The Active Site of Creatine Kinase - Journal of Biological Chemistry

Vol. 267. No. 4, Issue of February 5, pp. 2173-2178, 1992 Printed in U . S . A .

THE.JOURNAL ‘c:1992 by

OF‘BIOLOGICAL CHEMISTRY The Amencan Society for Biochemistry and Molecular Biology, Inc.

The Active Site of Creatine Kinase AFFINITYLABELING

OF CYSTEINE 282 WITH N-(2,3-EPOXYPROPYL)-N-AMIDINOGLYCINE* (Received for publication, August 16, 1991)

Douglas D. BuechterS, Katalin F. Medzihradszky, Alma L. Burlingame, andGeorge L. Kenyons From the DeDartment of Pharmaceutical Chemistrv. School of Pharmacy, University of California, San Francislo, California 94143

Epoxycreatine (N-(2,3-epoxypropyl)-N-amidino-1987). There are two known cytosolic creatine kinase subglycine) is an affinity label of creatine kinase that units: brain (B) and muscle (M), each of molecular weight irreversiblyand completely inactivates the enzyme approximately 43,000. These associate to form the dimeric (Marletta, M. A., and Kenyon, G. L. (1979) J . Biol. muscle (MM), brain (BB), and heterodimer (MB) isozymes. Chem. 254, 1879-1886). To identify active site resiIn addition to the cytosolic isozymes, there is also an isozyme dues of rabbit muscle creatine kinase, the site of mod- associated with the inner mitochondrial membrane. Creatine ification of it by epoxycreatine has been determined. kinasecDNA from a number of species, including rabbit Separation by high performance liquid chromatogra- muscle (Putney et al., 1984), hasbeen cloned and sequenced phy of a tryptic digestof [‘4C]epoxycreatine-modified (Babbitt et al., 1986; Chen et al., 1991). creatine kinaseyielded two radiolabeled peptides. The The considerable amount of information available regardlarger of theseconsisted of aminoacidsAla-266 ing the kinetic and mechanistic aspects of the reaction catathrough Arg-291 and was labeled with epoxycreatine a t Cys-282. Attempts to purify completely the other lyzed by creatine kinase has been reviewed (Kenyon andReed, labeled peptide were not successful; however, it was 1983). Attempts to delineate the roles played by individual a amino acid residuesin catalysishave beennumerous andhave possible to obtain, by tandem mass spectrometry, collision-induced dissociation spectrum of it from a resulted in the identification of several residues that may be mixture of several peptides. This peptide was a frag- involved in catalysis and/or binding. Chemical modification ment (amino acids Val-279 through Arg-291) of the and NMRspectroscopy studies have implicatedarginine (Bor1975; James,1976), lysine (Jamesand previously identified peptide and was also labeled at dersandRiordan, Cys-282. Model studies with cysteine and epoxycrea- Cohn, 1974; Kassab et al., 1968), and histidine (Clarke and tine have demonstrated that opening of the oxirane Price, 1979; Cook et al., 1981; Rosevear et al., 1981) residues. ring occurs by attack of the cysteine thiolate at the The arginine and lysine side chains are proposed to interact terminal carbon of the epoxide. These results are con- electrostatically with the phosphate groups of the nucleotide, sistent with previous studies on the base labilityof the and the histidineresidue is believed to act as a general acid/ label; however, a carboxyl group in the active site is base to protonate/deprotonate theguanidinium nitrogen that not labeled, as had been previously suggested. These is phosphorylated. A carboxyl group, which must be ionized results provide evidence that Cys-282 is located in or for the binding of both creatine and phosphocreatine, has near the creatine-binding site and will also be impor- been implicated from pH rate studies(Cook et al., 1981). None tant in identifying and delineating the boundaries of of the above groups has been located with any certainty in the active site of creatine kinase. the primary sequence of the enzyme. Creatine kinase also possesses one highly reactive sulfhydryl per subunit. This group can be modified with a number of sulfhydryl-specific reagents (Kenyon and Reed, 1983; MaCreatine kinase (ATP:creatine N-phosphotransferase, EC 2.7.3.2) catalyzes the reversible transfer of a phosphoryl group howald et ul., 1962; O’Sullivan, 1971), often, but not always, with complete loss of activity (der Terrossian and Kassab, from MgATP to creatine, leading to phosphocreatine and MgADP. Creatine kinase plays a critical role in cellular energy 1976; Smith and Kenyon,1974). EPR spectroscopy studie: of metabolism (Saks et al., 1975; Seraydarian and Abbott,1976) enzyme spin-labeled at this sulfhydryl place it within 10 A of and in the transport and utilization of “high energy phos- the metal-binding site (Cohnet al., 1971; Taylor et al., 1969). phate”withinthe cell (Jacobus, 1985). Inadditiontoits This residue has been identified as Cys-282 in the primary physiological role, creatine kinase is important in the diag- sequence of rabbit muscle creatine kinase (Mahowald, 1965; nosis of several diseasestates, including myocardialinfarction Thomson et al., 1964) and is conserved in all known creatine (Wu, 1989) and muscular dystrophy(GruemerandPrior, kinase sequences. The role that Cys-282 plays in catalysis is not known. * This work was supported by United States Public HealthService Epoxycreatine (N-(2,3-epoxypropyl)-N-amidinoglycine)’ Grant AR 17323 (to G. L. K.) and National Science Foundation Grant (Scheme 1, 1) is a creatine-based affinity label of creatine DIR 8700766 andUnitedStatesPublicHealth Service Grant 1979). In a process that follows RR01614 (to A. L. B.). The costs of publication of this article were kinase (Marletta and Kenyon, defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate thisfact. $Present address: Dept. of Biology, MassachusettsInstitute of Technology, Cambridge, MA 02139. § T o whom correspondence should be addressed: Box 0446, Dept. of Pharmaceutical Chemistry, University of California, San Francisco, CA 94143.

