5’-dithiobis(Z-nitrobenzoic acid); DTT, dithiothreitol; NEM, N-ethylmaleimide.
90 All
0003.9861/91 $3.00 Copyright 0 1991 by Academic Press, Inc. rights of reproduction in any form reserved.
INACTIVATION
OF GUANIDINOACETATE
pig GAMTs is their slow inactivation upon storage. The lost activity is readily restored by the addition of thiol compounds (3,7), indicating the involvement of sulfhydryl groups in the inactivation process. Indeed, recent chemical modification and site-directed mutagenesis studies have shown that rat GAMT has three reactive cysteine residues, Cys-15, Cys-90, and Cys-219, apparently in close proximity in the three-di:mensional structure; upon incubation of the enzyme with 1 eq of DTNB a disulfide is readily formed between Cys-15 and Cys-90 (6), and the same treatment of mutant GAMT in which Cys-90 is replaced with alanine effects the disulfide cross-linking between Cys-15 and Cys-219 (8). In each case, the disulfide bond formation is accompanied by a large but not complete loss of activity. An early study in this laboratory has shown that rat GAMT is subjected to complete inactivation by the naturally occurring disulfide GSSG and the inactivation is fully reversed by sulfhydryl compounds including GSH (3). GAMT has four cysteine residues (Cys-15, Cys-90, Cys-207, and Cys-219) that react with DTNB, iodoacetate (6), and 2-nitro-5-thiocyanobenzoate (8). None of these cysteines is essential for activity (8), and simultaneous modification of more than two residues is apparently required to eliminate activity by DTNB or iodoacetate (6). Since it seems difficult to imagine that a charged, bulky molecule like GSSG is sterically accessible to multiple thiol residues of the enzyme, the complete inactivation observed with GSSG would be of mechanistic interest. In addition, the intracellular concentrations of GSH and GSSG are known to vary under certain physiological and pathological conditions, and a number of enzymes are reported to undergo regulation by the in viva redox status of GSH/GSSG (9). The present study was undertaken to elucidate the mechanism by which GSSG causes complete inactivation of GAMT, and to determine if the intracellular activity of GAMT could be modulated by the change in the in uiuo ratio of [GSH]/[GSSG]. MATERIALS
AND
METH.ODS
Materials. Reagents were obtained from the sources cited in parentheses: AdoMet (chloride salt), DTNB, sinefungin, and adenosine deaminase (type VI) (Sigma); a-chymotrypsin (Worthington); GSSG (PL Biochemicals); GSH, NEM, and the amino acid calibration mixture (Wako Pure Chemical Industries, Osaka, Japan); iodo[2-‘4C]acetic acid (55 mCi/mmol) and S-adenosyl-L-[methyl-‘4C]methionine ([14C]AdoMet) (42 mCi/mmol) (Amersham). [2-3H]Guanidinoacetic acid was synthesized from S-ethylthiourea hydrobromide and [2-3H]glycine (Du Pont-New England Nuclear) according to Brand and Brand (10). AdoMet was purified before use by passage through a Cls cartridge (SepPak, Waters Associates) as described (ll), and its concentration was determined spectrophotometrically with a value of tZ6,,= 15.4 X lo3 M-’ cm-‘. Iodoacetic acid (Merck) was recrystallized from hot chloroform. Other chemicals were of the highest purity available from local commercial sources. AdoHcy hydrolase was obtained from rat liver as described (12). GAMT used in the present study was a recombinant enzyme produced in Escherichia coli JM109 transformed with the plasmid pUCGATS-1 that contained the coding region of the cDNA for rat liver GAMT (5).
