of vol. issue of 25, pp. of in u.s.a. affinity labeling of nucleotide … · 1999-02-01 · potent...

THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1986 by The American Society of Biological Chemists, Inc.

Vol. 261, No. 9, Issue of March 25, pp. 4126-4133,1986 Printed in U.S.A.

Affinity Labeling of Nucleotide-binding Sites on Kinases and Dehydrogenases by Pyridoxal 5’-Diphospho-5’-adenosine*

(Received for publication, September 3, 1985)

James K. TamuraS, Robert D. Rakov, and Richard L. Cross From the Department of Biochemistry and Molecular Bwlogy, State University of New York, Health Science Center at Syracuse, Syracuse, New York 13210

A new adenine nucleotide analog, [’Hlpyridoxal 5’- diphospho-5‘-adenosine (PLP-AMP), has been synthesized. The effectiveness of PLP-AMP as an affinity probe has been tested using a number of nucleotide- binding enzymes. In comparison to reaction with pyridoxal 5’-phosphate, PLP-AMP binds more tightly and exhibits greater specificity of labeling for most enzymes tested.

PLP-AMP is a very potent inhibitor of yeast alcohol dehydrogenase and rabbit muscle pyruvate kinase, with complete inhibition obtained upon incorporation of 1 mol of reagent/mol of catalytic subunit. The reagent is also a potent inhibitor of yeast hexokinase and phosphoglycerate kinase. When modified in the absence of substrates, these enzymes require 2 mol of reagent/mol of active site for complete inhibition. How- ever, when modified in the presence of sugar substrates, this stoichiometry decreases to 1.1 for the hexokinase-glucose complex and 1.4 for the phosphoglycerate kinase. 3-phosphoglycerate complex. The most potent inhibition by PLP-AMP was observed with rabbit muscle adenylate kinase. Half-maximal inhibition was obtained at a concentration of approximately 1 NM. In contrast to these examples, PLP-AMP, as well as pyridoxal 5‘-phosphate, fails to act as a potent or specific inhibitor of beef heart mitochondrial FI-ATP- ase.

The high specificity of labeling and the ability of nucleotide substrates to decrease the rate of inactivation of the kinases and dehydrogenase are consistent with the modification of active site residues. The complete reversibility of both modification and inactivation in the absence of reduction by NaBH4 and the absorption spectra of modified enzymes prior to and following reduction indicate reaction with lysyl residues.

We conclude that PLP-AMP holds considerable promise as an affinity label for exploring the structure and mechanism of nucleotide-binding enzymes.

Reactive nucleotide analogs have provided valuable information on the structure and mechanism of a number of enzymes. The structural similarities of these analogs to natural ligands have allowed specific labeling of essential amino

* This research was supported by Research Grant GM 23152 from the National Institutes of Health, United States Public Health Serv- ice, and by a grant from the Upstate New York Chapter of the American Heart Association. 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 U.S.C. Section 1734 solely to indicate this fact.

$ Postdoctoral fellow of the American Heart Association.

acid residues that participate in substrate binding, catalysis, or allosteric regulation (see Yount, 1975; Baker, 1976; Jakoby and Wilchek, 1977; Colman, 1983). When combined with other primary and tertiary structural information, this type of data can provide a detailed view of the topography of nucleotide-binding sites (Rossmann and Argos, 1981; Min- chiotti et al., 1981; Hollemans et al., 1983; Farley and Faller, 1985).

It has been suggested that positively charged loci, i.e. lysyl (Minchiotti et al., 1981) and/or arginyl residues (Riordan, 1979), may be present at all catalytic and regulatory sites that bind anionic phosphorylated ligands. Most reagents developed for modifying arginyl residues give derivatives that are unsta- ble under conditions used to sequence peptides. However, there are a number of electrophilic reagents known to form stable complexes with lysyl residues. Pyridoxal 5”phosphate has been used to modify lysyl residues on a variety of enzymes that bind phosphorylated substrates (Domschke and Domagk, 1969; Colombo and Marcus, 1974). Reaction of the pyridoxyl carbonyl with the €-amino group of lysine results in formation of a Schiff-base complex which can be reduced by borohydride to a stable product suitable for sequencing work (Schnackerz and Noltmann, 1971, Minchiotti et al., 1981). Although pyridoxal 5”phosphate has proven to be a good affinity analog for sugar phosphates, its specificity for nucleotide-binding sites is often poor, resulting in the need to use high concentrations to achieve inactivation (McKinley-McKee and Mor- ris, 1972; Markland et al., 1975; Koga and Cross, 1982).

In an attempt to produce a better lysyl-specific affinity analog for nucleotide sites, we have:designed and synthesized a new reagent, [3H]pyridoxal 5’-diphospho-5‘-adenosine (PLP-AMP,l see Fig. 1). A general survey of its reactivity with various kinases and dehydrogenases is reported here. From the results, we conclude that PLP-AMP will grove useful for labeling adenine nucleotide- and NAD(H)-binding sites on a large number of enzymes. A preliminary report of this work has been published (Tamura et aL, 1985).

EXPERIMENTAL PROCEDURES

Materials

Adenosine 5’-phosphomorpholidate, pyridoxal 5’-phosphate, pyridine, Sephadex G-10-120 and G-50-80, tri-n-octylamine, 3-phosphoglycerate, NAD+, NADH, AMP, ADP, ATP, Hepes, glutathione, and defatted bovine serum albumin were purchased from Sigma. Silica gel 6060 plates were purchased from Eastman Kodak, and polyethyleneimine cellulose-F plates and CaHz were obtained from Merck.

The abbreviations used are: PLP-AMP, pyridoxal 5’-diphospho- 5”adenosine; PLP, pyridoxal 5”phosphate; Hepes, 4-(2-hydroxy- ethyl)-1-piperazineethanesulfonic acid; F1, soluble, beef heart mitochondrial adenosine triphosphatase; PLP-UMP, pyridoxal 5”diphospho-f~’-uridine; Pyr-AMP, pyridoxal 5‘-phospho-5’-adenosine.

4126

Affinity Labeling of Nucleotide Sites 4127

Dowex 1-X8 (200400 mesh) was supplied by Bio-Rad. NaB[3H]4 was purchased from Research Products International. Aqueous Counting Scintillant (ACS) was purchased from Amersham/Searle. All other chemicals were reagent grade quality.

Crystalline suspensions of rabbit muscle adenylate kinase, yeast 3- phosphoglycerate kinase, yeast hexokinase, and glucose-6-phosphate dehydrogenase from Leuconostoc mesentemides were supplied by Sigma. Crystalline suspensions of rabbit muscle glyceralde-3-phosphate dehydrogenase, yeast alcohol dehydrogenase, and 50% glycerol solutions of hog muscle lactate dehydrogenase and rabbit muscle pyruvate kinase were obtained from Boehringer Mannheim. Beef heart mitochondrial F1-ATPase was prepared by the procedure of Knowles and Penefsky (1972a). With the exception of adenylate kinase, all preparations of enzymes used in the chemical modification studies were found to be free of major contaminants on sodium dodecyl sulfate-polyacrylamide gels.