’ The abbreviations used are: epoxycreatine, N-(2,3-epoxypropyl)N-amidinoglycine; TPCK, N-tosyl-L-phenylalaninechloromethyl ketone; TLCK, W-p-tosyl-L-lysine chloromethyl ketone; DTT, dithiothreitol;LSIMS, liquidsecondary ion massspectrometry;HPLC, high performance liquid chromatography; CID, collision-induced dissociation; HRMS, high resolution mass spectrometry; HEPES, 4-(2hydroxyethy1)-1-piperazineethanesulfonicacid.

2173

2174

Affinity Labeling of Creatine Kinase

2

4

SCHEME1

5

saturation kinetics, epoxycreatine irreversiblyand completely inactivatescreatinekinase from rabbit muscle with a 1:l stoichiometry. Creatine, MgATP, or a MgADP-creatine-nitrate mixture protect against inactivation. The siteof modification has notbeen determined, although the ease of removal of the label under relatively mild basic conditions suggested that a carboxyl group had been modified (Marletta and Kenyon, 1979). A crystal structure of creatine kinase is not yet available. The identification of active site residues, particularly their location with respect to bound substrates, is an important goal in mechanistic studies of creatine kinase and will also be important in determining the location and delineating the is boundaries of the active sitewhenanx-raystructure obtained. In this paper, we report the identification of the active site residue of rabbit muscle creatine kinase that is labeled by epoxycreatine. Mass spectrometric techniques, including tandem mass spectrometry, were used to determine that epoxycreatine labels Cys-282. Model studies of the reaction of cysteine and epoxycreatine suggest that attackof Cys282 on the oxirane ring is likely to occur at the terminal carbon of the epoxide. The results presented here help to localize Cys-282 with respect to the substrate-binding sites of creatine kinase.

