Catalytic Mechanism of Thioltransferase - Semantic Scholar

THEJOURNAL OF BIOLOGICAL CHEMISTRY (Q 1991 by The American Society for Biochemistry and Molecular Biology, Inc

Val. 266, No. 19, Issue of July 5, pp. 12766-12771, 1991 Printed in U. S . A.

Catalytic Mechanism of Thioltransferase* (Received for publication, October 18, 1990, and in revised form, February 12, 1991)

Yanfeng Yang and William W. Wells$ From the Department of Biochemistry, Michigan State University, East Lansing, Michigan 48824

To evaluate potential catalytic mechanism for thiol- ation of a thiol occurs only in the deprotonated thiolateform transferase thiol-disulfide exchange reactions, seven (12, 13), the formation of a mixed disulfide or an intramolecpig liver mutants were constructed by site-directed ular disulfide at theactive siteof thioltransferase should make mutagenesis. All the expressed enzymes, including the enzyme insensitive to alkylating reagents. wild-type and mutants with the exception of the inac(21), we described the construcIn an accompanying article tive mutant, ETT-C22S, were variably inhibited by tion of seven mutant pig liver thioltransferases by site-diiodoacetamide, and similar results were obtained when rected mutagenesis, directly identified the essential amino these enzymes were preincubated with GSH.However, acids at the active center, and characterized these mutants. when preincubated with S-sulfocysteine or hydroxyExchange of the cysteine with a serine a t position 25 caused ethyl disulfide, the activity of the enzymes was totally or partially protected against inhibition by iodoaceta- a n increase rather than adecrease inboth thiol-disulfide mide, with the exception of the mutants, ETT-C25S exchange activity and dehydroascorbic acid (DHA) reductase and ETT-C25A. When simultaneously pretreated with activity of the enzyme and prompted a more complete study GSH and S-sulfocysteine, allenzymes were highlypro- of the catalytic mechanism of native and mutant thioltranstected. Isoelectric focusing analysis of the above prein- ferases. In this article,we report the resultsof studies to investigate cubation mixtures showed that different enzyme-substrate intermediates occurred. Using radioactively the catalytic reactions of thioltransferase by (i) alkylation (ii) isoelectric focusing of enzyme-substrate comlabeled substrates, [U-14C]cystine and [ g l y c i ~ e - 2 - ~ Hinhibition, ] GSH, enzyme-substrate intermediates were detected. plexes, and by (iii) reactions with radioactivelabeled subThese data indicate that reduced thioltransferase strates. reacts first with disulfide substrates, then with a thiol EXPERIMENTALPROCEDURES substrate, e.g. GSH. The formation of either enzymesubstrate mixed disulfide or protein intramolecular Materials disulfide protected the enzyme from inactivation by L-Cystine, dithiothreitol (DTT), GSH, iodoacetamide, and iodoaiodoacetamide. Based on the experimental results, alternative methods ofthe catalytic mechanism for thiol- cetic acid were from Sigma; 2-hydroxylethyldisulfide (HED) was from Aldrich; ~-[U-’~C]cystine (300 mCi/mmol) and [glycine-Z-’H] transferases are proposed. GSH (1 Ci/mmol) were purchasedfromDuPont-NewEngland Nuclear; isoelectric focusing gel and PI marker proteins were from Serva; Safety-solve and 2,5-diphenyloxazole (PPO) were from Research Products InternationalCorp.; dimethyl sulfoxide was obtained from J. T. Baker, Inc.

