Chemical Modification of Chalcone Isomerase by Mercurials and Tetrathionate EVIDENCE FOR A SINGLE CYSTEINE RESIDUE IN THE ACTIVE SITE*

isomerase from soybean is inactivated by stoichiometric amounts ofp-mercuribenzoate HgCI2. Spectral titration of the enzyme with p-mercuriben-zoate indicates that a single thiol group is modified. Treatment of modified enzyme with KCN or thiols results in a complete restoration of enzyme activity demonstrating that the inactivation is not due to irreversible protein denaturation. A product of the enzymatic reaction, naringenin, provides complete kinetic protection against inactivation by both mercurials. The binding constant (33 PM) for naringenin determined from the concentration dependence of the protection agrees with the inhibition constant (34 p ~ ) for naringenin as a competitive inhibitor of the catalytic reaction. This agreement demonstrates that the observed kinetic protection results from the specific binding of naringenin to the active site. Incubation of native chalcone isomerase with sodium tetrathionate (0.1 M) re- slow time-dependent loss of enzymatic activity. The isomerase the 6 that native tertiary structure low reactivity the enzymatic thiol. The modification background counts could be reduced to several hundred counts/min by filtering the stock [3H]NEM solution through a small piece of GF/C filter in a Pasteur pipette prior to the studies. The concentration of [3H]NEM was determined by its absorbance at 302 nm (t = 620 M" cm" (13)). In determining the specific radioactivity, an excess of glutathione was added to the ['HJNEM before spotting an aliquot of this solution onto a filter disk and counting as described above. The reaction with glutathione prevents loss of NEM on drying with the heat lamp and stirring overnight on the orbital shaker is essential in order to obtain reproducible values (14).

Chalcone isomerase from soybean is inactivated by stoichiometric amounts ofp-mercuribenzoate or HgCI2. Spectral titration of the enzyme with p-mercuribenzoate indicates that a single thiol group is modified. Treatment of modified enzyme with KCN or thiols results in a complete restoration of enzyme activity demonstrating that the inactivation is not due to irreversible protein denaturation. A product of the enzymatic reaction, naringenin, provides complete kinetic protection against inactivation by both mercurials. The binding constant (33 PM) for naringenin determined from the concentration dependence of the protection agrees with the inhibition constant (34 p~) for naringenin as a competitive inhibitor of the catalytic reaction. This agreement demonstrates that the observed kinetic protection results from the specific binding of naringenin to the active site. Incubation of native chalcone isomerase with sodium tetrathionate (0.1 M) results in a slow time-dependent loss of enzymatic activity. The inactivation of chalcone isomerase by tetrathionate and N-ethylmaleimide becomes very rapid in the presence of 6 M urea, indicating that the native tertiary structure is responsible for the low reactivity of the enzymatic thiol. The stiochiometric modification of reduced and denatured chalcone isomerase by [3H] N-ethylm~eimide indicates that the enzyme contains only a single cysteine residue and does not contain any disulfides. The evidence presented suggests that the only half-cystine residue in chalcone isomerase is located in the active site and thereby provides the first clue to the location of the active site in chalcone isomerase.
*This research was supported by Grant GM 34832 from the National Institutes of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18

01)
Chalcone isomerase increases the rate constant for cyclization of chalcone (I) by 36 million-fold over that of the spontaneous reaction (5). However, little is known about the residues present in the enzymatic active site or about the mechanism(s) responsible for this rate enhancement. Chalcone isomerase has been shown to lose activity upon treatment with diethylpyrocarbonate (6) or iodoacetamide (7). Although, the enzymes from G. max (8) and P. vulgaris (9) are inhibited by mercurials, the absence of inhibition by sodium tetrathionate was taken as evidence that a s u l~y d r y l group is not essential for activity (1,7,9). The recent availability of homogeneous enzyme (5) has made it possible to probe the active site structure of this enzyme using chemical modification techniques.