Methods

Synthesis and Purification of pH]PLP-3H-Labeledpyridoxine 5'- phosphate was prepared by reducing 25 mg of pyridoxal 5'-phosphate (PLP) with 100 mCi NaB[3H]4 (16 Ci/mmol) essentially as described by Stock et al. (1966). All steps were performed under low light conditions. Reoxidation of the product to [3H]PLP was performed according to Johansson et al. (1974), except that the reaction mixture contained 0.7 N HCl in place of 4.4 M HC104, and the pH was adjusted after 15 h to 6.5 with 1 N KOH instead of KHC03 in order to eliminate gas formation in the subsequent column chromatography step (Koga and Cross, 1982). The final mixture was diluted to 200 ml with deionized water and applied to a 1.5 X 16.5-cm column of Dowex 1-X8 (200-400 mesh, acetate form). The [3H]PLP was eluted as described by Raibaud and Goldberg (1974) using 150 ml of a 0-4.4 N linear gradient of acetic acid. The specific activity of the PLP in the pooled fractions was 0.99 Ci/mmol. From the initial 69.5 mCi of 3H- labeled pyridoxine phosphate, 22.5 mCi of [3H]PLP was collected from the column.

Synthesis of ~H]Pyridoxal5'-Diphospho-5'-ademsine-The compound PLP-AMP shown in Fig. 1 was prepared as follows. Carrier PLP (1 mmol) was added to the pooled fractions containing the [3H] PLP and then lyophilized. The residue was suspended in 20 ml of anhydrous pyridine (dehydrated by distillation with storage over CaHz), and 1 mmol tri-n-octylamine was added to aid in solubiliza- tion. The yellow solution was rendered anhydrous by three consecutive evaporations with pyridine. In a separate flask, 0.5 mmol adenosine 5'-phosphomorpholidate was dissolved in 20 ml of pyridine and also rendered anhydrous .by three consecutive evaporations. The adenosine 5'-phosphomorpholidate was dissolved in 20 ml of pyridine, transferred to the flask containing the [3H]PLP, and evaporated once more. Finally, the yellow oily residue was dissolved in 10 ml of pyridine and the reaction continued at room temperature in the dark in a stoppered 100-ml round bottom flask. The progress of the reaction was monitored by thin-layer chromatography on Silica 6060 plates with fluorescent indicator using acetone/ethanol/N&OH/ Hz0 (3:l:l:l) as developer. In this system, PLP, adenosine 5'-phosphomorpholidate, and AMP standards migrated with RF values of 0.42, 0.76, and 0.34, respectively. When a small aliquot of the initial reaction mixture was analyzed by this procedure, only two spots, corresponding to [3H]PLP and adenosine 5'-phosphomorpholidate, were detectable under a UV Mineralight. Two additional spots ap- peared as the reaction proceeded. One lacked 3H label and co-migrated with an AMP standard both on the Silica plate and on a polyethyleneimine cellulose-F plate using 1 M LiCl as the mobile phase. The other spot accounted for 35% of the total 3H label and migrated on silica plates with an RF of 0.69. The remainder of the radioactivity was associated with unreacted [3H]PLP. The ratio of counts in the two radioactive spots remained unchanged beyond 24 h. In preparation for column chromatography, pyridine was evaporated and 8 ml of Hz0 containing 150 mg of lithium acetate dehydrate was added to the orange oily residue. Tri-n-octylamine was extracted by swirling with ether, the upper layer was discarded, and the orange aqueous phase was diluted to 135 ml with HzO.

Purification of PLP-AMP-The diluted aqueous fraction was applied to a 1.5 X 20-cm column of Dowex 2-X8 (Cl- form) and washed with 60 ml of HzO. The products were eluted with 740 ml of 0.003 N HC1 containing a linear gradient of 0-0.31 M LiC1. In order to identify fractions containing PLP-AMP, a small aliquot of each fraction was diluted with 0.1 M potassium phosphate (pH 7), and the absorbance at both 260 nm (adenine) and 388 nm (pyridoxyl) was measured. The

earlier fractions contained two adenine-containing compounds which co-migrated with AMP and adenosine 5'-phosphomorpholidate on polyethyleneimine cellulose-F thin-layer plates. This was followed by PLP, identified on the basis of its UV-visible spectrum and RF values on thin-layer chromatography plates. The last peak contained the product of interest. It had an A2so/A388 ratio of 3.2 and a UV-visible spectrum which closely resembled that of an equimolar mixture of AMP and PLP. The fractions collected between 0.18 and 0.25 M salt (430 and 600 ml of the gradient) were combined and concentrated. As in the case of [3H]PLP and PLP (Raibaud and Goldberg, 19741, the 3H-labeled material eluted from the anion exchange column slightly ahead of the major 260- and 388-nm absorbing material. This necessitated pooling fractions that included some PLP. The mixture was further resolved on a 2.7 X 100-cm column of Sephadex G-10- 120 using 10 mM KC1. In this case the radioactivity co-eluted with the bulk of the 260- and 388-nm absorbing material as expected, since there should be no isotope effect on gel filtration. A shoulder in the 388 nm elution profile, centered near 280 ml and following the major peak at 250 ml, was identified as PLP.

Characterization of PLP-AMP-In order to exclude PLP as a contaminant, only the material eluting between 230 and 250 ml was used. This material has an AZso/Am ratio of 2.8 and absorption maxima at 234, 257, 325 (shoulder) and 392 nm in 0.1 M potassium phosphate (pH 7) (Fig. 1). Thin-layer chromatography on silica plates indicates the presence of a single compound with an RF of 0.70 using acetone/ethanol/N&OH/H20 (3:l:l:l) and an RF of 0.85 with pro- panol/acetic acid/HzO (3:2:2). The purity was also assessed by high- performance liquid chromatograpy using a Cls column, yielding a single 260- and 392-nm absorbing peak.

The presence of an aldehyde group was demonstrated by the elimination of the 392-nm absorption maxima at pH 7 upon reduction with NaBH4, giving a spectrum characteristic of pyridoxine phosphate a t wavelengths longer than 290 nm. Spectral measurements were also made of reaction mixtures containing the compound and various enzymes. For example, under conditions where most added reagent was bound to phosphoglycerate kinase, the compound's absorbance peak at 392 nm (Fig. 2, PLP-AMP) shifted to 412 nm (Fig. 2, PLP- AMP + PGK) indicative of the formation of a Schiff-base complex (Uyeda, 1969; Milhausen and Levy, 1975). Furthermore, reduction with borohydride caused a loss of this peak with the appearance of a new peak at 323 nm (Fig. 2, PLP-AMP + PGK + BHT) characteristic of a pyridoxamine derivative (Uyeda, 1969; Schnackerz and Nolt- mann, 1971). Similar spectra, with peaks centering around 325 nm, were observed for all enzymes tested following reaction with the compound, treatment with borohydride, and removal of unbound ligand. The results indicate reaction of a carbonyl on the pyridoxal ring with protein amino groups.