Thioltransferase, also called glutaredoxin, has beenknown for more than 35 years (I). This low molecular weight, heatPreincubation and Inhibition Studies of Thioltransferases stable cytosolic enzyme is widely distributed in bacteria, yeast, The reduced wild-type and mutant pig liver thioltransferases were and animals, and its thiol-disulfide exchange catalytic properties and primary structure have been studied extensively prepared as described previously (see the accompanying paper (21)). NADPH, thiol- Each of the reduced enzymes (0.06 mM) was incubated with 0.12 mM (2-7). Coupled with glutathione reductase and IAM in the presence of 100 mM sodium phosphate buffer, pH 7.5, a t transferase catalyzesreversible thiol-disulfide exchange re- room temperature. At various times, samples were withdrawn, and actions in the presence of GSH. The whole reaction system thethiol-disulfide exchange activity of each enzyme was assayed mayplay animportant role inthe regulation of enzyme using the standard system (5). The IAM inhibition experiments were activity and/or in maintaining a normal cellular thiol/disul- also performed after preincubating eachenzyme (0.06 mM) with GSH (0.5 mM), CYS-SO, (2.5 mM), CyS-SO; (2.5 mM) + GSH (0.5 mM), fide ratio (8-11). The alkylationreagents, iodoaceticacid (IAA)’ and iodoacetamide (IAM),irreversibly inactivate thiol- and HED (2.5 mM), separately, in the presence of 100 mM sodium phosphate buffer, pH 7.5, for 15 min a t room temperature. transferase by reaction with a single sulfhydryl group at its active site,C y P , i na pH-dependent manner (7). Since alkylIsoelectric Focusing Analysis * This work was supported by United States Public HealthService Grants GM-38634 and CA-51937. 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. $ T o whom correspondence and reprint requests should be addressed. Tel.: 517-353-3978. ’ The abbreviations used are: IAA, iodoacetic acid; IAM, iodoacetamide; Cys-SO;, S-sulfocysteine; enzyme-SO;, S-sulfoenzyme; DTT, dithiothreitol; ETT, expressed wild type thioltransferase; ETT-C22S, expressedmutantthioltransferasewithan S substitution for the original C at position 22 (other mutants are represented similarly); HED, 8-hydroxyethyldisulfide; DHA, dehydroascorbic acid.

T h e wild-type thioltransferase (ETT) and mutants ETT-C25S and ETT-C78S:C82S, 3 pg each, were treated with 2.5 mM HED, 2.5 mM cystine, 2.5 mM Cys-SO, and 0.5 mM GSH + 2.5 mM cys-so;, separately, in 100 mM sodium phosphate buffer, pH 7.5, for 15 min a t room temperature. Each of the incubation mixtures (in a total volume of 10 pl)was then analyzeddirectly by isoelectric focusing gel electrophoresis asdescribed previously (14). Radioactive Labeling Studies Two radioactivelylabeled substrates, [14C]cystine and [“HIGSH wereusedfor the purpose of tracking the formation of enzymesubstrate intermediates. [’4C/Cystine-Each of the reducedwild-type and mutant thiol-

12766

12767

Thioltransferase Mechanism

100 transferases (0.3 mM)was incubated in a volume of 100 pl with a mixture of ["Clcystine (approximately 0.3 pci) and 0.6 mM noniso90 topic cystine in 100 mM sodium phosphate buffer, pH 7.5, for 20 min a t room temperature. Excess labeled substrate was separated from the enzymes on a Sephadex G-25 column (1 X 45 cm) whichwas :80 e equilibrated with 20 mM sodium phosphate buffer, pH 7.5. The catalytically active fractions (or the AZmpeak for mutant, ETT-C22S) E 70 g were collected, and the total radioactivity was measured by liquid scintillation spectrometry. Samples were then concentrated by Cen60 IHETT-CZSS triprep-10concentrators (Amicon), and the proteinconcentration H ETT- C251 r 00 ETT- RZOV W ETT- K27Q and radioactivity were determined. Protein was determined by the A 50 W €11- RE8V. K27Q .METTc7nsazs bicinchoninic acid protein assay protocol according to the manufac.-w turer's direction (Pierce Chemical Co.) with bovine serum albumin 2 40 as standard. ._ ["HIGSH-Eachreduced enzyme (0.3 mM) was incubated ina & 30 volume of 100 p1 with a mixture of 12.5 mM cys-so:, approximately 0.1 pCi of ['HIGSH, and 2.5 mM nonisotopic GSH in the presence of 20 100 mM sodium phosphate buffer, pH 7.5, for 20 min at room temperature. Sephadex G-25 chromatography, concentration of the IO catalytically active fractions, protein analysis, and radioactivity determinations were done as described above. For control experiments, reduced wild-type enzyme (ETT) and the 5 IO 15 20 25 30 35 4 0 45 mutant ETT-C25S were incubated with ['HIGSH in the absence of Time ( m i d the substrate Cys-SO;, or the C25S mutant was pretreated with IAM FIG. 1. IAM inhibition of pig liver thioltransferases. Wild before incubation with ['4C]cystine. The rest of the procedures was type and mutant thioltransferases (0.06 mM each) were incubated the same as described above. with IAM (0.12 mM),separately, in 100 mM sodium phosphate buffer, pH 7.5, at room temperature, andthe thiol-disulfide exchange activity Nonreducing Sodium Dodecyl Sulfate-Polyacrylamide Gel of each enzyme was measured at various times. The activity measured Electrophoresis and Autoradiography The wild-type enzyme (ETT) and the mutant, ETT-C25S, treated at zero time for each enzyme was defined as loo%, and the relative activities as a function of time were compared. The symbol for each with ["Clcystine or [3H]GSH as described above, were analyzed by enzyme is designated in the figure. Each value is the average of two electrophoresis on a 15% sodium dodecyl sulfate-polyacrylamide gel separate experiments. under nonreducing conditions. The gel wasfluorographed as described by Banner and Laskey (15) with an intensifying screen at -70 "C for GSH Cys-SO, GSH and Cys-SO, 2 days.