METHYLTRANSFERASE
BY GSSG
91
The recombinant rat GAMT (rGAMT) lacks the N-terminal acyl group present in the liver enzyme. Except for this difference, both enzymes show the same catalytic and physicochemical properties so far examined (5). The enzyme was purified to homogeneity as described previously using buffers containing DTT (5). Prior to experiments, the enzyme was dialyzed extensively against 50 mM Tris-HCl, pH 7.5/l mM EDTA to remove the DTT that had been added during purification. The purified enzyme was in the fully reduced state as determined by titration with Storage of the DTNB in the presence of 6 M guanidine hydrochloride. dialyzed enzyme at 0°C with no added thiols did not result in significant oxidation of sulfhydryl groups nor a decrease in enzyme activity at least for 1 week. Molar concentrations of the enzyme were calculated on the basis of M, 26,140 (5). Protein concentration was determined by the method of Lowry et al. (13) with purified rGAMT as the standard. Activity assay. The GAMT activity was determined spectrophotometrically by a coupled assay with AdoHcy hydrolase and adenosine deaminase (6). rGAMT was incubated with 20.0 pM AdoMet and 0.5 mM guanidinoacetate in 2.0 ml of 50 mM potassium phosphate, pH 8.0, in the presence of sufficient amounts of AdoHcy hydrolase and adenosine deaminase, and the decrease in absorbance at 265 nm due to the conversion of adenosine to inosine was followed continuously in a spectrophotometer. Reaction of guanidinoacetate methyltransferase with GSSG. rGAMT pH 7.5/l mM EDTA at was incubated with GSSG in 50 mM Tris-HCl, 30°C. The buffer was prepared with nitrogen-degassed water that had been boiled for 20 min and then bubbled with nitrogen while cooling, and incubation was carried out under nitrogen atmosphere. The extent of inactivation was determined by measuring the residual enzyme activity. An aliquot (~20 ~1) removed from the incubation mixture was directly added to the assay mixture and the decrease in absorbance at 265 nm was followed. The concentrations of GSSG present in the assay mixture did not interfere with the assay; linear decreases in absorbance were observed during the assay period (<5 min). Reactivation of the GSSG-inactiuated enzyme with GSH. Aqueous solutions of GSH were prepared using nitrogen-degassed water before each experiment. rGAMT was treated with 2 mM GSSG as described above until no enzyme activity was detected. The reaction mixture was then subjected to gel filtration through a column of TSK G2000SW (0.75 X 30 cm) (Tosoh; Tokyo, Japan) equilibrated and eluted with 5 mM Tris-HCl, pH 6.8/l mM EDTA. The inactivated enzyme thus prepared was incubated with various concentrations of GSH in 50 mM Tris-HCl, pH 7.5/l mM EDTA under nitrogen. Reactivation was monitored by activity measurements on aliquots withdrawn from the reaction mixture. Labeling of cysteine residues modified by treatment with GSSG. rGAMT (0.5 mg/ml) was inactivated by GSSG to <5% residual activity. The inactivated enzyme was freed from excess reagent by gel filtration through TSK G20OOSW equilibrated and eluted with 50 mM potassium phosphate, pH 6.8/l mM EDTA, and concentrated by ultrafiltration using a collodion bag (Sartorius SM-13200) to -1 mg/ml. The enzyme solution was made 4 M in guanidine hydrochloride, and incubated with 6 mM NEM for 30 min at 30°C. Following dialysis against 20 mM potassium phosphate, pH 6.8/l mM EDTA/3 M guanidine hydrochloride, and then against water, the modified enzyme was lyophilized. The lyophilized sample was dissolved in 0.5 ml of 0.5 M Tris-HCl, pH 8.0/4 M guanidine hydrochloride/l mM EDTA, and treated with 14 mM 2mercaptoethanol at 50°C. After 4 h, [‘4C]iodoacetate (420 cpm/nmol) was added to a final concentration of 14 mM, and the mixture was further incubated for 1.5 h at 37°C. The reaction mixture was dialyzed as above, and the 14C-labeled enzyme was lyophilized. Proteolytic digestion and isolation ofpeptides. The GSSG-inactivated, NEM- and [‘4C]iodoacetate-treated rGAMT was incubated with a-chymotrypsin (100~1, w/w) in 0.1 M NH4HCOB for 12 h at 37°C. The resulting peptides were separated by HPLC on a Hitachi 638-30 liquid chromatograph with a TSK ODS-12OT reverse-phase column (0.46 X 25 cm) (Tosoh), using a linear gradient of acetonitrile in 0.05% trifluoroacetic acid. The effluent was monitored by absorbance at 220 nm and
92
KONISHI
AND
GSSG (m-M)
0 ? a 0.1 0.05
0.0 I
I
20 Time
40 (min)
FIG. 1. Time course of inactivation of rGAMT by GSSG. rGAMT (0.5 mg/ml) was incubated with GSSG at pH 7.5 and 30°C as described under Materials and Methods. The GSSG concentrations from top to bottom are: 0, 1, 2, and 4 mM. A0 represents the activity at Time 0, and A is the activity at a given time. The zero-time activity was 139 U/mg protein. One unit of enzyme activity is defined as the amount of enzyme catalyzing the disappearance of 1 nmol of AdoMet per minute.