31P NMR analysis of the compound failed to detect the presence of a primary phosphate. PLP gives a peak at -2.8 ppm while the compound gives a peak at -14.7 ppm relative to the standard, tri- methyl phosphate. The latter resonance peak is characteristic of a diesterfied pyrophosphate group such as that found in NADH and

I .5

1.2

0 0.9 0 c 0 9 8

0.6

0.3

0 2

PLP-AMP

I , 260 320 380 440 )O

Wavelenglh, nm FIG. 1. Structure and absorption spectrum of PLP-AMP in

0.1 M potassium phosphate (pH 7).

4128 Affinity Labeling of Nucleotide Sites

PLP-AMP+PGK

275 325 375 425 475 Wavelength, nm

FIG. 2. Spectral changes accompanying the interaction of PLP-AMP with phosphoglycerate kinase. The reaction mixture contained in a final volume of 0.6 ml at 25 "C in 50 mM Hepes-KOH (pH 8): Curve PGK, 1.3 mg/ml phosphoglycerate kinase; Curve PLP- AMP, 35 p~ PLP-AMP; Curve PLP-AMP + PGK, 1.3 mg/ml phosphoglycerate kinase plus 35 p~ PLP-AMP incubated for 30 min. Curve PLP-AMP + PGK + BH,, the incubation described for the previous curve was stopped at 30 min by addition of 60 mM NaBH4. This was followed 15 s later by passage through a centrifuge column with exchange into Hepes buffer (pH 7.0) and a dilution of the enzyme to 1.2 mg/ml.

UDP-glucose. We conclude that the purified compound is PLP-AMP (Fig. 1).

Stock solutions of [3H]PLP and [3H]PLP-AMP were stored in 50% ethanol a t -20 "C. Prior to each experiment, an aliquots was lyophilized to dryness in a Speed Vac Concentrator (Savant) and dissolved in a small volume of 10 mM Hepes-KOH (pH 8) to give a final concentration of approximately 3 mM. The concentration of both reagents was determined by the absorbance at 388 nm in 0.1 N NaOH, assuming a molar extinction coefficient of 6600 (Peterson and Sober, 1954).

Enzyme Preparation-Prior to use, aliquots of ammonium sulfate suspensions of enzymes were centrifuged and the protein pellets dissolved at approximately 4 mg/ml in 50 mM Hepes-KOH (pH 8) or 50 mM potassium phosphate (pH 8.5). The enzyme solutions were desalted on centrifuge columns (Penefsky, 1977) that were equili- brated with the same buffer, and the effluents were diluted to between 0.5 and 1 mg/ml. F1 was treated in a similar manner except that the buffer was 25 mM Hepes-KOH (pH 8), 1 mM MgSO,, and 0.15 M sucrose. Initially, the enzyme concentrations were determined from the absorbance at 280 nm using et&' values of 0.53 for rabbit muscle adenylate kinase (Noda and Kuby, 1957), 1.26 for yeast alcohol dehydrogenase (Hayes and Velick, 1954), 0.92 for yeast hexokinase (Lazarus et al., 1966), 0.50 for yeast 3-phosphoglycerate kinase (Bucher, 1955), and 0.54 for rabbit muscle pyruvate kinase (Bucher and Pfeiderer, 1955). The enzyme stock solutions were then used to construct standard curves in a modified Lowry assay (Peterson, 1977) for use in determining protein concentrations in experimental samples. F, concentration was determined by the same assay using defatted bovine serum albumin as standard and with division by 1.18 (Penefsky and Warner, 1965) to convert to dry weight. The molecular weights used in calculations were 21,300 for rabbit muscle adenylate kinase (Mahowald et al., 1962), 141,200 for yeast alcohol dehydrogenase (Jornvall, 1977), 100,000 for yeast hexokinase (Colowick, 1973), 44,556 for yeast 3-phosphoglycerate kinase (Perkins et al., 1983), 228,000 for rabbit muscle pyruvate kinase (Cottman et al., 1969), and 347,000 for F1-ATPase (Knowles and Penefsky, 1972b).

Chemical Modification of Enzymes by PLP-AMP and PLP-En- zymes were incubated at 0.5 to 1 mg/ml with PLP-AMP or PLP in the same buffers used in the desalting step. All reactions were performed at 25 "C under low light conditions. At various times of incubation, the inhibitory complexes formed were stabilized by reduction with 60 mM NaBH4 for 15 s. Samples were then applied to centrifuge columns to remove unbound ligand and excess NaBH,. Column effluents were assayed for protein concentration, enzyme activity, and 3H counts.

In most cases, it was necessary to further resolve covalently bound counts in the column effluents by precipitating the protein using the following procedure. Aliquots were added to 1 ml of HZ0 in 4-ml Omni scintillation vials, 100 pl of 0.15% deoxycholate were added, followed after 15 min by 100 pl of 72% trichloroacetic acid. The vials were centrifuged and the supernatant discarded. Protein pellets were dissolved in 150 pl of 0.1 N NaOH and then acidified by addition of 150 pl of 0.15 N acetic acid. Lastly, 3 ml of scintillation mixture was added and the radioactivity counted on a Packard 2425 counter.

In preparation of samples for enzyme activity measurements, an aliquot of each column effluent was diluted approximately 25-fold in 50 mM Hepes-KOH (pH 8) containing 1 mg/ml bovine serum albumin. Due to its cold lability (Penefsky and Warner, 1965), F1-ATPase was assayed immediately while other enzymes could be frozen at -70 "C for later analysis.

Enzyme Assays-All assays were performed in a total volume of 1 ml at 30 "C and initiated by the addition of enzyme.

The activity of yeast alcohol dehydrogenase was measured by following the formation of NADH at 340 nm. The assay medium contained 0.1 M sodium pyrophosphate, 20 mM glycine (pH 8.9), 0.3 mg/ml glutathione, 1 mg/ml bovine serum albumin, 0.5 M ethanol, and 2.5 mM NAD+.

Yeast 3-phosphoglycerate kinase activity was measured by coupling the production of 1,3-diphosphoglycerate to the oxidation of NADH. The assay contained 50 mM Hepes-KOH (pH 7.6), 2 mM ATP, 2.6 mM MgSO,, 4 mM 3-phosphoglycerate, 6 units/ml glyceraldehyde-3- phosphate dehydrogenase, 10 units/ml pyruvate kinase, 2 mM phosphoenolpyruvate, and 0.2 mM NADH.

Rabbit muscle adenylate kinase activity was measured by coupling the production of ADP to the oxidation of NADH in 50 mM Hepes- KOH (pH 81, 0.1 M KCl, 2.5 mM phosphoenolpyruvate, 1 mM ATP, 1 mM AMP, 2 mM MgS04,25 units/ml pyruvate kinase, 7.5 units/ml lactate dehydrogenase, and 0.2 mM NADH.