-

L

0)

z

RESULTS

Preincubation with Inhibition Studies-For pig liver thioltransferase, only the sulhydryl group of CysZ2is the target of the alkylating reagents at pH 7.5 (7, also see the accompanying paper (21)). When incubated with IAM, more than 85% of the thiol-disulfide exchange activities of the wild-type enzyme (ETT)andthemutants, ETT-C25S, ETT-C25A, ETT-K27Q,and ETT-C78S:C82S were inhibited, whereas the mutants of ETT-R26V and ETT-R26V:K27Q still had approximately 70% and 90% activity remaining, respectively (Fig. 1).The latter two mutants, each with a Val exchanged for an Arg at position 26, were not sensitive to IAM inhibition at pH 7.5 since theyhad lost the ability to facilitate the deprotonation of their Cys2' side chain. The inactivation of other expressed enzymes by IAM wasa function of the extent of the available thiolate side chain (-CHZS-) of CysZ2.These results were consistent with those of the previous pK,, measurement experiments described in an accompanying paper (21). A typical thioltransferase-catalyzed thiol-disulfide exchange reaction involves the enzyme, a disulfide substrate (cystine, HED, or a thiosulfate ester,e.g. Cys-SO;) and GSH. The inhibition of such enzyme activity by IAM is dependent upon the availability of the free reduced enzyme with the thiolate form of CysZ2.In order to test the possible catalytic mechanism, each of the wild-type and mutant thioltransferases, except ETT-C22S, was pretreated with eitherGSH, Cys-SO;, or Cys-SO; + GSH, separately, and thenincubated with IAM (Fig. 2). Preincubation of these enzymes with GSH led to results similar to those when reduced enzymes were incubated directly with IAM (Fig. 2, left). These results implied that GSH alone had little effect on reduced enzyme inactivation except for slight promotion of the alkylation reaction, presumably by reducing previously air-oxidized

u 5 IO 15x) 2530354045

Time(min)

FIG. 2. Differential preincubation of pig liver thioltransferases. Each of the wild-type and mutant thioltransferases (0.06 mM) was preincubated with 0.05 mM GSH (left),2.5 mM Cys-SO1 (middle), and Cys-SO; (2.5 mM) GSH (0.5 mM) (right), separately, in the presence of sodium phosphate buffer, pH 7.5, for 15 min at room temperature, then 0.12 mM IAM was added. For the remaining steps see the legend to Fig. 1. Each value is the average of two separate experiments.

+

forms of the enzymes. However, when pretreated with the disulfide-like substrate, Cys-SO:, all enzymes were totally or partially protected from inactivation by IAM except the two mutants, ETT-C25S and ETT-C25A which lack cysteine 25 (Fig. 2, middle). Identical results were obtained using Lcystine instead of Cys-SO: (data not shown). The mutants, ETT-R26V, ETT-R26V:K27Q, and ETT-C78S:C82S retained their total activity by pretreatment with Cys-SO; or L-cystine, possibly by the formation of an intramolecular disulfide between cysteines 22 and 25, which prevented IAM reaction with Cys2*.The wild-type and mutant, ETT-K27Q were partially protectedby the disulfide substrate, suggesting that there were still some free reduced enzyme molecules available to IAM. The mutants, ETT-C25S and ETT-C25A, lacking the ability to establish an intramolecular disulfide bond within their active center, were inhibited by IAM even after the initial formation of a mixed disulfide bond between