13.9 Hz), 2.51-2.61 (lH, m).LSIMSHRMS observed (MH+), 295.1074; calculated for CsHlsNrOsS, 295.1076. N-(2-HydroxypropyU-N-amidinoglycine (Scheme 1, 4)"Raney nickel (250 ~1of a 50% suspension in water) was washed eight times with distilled water, twice with ethanol, and four additional times with water and was finally resuspended in 100 pl of water. To this was added 6 mg(0.02 mmol) of Structure 2 dissolved in 1 ml of water. The resultant slurry was shaken under 20 p.s.i. of H2 for 2 h. The slurry was centrifuged for 2 min at 15,000 X g, the supernatant was saved, and the pellet was washedtwice with 1 ml of water. The combined washings and supernatant were filtered through washed cotton and taken to dryness i n Vacuo to give 3.2 mg of a white solid containing both alanine and N-(2-hydroxypropyl)-N-amidinoglycine (4) (53% yield). The two products were separated by CIRHPLC, as described above for Structure 2. 'H NMR (for N-(2-hydroxypropyl)-N-amidinoglycine)6 4.0-4.1 ( l H , m), 3.90 (2H, s), 3.2-3.36 (2H, m), 1.16 ('H, d, J = 6.0 Hz). LSIMS HRMS observed (MH+), 176.1035; calculated for C6H1:%N303, 176.095. Purification of Rabbit Muscle Creatine Kinme-Rabbit muscle creatine kinase waspurified by a combination of previously reported methods, including ethanol fractionation (Kuby et al., 1954), Blue Sepharose CL-GB chromatography (Miller et al., 1982), and anion exchange chromatography (Quest et al., 1989). The complete purification is reported elsewhere (Buechter, 1990). Purified enzyme was >95%homogeneous by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and had a specific activity of 90 units/mg. Assay of CreatineKinaseActivityandProteinConcentrationCreatine kinaseactivit,y wasdetermined by the pyruvatekinase/lactic EXPERIMENTALPROCEDURES dehydrogenase coupled assay (Tanzer and Gilvarg, 1959). The couMateriaL~-['~CJCyanamide (57 mCi/mmol) was from American pling enzymes in thisassay were not affected by epoxycreatine under Radiolabeled Chemicals, Inc., St. Louis, MO. Sequencing grade en- the assay conditions. Protein concentrations were determined by the method of Lowry et al. (1951). doproteinase Glu-C was from Boehringer-Mannheim. Centricon-30 Inactivation of Creatine Kinase with Epoxycreatine-Inactivations microconcentrators and Amicon ultrafiltration cells were from Amiwere carried out at 0 "C in 10 mM NaHEPES, pH 7.4, following the con Corp., Danvers, MA. TPCK-treated bovine pancreatic trypsin, TLCK, iodoacetic acid, and Raney nickel were from Sigma. All other procedure of Marletta and Kenyon (1979). The epoxycreatine conreagents andchemicals were from either Sigma or Aldrich. Iodoacetic centration was typically 20 mM, and a creatine kinase concentration of 1 mg/ml was used. Inactivations were allowed to proceed for 10 acid was recrystallized from benzene prior to use. Epoxycreatine andf4C]Epoxycreatine-Epoxycreatine was synthe- half-lives (10 h a t 20 mM epoxycreatine). The activity at this time sized by the method of Marletta and Kenyon (1979). Purified epox- was -1% that of control enzyme that had not been treated with ycreatine was stable for extended periods of time (>6 months) if epoxycreatine. Excess epoxycreatine was removed by repeated ultrastored desiccated at or below -20 "C. ["CIEpoxycreatine was synthe- filtration through either a Centricon-30 or an Amicon YM-30 memwith ('4C]epoxycreatine, ultrafiltration was sized as described (Marletta and Kenyon, 1979), except that thelabel brane. In the inactivation was introduced as ['4C]cyanamide. The purified product hada specific continued until the filtrate showed less than twice the number of background counts (less than 45 cpm). The inactivated enzyme was activity of 0.24 mCi/mmol. lyophilized and stored a t -70 "C. S-[2-Hydroxy-~N-carboxymethyl-N-amidino)-3-aminoprop~~-~Tryptic Digestion of Rabbit Muscle Creatine Kinase-Creatine kicysteine (Scheme I , 2)-~-Cysteine (5.6 mg, 0.05 mmol) was dissolved in 100 pl of water to which 20 pl of 1 N NaOH had been added. To nase (both epoxycreatine-modified and unmodified) was dissolved a t this was added, in 10-pl aliquots,5.0 mg (0.03 mmol) of epoxycreatine 10 mg/ml in alkylation buffer (6 M guanidine HC1, 100 mM Tris-HC1, dissolved in 150 p1of water. An additional 30 p1of 1 N NaOH was 1 mM EDTA, pH 8.3) containing 2 mM freshly added dithiothreitol. added, and the solutionwas stirred at room temperature for 2 h. The After sittingat room temperature for 60 min, sufficient 50 mM reaction was stopped by precipitation of thecrudeproduct with iodoacetic acid in water was added to give a 1.2 X molar excess over a t room acetone. The resultingwhite solid was washed twice withacetone and total thiols. This solution was allowed to sit in the dark dried in vacuo. The crude product was dissolved in water, the p H temperature for 1.5 h, a t which time the alkylationreaction was adjusted to -7 with 0.1 N HC1 and further purified by HPLC on a quenched by the addition of p-mercaptoethanol to 1% (v/v). The mM C18column (Beckman Ultrasphere, H,O, 1 ml/min) to give 6 mg of a quenchedsolution was switched into digestionbuffer(100 NH4HC03,0.1 mM CaC12,p H 8.0) by repeated ultrafiltration through white solid (70% yield). either a Centricon-30 or an Amicon YM-30 membrane. The volume 'H NMR 6 3.92-4.01 ( l H , m), 3.85 (2H, s), 3.79-3.84 (lH, m),3.353.44 (lH, d of d, J = 8.5 Hz, 15.6 Hz), 3.25-3.33 ( l H , d of d, J = 3.3 of the solution was adjusted with digestion buffer to give a protein Hz, 15.6 Hz), 2.90-3.10 (2H, m), 2.66-2.73 ( l H , d of d, J = 4.5 Hz, concentration of 5 mg/ml, and TPCK-treated trypsin (5 mg/ml in

2175

Affinity Labeling of Creatine Kinase 0.1 mM HC1) was added to give a creatine kinase/trypsin ratio of 50/

1. Digestion was allowed to progress at room temperature for 5 h. The reaction was quenched by the addition of TLCK (18pg/ml in water) to give a1.1 X molar excess over trypsin. Digests were lyophilized and storedat -70 "C. HPLC of Tryptic Digests-Samples from tryptic digests of creatine kinase were dissolved in water containing 0.1% trifluoroacetic acid and chromatographed on a Vydac protein/peptide Cls column (4.6 X 250 mm). A linear gradient from 100% solvent A (0.1% trifluoroacetic acid in water) to 70% solvent B (0.08% trifluoroacetic acid in acetonitrile) at a rate of O.7%/min was used. The flow rate was l ml/min, and detection was at 215 nm. Collected fractions were takento dryness on a Speed Vac concentrator (Savant Instruments, Inc.) and stored at -70 "C. Mass Spectrometry-Liquid secondary ion mass spectrometry (LSIMS)and high energy collision-induced dissociation (CID) analysis were performed as described previously (Buechter et al., 1991). N M R Spectroscopy-NMR spectra were recorded at 300 mHz on a General Electric QE-300 spectrometer. All spectra were obtained in D20with reference to the water peak at 4.68 ppm.