Rabbit muscle pyruvate kinase activity was measured by coupling the production of pyruvate to the oxidation of NADH in 50 mM Hepes-KOH (pH 8), 0.1 mM KC1, 2.5 mM phosphoenolpyruvate, 3 mM MgSO,, 2 mM ADP, 15 units/ml lactate dehydrogenase, and 0.2 mM NADH.

Yeast hexokinase activity was determined using two different methods. The first coupled the formation of glucose 6-phosphate to the reduction of NADP+. The assay contained 50 mM Hepes-KOH (pH 8), 40 mM glucose, 8 mM MgSO4,2 mM NADP', 1 mM ATP, and 5 units/ml glucose-6-phosphate dehydrogenase. The second method linked ADP production to the oxidation of NADH in 50 mM Hepes- KOH (pH 81, 2.5 mM phosphoenolpyruvate, 25 units/ml pyruvate kinase, 7.5 units/ml lactate dehydrogenase, 3 mM MgS04,l mM ATP, 40 mM glucose, and 0.2 mM NADH. Both assays gave the same results.

The activity of F1-ATPase was determined as previously described (Cross and Nalin, 1982), except that 2 mM ATP and 2.5 mM MgSOl were used.

Analysis of Equilibrium Data-The complete reversibility of both reagent binding and inhibition with omission of the borohydride reduction step and the spectral data presented in Fig. 2 indicate formation of a Schiff-base complex between PLP-AMP and the enzymes tested. In their studies of PLP modification of alcohol dehydrogenase, Chen and Engel (1975) described a procedure for resolving the dissociation ( K d ) and equilibrium (K,) constants for a two-step process in which reagent binding is followed by reversible formation of a covalent derivative. The procedure involves measuring the fractional residual activity a t equilibrium ( R ) as a function of reagent concentration. In our experiments, plots of [PLP-AMPI-l versus (1 - R)-' were linear and yielded values for Kd and K- using the following equation:

RESULTS

Yeast Alcohol Dehydrogenase-Incubation of yeast alcohol dehydrogenase with PLP-AMP in 50 mM Hepes-KOH (pH 8) results in a time-dependent loss of activity that reaches equilibrium within 200 min for all concentrations tested. Maximal inhibition is approached at PLP-AMP concentrations above 80 PM (Fig. 3). Analysis of the equilibrium data

Affinity Labeling of Nucleotide Sites 4129 I I I I

100 I I I I PLP

0 40 80 120 160 200

[Inhibitor], UM

I

0 h IO 20 30 40

Time, rnin Inhibitor IncorporatediEnzyme (mol/mol) FIG. 3. Inactivation of yeast alcohol dehydrogenase by PLP-AMP and PLP. The enzyme was incubated

at 6.5 PM with varying concentrations of [3H]PLP-AMP (0) or [3H]PLP (H) in 50 mM Hepes-KOH (pH 8) for 200 min. The reaction was stopped by addition of 60 mM NaFX&, and after 15 s unbound ligand was removed on centrifuge columns. The column effluents were assayed for enzyme activity, protein concentration, and radioactivity as described under “Methods.” Control samples, treated identically except for the omission of PLP and PLP- AMP, had an activity of 480 pmol.min-’. mg-’ (100% activity). The concentration of free inhibitor plotted on the abscissa was calculated by subtracting the amount of reagent covalently bound to enzyme from the total amount added.

FIG. 4. Substrate protection of yeast alcohol dehydrogenase against PLP-AMP-dependent inactivation. The enzyme was preincubated at 6.5 PM in 50 mM Hepes-KOH (pH 8) for 30 min with 1 mM NADH (A), 1.5 mM NAD+ (V), or without addition (0). The reaction was started by addition of 150 PM [3H]PLP-AMP and terminated by addition of NaBH4 at the times indicated. Unbound ligand was removed and the samples assayed as described under “Methods” and in Fig. 3. The specific activity of control samples lacking PLP-AMP was not affected by preincubation with NAD+ or NADH.

FIG. 5. Specificity of modification and inactivation of yeast alcohol dehydrogenase. Residual activity is plotted against the mol of PLP-AMP (0) or PLP (W) incorporated per mol of enzyme (or per 4 mol of catalytic subunit). Conditions were as described in Fig. 3.

by the procedure described under “Methods” gives a Kd (PLP- AMP) = 130 p~ and an equilibrium constant for interconver- sion of bound PLP-AMP and a Schiff-base complex of 9.3. The equilibrium constant was also measured by comparing the amount of reagent covalently bound to protein after borohydride reduction (a measure of the Schiff-base complex) to the total amount of reagent remaining bound to the dehydrogenase during rapid passage through a centrifuge column (a measure of the complex plus noncovalently bound reagent). A value of 10 was calculated, in good agreement with the former method. The maximal level of inhibition by PLP- AMP in a single treatment is predicted from the equilibrium constant to be 90%. Half-maximal inactivation is obtained at a concentration of free PLP-AMP of 14 p~ (Fig. 3). When enzyme inhibited by 80% is subjected to a second incubation with PLP-AMP followed by borohydride reduction, the residual activity is inhibited by a similar amount (72%) to give a final rate that is only 5.7% of the initial value.

In contrast, PLP is not a potent inhibitor (Fig. 3). At the highest concentration tested (200 p ~ ) , only 6% inhibition is observed. This apparent weak binding is consistent with the results of Chen and Engel (1975) for the horse liver enzyme which has a Kd (PLP) = 2.8 mM and an equilibrium constant = 3.4 for formation of the Schiff-base complex.

The ability of coenzyme to compete with PLP-AMP for binding at the nucleotide site was tested (Fig. 4). NADH at 1 mM provides considerable protection against inactivation while NAD+ at 1.5 mM is less effective. These differences are likely due to the fact that NADH was used at a concentration 24 times its Kd while the concentration of NAD+ was approximately equal to its Kd (Whitehead and Rabin, 1964; and Rashed and Rabin, 1968). Enzyme having as little as 20% residual activity after reaction with PLP-AMP had the same K , for NAD+ (0.4 mM at pH 9) as unmodified enzyme, in

accordance with an all-or-none inactivation by PLP-AMP (data not shown). These results are consistent with a common site of binding for PLP-AMP and the coenzymes.

Further support for specific labeling of the enzyme by PLP- AMP, but not by PLP, is shown in Fig. 5. The stoiohiometry of covalently bound PLP-AMP at complete inactivation extrapolates to 4.0 mol/mol of enzyme. There is considerable evidence that yeast alcohol dehydrogenase is a tetramer of identical polypeptide chains that function independently during catalysis (Briindhn et al., 1975). Therefore, it appears likely that inactivation results from the specific labeling of 1 residue/active site.