12768

Mechanism

Thioltransferase Dm +

""_

-+"

"-"

---

-+ -- +"the enzyme and the substrate.However, when simultaneously cy8," +-+ --Cm""+-+"- +""+ ++-" ++preincubated with Cys-SO; and GSH, all thioltransferases GSH ?"-++ ""+"" -+wererelatively well protected,includingthetwomutants IAM _ _ _ _ _ _ + - _ _ - - ----+ IAA -" - -- - ----+"-+ lacking Cys'" (Fig. 2, right). We believe that the co-existence p I 1 2 3 4 5 6 7 8910111213141518171819 3.50 7,of saturation levels of both substrates promoted the formation 4.40 of the relatively stable enzyme-SG mixed disulfide a t Cys2' 5.30 (see discussion below). 5.90 The enzymes were also pretreated with HED, and, except 6.90 7.3c for the two Cys2"-substituted mutants, all others were fully 7.70 8.3C protected against inactivation by IAM (Fig. 3). Because of the 9.4: 10.6: two exceptions, this protection appears tobe acquired by the ETT ETT- C25S ETT- C78St C82S formation of intramolecular disulfides. Similar studies were doneusingthenative pigliver thioltransferase,andthis FIG. 4. Isoelectric focusing analysis of the ES intermediates enzyme was fully protected when preincubated with cystine, of thioltransferase. The wild-typeenzyme, ETT,andmutants ETT-C25S and ETT-C78SC82Swere separately treated with DTT, Cys-SO;, or HED, but not with GSH (data not shown) in HED, cystine, Cys-SO;, Cys-SO; + GSH, cystine + GSH, and IAM agreementwithGanandWells (16). The only difference as indicated by + (added) and - (not added). The PI value for each between the native and the recombinant enzyme is that the of the enzymes, 3 pg, (either in thefree state or in themodified state) former has an N-acetylated N terminus (14). The different was measured on a Servalyt Precoatisoelectric focusinggel, according sensitivity to IAM of the two enzymes when pretreated with to the manufacturer's instructions. The sample in each lane is indicystine or Cys-SO; suggests that one possible role of acety- cated. lation at the N terminus is protection of the native enzyme IAM-treatedmutant ETT-C78S:C82S, respectively. Noragainst unwantedphysiological thiol alkylation reactions. Isoelectric Focusing Analysis-The inhibition studies de- mally, when treated with 10 mM DTT or HED, the enzyme scribed above couldnot reveal what kindsof enzyme-substrate is fully reduced (thiolate) or oxidized (intramolecular disulfide), respectively, and the PI values are widely different from intermediates were formed. Wehavedemonstratedinan accompanying paper (21) that the exchange of Cys?' with a each other (Fig. 4, lanes 2 versus lane 3, lane 9 versus lane serine caused a significant protein PI shift of 0.5 pH unit. The I O ) . However, when treated with cystine or Cys-SO;, the PI values of ETT and ETT-C25S lay between those of their wild-typeenzyme ETT, mutantETT-C25S,andmutant ETT-C78S:C82S were differentially treated with DTT, HED,reduced and oxidized forms, i.e. the enzymes were neither in disulfide forms.Instead, they cystine, Cys-SO;, and IAM as described under "Experimental their thiolate nor intramolecular appeared to be in mixed disulfide forms, i.e. the charged side Procedures" and analyzed by an isoelectric focusing gel (Fig. 4). Lanes 2 to 8 were the wild-type enzyme,ETT, treated with chain (-CH,S-) of cysteine 22 had been modified by the DTT, HED, cystine, Cys-SO;, cystine + GSH, Cys-SO; + chemical natureof the mixed disulfide. In addition,more then Cys-SOT-treated samples (lane 5 and lane GSH, andIAM, respectively; lanes 9 to I5 were mutant ETT- one band existed in I2), suggesting that both substrate components,i.e. cysteine C25S incubated with DTT, HED, cystine, Cys-SO;, Cys-SO; + GSH, IAM, and IAA, respectively; and lanes 16 to and bisulfite, can form a mixed disulfide or thiosulfate ester 19 were loaded with cystine, Cys-SO;, Cys-SO;/GSH, and with ETT and ETT-CZ5S. The top, middle,and bottom bands represent reducedfreeenzyme, enzyme-SO;, and enzymecysteine disulfideforms, respectively. The mixeddisulfide formed between enzyme and GSHwas also observed in ETT, ETT-C25S, andETT-C78S:C82S in thepresence of disulfide substrates (theupper bands of lanes 6 , 7 , and 18, and lane 13), and theirPI values were fortuitously the same as those of the reduced enzymes due to the net negatively charged peptide, 0-0 ETT GSH.ThePIvalues of IAM-treated enzymes were quite HETT- C25S HETT- C25A similar to those of oxidized ones. This gave additional eviC-0 ETT- R26V dence that IAM reacted with Cys22eliminating its negative 615 ETT- K270 W ETT- R26V~K279 charge as in the case of intramolecular disulfide formation. @O ETT- C78SC82S For mutant ETT-C78S:C82S, the PI value was equal to that of the oxidized formwhentreatedwitheithercystine or Cys-SO;. Therefore, this mutant seems to favor the intramolecular disulfide form. Radioactive Labeling Studies-For further testing of the mixed disulfides between the enzyme and its substrates,two radioactively labeled substrates, ['4C]cystine and ["HIGSH, were used to track the reaction progress. The details of the labeling experiments weredescribed under"Experimental Procedures," and the results are listed in Table I. Enzymesubstrate intermediateswere detected, sinceradioactivity was measured in the collected G-25 protein fractions both before Time (mid and after the concentration of the enzymes. The mutant, FIG.3. Pig liver thioltransferase protection by HED. The ETT-C25S, had the highest specific radioactivity (counts/ protection experiments were performed by pretreating each of the min/pg) both in ['4C]cystine-labeled samples or in["HIGSHwild-type and mutant thioltransferases (0.06 mM) with 2.5 mM HED before the addition of 0.12 mM IAM. The test conditions were the labeledsamples. No labeling by ['HH]GSH wasdetectedin of the active site Cys2'. same as described in the legend of Fig. 2. Each value is the average mutant ETT-C22S due to the absence When labeled by [I4C]cystine, there was no radioactivity of two separate experiments. " "