300

i"

I B

2501

RESULTS

Site of Modification of Creatine Kinase by EpoxycreatineThe reaction of rabbit muscle creatine kinase withepoxycreatine proceeded with a comparable half-life and to the same extent of inactivation as previously reported (Marletta and Kenyon, 1979). Although the stoichiometry of inactivation wasnotindependentlydetermined,allinactivations were carried out under conditionspreviously determined tolead to 1mol of epoxycreatine bound/mol of creatine kinase subunit. I n addition,approximately 46% of the expectedenzymebound radioactivity (based upon the specific activity of the [''C]epoxycreatine) was recovered in the HPLC separation of the trypticpeptides. This is a reasonable recovery, considering the known instability of the enzyme-bound label. The 5-h tryptic digestion of reduced and carboxymethylated creatine kinase resulted in a completely reproducible peptide map (Fig. 1A) in which over 47 peptides (52 expected) could be separated by reverse phase HPLC. Peptides representing 88% of the sequence of creatine kinase were identified by LSIMS of the separated fractions (data not shown). The chromatogram from the separation of the trypticdigest of creatine kinase modified with ["C]epoxycreatine (Fig. 1B) was identical in appearance with that obtainedfrom unmodified enzyme, with one exception. The peak at a retention time of 46 min in the chromatogram of unmodified creatine kinase (shownby the arrow in Fig. 1A) was shiftedto a shorter retention time (44.5 min) in the chromatogram of modified creatine kinase (shown by an arrow in Fig. 1B). It was expected that this change resulted from the labeling of this peptide by epoxycreatine.However, scintillation counting of the collected fractions from the digest of labeled enzyme revealed not one but two fractions that containedradioactivity. As expected, the peak at44.5 min contained the majority of the counts (75%), whereas the fraction at31.5 min (identified by an arrow in Fig 1B) contained the remainder. There was no apparent change in the appearance of this peak in the two chromatograms. No other significant amounts of radioactivity were detected in the HPLC fractions. LSIMS of the fraction at 46 min in Fig. L4 and of that at 44.5 minin Fig. 1B showed thateachcontained A266GHPFMWNEHLGYVLTC'82PSNLGTGLR(Peptide I) (Buechter et al., 1991). In the case of the peptide from the labeled digest, LSIMS analysis was consistent with modification of this peptide with epoxycreatine. CID analysis of the modified peptide from an endoproteinase Glu-Csubdigestion of this peptideconfirmed its sequence and the fact thatwas it labeled exclusively at Cys-282 (Buechter et al., 1991). These results demonstrated that themajor site of labeling

10

20 Ttme

30

40

50

(rnln.1

FIG. 1. HPLC chromatograms from the separation of the peptides resulting from 5-h tryptic digestions of 250 pgof rabbit muscle creatine kinase. A , unmodified creatine kinase; B, [14C]ep~xycreatine-modifiedcreatine kinase. mA U, milliabsorbance units.

of creatine kinase by epoxycreatine is Cys-282; however, still unaccounted for were the significant number of bound counts in the second radioactivity-containing fraction in the digest of labeled enzyme (Fig. 1B). LSIMSof this fractionindicated that it contained at least three major peptides, all of which could also be found in the corresponding fraction in the digest of unmodified enzyme. There was, however, an ion of low abundance at m/z 1503.8 that had not been observed in the corresponding fraction in the digest of unlabeled enzyme. Its molecular weight was consistent with the epoxycreatine-labeled and noncarboxymethylated peptide 11). V279LTC282PSNLGTGLR (Peptide Several attempts were made to separate further by HPLC the peptides contained in this fraction and to improve the abundance of the ion at m/z 1503.8. These attempts metwith only partial success, andit was not possible toseparate completely the desired peptide from the othermajor peptides. It was, however, possible to obtaina fraction thatgave an ion at m/z 1503.8 of sufficient abundance for CID analysis. The high energy CID spectrum of epoxycreatine-labeled Peptide I1 (Fig. 2) showed abundant C-terminal fragmentions, as well as several N-terminal fragmentions. Peptides thathave basic residues at the C terminus often show preferential charge retention at the C terminus (Johnson et al., 1988). Several ions consistent with the fragmentation of bound epoxycreatine are present, as diagramed in Fig. 2. The presence of epoxycreatine boundat Cys-282 is clearly evident from several of the ion series. For example, the expected a4 ion for this sequence, if Cys-282 were not modified, would be at m/z 389. Instead, the a4 ion is at m / t 562, exactly the shift expected for bound epoxycreatine. The modification is at Cys-282; ions representing the sequence from both the C and N termini up

2176

Affinity Labeling of Creatine Kinase i246 W

x

1346 843 1217 1431

968

Y 1389 1275

2

785

611 610

E17 801

730 714

811

558 557 642 616 599

529 503 481

400 413 472 446 430

311 345 329

229 314 288 272

201 175 159

Val-Le~-Thr-Cy~”Pra-Ser-A~n-Leu-Gly-Thr-Gly-Leu-~r~ a

b

72 213 100

185

286 314

562 590

146 659 687 704

C

d

FIG. 2. CID spectrum of Peptide I1 (see text) from the tryptic digest of epoxycreatine-labeled creatine kinase. Observed fragment ions are listed

931

714 191 130

860 E88 905 817

973

1030 1132 1189 1302 1058 1330 1160 1016 1234 1116 1260

X25

h

100l IL

above the spectrum. Ion series are listed in rows. The inset shows the proposed structure of the label bound at Cys-282 and the observed ions from fragmentation of the label. The peptide fragment ion nomenclature is according to Roepstorff and Fohlman (1984), as modified by Biemann (1988).