In this paper, we demonstrate stoichiometric modification of a single active site thiol of chalcone isomerase by mercurials. Further, the previously reported failure of tetrathionate to inhibit the enzyme (7,9) is shown to result from a slow rate of reaction with the active site thiol rather than the formation of a modified enzyme which is active. The assignment of the only half-cystine residue in the enzyme to the active site provides the first tag for the location of the active site in the primary sequence of the enzyme. isomerase activity was performed at 25 "C by monitoring the loss of the chalcone absorbance at 390 nm using 40 p M 2',4',4-trihydroxychalcone in 50 mM Tris chloride buffer (pH 7.6) containing 1% EtOH. One international unit of enzyme activity is defined as the amount of enzyme which catalyzes the loss of 1 pmol of 2',4',4-trihydroxychalcone in 1 min at 25 "C in 50 mM Tris chloride, pH 7.6 (5).
Kinetics of Inactivation-Chalcone isomerase (20-300 nM) was incubated with pMB (200-2000 nM), HgCl, (20-100 nM), or sodium tetrathionate (1 mM, 100 mM) at 25 "C and at various pH values. Aliquots (2-100 pl) were withdrawn over time and assayed for chalcone isomerase activity. Control incubations lacking only the inactivating reagent were run in parallel. The activity of the experimental incubation solution ( E ) was divided by the activity of the control incubation (E,) to correct for any small activity losses not due to the inactivating reagent. When the concentration of the inactivator was in a large excess over the enzyme, the pseudo-first order rate constant for inactivation was obtained by a linear least squares fit of the plot of ln(E/E,f versus time. When the concentration of inactivator was not in great excess over the enzyme concentration, a second-order rate constant (k,,,J for inactivation of the enzyme was obtained from a nonlinear least squares fit to Equation 2 (12). The parameters a and b are the initial concentrations of the enzyme and inactivator, respectively. D~~e r m~n a~~o n of Covalent Inco~oration ~esulting from ~r e a~m e n~ of Chalcone Isomerase with Radioactive NEM-Chalcone isomerase was denatured in either 6 M urea or 5.1 M guanidine thiocyanate. At selected times after addition of 1.4 mM [3H]NEM (60 cpm/pmol), aliquots (2-8 pg of chalcone isomerase) were removed and added to test tubes containing 20% C1,CCOOH (1 ml) at 0 "C. Approximately 50 fig of bovine serum albumin was added to each tube prior to vacuum filtration through 24-mm Whatman GF/C glass microfiber filters. The test tubes and the filter disks were washed with cold 10% C1,CCOOH (50 ml total? to effect quantitative transfer and to remove noncovalently bound radioactivity. The filter disks were dried under a heat lamp, added to 5 ml of scintillation mixture, and allowed to stand overnight on an orbital shaker before counting with a liquid scintillation counter. The scintillation mixture was prepared by mixing toluene (4 liters), Protosol (100 ml?, Omnifluor (15.2 g), and water (14 ml). Control incubations that lacked enzyme were run in parallel and used to correct for any nonspecific adsorption of radioactivity to the filter disks. These background counts could be reduced to several hundred counts/min by filtering the stock [3H]NEM solution through a small piece of GF/C filter in a Pasteur pipette prior to the studies. The concentration of [3H]NEM was determined by its absorbance at 302 nm ( t = 620 M" cm" (13)). In determining the specific radioactivity, an excess of glutathione was added to the ['HJNEM before spotting an aliquot of this solution onto a filter disk and counting as described above. The reaction with glutathione prevents loss of NEM on drying with the heat lamp and stirring overnight on the orbital shaker is essential in order to obtain reproducible values (14).