" "

" "

"

"-

12769

Thioltransferase Mechanism TABLEI Radioactive labelingof thioltransferases Wild-type and mutant pig liver thioltransferases were incubated with ['"CJcystine,[glycine-2-:'H]GSH,or Cys-SO: + ['HIGSH in the presence of 100 mM sodium phosphate buffer, pH 7.5, for 20 min at roomtemperature, theexcess radioactive-labeled substrates were removed by Sephadex G-25, and the samples were concentrated under conditions described in the text. The radioactivity of each enzyme was counted by liquid scintillation spectrometry, and the specific radioactivity was calculated.

directly incubated with ['HJGSH alone(Fig. 5, lunes 3,6, and

7). DISCUSSION

A thiol-disulfideexchange reaction is actually a nucleophilic ionic displacement that takes place either spontaneously or enzymatically, in vitro or in vivo (17, 18).The key aspect of this reaction is the existence of a thiolate anion as an attacking nucleophile (11). The extremely low pK, a t t h eactive site, Substrates cysteine 22, facilitates the catalytic action of thioltransferase. Alkylating reagents (e.g.IAA and IAM) inhibit thioltransferase-catalyzed exchange reactions,by binding to thesulfhydryl group of cysteine 22 (7), but this kind of inhibition only cpmlpg happens when cysteine 22 is in its thiolate form. Whenever cysteine 22 is in either the -SH,mixed disulfide, or intramoETT 140 0 293 lecular disulfide forms, the enzyme is no longer sensitive to 0 ND" 0 ETT-C22S the alkylating reagent. Thus, substrate protection of the enETT-C25S 386 0 2020 zyme from the inhibitionby a n alkylating reagent, aswell as ETT-C25S + IAMh ND 0 ND ETT-R26V 74 ND 114 radioactive labeling of the enzyme, makes it possible to test 137 ND 242 ETT-K27Q the nature of enzyme-substrate intermediates and further to 0 ND 83 ETT-C78S:C82S provide evidence forthe natureof the catalytic mechanismof ND, not determined. thioltransferase. ETT-C25S was pretreatedwith IAM, then with L-[U-'"C]cysteine. Based on the data presented in this study, a model of the catalytic mechanism for thioltransferase is proposed as folETT-C25S E TT lows in Scheme 1. In scheme 1,E, S-, SH, and OH represent " MW 1 2 3 4 5 6 7 thioltransferase, the thiolate anionof cysteine 22, the sulhy(K) a dry1 group of cysteine 25, and the hydroxyl group of S e P of mutant ETT-C25S,respectively, whereas RSSR, RSH, GSH, 11.7 andGSSGrepresent disulfide substrate, reduced product, 1 glutathione, and glutathionedisulfide, respectively. FIG. 5. Autoradiography of ETT and ETT-C25S. Wild-type In this model, the thiolate anion of the cysteine 22 of the (ETT) and mutant (ETT-C25S) thioltransferases were treated with reduced enzyme attacks thedisulfide bond of RSSR to form radioactively labeled substrates, ['"Clcystine or ["HJGSH,as described under "Experimental Procedures." Lanes l to 3 were ETT- a n enzyme-mixed disulfide and a substrate thiol derivative treated with ["C]cystine, Cys-SO; + ["HIGSH, and ["HIGSH, re- (reaction 1).At this point, the enzyme may form an intraspectively. Lanes 4,5, and 7 were ETT-C25S treated with ["Clcystine, molecular disulfide bond between cysteines 22 and 25 releasCys-SOs + ["HIGSH,and ["HJGSH,respectively, whereas lane 6 was ing the reduced half-substrate thiol, and the enzyme intraETT-C25S pretreated with IAM followed by ['"Clcystine. moleculardisulfide bond is displaced