20 10 0

200

400

600

800

1000

1200

1400

M W

DISCUSSION to Cys-282 are consistent with the sequence shown, with no additional modified residues. Massspectrometryis a highly efficient method for the The Reaction of Cysteine and Epoxycreatine-The two pos- analysis of samples of biological significance (Burlingame et sible products resulting from the attack of cysteine on the al., 1990). In particular, LSIMS and tandem mass spectromoxirane ring of epoxycreatine (Structures 2 and 3)are shown etryarethemethods of choice for thedetermination of in Scheme I. In solution, a t neutral and basic pH, cysteine covalent modifications of proteins and peptides, including reacts withepoxycreatine to give exclusively Structure 2. post-translational modifications (Settineri et al., 1990) and That Structure 2 has the structure shown and is not the the locations and structuresof other covalently bound small productresulting from attack of cysteine atthemethine molecules (DeWolf et al., 1988; Kaur et al., 1989). Together, carbon of the epoxide (Structure 3 in Scheme 1) rests upon LSIMS and tandem mass spectrometry provide a powerful three considerations. First, nucleophilic addition to epoxides method for the characterization andsequencing of proteins. under basic conditions is expected to proceed via attack of These two techniquescomplementedeachotherinthe the nucleophile at the less hindered position (Pocker et al., analysis of the peptides resulting from the tryptic digestion 1988; Ross, 1950). This would be the terminal carbon of the of epoxycreatine-labeled and unlabeled creatine kinase. The epoxypropyl moiety of epoxycreatine. Second, when the syn- majority of the primarysequence could be mapped by LSIMS thesis of Structure 2 was repeated using ~-[3-’~C]cysteine of the HPLC-separated peptides. Several HPLC fractions (99% enrichment), the ‘H NMR of the product showed en- obviously contained more than one peptide; however, it was hanced W - H coupling to the methylene and methine protonspossible with LSIMS to obtain a molecular ion for each of of the cysteine moiety, as well as longer range W - H coupling them. The shift in retention time of the single peak in the to an additional pair of methyleneprotons;there was no chromatogram of epoxycreatine-labeled creatine kinase (comadditional coupling seen to the methine protonof the epoxy- pare Figs. lA and 1B) suggested that the peptide in this creatine portion of Structure 2. This suggests that two meth- fraction was labeled with epoxycreatine. LSIMS of this fracylene carbons are bonded to the sulfur and therefore nearer tionsupportedthis conclusion. The observation thatthe to the %-labeled carbon. This is the case for Structure 2. labeled peptide did not appear tobe carboxymethylated proAttack at the other carbon of the epoxide would result in the vided the first evidence that modification of Cys-282 had methine carbonbeing nearer to the 13C-labeledcarbon. Third, occurred. This was confirmed from the CID spectrum (Buechdesulfurization of Structure 2 with Raney nickel results in a ter et al., 1991). Because no additional changes were seen in mixture of two products (see “Experimental Procedures”) that the chromatogram, it initially appeared that only this peptide exhibits two methyl group resonances in the ‘H NMR. One was labeled. This would not be unexpected since epoxycreaof these is due to alanine. The other be must due tosecondary tine labels creatine kinase witha 1:l stoichiometry (Marletta alcohol (Structure 4); the isomeric primary alcohol (Structure and Kenyon, 1979). It was a surprise, therefore, to find that 6) expected from desulfurizationof the product resulting from a second HPLC fraction contained a significant amount of attack at the methine carbon of epoxycreatine does not have radioactivity. Attempts to separate further the peptides in a methyl group. Only desulfurization of Structure 2 would be this fraction were onlymarginally successful. However, as shown in Fig.2, this peptide was amenable to analysis by expected to give Structure 4.

Affinity Labeling of Creatine Kinase tandem mass spectrometry and gave an excellent CID spectrum. This allowed the determination of an unambiguous sequence for it andalso identified the site of labeling as Cys282. The use of tandem mass spectrometry was especially critical in this case; the sequencing of peptide mixtures by conventional methods, such as Edman degradation, sometimes gives ambiguous results, especially if the peptide of interest is a minor component. Peptide I1 appears to be a chymotryptic fragmentof Peptide I; it most likely arises from chymotryptic activity associated with the trypsin(or, less likely, from the nearby basic guanidinium moiety of epoxycreatine being recognized bytrypsin). Having determined the amino acid residue that was labeled by epoxycreatine, it was of interest toinvestigate the possible regiochemistry of the opening of the epoxide by Cys-282.This information candefine further thegeometry of the active site and assist in determining the spatial relationshipbetween the epoxycreatine/creatine binding site and thesulfhydryl of Cys282. Unfortunately, attempts todetermine directly the structure of the enzyme-bound label (such as reduction of the putative thioether linkage with Raney nickel) were not successful. The reaction of cysteine and epoxycreatine in solution was therefore chosen as a model of the reaction in the active site of the enzyme. In solution, at neutral and basic pH, opening of the oxirane ring of epoxycreatine occurs by attack of cysteine at the terminal, less hindered carbon of the epoxide. It is likely that the terminal carbon is also the preferred site of nucleophilic attack by Cys-282. However, it is also possible that the regiochemistry of attack by Cys-282 is dictated not only by the steric properties of epoxycreatine but also by the constraints imposed upon the system by the enzyme. Thus, theregiochemistry of the modification may reflect the positioning of the two reacting groups in the active site or may reflect a situation where few constraints areimposed by the enzyme and reaction simply occurs at themore favored site. It should also be noted that theepoxycreatine used in these experimentsis a racemic mixture. These results donot address the question of whether only one enantiomer preferentially binds to and covalently modifies the enzyme. Contrary to the tentative conclusion from previous work (Marletta and Kenyon, 1979), epoxycreatine does not label a carboxyl group in the active site of creatine kinase. The suggestion that itdoes was based on the observation that the bound label was removed from the enzyme under relatively mild basic conditions and thatloss of the label was enhanced by hydroxylamine. Removal of the label from Cys-282 by base can be explained by a simple reversal of the alkylation reaction. This would lead to the free thiolate at Cys-282 and free epoxycreatine. Under basic conditions, epoxycreatine would be hydrolyzed to the diol. The diol was detected upon treatment of the inactivated enzyme with base (Marletta and Kenyon, 1979). This is not only consistent with hydrolysis of an ester linkage as proposed by Marletta and Kenyon but is also consistent with the mechanism of base-catalyzed loss of the thioether-linked label proposed here. The mechanism by which hydroxylamine enhances loss of the label is not known but may be the result of a unique environment within the active site. The role that Cys-282 plays in catalysis by creatine kinase continues to be controversial. A mechanism that hasCys-282 playing an integral part in the chemical steps of phosphoryl transfer has been proposed (Rabin and Watts, 1960; Watts and Rabin, 1962).There hasbeen little experimental evidence in support of this mechanism. Other work has suggested that Cys-282 is either involved in maintaining the correct confor-