collected in 1.6-ml fractions. Aliquots from each fraction were monitored for radioactivity. The radioactive fraction was rechromatographed on the same column with a gradient of acetonitrile in 10 mM ammonium acetate, pH 6.8. Isolated peptides Amino acid analysis and sequence determination. were hydrolyzed in 5.7 M HCl, containing 0.1% (v/v) phenol and 10 mM DTT for 24 h at 108°C. Amino acid compositions were determined based on reverse-phase separation of phenylthiocarbamoyl derivatives as described previously (14). Sequence determination was carried out by automated Edman degradation on an Applied Biosystems 470A gas-phase sequenator, equipped with a 120A high-performance liquid chromatograph system. Determination of total sulfhydryl groups. The contents of sulfhydryl groups of the native and GSSG-inactivated enzymes were determined by the reaction with DTNB in 6 M guanidine hydrochloride as described by Riddles et al. (15). Binding assay. The binding of [“C]AdoMet and [3H]guanidinoacetate to the GSSG-modified GAMT was determined by the equilibrium dialysis method. The enzyme (lo-20 pM) was allowed to equilibrate with 0.01-0.2 mM [W]AdoMet or with 0.01-0.5 mM [3H]guanidinoacetate in the presence and absence of 0.2 mM sinefungin in 50 mM potassium phosphate, pH 7.7/l mM EDTA at 4°C for 16 h. Other analytical procedures. Spectrophotometric and absorbance measurements were made with a Hitachi 320 spectrophotometer, and fluorescence measurements with a Farrand MK-2 spectrofluorometer. CD spectra were recorded with a Jasco J-500 spectropolarimeter.
FUJIOKA
tivation obeyed pseudo-first-order kinetics, the rate of which was a linear function of the GSSG concentration. A value of 20.8 MP1 min-’ was obtained for the secondorder rate constant. The presence of AdoMet (0.2 mM; >55 K,,,) or guanidinoacetate (1.0 mM; >30 K,) in the reaction mixture did not affect the rate of inactivation significantly. Sinefungin, a competitive inhibitor with respect to AdoMet, was also without effect at a concentration of 0.2 mM. However, in the presence of both sinefungin (0.2 mM) and guanidinoacetate (1.0 mM), complete protection against inactivation was observed (not shown). The GSSG-inactivated rGAMT exhibited the same behavior as the untreated enzyme on a gel filtration column TSK G2000SW (data not shown), indicating that the inactivation does not involve intermolecular disulfide formation. Reactivation of the GSSG-inactivated enzyme by GSH. Whereas no enzyme activity was recovered from the GSSG-inactivated enzyme by dilution or dialysis, reactivation was readily achieved by incubation with thiol compounds. With 20 mM DTT, recovery of activity to the original value occurred within 5 min of incubation at pH 7.5 and 30°C. Figure 2 shows the time course of regeneration of activity when the GSSG-inactivated enzyme was incubated with various concentrations of GSH. Plots of log [A,/(A, - A,)] against time yielded straight lines. (A, represents the maximum enzyme activity obtained by incubation with 20 mM DTT, and At is the activity at a given time of incubation with GSH.) A linear relationship was obtained between the slope and the GSH concentra-
I
Kinetics of inactivation of guanidinoacetate methyltransferase by GSSG. Incubation of rGAMT with excess GSSG at pH 7.5 resulted in a time- and concentrationdependent loss of enzyme activity (Fig. 1). No detectable activity remained after prolonged incubation. The inac-
1
1
GSH
(mM
/
20
IO RESULTS
1
20 30 Time (min)
40
FIG. 2. Reactivation of the GSSG-inactivated rGAMT by GSH. The GSSG-inactivated enzyme was prepared and incubated with GSH as described under Materials and Methods. The GSH concentrations from top to bottom are: 20, 10, 5, and 2.5 mM. A0 is the maximum activity attainable by incubation with 20 mM DTT (135 U/mg protein), and A, is the activity at a given time of incubation with GSH.
