Betaine:Homocysteine methyltransferase from rat liver: Purification and inhibition by a boronic acid substrate analog

ARCHIVES

OF BIOCHEMISTRY

Vol. 292, No. 1, January,

AND

BIOPHYSICS

pp. 77-86,

1992

Betaine:Homocysteine Methyltransferase Rat Liver: Purification and Inhibition by a Boronic Acid !Substrate Analog’ Kyung-Hee

Lee, Mark

Department

of Chemistry

Received

22, 1991

July

Lava, Payman ana! Biochemistry,

Amiri,

Tom

San Francisco

Betaine:homocysteine methyltransferase (BHMT) from rat liver has been highly purified by an efficient procedure requiring only two chromatographic steps: Sephadex G-100 chromatography and fast protein liquid chromatography chromatofocusing. A 170-fold purification and 7.5% overall yield were achieved. Chromatofocusing yielded three active forms of BHMT with pl values near 8.0, 7.6, and 7.0. The subunit molecular weight of each active form is 45,000 Da as determined by sodium dodecyl sulfate-polyacrylamlde gel electrophoresis, and the native enzyme has a molecular weight of 270,000 as determined by exclusion chromatography. The stability of the purified enzyme was found to be potentiated by the presence of 1 mM dimethylglycine and 1 mM homocysteine. Boronate analogs of betaine (pinanyl N,N,N-trimethylaminomethaneboronate) (4) and dimethylglycine (pinanyl N,N-dimethylaminomethaneboronate) were synthesized from pinanyl iodomethaneboronate (3) and trimethylamine or dimethylamine, respectively. The free acid of the betaine analog (5) was reversibly generated from (4). The inhibition of BHMT by (5) appears competitive with a Ki = 45 pM. Since the K, for betaine measured with the purified enzyme is near 0.1 mM, the boronic acid analog of betaine appears to function effectively as a substrate a:nalog inhibitor of BHMT. The analog does not appear to act as a methyl donor to homocysteine when (5) is substituted for betaine in the enzyme reaction. In addition, an enzyme assay based upon &-cyan0 reverse phase HI’LC detection of the o-phthalaldeyde derivative of methionine was developed as an alternative to the standard radiochemical assay. Betaine: homocysteine methyltransferase in the picomole range can be quantitated using this assay as indicated by a linear response of enzyme activity to protein concentration. 8 1992

Academic

Press,

Inc.

i This study was supported ’ To whom correspondence

in part by NIH Grant should be addressed.

0003.9861/92 $3.00 Copyright 0 1992 by Academic Press, All rights of reproduction in any form

GM-42020.

Ottoboni,

State University,

from

and Robert

N. Lindquist2

1600 Holloway,

San Francisco,

California

94132

Betaine:homocysteine methyltransferase (BHMT)3 (EC 2.1.1.5) and 5-methyltetrahydrofolate:homocysteine transmethylase (methionine synthase) (EC 2.1.1.13) catalyze the biosynthesis of methionine from homocysteine and betaine. Although the activity of BHMT is dependent upon nutritional conditions (l-5), age, and tumor conditions (6, 7), it is generally 5-10 times higher than the activity of methionine synthase. More recent studies using an in vitro system have suggested that BHMT may play a role in methionine conservation under conditions of low dietary intake of methionine (8, 9). The interconversion of homocysteine and methionine has been shown to be one of the key reactions at a major regulatory locus for methionine metabolism in rat liver (1,7-10). Methionine has an important biochemical significance among amino acids due to its function as a methyl group donor in the biosynthesis of S-adenosylmethionine (AdoMet). This latter molecule is a key methyl group donor in many biological transmethylation reactions. Methionine can also serve as a precursor to cysteine and its derivatives via a transsulfuration pathway (8). Increased levels of methionine are required for tumor cells which also exhibit characteristically high rates of transmethylation (6, ll13). In addition to the key role of BHMT in methionine metabolism, the utilization of betaine by BHMT is an obligatory reaction in the catabolism of choline in mammalian tissues. Since a purified preparation of rat liver BHMT is desirable for enzyme kinetic and mechanistic studies, one of the goals of this investigation was to develop an efficient purification scheme using FPLC chromatofocusing. Sev3 Abbreviations used: BHMT, betaine:homocysteine methyltransferase; TAMB, N,N,N-trimethylaminomethaneboronate; PTAMB, pinanyl N,N,N-trimethylaminomethaneboronate; SDS, sodium dodecyl sulfate; PAGE, polyacrylamide gel electrophoresis; OPA, o-phthalaldehyde; THF, tetrahydrofuran; EtpO, diethyl ether; AdoMet, S-adenosylmethionine; FPLC, fast protein liquid chromatography; AlaB, alanine (l-aminoethyl)boronic acid. 77