Yeast Hexokinase-The concentration dependence for inactivation of yeast hexokinase by PLP-AMP and PLP is shown in Fig. 6. Aliquots of enzyme containing either inhibitor were incubated for sufficient time to allow the level of inhibition to reach equilibrium. In the absence of glucose, 90 p~ PLP-AMP is required for half-maximal inhibition. In comparison, PLP is a much weaker inhibitor with only 40% inhibition observed at the highest concentration tested (240 p ~ ) . When 20 mM glucose is also included, PLP-AMP, but not PLP, becomes a much more potent inhibitor. Analysis of the equilibrium data (see “Methods”) for the hexokinase- glucose complex gives a Kd (PLP-AMP) = 210 p~ and an equilibrium constant for formation of the Schiff-base complex of 10.7. Half-maximal inhibition (55% residual activity) is obtained at 20 ~ L M PLP-AMP (Fig. 6). The ability of glucose to enhance reaction of hexokinase with PLP-AMP was not duplicated by glucose 6-phosphate (data not shown). These results suggest that PLP-AMP binds at the ATP site, since substrate binding is ordered with glucose binding first (Fromm et al., 1964; Ricard et aL, 1972; Wilkinson and Rose, 1979).

Other evidence is consistent with this interpretation. Rates of catalysis by native and PLP-AMP-modified enzyme were


100

80

60

40

20

I I I I I

0 60 120 IS0 240

[Inhibitor], NM

FIG. 6. Inactivation of yeast hexokinase by PLP-AMP and PLP and the effect of glucose. The enzyme was incubated at 6 pM in 50 mM Hepes-KOH (pH 8) with varying concentrations of [3H] PLP-AMP (0, 0) or [3H]PLP (0, W) in the presence (0, 0) or absence (0, H) of 20 mM glucose. Reaction was continued until the degree of inhibition reached equilibrium (30 min with and 90 min without glucose) at which time NaBH4 was added. Unbound ligand was removed and the samples assayed as described. Control samples lacking PLP-AMP and PLP had an activity of 740 pmol . min" . mg" (100% activity). The concentration of free inhibitor is plotted on the abscissa.

I I \ I '. I

0 I 2 3 4

Inhibitor Incorporated/Enzyrne(rnol/mol)

FIG. 7. Specificity of modification and inactivation of yeast hexokinase by PLP-AMP and PLP. Residual activity is plotted against the mol of reagent incorporated per mol of the yeast enzyme (or per 2 mol of catalytic subunit). Data from the experiment reported in Fig. 6, obtained with (0,O) or without (0, W) glucose, are given.

measured using 40 mM glucose and ATP concentrations varying up to 0.5 mM. Enzyme inhibited by 55% had the same K, for ATP (220 p ~ ) as unmodified enzyme. Furthermore, both ATP and ADP provided substrate protection against inactivation by PLP-AMP (data not shown).

The stoichiometry of labeling hexokinase with PLP-AMP and PLP is given in Fig. 7. Glucose not only enhances the rate of reaction between PLP-AMP and hexokinase but also increases the specificity of the reaction. From extrapolation of the linear portion of the inactivation plot, the stoichiometry of incorporated label at 100% inhibition is 3.8 mol/mol of enzyme in the absence of glucose, but only 2.2 in its presence. Since yeast hexokinase is a dimer of catalytic subunits, inactivation in the presence of glucose appears to result from the labeling of a single residue/active site. In contrast, PLP labeling is bss specific and is unchanged by the addition of glucose.

Yeast 3-Phosphoglycerate Kinuse-The reaction of PLP- AMP and PLP with phosphoglycerate kinase has been measured (Fig. 8). In contrast to both alcohol dehydrogenase and hexokinase, the rate of approach to equilibrium inactivation by PLP-AMP in the absence of substrates is relatively fast, with a constant level of inhibition attained by 30 min. The concentration of PLP-AMP giving half-maximal inhibition is

4 p ~ . PLP is much less potent, giving only 30% inhibition at 120 p~ (Fig. 8). Enzyme inhibited to various degrees (up to 75%) by PLP-AMP exhibits the same K, for ATP (0.42 mM) as the native enzyme (data not shown). Addition of ATP or, to a lesser extent, 3-phosphoglycerate at concentrations more than 10-fold higher than their K, values decreases the rate of inactivation by PLP-AMP (Fig. 9). When the stoichiometry of labeling by PLP-AMP is plotted against the remaining activity, the results are linear up to 75% inhibition (Fig. 10). At 100% inactivation, the extrapolated stoichiometry gives 2.0 mol of PLP-AMP/mol of enzyme. When similar studies were performed with PLP in the same concentration range, the data could not be fit to a straight line (Fig. 10). However, it is clear that PLP modification is less specific than PLP- AMP. The specificities of labeling by both inhibitors are unchanged when the reaction is performed in 50 mM potassium phosphate buffer (pH 8.5) for 10 min, conditions previously employed by Markland et al. (1975).

When the modification reaction by PLP-AMP is performed in the presence of 10 mM 3-phosphoglycerate, extrapolation to complete inactivation gives 1.4 mol of PLP-AMP/mol of enzyme (Fig. 10). This increased specificity suggests the possible presence of 2 reactive residues at the catalytic site. Although the other substrate, ATP, also decreases the rate of inactivation by PLP-AMP (Fig. 9), the specificity for inactivation is not improved (data not shown).

Rabbit Muscle Adenylate Kinase-Fig. 11 shows the concentration dependence of inactivation of rabbit muscle adenylate kinase by PLP-AMP and PLP. Half-maximal inhibition oc- curs with either 1 p~ PLP-AMP or 11 p~ PLP. Unlike the other enzymes tested, the maximal level of inhibition ap- proaches lOO%."At a concentration of 0.4 mM, ATP or ADP provides near complete protection against inactivation, whereas AMP causes only a 2.5-fold decrease in rate (Fig. 12). These results are consistent with the modification of essential residues at the catalytic site. The stoichiometry of incorporated label could not be determined due to the presence of contaminating proteins in the adenylate kinase preparation.

Other ' Enzymes-Rabbit muscle pyruvate kinase and the mitochondrial F,-ATPase from beef heart are two examples in which the use of PLP-AMP does not appear to offer an advantage over PLP. Pyruvate kinase is inhibited by both reagents and, in each case, complete inactivation extrapolates to 4 mol/mol of the tetrameric enzyme. These results confirm earlier studies which employed PLP (Johnson and Deal, 1970) and trinitrobenzene sulfonate (Hollenberg et al., 1971).

The mitochondrial F1-ATPase contains a total of 6adenine nucleotide-binding sites (Cross and Nalin, 1982), of which at least two (Choate et al., 1979; Grubmeyer and Penefsky, 1981) and possibly three (Cross et al., 1982; Gresser et al., 1982) are catalytic. Results from the modification reaction using PLP- AMP give an extrapolated value of about 10 mol/mol of enzyme for complete inactivation. The same stoichiometry was found for PLP, confirming an earlier study from this laboratory (Koga and Cross, 1982).