INACTIVATION
OF GUANIDINOACETATE
Time (min)
FIG. 3. Time course of change in GAMT activity in the buffer containing GSH and GSSG. The active (m) and the GSSG-inactivated (0) pH rGAMT (0.4 mg/ml) were incubated at 30°C in 50 mM Tris-HCl, 7.5/l mM EDTA, containing 5 mM GSH and 2.5 mM GSSG. The fractional activity recovered from the inactivated enzyme is the value obtained by comparison of the enzyme activity at a given time with the maximum activity attained by treatment with 20 mM DTT (131 U/mg protein).
tion. Thus, the reactivation is first order in GSH concentration, and the second-order rate constant for reactivation is calculated to be 11.1 M-l min-‘. The reactivation was not influenced by AdoMet or by guanidinoacetate in the presence and absence of sinefungin. Equilibrium study. When rGAMT was incubated in the buffer containing both GSH and GSSG, it exhibited the same final activity reigardless of whether the reaction was started with the active enzyme or the GSSG-inactivated enzyme (Fig. 3), indicating an equilibrium between the active and the inactive enzymes. If the loss of enzyme activity is due to the oxidation of a cysteine residue to mixed disulfide, the reaction may be written as E-SH + GSSG 2 E-SSG + GSH K eq Equation
= p-SSG][GSH] [ES-SH][GSSG]
’
PI
METHYLTRANSFERASE
93
BY GSSG
the plot is not a hyperbola and the [GSH]/[GSSG] required for half maximal activity shifts with [GSH]. Figure 4 shows the experiment performed to distinguish between these alternatives. The active rGAMT was incubated in the buffers containing 5, 10, and 20 mM GSH and various amounts of GSSG to give the indicated ratios, and the enzyme activity was measured on aliquots removed from the reaction mixtures. Three or more determinations of activity were carried out for each point to ensure that equilibrium had been established. In all cases, equilibrium was reached within 60 min. As shown in the figure, the points obtained at each fixed level of GSH fell on the same curve, indicating that the inactivation results from the formation of a protein-SSG mixed disulfide. A Keg value of 1.69 t 0.09 was obtained by fit of the data to Eq. [3]. A similar Keq value was obtained in the experiment started with the GSSG-inactivated enzyme (data not shown). Relationship between loss of enzyme activity and number of sulfhydryl groups disappearing. To confirm that the GSSG inactivation is the result of oxidation of a single cysteine residue, we determined the extent of inactivation as a function of sulfhydryl group oxidation. rGAMT was incubated with GSH and GSSG at various ratios. After measuring the equilibrium enzyme activities, the modified enzymes were isolated by gel filtration and their sulfhydryl contents were determined by the reaction with DTNB under denaturing conditions (15). Complete loss of enzyme activity was associated with the disappearance of 1.17 mol of sulfhydryl group per mole of enzyme (Fig. 5). Identification of cysteine residue modified by treatment with GSSG. To identify the residue modified by GSSG,
‘*or----l
[21
[2] is rearranged to [E-SH] -=-
[&I
[GSH]/[GSSG]
KS + W=W[GSSGl) ’
[31
where [E,] is the total concentration of the enzyme. Therefore, plot of frac:tional enzyme activity versus [GSH]/[GSSG] should give a hyperbola which is dependent on the ratio but not on the total concentration of GSH and GSSG. If a protein disulfide rather than a mixed disulfide is formed, two molecules of GSH are produced for one molecule of GSSG consumed, and the [GSH] terms in Eqs. [2] and [3] must be squared. In this case
OoY
I
2I
31
4
5
1
CGSHl/ CGSSGI FIG. 4. GAMT activity as a function of [GSH]/[GSSG] ratio. rGAMT (0.5 mg/ml; 19.1 pM) was incubated at 30°C in 50 mM Tris-HCl, pH 7.5/l mM EDTA, containing 5 (O), 10 (A), and 20 mM (m) GSH and various amounts of GSSG to give the indicated ratios. The activities obtained after 60-min incubation were taken as equilibrium activities. Fractional activity represents the value obtained by comparison with the maximum activity attained by incubation with 20 mM DTT (136 U/mg protein). The curve was drawn according to Eq. [3] with K, = 1.69.