~n~t i v a t~o n
by Mercurials-Incubation of chalcone isomerase with mercurials pMB or H&12 results in a time-dependent loss of the enzyme activity ( Fig. 1). In order to obtain rates slow enough to measure conveniently, low concentrations of enzyme and concentrations of the mercurial in only a 2-3fold excess of the enzyme were required. The inactivation of chalcone isomerase followed second-order kinetics, with the rate of inactivation proportional to both the concentration of enzyme and the Concentration of the mercurial (data not shown). The second-order rate constants for inactivation, kinact, are given in Table I. Treatment of chalcone isomerase with a substoichiometric amount of either of the mercurial reagents results in a partial loss of enzyme activity. Fig. 2 shows a direct relationship between the fraction of activity lost and the fraction of a molar equivalent of mercurial added. The addition of one or more molar equivalents of pMB or HgC12 results in a loss of

TABLE I
Rate constants for inactivation of chalcone isomerase by thiol reagents and protection against inactivation by naringenin The second-order rate constants for inactivation of chalcone isomerase by the indicated thiol reagent were determined at 25 "C as described under "Experimental Procedures." The dissociation constant for naringenin at the site from which it protects against inactivation by the mercurials was determined by varying the concentration of naringenin (0-250 pM) and fitting the observed second-order rate constant for inactivation, hob&, by a nonlinear least squares fit to Equation 3. The inhibition constant for naringenin as a competitive inhibitor against the substrate 2',4',4-trihydroxychalcone was determined from a series of initial velocity data at six different substrate concentrations and five different fixed concentrations of naringenin. Titration of chalcone isomerase with xnercurials. A, chalcone isomerase (0.9 p M ) was incubated with the indicated equivalence of HgClz at pH 6.8 in 5 mM phosphate buffer containing 50 mM K,SO,. The activity measured after 3,6, and 9 min was constant, and the average activity is presented as the fraction of the activity of enzyme to which no HgC12 was added. The activities were measured in the standard assay omitting KCN. B, chalcone isomerase (17 p~) was titrated with increasing amounts of pMB at pH 6.8 in 5 mM potassium phosphate buffer containing 50 mM KzSO,. The UVspectrum was measured and the increase in the absorbance at 250 nm resulting from complex formation (0) was determined.
The extinction coefficient for this enzyme mercaptide is 8 X los M" cm". The titration of the enzyme activity with pMB (0) was performed with 3 p~ enzyme as described above for HgC12. more than 99.96 and 99.6% of enzyme activity, respectively (Fig. 2).
The formation of a complex between a cysteine residue and the mercury atom of pMB results in an increase in the absorbance at 250 nm due to formation of a mercaptide (11, 15, 16). This spectral change provides a probe to determine the number of cysteine residues which interact with pMB. Titration of the enzyme shows an increase in the absorbance at 250 nm for only the first equivalent of pMB added, indicating only 1 cysteine residue is modified per mol of enzyme ( Fig. 2B).
Restoration of Catalytic Activity-The standard assay for chalcone isomerase historically contains 50 mM KCN to inhibit possible contaminating chalcone peroxidase activity (7,17). Treatment of chalcone isomerase with either p M 3 or HgC12 leads to a time-dependent loss of all enzyme activity if the assay is conducted without KCN (Fig. 1, solid symbols). However, when the assay of the mercurial-inactivated enzyme is conducted in the presence of 50 mM KCN, the catalysis is initially slow but increases to a value equivalent to that of untreated enzyme. The open symbols in Fig. 1 show that complete activity is recovered by pretreatment of mercurial inactivated enzyme with KCN prior to addition of substrate. Complete recovery of enzyme activity was also afforded by addition of the thiols P-mercaptoethanol and D T T (data not shown).
Kinetic Protec~ion against Inactiva~ion-Addition of narin-genin, a product of the enzyme catalyzed cyclization, to an incubation containing either pMB or HgC12 results in a decrease in the rate of inactivation. A high concentration of naringenin (1.5 mM) reduces the rate constant for inactivation by over 35-fold. A dissociation constant for naringenin at the binding site from which it protects against both mercurials was obtained by varying the concentration of naringenin (0-250 p~) and fitting the observed second-order rate constants for inactivation, kabsdr to Equation 3. Table I summarizes these dissociation constants along with the inhibition constant for naringenin determined from its competitive inhibition of the catalytic reaction under the same experimental conditions. Treatment of Chalcone Isomerase with Sodium Tetrathionate ~~~~S * O~-I n c u b a t i o n of chalcone isomerase with 0.1 M sodium tetrathionate shows no inactivation at pH 5.2 and very slow time-dependent inactivation (tlA = 13 h) at pH 7.5. Over 10-fold faster inactivation is seen when the incubation is performed at pH 9.4, consistent with the modification of a cysteine residue. The second-order rate constants for tetrathionate mediated inactivation of chalcone isomerase are given in Table I.