2177

mation at the active site (Hooton, 1968) or that its integrity is necessary for the conformational change to the “working” enzyme that occurs upon substrate binding (Markham et al., 1977; O’Sullivan and Cohn, 1968; Reed and Cohn, 1972; Taylor et al., 1969). Modification of Cys-282 with many sulfhydryl-specific reagents leads to the complete loss of activity. However, modification with either a thiomethyl group (Smith andKenyon, 1974) or with a cyano group (der Terrossian and Kassab, 1976) appears to result in the retention of some activity: -20% in the case of the -SSCH3 and -73% for the S-CN derivative. From these results,it hasbeen proposed that modification of the sulfhydryl with bulky or charged groups imposes steric and/or electronic constraints on the enzyme and therefore leads to a loss of activity, but that the smaller, uncharged -SCH3 or -CN groups can be better accommodated by the enzyme without loss of all activity (Smith and Kenyon, 1974; Smith et al., 1975). This proposal has been questioned (Wu et al., 1989; Zhouand Tsou, 1987), and itmay be that the residual activity seen after modification with the cyano or thiomethyl groups is due to the loss of the label, possibly by an intramolecular transfer of the modifying group to a second sulfhydryl (Fawcett et al., 1982; Reddyand Watts, 1979). Site-directed mutagenesis of Cys-282 in rabbit muscle creatine kinase to alanine and serine has been carried out (Chen et al., 1990). Preliminary data show that the alanine mutant (C282A)has -%WO of the activity of the native enzyme and that theserine mutant (C282S) has -%OO. Further characterization of these and otherCys-282 mutants is ongoing. The location of Cys-282 with respect to the substratebinding sites is equally controversial. Some studies have indicated that modification of the sulfhydryl has little effect on nucleotide binding (O’Sullivan and Cohn, 1968; O’Sullivan et al., 1966; Roustan et al., 1970) and therefore that it is likely to be near the creatine-binding site. Alternative proposals suggest that Cys-282 has a significant effect on nucleotide binding (Maggio et al., 1977; Vandest et al., 1980) and is not involved in creatine binding (Wang et al., 1988). Epoxycreatine fulfills all of the requirements of a classical affinity label and is also turned over by the enzyme 15 times for every inactivation event (Marletta and Kenyon, 1979). This suggests that it binds to the creatine-binding site, in a manner similar to creatine itself and, therefore, that Cys-282 lies near the creatine-binding site. The epoxypropyl group is consid; erably 1onger;han the N-methyl group of creatine (-4-4.5 A uersus -1.4 A for total C-C and C-N bond lengths). It is conceivable that this group has a certain amount of conformational flexibility and may be able to overlap partially the nucleotide-binding site. If this is the case, then Cys-282 may lie more at the interface of the two sites, perhapsin an appropriate position to mediate both the conformational change and the “closing down” of the active site that apparently occur upon substrate binding (McLaughlin et al., 1976; Reed and Leyh, 1980). This may help to explain some of the ambiguous results found in previous attempts to localize this residue. In eithercase, the thiol group of Cys-282 is evidently suitably positioned with respect to enzyme-bound epoxycreatine tobe within “striking” distanceof the epoxide ring. Despite these ambiguities, Cys-282 is likely to be important for efficient catalysis, even if it is not directly involved in the chemical steps. The results presented here suggest that Cys282 is near the creatine-binding site. The importance of such information is aptly illustrated in the case of adenylate kinase. A crystal structure of this enzyme has been available for a number of years (Egner et al., 1987; Schulz et al., 1974), and extensive NMR studies have been carried out to investigate the binding of substrates (Fry et al., 1985; Honggao et al.,