94
KONISHI
AND
the inactivated enzyme was first treated with NEM under denaturing conditions to block the unreacted, free sulfhydryl groups, and the cysteine residue engaged in the mixed disulfide was labeled with [14C]iodoacetate after reduction with 2-mercaptoethanol. Figure 6 shows the HPLC elution profile of chymotryptic peptides derived from the labeled enzyme. About 80% of the radioactivity applied to the column was found in only one fraction. Rechromatography of the radioactive fraction on the same column using a different solvent system revealed the presence of a number of peptides, but radioactivity was associated with only one peptide (Fig. 7). Amino acid analysis and sequence determination (Table I) showed that the peptide is the cysteine-containing peptide comprising residues 9-19. Thus, the results described above indicate that the inactivation of rGAMT by GSSG is the consequence of formation of a mixed disulfide between Cys-15 and glutathione. Binding of substrates to the GSSG-inactivated enzyme. The ability of the GSSG-inactivated enzyme to bind substrates was examined by equilibrium dialysis as described under Materials and Methods. The inactivated enzyme bound [14C]AdoMet. A Scatchard plot of the binding data indicated a stoichiometric binding of AdoMet with a dissociation constant of 17.0 + 3.9 PM at pH 7.7 and 4°C (plot not shown). A similar value of 10 PM is
0-v
lhmber df -SH groups disappeared
1.5
FIG. 5. Relationship between residual enzyme activity and the number of sulfhydryl groups disappeared. rGAMT (2 mg/ml) was incubated with GSH and GSSG at ratios of 4, 2.5, 1, 0.5, and 0.2 in 50 mM Tris-HCl, pH 7.5/l mM EDTA at 30°C until equilibrium was reached. After measurement of enzyme activity, the reaction mixture was subjected to gel filtration through a column of TSK G2000SW equilibrated and eluted with 5 mM Tris-HCl, pH 6.8/l mM EDTA, and the sultbydryl content of the modified enzyme was determined as described under Materials and Methods. The number of sulfhydryl groups that had disappeared was obtained by subtracting the experimental value from a value of 5.03 mol/mol of enzyme found for the untreated enzyme. The line is drawn by a least-squares linear regression.
FUJIOKA
FIG. 6. HPLC profile of a chymotryptic digest of the GSSG-inactivated rGAMT treated with NEM and with iodo[2-“Clacetate after reduction with 2-mercaptoethanol. About 5 nmol of chymotryptic peptides were fractionated on a TSK ODS-12OT column using a linear gradient of acetonitrile in 0.05% trifluoroacetic acid. The concentration of acetonitrile was varied from 0 to 50% over a period of 60 min starting at 10 min. Flow rate, 0.8 ml/min. Shadowed bars represent total radioactivity contained in each fraction (1.6 ml).
reported for the unmodified enzyme under the same conditions (6). GAMT binds substrates in an obligatory order: AdoMet is bound first followed by guanidinoacetate (16). Since sinefungin, a structural analog of AdoMet, can bind to the free enzyme, the effectiveness of the compound to support binding of guanidinoacetate was tested. Incubation of the unmodified enzyme with sinefungin (0.2 mM) and [3H]guanidinoacetate (0.01-0.5 mM) under the same conditions as above revealed binding of guanidinoacetate to the rGAMT-sinefungin complex with a Kd value of 70.1 f 9.8 PM. However, no appreciable binding of [3H]guanidinoacetate to the GSSG-modified enzyme was observed in the presence or absence of sinefungin. That the modified enzyme could bind sinefungin was demonstrated by its competition with [14C]AdoMet binding. enSpectroscopic properties of the GSSG-inactivated zyme. The spectroscopic properties of rGAMT remained unchanged after GSSG inactivation. The UV absorption, CD, and fluorescence spectra of the native and modified enzymes were almost identical in shape and intensity (spectra not shown). Therefore, no gross change in the secondary and tertiary structure appears to accompany the formation of glutathione mixed disulfide at Cys-15. DISCUSSION
The complete inactivation of rat liver GAMT by GSSG observed previously (3) has now been confirmed and shown to be the consequence of mixed disulfide formation between glutathione and Cys-15 as follows. (a) In the
INAlCTIVATION
0
OF GUANIDINOACETATE
IO 20 30 Time (min 1
FIG. 7. Rechromatography of the radioactive fraction of Fig. 6. The radioactive peptides eluting at 38-40 min were rechromatographed on a TSK ODS-12OT column with an acetonitrile gradient in 10 mM ammonium acetate, pH 6.8. Acetonitrile gradient: O-16% between 10 and 15 min; 16-57% between 15 and 65 min. The arrow shows the radioactive peptide.