DISCUSSION

The major goal of this study was to design and synthesize a lysyl-specific affinity analog for characterizing nucleotide- binding sites on enzymes. Such a reagent should exploit normal enzyme-nucleotide interactions to maximize its local concentration at the binding site. Providing that the essential residue is properly oriented with respect to the reactive group on the analog, the reaction rate may be considerably faster than with nonessential residues located at the enzyme surface or in other accessible spaces.


I f I

I PLP I a

-

PLP-AMP 20- -

I I I I I I I I I 0 25 50 75 100 125 0 5 IO 15 20 25 0 I 2 3 4

FIG. 8. Inactivation of yeast phosphoglycerate kinase by PLP-AMP and PLP. The enzyme was incubated at 20 p~ with varying concentrations of [3H]PLP-AMP (0) or [3H]PLP (W) in 50 mM Hepes-KOH (pH 8). The reaction was continued until the degree of inhibition approached equilibrium (30 min) at which time NaBH4 was added. Unbound ligand was removed and the samples assayed as described. Control samples lacking PLP-AMP and PLP had an activity of 660 pmol. min-' . mg" (100%). The concentration of free inhibitor is plotted on the abscissa.

FIG. 9. Substrate protection of yeast phosphoglycerate kinase against PLP-AMP-dependent inactivation. The enzyme was preincubated at 20 p~ in 50 mM Hepes-KOH (pH 8), 2 mM MgS04 for 30 min with 10 mM MgATP (A), 10 mM 3-phosphoglycerate (V), or without addition (0). Reaction was started by addition of 50 p~ [3H]PLP-AMP and terminated by addition of N a B b a t the times indicated. Unbound ligand was removed and the samples assayed as described.

FIG. 10. Specificity of modification and inactivation of yeast phosphoglycerate kinase by PLP-AMP

were incubated with varying concentrations of [3H]PLP (W, 0) or [3H]PLP-AMP ( 0 , O ) or [3H]PLP-AMP with and PLP. Residual activity is plotted against the mol reagent incorporated per mol of enzyme. Enzyme samples

10 mM 3-phosphoglycerate (V) in 50 mM Hepes-KOH (pH 8) (W, 0, V) or 50 mM potassium phosphate (pH 8.5) (0,O) as described in Figs. 8 and 9.

[Inhibitor], pM Time, rnin Inhibitor Incorporated/Enzyme (mol/mol)

I I

+ATP or ADP 4

-

-

-

-

0 IO 20 30 40 50 6( [Inhibitor], gM

1 0 5 IO 15 20

Time, rnin FIG. 11. Inactivation of rabbit muscle adenylate kinase by PLP-AMP and PLP. The enzyme was

incubated at 0.5 mg/ml with varying concentrations of [3H]PLP-AMP (0) or [3H]PLP (W) in 50 mM Hepes-KOH (pH 8) for 30 min. The reaction was stopped by addition of 60 mM NaBH4, and after 15 s unbound ligand was removed on centrifuge columns. The column effluents were assayed as described. Control samples lacking PLP- AMP and PLP had an activity of 1200 pmol.min".mg" (100%). The concentration of free inhibitor is plotted on the abscissa.

FIG. 12. Substrate protection of rabbit muscle adenylate kinase against PLP-AMP-dependent inactivation. The enzyme was preincubated at 0.4 mg/ml in 50 mM Hepes-KOH (pH 8) for 10 min with 0.4 mM ATP (A), 0.4 mM ADP (A), 0.4 mM AMP (V), or without addition (0). Reaction was started by addition of 20 p~ PLP- AMP and terminated by addition of NaBH4 at the times indicated. Unbound ligand was removed and the samples assayed as described.

[3H]PLP-AMP has several attractive features for its appli- cation in enzyme modification studies. First, the molecule is identical to ADP up to the point of attachment of a pyridoxyl ring on the P-phosphoryl group. Its unmodified adenosine moiety and negatively charged pyrophosphate group should enhance its specificity for adenine nucleotide-binding sites. Second, the placement of the reactive pyridoxyl group in the phosphate region of the molecule may increase the probability of Schiff-base formation with lysyl residues that normally interact with the phosphoryl groups. Third, the placement of the 3H label on the pyridoxyl moiety will ensure retention of the label even if conditions used to isolate and sequence peptides cause cleavage of the pyrophosphate bond with re-

lease of AMP. Introducing the 3H label with the pyridoxyl group has advantages over reducing a nonradioactive Schiff- base complex with NaEi[3H]4. The latter method can result in nonspecific 3H incorporation into protein and, with frequent large-scale use, can pose a safety hazard. Lastly, [3H]PLP- AMP has the advantage of being highly soluble and fairly stable in aqueous media.

Following our preliminary report on [3H]PLP-AMP (Ta- mura et d., 1985), Tagaya et al. (1985) independently reported the synthesis of nonradioactive PLP-UMP. Instead of using morpholine as the leaving group in the condensation reaction (Moffatt, 1966), they employed diphenyl phosphate (Michel- son, 1964). PLP-UMP was shown to be a potent affinity label


for glycogen synthase and an essential lysyl residue was identified by amino acid sequencing. These results, combined with ours, demonstrate the general utility of this type of structure in chemical modification of nucleotide sites.

The specificity of PLP-AMP for lysyl residues has not been fully assessed in this study since chemical identification of the derivatives has not been performed. However, we anticip- tate that, like PLP, the new reagent will be highly specific for lysine as long as photoactivation is avoided (Rippa and Pon- tremoli, 1969). Several results are consistent with modification of lysyl residues on the enzymes tested in this survey. For each enzyme, both inactivation and labeling by PLP- AMP were fully reversible upon removal of unbound reagent if the borohybride reduction step was omitted. Also, the level of inhibition for most enzymes tested approached a maximum value less than 100%. At saturating levels of PLP-AMP, this observation could be due to an equilibrium between noncovalently and covalently bound reagent, with the latter form stabilized by borohydride treatment. This possibility is supported by the finding that when enzyme is exposed to high conCentrations of PLP-AMP in two sequential steps, each followed by borohydride reduction, the per cent loss of residual activity during the second treatment is similar to the per cent loss of initial activity during the first treatment. Finally, spectral measurements of modified enzymes indicated the presence of a Schiff-base complex before, and a pyridoxamine derivative after, borohydride reduction. In each case, the amount of pyridoxamine derivative estimated from the absorbance at 325 nm was in agreement with the level of tritium- labeled reagent incorporated. Hence, formation of a pyridoxamine derivative appears to be responsible for inhibition. However, these results do not eliminate the possibility of reaction with an N-terminal a-amino group.

Several observations suggest that PLP-AMP is acting as an affinity analog for modifying nucleotide-binding sites on the enzymes tested. Comparisons of K,,, values for nucleotide substrates, measured for native and partially inactivated enzymes, indicate an all-or-none inhibition, consistent with modification of active site residues. Also, for most enzymes tested, PLP-AMP proved to be much more potent and specific than PLP, suggesting enhanced binding due to the presence of the adenosine moiety. Finally, competition for binding between PLP-AMP and the natural nucleotide substrates is demonstrated by the ability of substrates to protect against modification and inactivation.