2178

Affinity LabelingKinase of Creatine

Hooton, B. T. (1968) Biochemistry 7 , 2063-2071 1990; Mildvan and Fry, 1987). Despite this wealth of infor- Jacobus, W. E. (1985) Annu. Reu. Physiol47, 707-725 mation, thereis still controversy regarding the exactlocation James, T. L. (1976) Biochemistry 15,4724-4730 James, T. L., and Cohn, M. (1974) J. Biol. Chem. 249,2599-2604 of the substrate-binding sites (Fry et al., 1988; Shyy et al., Johnson, R. S., Martin, S. A., and Biemann, K. (1988) Intl. J. Mass Spectrom. 1987; Vetter et al., 1990; Tsai and Yan, 1991) and the roles Ion Processes 86,137-154 Kassab, R., Roustan, C., and Pradel, L. (1968) Biochim. Biophys. Acta 167, that individual amino acid residues play in catalysis (Diederm-21 G ichs and Schulz, 1990; Honggao et al., 1990). One reason for Kaur, S., Hollander, D., Haas, R., and Burlingame, A. L. (1989) J. Biol. Chem. 2 6 4 , 16981-16984 this is that it has been difficult to co-crystallize adenylate Kenyon, G . L., and Reed, G. H. (1983) Adu. Enzymol. Re&. Areas Mol. Biol. kinase with substrates. The most informativework has been 54.367-426 Kuby, S. L., and Lardy, H. A. (1954) J. Biol. Chem. 2 0 9 , 191-201 done with the enzyme complexed with Ap5A, a bisubstrate Lowry, 0.A.,H.,Noda, Rosebrough, N. J., Farr, A. L., and Randall, R. J. (1951) J. Biol. analogue inhibitor. This example highlights the importance Chem. 193,265-275 E. T., Kenyon, G. L., Markham, G. D., and Reed, G. H. (1977) J. Bid. of substrate analogues, such as epoxycreatine and Ap,A, in Maggio, Chem. 2 5 2 , 1202-1207 the evaluation of the bindingof substrates from crystal struc- Mahowald, T. A. (1965) Biochemistry 4 , 732-740 T. A., Noltmann, E. A., and Kuby, S. A. (1962) J. Biol. Chem. 2 3 7 , tures. Thus, the results presented here arelikely to be useful Mahowald, 1535-1548 when a crystal structureof creatine kinasebecomes available. Markham, G. D., Reed, G. H., Maggio, E. T., and Kenyon, G. L. (1977) J. Biol. Chem. 252,1197-1201 of the enzyme with epoxycreatine Marletta, In particular, crystallization M. A,, and Kenyon, G. L. (1979) J. Biol. Chem. 2 5 4 , 1879-1886 covalently bound in the active site may give important in- McLaughlin, A. C., Leigh, J. S., Jr., and Cohn, M. (1976) J. Biol. Chem. 2 5 1 , 2777-2787 sights into the bindingmode of creatine and the positionsof Mildvan, A. S., and Fry, D. C. (1987)Adu. Enzymol. Relnt. Areas Mol. Biol. 5 9 , active site residues relative to bound substrates. Work is in 241-313 J.. Johnson. M.. and Wei. R. (1982) Clin. Chim. Acta 118.67-76 progress, including x-ray crystallography, to further elucidate Miller. O'Sullkan, W. J. (1971) Int. J. Protek Res. 3, 139-148 the role of Cys-282 and otheractive site residues in catalysis O'Sullivan, W. J., and Cohn, M. (1968) J. Biol. Chem. 243,2737-2744 O'Sullivan, W. J., Diefenbach, H., and Cohn, M. (1966) Biochemistry 5 , 2666by creatine kinase. 2672 " Y

would like to thank Carolyn Koo and Dr. Lorenzo Chen for their advice and assistance and F. C. Walls for obtaining t h e CID spectra. Acknowledgments-We

REFERENCES Babbitt, P. C., Kenyon, G. L., Kuntz, I. D., Cohen, F. E., Baxter, J. D., Benfield, P. A., Buskin, J. D., Gilbert, W. A,, Hauschka, S. D., Hossle, J. P., Ordahl, C. P., Pearson, M. L., Perriard, J.-C., Pickering, L. A., Putney, S. D., West, B. L., and Ziven, R. A. (1986) J. Protein Chem. 5,1-14 Biemann, K. (1988) Biomed. Enuiron. Mass Spectrom. 16,99-111 Borders, C. L., Jr., and Riordan,J. F. (1975) Biochemistry 14,4699-4704 Buechter, D. D. (1990) Affinity Lobeling of Creatine Kinase. Ph.D. dissertation, University of California, San Francisco Buechter, D. D., Medzihradszky, K. F., Burlingame, A. L., and Kenyon, G. L. (1991) in Techniques in Protein Chemistry (Villafranca, J., ed) Vol. 11, pp. 537-542, Academic Press, New York Burlingame, A. L., Millington, D. S., Norwood, D. L., and Russell, D. H. (1990) Anal. Chem. 62,268R-303R Chen, L. H., Babbitt, P. C., and Kenyon, G. L. (1990) FASEB J. 4, A2119 Chen, L. H., Babbitt, P. C., Vasquez, J. R., West, B. L., and Kenyon, G.L. (1991) J. Biol. Chem. 2 6 6 , 12053-12057 Clarke, D. E., and Price, N. C. (1979) Biochem. J. 181,467-475 Cohn, M., Diefenbach, H., and Taylor, J. S. (1971) J. Biol. Chem. 2 4 6 , 60376042 Cook, P. F., Kenyon, G. L., and Cleland, W. W. (1981) Biochemistry 20,12041210 der Terrossian, E., and Kassab, R. (1976) Eur. J. Biochem. 70,623-628 DeWolf, W. E., Jr., Carr, S. A,, Varrichio, A. S., Goodhart, P. J., Mentzer, M. A,, Roberts, G. D., Southan, C., Dolle, R. E., andKruse, L. I. (1988) Biochemistry 27,9093-9101 Diederichs, K., and Schulz, G. E. (1990) Biochemistry 29,8138-8144 Egner, U., Tomasselli, A. G., and Schulz, A. G. (1987) J. Mol. Biol. 1 9 5 , 649658 Fawcett, A. H., Keto, A. I., Mackerras, P., Hamilton, S. E., and Zerner, B. (1982) Biochern. Biophys. Res. Commun. 107,302-306 Fry, D. C., Kuby, S. A,, and Mildvan,A. S. (1985) Biochemistry 24,4680-4694 Fry, D. C., Byler, D. M., Susi, H., Brown, E. M., Kuby, S. A., and Mildvan, A. S.(1988) Biochemistry 27,3588-3598 Gruemer, H., and Prior, T. (1987) Clin. Chim. Acta 1 6 2 , 1-18