presence of both GSH and GSSG, GAMT activity reaches an equilibrium value, whilch is dependent on the [GSH]/ [GSSG] ratio but not on the absolute concentrations of the reagents. Plotting equilibrium activities against [GSH]/[GSSG] ratios yields a hyperbolic curve (Fig. 4). These results show that the inactivation is due to the formation of a protein-SSG mixed disulfide, not to an intramolecular disulfide cross-linking. The equilibrium constant of 1.69 found in this experiment is in good agreement with that obtained as the ratio of the bimolecular rate constants for inactivation and reactivation reactions (k,/lz, = 1.87, Reaction [l]). (b) A 1:l stoichiometry between loss of enzyme activity and the number of sulfhydryl groups that had disappeared (Fig. 5) and the identification of Cys-15 as the sole residue modified in the inactivated enzyme indicate that the formation of one glutathione mixed disulfilde is sufficient to eliminate all activity. It has been shown that treatment of rGAMT with an equimolar amount of DT’NB results in the formation of a disulfide bond between ~Cys-15 and Cys-90 (6), and the same treatment of mutant GAMT in which Cys-90 is mutated to alanine leads to ldisulfide cross-linking between Cys-15 and Cys-219 (8). Therefore, in view of numerous examples of GSSG-mediated formation of protein disulfides, it is rather surprisi:ng to find that the glutathione moiety fixed on Cys-15 is not easily displaced. It is possible that the redox equilibrium involving the protein aryl mixed disulfide and protein thiol lies so far toward aryl thiol, while the reverse is true for the protein-SSG mixed disulfide and protein thiol. Another possibility is that co-
METHYLTRANSFERASE
95
BY GSSG
valent bonding of the charged tripeptide group onto Cys15 causes local structural perturbation thereby preventing the attack of Cys-90 or Cys-219 on the mixed disulfide. A drastic conformational change, however, is unlikely because the modified enzyme exhibits spectroscopic properties almost identical to those of the unmodified enzyme. The complete loss of activity that occurs upon disulfide formation between Cys-15 and glutathione appears to reflect the inability of the modified enzyme to bind guanidinoacetate. A previous study has shown that the removal of N-terminal eicosapeptide from rGAMT by trypsinolysis results in an enzyme that is catalytically inactive but weakly binds AdoMet and guanidinoacetate (16). Also, the enzyme in which Cys-15 is cross-linked to Cys-90 or Cys-219 appears to have affinities for substrates not greatly different from those of the native enzyme (8). Therefore, it is highly improbable that the peptide segment around Cys-15 is part of the active site. GAMT binds substrates in a compulsory order: guanidinoacetate binds only to the enzyme-AdoMet binary complex (16). Thus, the binding of AdoMet must induce a conformation change to generate or expose the binding site for guanidinoacetate. It is possible that this process is disrupted in some way in the modified enzyme. The protection against GSSG inactivation afforded by sinefungin plus guanidinoacetate may be interpreted as the result of substantial movement of the portion around Cys-15 upon ternary complex formation. It has been suggested that this portion of the enzyme lies on the protein surface and is flexible (16). The structural change induced by binding of both substrates might be a prerequisite for normal catalysis since the loss of flexibility caused by cross-linking of Cys-15 with either Cys-90 or Cys-219 leads to a large loss of activity (6,8), and the removal of 20 amino acids from the N-terminus results in complete inactivation (16). Based on the results of the present and previous studies it seems likely that the N-terminal segment around Cys-
TABLE
I
Edman Degradation of Radioactive Peptide from the GSSGInactivated Enzyme Labeled with [ 14C]Iodoacetate Cycle 1 2 3 4 5 6 7 8 9 10 11 ’ S-Carboxymethylcysteine.