Results with yeast alcohol dehydrogenase clearly emphasize the greater effectiveness of PLP-AMP compared to PLP. The concentration of unbound PLP-AMP resulting in half-maximal inhibition is 14 PM. PLP at approximately 15 times this concentration gives only 6% inhibition (Fig. 3). From analysis of the equilibrium data, the greater potency of PLP-AMP appears to be due primarily to a higher binding affinity at the active site.

The ability of the natural substrates, NAD' and NADH, to decrease the rate of inactivation and labeling of the dehydrogenase by PLP-AMP is consistent with a common site of binding. Furthermore, the covalent incorporation of 4 mol of PLP-AMP/mol of enzyme required for complete inactivation suggests that each of the four catalytic sites on yeast alcohol dehydrogenase (Briindh et al., 1975) is specifically labeled. In contrast, the specificity of PLP labeling of the yeast enzyme is poor (Fig. 5). This agrees with previous studies reported on the horse liver enzyme (McKinley-McKee and Morris, 1972). Differential labeling studies on the liver enzyme led to the identification of an essential lysyl residue at the active site (Sogin and Plapp, 1975; Eklund et aZ., 1984).

PLP-AMP appears to be a useful reagent for exploring the structure and mechanism of yeast hexokinase. X-ray crystal- lographic studies have revealed that each subunit contains a deep central cleft that divides two lobes (Fletterick et al., 1975). A substantial conformational change results upon binding glucose, causing partial closure of the cleft (Bennett and Steitz, 1978). The enzyme-glucose complex binds Cr(NH,),ATP with 6-fold higher affinity than enzyme alone (Peters and Neet, 1978). These findings, combined with the results of kinetic studies (Koshland, 1958; Fromm et al., 1964; Kaji and Colowick, 1965; DelaFuente et al., 1970; Wilkinson and Rose, 1979), have led to the proposal of an induced-fit mechanism with glucose binding first. Consistent with this type of mechanism, we find that PLP-AMP is approximately four times as potent in the presence of glucose (Fig. 6). Furthermore, glucose enhances the specificity of labeling by PLP-AMP to an extrapolated stoichiometry of 1.1 mol/mol of subunit for complete inactivation. In contrast, PLP is a less potent inhibitor and its reactivity is unaffected by glucose. This clearly demonstrates the contribution of the AMP moiety to the specificity of the PLP-AMP reaction.

These results combined with the ability of ATP and ADP to protect against inactivation suggest the possible presence of a lysyl residue in the ATP binding domain on yeast hexokinase. However, a model of the active site deduced from crystal structures lacks both lysine and arginine side chains (Shoham and Steitz, 1980). This does not rule out participa- tion of a lysyl residue in the ternary complex since the model is based on the structures of binary complexes and a further conformational change is known to occur upon binding ATP to the binary complex (Shoham and Steitz, 1980). Since it has not been possible to obtain crystals of the active ternary complex with MgATP and glucose, PLP-AMP may offer a unique opportunity to study the structure of this complex.

Yeast 3-phosphoglycerate kinase is rapidly inactivated by reaction with PLP-AMP. Half-maximal inactivation is reached at 4 PM PLP-AMP (Fig. 8), and 2.0 mol/mol are required for complete inhibition (Fig. 10). It is possible that both residues labeled by PLP-AMP are at the active site. Binding studies have revealed a high and low affinity site for ATP, but only one site for 3-phosphoglycerate (Scopes, 1978). The low affinity site for ATP may represent binding to the 3-phosphoglycerate site. This would explain the ability of 3- phosphoglycerate to partially protect 1 residue while not preventing modification at the ATP site (Fig. 10).

The presence of two distinct adenine nucleotide-binding sites on adenylate kinase (Noda, 1958; Rhoads and Lowen- stein, 1968) makes this enzyme an attractive system for studying affinity labeling by PLP-AMP. Of all the enzymes tested in our initial survey, rabbit muscle adenylate kinase appears to be the most susceptible to inhibition by PLP-AMP with half-maximal inhibition occurring at 1 FM. In comparison, 11 PM PLP is required to yield a similar degree of inactivation. The near complete protection against inactivation by PLP-AMP afforded by ATP or ADP suggests that the labeling of one or both of the nucleotide-binding sites is responsible for the loss of activity. Since nearly 100% inhibition is achieved compared to a maximum of 80-90% for the other enzymes tested, it is possible that PLP-AMP binds to both sites. If this were the case, and the equilibrium for formation of a Schiff-base complex at each site were the same as for the other enzymes tested, then the probability of a t least one molecule of reagent being in a form that can be trapped by borohydride reduction would be very high. Alter- natively, PLP-AMP may interact with a single site where the equilibrium constant strongly favors formation of the Schiff-


base complex. Unfortunately, the stoichiometry of labeling could not be measured accurately due to the presence of impurities in the enzyme preparation used.

Results presented in this paper demonstrate the considerable promise of the new adenine nucleotide analog, PLP- AMP, for characterizing ATP/ADP- and NAD(H)-binding sites in proteins. However, it is anticipated that this reagent will not work with all nucleotide-binding enzymes. Its useful- ness will depend on whether it binds and, if bound, whether an essential residue is properly oriented with respect to the reactive pyridoxyl carbonyl. Some enzymes may not tolerate the pyridoTy1 moiety or the greater length of the reagent (22.5 uersw 19 A for ATP based on space filling models of the anti conformation). This may be the case for the mitochondrial F1-ATPase since PLP-AMP and PLP were found to be equally poor in terms of potency and specificity. We have recently synthesizedpyridoxal5’-phospho-5’-adenosine (Pyr- AMP), which is closer in length to ATP. Together, these two reagents may provide a convenient method of introducing a radioactive tag for isolating and sequencing peptides derived from nucleotide binding domains in enzymes.

Acknowledgments-We wish to thank Drs. George Levy, Anirban Banejee, and Tadeusz Holak of Syracuse University for their assist- ance in collecting “P NMR spectra.

REFERENCES

Baker, B. R. (1976) Design of Active-Site Directed Irreversible Enzyme

Bennett, W. S., Jr., and Steitz, T. A. (1978) Proc. Natl. Acad. Sci. U.

Briindhn, C.-I., Jornvall, H., Eklund, H., and Furugren, B. (1975) in The Enzymes (Boyer, P. D., ed) 3rd Ed., Vol. 11, pp. 103-190, Academic Press, New York

Inhibitors, Wiley, New York

S. A. 75,4848-4852

Biicher, T. (1955) Methods Enzymol. 1,415-422 Bucher, T., and Pfeiderer, G. (1955) Methods Enzymol. 1 , 435-440 Chen, S.-S., and Engel, P. C. (1975) Biochem. J. 149,627-635 Choate, G. L., Hutton, R. L., and Boyer, P. D. (1979) J. Bwl. Chem.