To test the hypothesis that tetrathionate rapidly modifies the cysteine residue of chalcone isomerase without complete loss of enzyme activity (see "Discussion"), the ability of other thiol reagents to completely inactive the tetrathionate-treated enzyme was explored. Enzyme incubated with sodium tetrathionate (0.1 M at pH 5.2, 7.5, and 9.4) was periodically assayed for enzyme activity in the presence or absence of 0.8 mM pMB. A slow exponential decrease in enzyme activity was observed when the assay was performed in the absence of pMB (Table I). In the presence of pMB, the tetrathionatetreated enzyme was always completely inactivate. HgC& and NEM were also capable of completely inactivating the tetrathionate-treated enzyme (data not shown).
Effect of Urea on Enzyme activity-Chalcone isomerase is not catalytically active when assayed with 2',4',4-trihydroxychalcone (40 pM) in 6 M urea (pH 7.6,50 mM Tris). However, when an aliquot of the enzyme is removed and diluted 100fold into the standard assay solution, rapid turnover of the substrate into product occurs. Approximately 50% of the original specific activity of the enzyme is rapidly recovered. Extended incubation of chalcone isomerase (4.7 pM) at 25 "C in 6 M urea, slowly leads to an irreversible loss of the ability to recover enzyme activity. The half-times for this irreversible loss of enzyme activity at pH 7.5 and 9.6 are 290 and 80 min, respectively. The rapid reactivation of urea-denatured enzyme on dilution into the standard assay provides a convenient method of monitoring the modification of the unfolded enzyme. Addition of sodium tetrathionate (1 mM, pH 7.5 or 9.6) to chalcone isomerase in 6 M urea, followed by a determination of activity on dilution into the standard assay, indicates that tetrathionate rapidly modifies (ts < 1 min) the enzyme leading to a loss of greater than 99% of the enzyme activity.
Cova~nt Incorporat~on of 3H from ~~~~~~ into C~l c o n e Isomerase-Incubation of chalcone isomerase (4.7 p M ) with f3H]NEM (1.4 mM) in the presence of 6 M urea resulted in a rapid loss of all enzyme activity in less than 1 min at pH 7.5 and 9.6. The covalent incorporation of radioactivity into denatured chalcone isomerase was studied under a variety of conditions (see Table 11). At neutral pH, incorporation of radioactivity occurs within a minute without further incorporation over an hour period. However, at pH 9.6 further incorporation occurs after an initial burst. While NEM is selective for thiol groups, it can form stable adducts with amines (18,19). At pH 7 the half-time for reaction of the thiol group of cysteine with NEM (1.4 mMi is only 0.5 s (201, while  (1.4 mM) at the indicated pH and 25 "C. The incorporation of radiolabel into protein was determined after the indicated time using a filter disk method described under "Experimental Procedures." Additions to reaction mixture the half-time for the reaction with the amino group of glycine is 9.3 days (21). As the pH is increased both reactions increase in rate. At pH of 9.6 the half-time for reaction with the amino group of glycine can be calculated' to be about 66 min. The initial very rapid incorporation seen on treatment of denatured enzyme with [3H]NEM is attributable to modification of a cysteine residue, while it is likely that the additional incorporation seen on long incubations at pH 9.6 is a consequence of nonspecific modification of protein amino groups. The slightly less than stoichiometric incorporation of radiolabel from [WJNEM into chalcone isomerase denatured in urea or guanidine thiocyanate (Table 11) may be due to a partial oxidation of the enzymatic sulfhydryl to a sulfenic acid (-SOH) or an intermolecular disulfide (19). Pretreatment of the enzyme with DTT in urea or guanidine thiocyanate prior to the addition of [3H]NEM results in the incorporation of 1 mol eq of radioactivity (Table 11).