Residue Phe Ala Pro GUY
Glu Asp Cys (cm)’ GUY Pro Ala Trp
pm01 41.0 96.4 11.7 32.3 22.8 9.5 15.4 12.1 21.5 7.5 3.1
96
KONISHI
AND
15 is functionally linked to the active site. Because of concurrent modification of Cys-15 and Cys-90 by DTNB and iodoacetate (6), it has not been possible to determine the consequence of modification of individual residues. The results obtained here indicate that modification of Cys-15 alone exerts a profound effect on enzyme catalysis. The predominant intracellular low-molecular-weight thiol and disulfide are GSH and GSSG. Recent measurements have shown that the levels of GSH and GSSG in normal liver, assuming their homogeneous distribution, are 8-10 mM and 0.02-0.5 mM, respectively (17, 18). The intracellular concentrations of GSH and GSSG, besides undergoing changes under conditions of induced oxidative stress (19,20), can vary depending on the metabolic state of an animal. For example, fasting causes a significant reduction in the concentration of GSH (to 3-4 mM) (17, 21). Thus, the value of [GSH]/[GSSG] in uiuo could vary between 100 and 400 under physiological conditions. For the activity of a given enzyme to be modulated by the change in the cellular oxidation-reduction state of glutathione, the thiol/disulfide redox equilibrium of the protein sulfhydryls should be near the ilz uiuo [GSH]/[GSSG] ratio, and the thiol/disulfide interchange reaction must be kinetically competent (22). Although it might be possible that the thiol/disulfide interchange is a catalyzed reaction in uiuo, the low overall redox equilibrium constant of GAMT (1.7-1.9) is consistent with the prediction that the enzyme exists predominantly in the reduced, active form under normal metabolic conditions. Thus, it appears unlikely that the GAMT activity is modulated in response to physiological alterations in the [GSH]/[GSSG] ratio. REFERENCES 1. Mudd, S. H., and Poole, J. R. (1975) Metab. Clin. Exp. 24, 721735. 2. Mudd, S. H., Ebert, M. H., and Striver, C. R. (1980) Metab. Clin. Exp. 29, 707-720.
FUJIOKA 3. Ogawa, H., Ishiguro, Y., and Fujioka, M. (1983) Arch. Biochem. Biophys. 226, 265-275. 4. Im, Y. S., Chiang, P. K., and Cantoni, G. L. (1979) J. Biol. Chem. 254, 11,047-11,050. 5. Ogawa, H., Date, T., Gomi, T., Konishi, K., Pitot, H. C., Cantoni, G. L., and Fujioka, M. (1988) Proc. N&l. Acad. Sci. USA 85, 694698. 6. Fujioka, M., Konishi, K., and Takata, Y. (1988) Biochemistry 27, 7658-7664. 7. Cantoni, G. L., and Vignos, P. J., Jr. (1954) J. Biol. Chem. 209, 647-659. 8. Takata, Y., Date, T., and Fujioka, M. (1991) &o&em. J., in press. 9. Gilbert, H. F. (1990) Ado. Enzymol. Relat. Areas Mol. Biol. 63,69172. 10. Brand, E., and Brand, F. C. (1955) in Organic Syntheses (Horning, E. C., Ed.), Collective Vol. 3, pp. 440-442, Wiley, New York. 11. Fujioka, M., and Ishiguro, Y. (1986) J. Biol. Chem. 261,6346-6351. 12. Fujioka, M., and Takata, Y. (1981) J. Biol. C/rem. 256, 1631-1635. 13. Lowry, 0. H., Rosebrough, N. J., Farr, A. L., and Randall, R. J. (1951) J. Biol. Chem. 193,265-275. 14. Gomi, T., Ogawa, H., and Fujioka, M. (1986) J. Biol. Chem. 261, 13,422-13,425. 15. Riddles, P. W., Blakeley, R. L., and Zerner, B. (1983) in Methods in Enzymology (Hire,, C. H. W., and Timasheff, S. N., Eds.), Vol. 91, pp. 49-60, Academic Press, New York. 16. Fujioka, M., Takata, Y., and Gomi, T. (1991) Arch. B&hem. Biophys. 285,181-186. 17. Alpert, A. J., and Gilbert, H. F. (1985) Anal. Biochzm. 144, 553562. 18. Akerboom, T. P. M., Bilzer, M., and Sies, H. (1982) J. Biol. Chem. 257, 4248-4252. 19. Collison, M. W., and Thomas, J. A. (1987) Biochim. Biophys. Acta 928,121-129. 20. Lauterburg, B. H., Smith, C. V., Hughes, H., and Mitchell, J. R. (1984) J. Clin. Znoest. 73, 124-133. 21. Tateishi, N., and Sakamoto, Y. (1983) in Glutathione: Storage, Transport, and Turnover in Mammals (Sakamoto, Y., Higashi, T., and Tateishi, N., Eds.), pp. 3-9, Japan Scientific Society Press, Tokyo. 22. Gilbert, H. F. (1984) in Methods in Enzymology (Weld, F., and Moldave, K., Eds.), Vol. 107, pp. 330-351, Academic Press, Orlando.