Inc. reserved

78

LEE

era1 different partially purified preparations of rat liver BHMT have been used in previously reported investigations (2, 3, 10, 14-17), and human liver BHMT has been purified to apparent homogeneity via a lengthy procedure requiring no less than six column chromatographic steps (18). A rapid procedure for the preparation of highly purified BHMT from rat liver is reported in this paper. In addition, as an alternative to the radiochemical assay of Finkelstein and Mudd (lo), a BHMT assay based upon reverse phase HPLC has been developed. A second goal of this study was to synthesize a boronic acid analog of betaine as a potential inhibitor of BHMT and to investigate the kinetics and interaction of this molecule with the enzyme. The interest in discovering inhibitors for BHMT arises not only from their potential usefulness in mechanistic studies of BHMT, but also from the important role of methionine in biological methylations, especially in tumor cell metabolism (6, 11-13). Only a few competitive inhibitors of BHMT have been reported, including several substrate/product analogs which possess carbon in place of nitrogen. Isovalerate (Ki = 0.3 mM), and 3,3-dimethylbutyrate (Ki = 0.45 mM) are analogs of betaine and dimethylglycine, respectively (19). Similarly, butyrate is an analog of sarcosine that is a somewhat weaker competitive inhibitor (Ki = 1.0 mM). The bisubstrate analog S-(&carboxybutyl)-DL-homocysteine binds more tightly to BHMT as expected (Ki = 6.5 PM) (19). A boronic acid analog was chosen for a number of reasons. A number of kinetic and mechanistic investigations involving boron-containing enzyme substrate analogs have been reported. Boronic (20-25) and borinic acids (26, 27) have proven to be effective enzyme inhibitors, and amino and peptide boronic acids are very potent specific inhibitors of proteinases (28-32). The presence of serine-boron or histidine-boron covalent interactions in enzyme-inhibitor complexes of a-lytic protease (30, 3336), cw-chymotrypsin (37), subtilisin (38), and porcine pancreatic elastase (39) is well documented. Covalent interactions between boronates and other groups within enzyme active sites have been suggested by several recent studies. The boronic acid analog of alanine (l-aminoethyl)boronic acid (AlaB) complexes with alanine racemase and D-Ala:D-Ala ligase (40). Proposed adducts for alanine racemase and AlaB involve either intramolecular addition of a boron hydroxyl to the electrophilic carbon of pyridoxal aldimine or a nucleophilic attack of the pyridoxal phosphate oxygen on the (-B(OH),) moiety of AlaB. A boronophosphate adduct is proposed for the AlaB-ligase inhibitory complex. In addition, P-lactamases are inhibited by boronate substrate analogs including trifluoroacetamido- and phenylacetamidomethaneboronic acids (41). Other boron-containing analogs have demonstrated interesting physiological properties. Spielvogel and others have synthesized an extensive series of amine cyanoboranes, amine carboxyboranes, and trimethylamine (N-B)-

ET

AL.

borane-carbohydroxamic acids as boron analogs of amino acids (42-48). These analogs all contain boron in place of the methylene carbon, and a number of them possess significant pharmacological activity including antitumor (47-50), antiarthritic (51), and hypolipidemic effects (52, 53) in animals. A boron analog of acetylcholine with BHB substituted for a methyl group has also been synthesized (49). This paper describes an efficient FPLC chromatofocusing procedure for obtaining highly purified BHMT from rat liver. The syntheses of boronic acid analogs of betaine and dimethylglycine, and the kinetic inhibition studies for the former compound in the presence of purified BHMT are also reported. In addition, the development of an alternative nonradioactive enzyme assay method is discussed. EXPERIMENTAL

PROCEDURES

Materials Betaine, L-methionine, L-cysteine, L-homocysteine thiolactone . HCI, Tris, silver nitrate, citric acid, o-phthaldialdehyde, gel filtration molecular weight standards, and sodium dodecyl sulfate were obtained from Sigma. Scintillation fluid (Scintiverse I) was purchased from Fisher Scientific. Amersham was the source of methyl[i4C]betaine hydrochloride (7.4 mCi/ mmol). Sephadex G-100, Pharmalyte, and polybuffers were obtained from Pharmacia/LKB. Ion-exchange resin AG-l-X4(0H-), acrylamide, his-acrylamide, ammonium persulfate, bromphenol blue, and PAGE molecular weight standards were obtained from Bio-Rad. Protein concentrations were determined using the method of Bradford (54) (BioRad protein assay kit). Fluoraldehyde reagent, OPA, and 30% Brij 35 were purchased from Pierce. All other chemicals were analytical grade.