DISCUSSION
Enzymatic reactions involving the addition of a nucleophile to a double bond conjugated to a carbonyl group are an important class of biochemical reactions (23). Chalcone isomerase is unique in this class because it catalyzes an intramolecular reaction in which the same substrate molecule contains both the nucleophile as well as the double bond (see Equation 1). Several mechanisms have been proposed for the catalysis of this apparently simple chemical reaction. Hahlbrock and co-workers (24) propose a mechanism involving general acid base catalysis through an intermediate flav-3-en-4-01. In contrast, Boland and Wong (6) suggest nucleophilic catalysis by an active site imidazole, followed by an SN2 displacement by the 2"phenolate of the substrate. A full description of enzymatic catalysis must await an inventory of active site residues that may interact with substrates, intermediates, and products. Chemical modification provides a viable method of identifying amino acid residues that may play an essential role in the catalytic reaction. Enzymatic sulfhydryl groups were the initial target of our investigation, since they are the most reactive nucleophilic amino acid side chain, and there is precedent for their involvement in acidbase catalysis and covalent catalysis in proline racemase (25) Calculated from the data of Leslie (21) using a pK of 9.6 for the amino group of glycine (22). and thymidyla~ synthase (261, respectively.
Cysteine forms very tight complexes with both organic and inorganic mercurials (16,271. Although mercurials are capable of interaction with protein ligands other than cysteine (28-30), they are highly selective for the thiol side chain of cysteine when incubated at low concentration and in the presence of salt (16,19,31). The ability to titrate the activity of chalcone isomerase with substoichiometric amounts of pMB or HgCL and an essentially complete loss of enzyme activity on the addition of stoichiometric amount of these mercurials (Fig. 2) suggests that modification of a single sulfhydryl group is responsible for the loss of enzyme activity. The increase in absorbance at 250 nm on titration of the enzyme with pMB (Fig. 2B) allows a direct spectral monitoring of the complex formation. The leveling off of the absorbance increase after 1 eq of pMB indicates that only a single site of the enzyme is modified. The extinction coefficient for the complex, 8 x lo3 M-' cm", is in agreement with a value of 7.5 X lo3 M" cm" reported for the glutathione mercaptide (321, offering support that the increase in absorbance at 250 nm is indeed due to the formation of an enzymatic mercaptid~. These results provide direct evidence that pMB interacts with a single sulfhydryl group of the enzyme.
The interaction of mercury with chalcone isomerase is very tight since there is complete inactivation when 10 nM HgClz is added to 7 nM enzyme. Therefore, the observed dissociation constant, [~g ] [ E ] /~~. H~] * must be less than 10-s,8 M. If the enzymatic thiolate displaces a chloride ligand of HgCl2, then the calculated3 dissociation constant for the ES"HgC1 complex would be less than M. This very tight interaction argues strongly that the inactivation by HgC1, results from an interaction with a sulfhydryl group.
The inactivation by pMB is not simply due to the steric bulk of the benzoate moiety, since the smaller HgCIz also inactivates the enzyme. Addition of KCN to enzyme that was previously inactivated by either pMB or HgClz results in a rapid and complete restoration of the catalytic activity (Fig.  1). Cyanide forms very tight complexes with inorganic and organic mercurials. The dissociation constants for Hg(CN)Z and MeHgCN are 10-34.7 M* and M which are almost as tight as the analogous complexes with a cysteine thiolate which have dissociation constants of 10-43.57 M' and 10"5.7 M, respectively (16). Presumably, cyanide binds to the mercury atom in the thiolate complex resulting in a release of the sulfhydryl ligand. This ligand exchange mechanism allows the kinetics of the reaction to be facile, and the high concentration of cyanide relative to the enzyme thiolate shifts the equilibrium to favor the free enzyme sulfhydryl and the cyanidechelated mercurial. The regeneration of full enzyme activity indicates that the activity loss on mercaptide formation is not due to an irreversible denaturation of the modified enzyme.
Naringenin, a product of the enzymatic reaction, provides kinetic protection against the loss of enzyme activity by pMB and HgCl,. The dissociation constant determined for naringenin at the site from which it protects agrees with the value of KI for naringenin as a competitive inhibitor of the catalytic reaction (Table I). This agreement indicates that, the protection results from the binding of naringenin at the active site of the enzyme. The complete kinetic protection afforded by binding of naringenin suggests that is provides a direct steric The dissociation constant €or a ES-*HgCl complex was calculated as described by Webb (16), using a pK of 6.74 for chloride dissociation from HgC12 and a pK of 9.0 for a protein thiol. This calculation only yields an estimate of the dissociation constant since a variety of complexes of inorganic mercury exist in aqueous solution (31) and addition to the thiol group.
the divalent H F may coordinate to other active site residues in block between the inactivating reagents and the enzymatic s u l~y~y l group. These results provide evidence that both mercurials modify the active site cysteine residue of chalcone isomerase.