Colman, R. F. (1983) Annu. Rev. Biochem. 52,67-91 Colombo, G., and Marcus, F. (1974) Biochemistry 13,3085-3091 Colowick, S. P. (1973) in The Enzymes (Boyer, P. D., ed) 3rd Ed.,

Cottman, G. L., Hollenberg, P. F., and Coon, M. J. (1969) J. Bwl.

Cross, R. L., and Nalin, C. M. (1982) J. Biol. Chem. 2 5 7 , 2874-2881 Cross, R. L., Grubmeyer, C., and Penefsky, H. S. (1982) J. Biol. Chem.

DelaFuente, G., Lagunas, R., and Sols, A. (1970) Eur. J. Biochem.

Domschke, V. W., and Domagk, G. F. (1969) Hoppe-Seyler’s 2.

Eklund, H., Samama, J.-P., and Jones, T. A. (1984) Biochemistry 2 3 ,

Farley, R. A., and Faller, L. D. (1985) J. Biol. Chem. 260,3899-3901 Fletterick, R. J., Bates, D. J., and Steitz, T. A. (1975) Proc. Natl.

Fromm, H. J., Silverstein, E., and Boyer, P. D. (1964) J. Bid. Chem.

Gresser, M. J., Myers, J. A., and Boyer, P. D. (1982) J. Biol. Chem.

Grubmeyer, C., and Penefsky, H. S. (1981) J. Bwl. Chem. 256,3718-

Hayes, G. E., and Velick, S. F. (1954) J. Bwl. Chem. 2 0 7 , 225-244 Hollemans, M., Runswick, M. J., Fearnley, I. M., and Walker, J. E.

254,286-290

Vol9, pp. 1-47, Academic Press, New York

Chem. 244,1481-1486

257,12101-12105

16,226-233

Physiol. Chem. 350,1111-1116

5982-5996

Acad. Sci. U. S. A. 7 2 , 38-42

239,3645-3652

257,12030-12038

3727

(1983) J. Biol. Chem. 258,9307-9313 Hollenberg, P. F., Flashner, M., and Coon, M. J. (1971) J. Biol. Chem.

Jakoby, W. B., and Wilchek, M. (1977) Methods Enzymol. 46,l-774 Johansson, S., Lindstedt, S., and Tiselius, H.-G. (1974) J. Biol. Chem.

Johnson, G. S., and Deal, W. C., Jr. (1970) J. Biol. Chem. 245,238-

Jornvall, H. (1977) Eur. J. Biochem. 72,443-452 Kaji, A., and Colowick, S. P. (1965) J. Biol. Chem. 240,4454-4462 Knowles, A. F., and Penefsky, H. S. (1972a) J. Biol. Chem. 2 4 7 ,

Knowles, A. F., and Penefsky, H. S. (1972b) J. Bwl. Chem. 247 ,

Koga, P. G., and Cross, R. L. (1982) Biochim. Biophys. Acta 679,

Koshland, D. E. (1958) Proc. Natl. Acad. Sci. U. S. A. 44,98-104 Lazarus, N. R., Ramel, A. H., Rustum, Y. M., and Barnard, E. A.

Mahowald, T. A., Noltmann, E. A., and Kuby, S. A. (1962) J. Bwl.

Markland, F. S., Bacharach, A. D. E., Weber, B. H., O’Grady, T. C., Saunders, G. C., and Umemura, N. (1975) J. Bwl. Chem. 2 5 0 ,

McKinley-McKee, J. S., and Morris, D. L. (1972) Eur. J. Biochem.

Michelson, A. M. (1964) Bwchim. Biophys. Acta 9 1 , 1-13 Milhausen, M., and Levy, H. R. (1975) Eur. J. Biochem. 50,453-461 Minchiotti, L., Ronchi, S., and Rippa, M. (1981) Biochim. Biophys.

Moffatt, J. G. (1966) Methods Enzymol. 8, 136-142 Noda, L. (1958) J. Biol. Chem. 233,237-247 Noda, L., and Kuby, S. A. (1957) J. Biol. Chem. 226,551-558 Penefsky, H. S. (1977) J. Biol. Chem. 252,2891-2899 Penefsky, H. S., and Warner, R. C. (1965) J. Bwl. Chem. 240,4694-

4702 Perkins, R. E., Conroy, S. C., Dunbar, B., Fothergill, L. A., Tuite, M.

F., Dobson, M. J., Kingsman, S. M., and Kingsman, A. J. (1983) Biochem. J. 211,199-218

Peters, B. A., and Neet, K. E. (1978) J. Biol. Chem. 253,6826-6831 Peterson, G. L. (1977) Anal. Biochem. 8 3 , 346-356 Peterson, E. A., and Sober, H. A. (1954) J. Am. Chem. SOC. 7 6 , 169-

Raibaud, O., and Goldberg, M. E. (1974) FEBS Lett. 40,41-44 Rashed, N., and Rabin, B. R. (1968) Eur. J. Biochem. 5, 147-150 Rhoads, D. G., and Lowenstein, J. M. (1968) J. Biol. Chem. 2 4 3 ,

Ricard, J., Noat, G., Got, C., and Borel, M. (1972) Eur. J. Biochem.

Riordan, J. F. (1979) Mol. Cell. Biochem. 2 6 , 71-92 Rippa, M., and Pontremoli, S. (1969) Arch. Biochem. Biophys. 103 ,

Rossmann, M. G., and Argos, P. (1981) Annu. Reu. Biochem. 50,

Schnackerz, K. D., and Noltmann, E. A. (1971) Biochemistry 10,

Scopes, R. K. (1978) Eur. J. Biochem. 91,119-129 Shoham, M., and Steitz, T. A. (1980) J. Mol. Bwl. 140 , 1-14 Sogin, D. C., and Plapp, B. V. (1975) J. Biol. Chem. 250,205-210 Stock, A., Ortanderl, F., and Pfleiderer, G. (1966) Biochem. 2. 344 ,

Tagaya, M., Nakano, K., and Fukui, T. (1985) J. Bwl. Chem. 2 6 0 ,

Tamura, J. K., Rakov, R. D., and Cross, R. L. (1985) Fed. Proc. 4 4 ,

Uyeda, K. (1969) Biochemistry 8,2366-2373 Whitehead, E. P., and Rabin, B. R. (1964) Bwchem. J. 90,532-539 Wilkinson, K. D., and Rose, I. A. (1979) J. Biol. Chem. 2 5 4 , 12567-

Yount, R. G. (1975) Adv. Enzymol. Relat. Areas Mol. Biol. 4 3 , 1-56

246,946-953

249,6040-6046

245

6617-6623

6624-6630

269-278

(1966) Biochemistry 5,4003-4016

Chem. 237,1138-1145

1301-1310

28 , l -11

Acta 657,232-242

175

3963-3972

31 , 14-24

112-118

497-532

4837-4843

353-360

6670-6676

1619

12572

of vol. issue of 25, pp. of in u.s.a. affinity labeling of nucleotide … · 1999-02-01 · potent...

Documents