by GSH to form an enzyme-glutathione-mixed disulfide (reactions 2 and 3). Aldetected in mutantETT-C78S:C82S. This resultagreed with ternatively, GSH directly displaces the half-substrate toform that of the inhibition studies and further suggested that an intramolecular disulfide was formedin this mutant. Pretreat- the enzyme-glutathione-mixeddisulfide (reaction 4). In either ing ETT-C25S with IAM blocked [I4C]cystine labeling. This case, a second GSH molecule attacks the mixed disulfide to reduced enzyme (reaction 5). is additional evidence for the identification of Cys" as the yield GSSG and regenerate the This model is consistent with the experimentalresults active site, and alkylationat thisposition prevented labeling described herein. First, the order of thethioltransferaseby either substrate. Except for the ETT-C22S mutant, all catalyzed reaction is support by the data, in which the enzyme other tested enzymes were labeled by[3H]GSH in the presence was not protected by GSH against IAM inhibition (Fig. 2, of Cys-SO;. Direct incubation of ETT or ETT-C25S with ["HJGSHwas unreactive, i.e. no isotope wasincorporated into left) and could not be labeled by ["HJGSH (Table I and Fig. of a disulfide substrate.Theseresults the enzyme. This result is also consistent with that of the 5),intheabsence suggest that the reduced enzyme is not protected from alkylinhibition studies, in which GSH alone did not protect the enzymeagainst IAM inactivation.Thus, reduced enzyme ation by GSH. However, GSH did bind to and protect the must react witha disulfide substrate first, then with GSH to enzyme against IAM inhibition in the presence of disulfide regenerate the dithiol or monothiol (ETT-C25S) forms of substrate (Fig. 2, right, Table I, andFig. 5) implying that the reduced enzyme reacts with a disulfide substrate first, folthioltransferase. ETT and the mutant, ETT-C25S, treated with the radio- lowed by GSH. The differential protection of the wild type actively labeled substrate as described above, were subjected and mutant enzymes by disulfide substrate against IAM int o SDS-PAGE under nonreducing conditions, and the dried hibition (Fig. 2, middle) can be explained with this model. In gel was subjected to autoradiography (Fig. 5). When treated the absence of GSH, the reaction progresses only through with ["CJsystine or Cys-SO: ['HJGSH, strong labeling was reactions 1 and 2. The full protection against IAM inhibition mutants, ETT-R26V, ETT-R26V:K27Q, and ETTfoundinthemutantETT-C25S (Fig. 5, lanes 4 and 5 ) , in C78S:C82S is likely due to the formation of an intramolecular indicating a covalent linkage between the enzyme and the disulfide bond, which is more stable than the enzyme-SRrespective labeled substrate. With the same treatment, only faint signals were observed in ETT (Fig. 5, lanes 1 and 2 ) mixeddisulfide bond. The complete inhibition of the two suggesting the formationof some intramolecular disulfide due mutants without cysteine25 (ETT-C25S and ETT-C25A)by 1is reversible, to air oxidation during the electrophoresis process. No labeled IAM can be explained on the basis that reaction and IAMcould remove the enzyme from the reaction by bands could be seen when ETT-C25S was pretreated with reacting with reduced free enzyme and forming a dead end IAM followed by [I4C]cystine or when either enzyme was "