Previous workers have concluded that chalcone isomerase does not contain an essential cysteine residue based on the absence of enzyme inhibition in the presence of up to 0.1 M sodium tetrathionate (1,7,9). The implicit assumption involved in drawing this conclusion was that tetrathionate should very rapidly modify cysteine and lead to loss in activity if a thiol group was essential for catalytic activity. In order to determine if tetrathionate might modify the enzymatic cysteine residue without causing complete loss in enzyme activity, we looked for the ability of tetrathionate to protect against ~nactivation by other thiol reagents which completely inactivate the enzyme. Chalcone isomerase with its cysteine residue modified by tetrathionate should no longer be modifiable by other thiol reagents. If this tetrathionate-modified enzyme retained some catalytic activity, then treatment with another thiol reagent should no longer be able to completely inactivate the enzyme. Enzyme treated with tetrathionate can still be completely inactivated by pMB, demonstrating that tetrathionate does not produce a thiol modified enzyme which retains catalytic activity. The enzyme modified by tetrathionate retains no ( 4 % ) catalytic activity.
Tetrathionate is a slow time-dependent inactivator of the chalcone isomerase. The rate constant for inactivation increases at high pH values as would be expected for reaction with an enzyme thiolate (Table I). The low reactivity of this active site thiol is also exhibited in its reaction with mercurials. Although pMB and HgClz are very potent inactivators of chalcone isomerase with large second-order rate constants for inactivation (Table I), at pH 6.8 pMB inactivates the enzyme 330-fold slower than the reaction of pMB with pmercaptoethanol (32). Incubation of chalcone isomerase in 6 M urea, increases the rate constant for tetrathionate-mediated inactivation of the enzyme by over 77,000-fold (Table I). The rapid inactivation of the urea-denatured enzyme indicates that the three-dimensional structure of the enzyme is responsible for the low reactivity of the cysteine residue in chalcone isomerase.
Several lines of evidence lead us to conclude that chalcone isomerase from G. max contains no disulfides and only a single cysteine residue. Titration of denatured enzyme with [3H]NEM results in a stoichiometric incorporation of radioactivity. The absence of additional incorporation when the enzyme is reduced and denatured strongly argues that the enzyme does not contain any disulfide bonds (Table 11). Additional evidence for only a single cysteine in chalcone isomerase comes from titration with 5,5'-dithiobis(nitrobenzoic acid) in 6 M guanidine hydrochloride (5). The presence of only a single half-cystine was confirmed by HPLC quantitation of cysteic acid following performic acid oxidation and acid hydrolysis of chalcone isomerase (5).
A survey of 207 protein sequences showed a 2.8% frequency for the amino acid cysteine (33), which is over 6-fold greater than that present in chalcone isomerase. Very few other enzymes contain only a single half-cystine residue in their sequence, human carbonic anhydrase (34), streptococcal proteinase (31), and yeast malate dehydrogenase (35) being the only examples of which we are aware. The recently reported cDNA sequence of chalcone isomerase from P. hybrida (3) also shows an unusually low cysteine content ( 2 out of 241 residues or 0.8%). The low s u l~y d r y l content of the enzyme may not be surprising from an evolutionary point of view considering that the chalcone substrate is a powerful electrophile and is known to react in solution with free thiols (36).
Studies presented in this paper demonstrate that soybean chalcone isomerase is one of the rare enzymes that only contains a single cysteine residue and no disulfides. The modification of this residue by thiol reagents leads to complete loss of enzymatic activity, and the native tertiary structure of the enzyme dampens its reactivity relative to free cysteine. The assignment of the only half-cystine residue to the active site provides the first clue to its location in the linear sequence of the enzyme and makes it tempting to postulate that this residue has a role in either the catalysis or regulation of this enzyme. Current studies are directed toward elucidating the function of the active site cysteine residue in chalcone isomerase.