Yl-

Pockei, Y., Ronald, B. P., and Anderson,K. W. (1988) J . Am. Chem. SOC. 110, 6492-6497 Putney, S., Herlihy, W., Royal, N., Pang, H., Aposhian, H. V., Pickering, L., Belagaje, R., Biemann, K., Page, D., Kuby, S., and Schimmel, P. (1984) J. Biol. Chem. 2 5 9 , 14317-14320 Quest, A. F. G., Eppenberger, H. M., and Wallimann,T. (1989) Enzyme (Basel) 41,33-42 Rabin, B. R., and Watts, D. C. (1960) Nature 1 8 8 , 1163-1165 Reddy, S. R. R., and Watts, D. C. (1979) Biochirn. Biophys. Acta 569,109-113 Reed, G. H., and Cohn, M. (1972) J. Biol. Chern. 247,3073-3081 Reed, G. H., and Leyh, T. S.(1980) Biochemistry 19,5472-5480 Roepstorff, P., and Fohlman, J. (1984) Biomed. Mass S ectrom 11, 601 Rosevear, P. R., Desmeules, P., Kenyon, G., and Milfvan, A. S.(1981) Biochemistry 2 1 , 6155-6164 Ross, W. C. J. (1950) J. Chem. SOC. (Lond.) 7 2 , 2257-2272 Roustan, C., Kassab, R., Pradel, L. A,, andThoai, N. V. (1970) Biochim. Biophys. Acta 167,326-338 Saks, V. A., Chernousova, G. B., Gukovsky, D. E., Smirnov, V. N., andChazov, E. I. (1975) Eur. J. Biochem. 57,273-290 Schulz, G. E., Elzinga, M., Marx, F., and Schirmer, R. H. (1974) Nature 2 5 0 , 120-123 Seraydarian, M. W., and Abbott, B. C. (1976) J. Mol. Cell. Cardiol. 8, 741-746 Settineri, C. A,, Medzihradszky, K. F., Masiarz, F. R., Burlingame, A. L., Chu, C., and George-Nascimento,C. (1990) Biomed. Enuiron. Mass Spectrom. 1 9 , 665-676 Shyy, Y., Tian, G., and Tsai, M.-D. (1987) Biochemistry 2 6 , 6411-6415 Smith, D. J., and Kenyon, G. L. (1974) J. Biol. Chem. 249,3317-3318 Smith, D. J., Maggio, E. T., and Kenyon, G. L. (1975) Biochemistry 14, 766771 Tanzer, M. L., and Gilvarg, C. (1959) J. Biol. Chem. 234,3201-3204 Taylor, J. S., Leigh, J. S., Jr., and Cohn, M. (1969) Proc.Natl. Acad. Sci. U. S. A. 6 4 , 219-226 Thomson, A. R., Eveleigh, J. W., and Miles, B. J. (1964) Nature 2 0 3 , 267-269 Tsai, M.-D., and Yan, H. (1991) Biochemistry 30, 6806;6818 Vandest, P., Labbe, J., and Kassab, R.(1980) Eur..J. Blochem. 104,433-442 Vetter, I. R., Reinstein, J., and Rosch,P. (1990) Blochemstry 2 9 , 7459-7467 Wang, Z., Preiss, B., and Tsou, C. (1988) Biochemistry 27,5095-5100 Watts, D. C., and Rabin, B. R. (1962) Blochem. J. 85,507-516 Wu, A. H. B. (1989) Clin. Chem. 3 5 , 7-13 Wu, H., Yao, Q., and Tsou, C.(1989) Biochim. Biophys. Acta 9 9 7 , 78-82 Yan, H., Shi, Z., and Tsai, M.-D. (1990) Biochemistry 29,6385-6392 Zhou, H., and Tsou, C. (1987) Biochim. Biophys. Acta 9 1 1 136-143 ,