+

12770

Thioltransferase Mechanism

/s

,S C H 2 C 0 ~ $

,SCH2CONH2

&OH) HI

SCHEME I

":[

/'I

IAM

(6)

I

k S .

'sH(oH)

J

3Lz (1)

complex, i.e. pull the reaction to the left (reaction 6). For ETT and ETT-K27Q, the mixed disulfide formed in reaction 1was competitively replaced either by the -SHgroup of CysZ5 to form an intramolecular disulfide (reaction 2) or by the products to reverse the direction of reaction 1. In this situation, the rate of the reversal of reaction 1 was much slower, i.e. less free enzyme was available to IAM (reaction 6), and these two enzymes were only partially inactivated by IAM. Simultaneous preincubation of the enzymes with Cys-SO; and GSH yielded much stronger protection of the enzymes, including the mutants ETT-C25S and ETT-C25A (Fig. 2, right). In addition, the enzymes could be labeled by [3H]GSH in the presence of Cys-SO;. These results indicated the formation of enzyme-SG-mixed disulfide and provide credence to reactions 3 and 5 or 4 and 5 of the proposed model. This newly formed enzyme-SG-mixed disulfide bond may be relatively more stable than the enzyme-half-substrate disulfide bond (i.e. enzyme-SR). It is likely that during the preincubation, the enzyme-catalyzed reaction had reached a steady state in which the enzyme was virtually saturated by substrates, forming predominantly mixed and intramolecular disulfides, so that when IAM was added, little free reduced enzyme was available. However, IAMgradually pulled some of the enzyme molecules from the system at a very slow rate (reaction 6) since partial inhibition of the activity was observed. The existence of the enzyme-substrateintermediateshas been confirmed by isoelectric focusing analysis of the thioltransferases pretreated with substrates and by radioactive labeling of the enzymes. For a native protein, a PI value is normally the sum of its total molecular surface charges, and, therefore, the modification of any surface charge would change the PI value of the protein (19). In a companion paper (21), we have shown the PI differences between the reduced (treated with DTT) and oxidized (treated with HED) thioltransferases. In the current study, when treated with cystine, the wild-type enzyme (ETT) had a PI different from either the reduced (thiolate) form or oxidized (intramolecular disulfide) form (Fig. 4, lane 4 ) and there were twosuch PI forms for the Cys-SOY-treatedenzyme (Fig. 4, lane 5). In addition, the enzyme was labeled by ["C] cystine (Table I and Fig. 5 ) , whichgave evidence for the presence of an enzyme-SR intermediate. The existence of an enzyme-SG complex was also provided in the same way. For mutant, ETT-C78S:C82S, an intramolecular disulfide is formed when treated with disulfide substrates, since it has the same PI as that of the oxidized form (HED-treated), is not sensitive to IAM and is not labeled by [14C]cystine.Thus, we believe that, during the catalytic reactions, the formation of an intramolecular disulfide is an optional reaction dependent on the substrateinvolved. In the absence of GSH, whether reaction 2 proceeds or not depends on the strength of the -CH& nucleophilicity of the sulhydryl group of CysZ5and the thiol of the first product, RSH or RS- (16, 18). If the former is stronger (e.g. in the case of HED as substrate), an intramolecular disulfide is formed (reaction 2). If the latteris stronger (i.e. cysteine or H-SO; derived fromCys-SO; as

/"-

E,

substrate), the reaction will be equilibrated as in reaction 1. In the presence of GSH, the reaction proceeds to reaction 5 from reaction 1 via either reactions 2 and 3 or reaction 4 to complete a catalyticcycle. When treatedsimultaneously with Cys-SO; and GSH, there were two PI-mixed disulfide forms ( i e . enzyme-cysteine and enzyme-SG) for ETT (Fig. 4, lane 7),whereas there was only one PI form (i.e. enzyme-SG) for ETT-C25S (Fig. 4, lune 13). Therefore, the efficiency for displacement of the half-cystine from the first mixed disulfide by GSH is greater for ETT-C25S than for ETT. This might explain why this mutantenzyme has increased thiol-disulfide exchange and DHA reductase activities. Thus, in this model, reaction 3 or 4 may be the rate-limiting reaction. The mechanism of the DHA reductase activity of thioltransferase is not established. However, CysZ2is likely to be the active site for both intrinsic enzymatic activities, and the mechanism of DHA reductase activity is presumably similar to that of a thiol-disulfide exchange reaction, i.e. athiohemiketal intermediate, insteadof a mixed disulfide, followed by displacement with GSH as shown in Scheme 2 (20). Studies are underway to further explore this hypotheses. $H20H HO-C-H

YH,OH

no-c-n

CH2W HO-C-H

& . . S I

GSSG

EWi (OH)

SCHEME I1 Acknowledgments-We thank Drs. Thomas L. Deits and Robert P. Hausinger for helpful discussions and Carol McCutcheon for typing the manuscript. REFERENCES 1. Racker, E. (1985) J. Biol. Chem. 217, 867-874 2. Mannervik, B. (1986) in Thioredoxin and Giutaredoxin System: Structure and Function (Holmgren, A., Branden, C.-T., Jornvall, H., and Sjoberg, B. eds) pp. 349-385, Raven Press, New York 3. Fuchs, J. (1989) in Glutathione: Chemical, Biochemical, and Medical Aspect (Dolphin, D., Poulson, R., and Avramovic, O., eds) Part B, pp. 551-570, John Wiley & Sons, New York 4. Hopper, S., Johnson, R. S., Vath, J. E., and Biemann, K. (1989) J. Biol. Chem. 264,20438-20447 5. Gan, Z.-R., and Wells, W. W. (1987) J. Biol. Chem. 262, 66996703 6. Yang, Y., Gan, Z.-R., and Wells, W. W. (1989) Gene (Amst.) 83, 339-346 7. Gan, Z.-R., and Wells, W. W. (1987) J. Bioi. Chem. 262, 67046707 8. Mannervik, B., and Axelsson, K. (1980) Biochem. J. 190, 125130 9. Mannervik, B. (1980) in Enzymatic Basis ofDetoxication(Jakoby, W. B., ed) Vol. 2, pp. 229-244, Academic Press, New York 10. Ziegler, D. M. (1985) Annu. Reu. Biochem. 54, 305-329

Mechanism

Thioltransferase

11. Gilbert, H. F. (1990) Adu. Enzymol. Relat. Areas Mol. Biol. 63, 69-172 12. Lindley, H. (1960) Biochem. J. 74,577-584 eds) Part 13, Creighton, T. E, (1978) prog, Biophys, Mol, Bioi. 33,231-297 14. Yang, Y., and Wells, W. W. (1990) J. Biol. Chem. 265,589-593 15. Banner,w. M., and Laskey, R. A. (1974) J . Biochem. 469 83-88 16. Gan, Z.-R., and Wells, w. W. (1987) Anal. Biochem. 162,265273

12771

17. Kosower, E. M. (1989) in Glutathione: Chemical, Biochemical, and AspectMedical (Dolphin, D., Poulson, R., and Avramovic, O., A, pp. 104-146, John Wiley & Sons, New York 18. Parker, A. J., and Kharasch, N. (1959) Chem. Reu. 59,583-628 19. Righetti, P. G. (1989) in Protein Structure: A Practical Approach (Creighton, T. E., ed) pp. 23-63, IRL Press, New York 20. Lewis, s. (1976) Vitamin C; Its Molecular Biology and Medical Potential, p. 38, Academic Press, New York 21. Yang, Y., and Wells,W. W. (1991) J . Bioi. Chen. 266, 1275912765