Animals

and Diets

Sprague-Dawley male rats weighing 150-200 g (Simonson Laboratories, Gilroy, CA) were fed over a period of 10 days to an average weight of approximately 300 g. The diet was supplemented with L-methionine and L-cysteine to increase levels of liver BHMT (1, 2, 10). A solution containing 0.57% of L-methionine and 0.23% L-cysteine was added to standard rodent chow (Wayne Lab-Blox) in the ratio of 6.5 g/ml and oven dried. Excised livers were either used immediately or frozen at -70°C. Livers from two rats were used in each purification run.

Enzyme

Assay

Assays were performed according to a modification of the method of Finkelstein and Martin (17). The assay solution contained 35 mM potassium phosphate, pH 7.4, 2 mM unlabeled betaine, and 7 mM homocysteine, and included 10 ~1 of a solution of 0.67 mM methyl[i4C]betaine (specific activity, 7.40 mCi/mmol). The reaction was initiated by addition of 30-90 ~1 of enzyme to give a final volume of 1.0 ml. Homocysteine was prepared prior to use from homocysteine thiolactone according to the method of Mudd et al. (55). Samples were incubated at 37’C for 1 to 24 h depending upon the activity of the preparation. The reaction was linear if less than 10% of the betaine was utilized, and was stopped by rapid freezing in isopropanol/dry ice. The incubation mixture was applied to an AG-lX4(0H-) column (0.9 X 5 cm) and washed with 12 ml of cold water to remove unreacted methyl[i4C]betaine. Methionine and dimethylglycine were subsequently eluted from the column as a single peak with 10 ml of 1.5 M HCl. A solution consisting of a l-ml aliquot from the HCl effluent and 10 ml Scintiverse was counted. Radioactive [‘4C]methionine and [‘Y!]dimethylglycine are produced in equal amounts. The methionine contributes one-third of the total counts while

RAT dimethylglycine contributes done in all cases.

Kinetic

Inhibition

two-thirds.

LIVER

BETAINE:HOMOCYSTEINE

Nonenzymatic

control

runs were

Studies

The assay procedure described. in the previous section was used with the exception that the 35 mM potassium phosphate buffer was pH 7.20, and 100 ~1 of enzyme solution was added to give a final volume of 1.0 ml. The enzyme reaction was foIlowed in the presence and absence of N,N,N-trimethylaminomethaneboronic acid (TAMB), which was generated prior to use from pinanyl1V,N,N-trimethylaminomethaneboronate (PTAMB). A stock solution of 10 mM PTAMB in 35 mM potassium phosphate buffer (pH 7.20) was prepared and incubated at room temperature at least 1 h before the kinetic runs to generate the free boronic acid. Final inhibitor concentrations in the reaction solutions were in the range 1 to 100 fiM. As a control for the iodide ion in PTAMB, the presence of 100 pM potassium iodide was shown not to inhibit the enzyme.

Enzyme Purification All procedures were performed at 4°C unless otherwise indicated. Previous studies on the isolation of BHMT from human liver indicated that a marked improvement in enzyme stability could be achieved if one substrate (1 mM homocysteine) and one product (1 mM dimethylglycine) were included in all solutions during purification (18). These were found to stabilize the rat livmer BHMT in our studies also, and both reagents were included in all buffer solutions except during FPLC chromatofocusing. Rat liver homogenate was centrifuged for 30 min at 15,000 rpm. The supernatant was heated in a 75°C water bath with constant stirring until a temperature of 70°C was reached and maintained for 2 min. The solution was immediately cooled to 4°C in an ice bath, and the particulate matter removed by centrifugation for 20 minutes at 15,000 rpm. Onethird of the supernatant (approximately 15-20 ml) was applied to a Sephadex G-100,2.6X SO-cm co:lumn, previously equilibrated with 0.01 M potassium phosphate buffer, pH 7.5, containing 1 mM dimethylglycine and 1 mM homocysteine. Tubes containing BHMT activity were pooled and dialyzed/concentrated overnight to l-3 ml against the chromatofocusing start buffer (0.025 M Tris/acetate, pH 8.70) using a MicroProDiCon (BioMolecular Dynamics, Beaverton, OR) with a PA-30 membrane. The concentrated enzyme solution was filtered through a 0.45.wrn filter disk (Acrodisc, Gelman Sciences) and diluted to 2-5 ml with the Tris/acetate, pH 8.70, buffer. This enzyme preparation was applied to a MonoP chromatofocusing column, 5 mm X 20 cm (Pharmacia) which had been preequilibrated with the start buffer. A Rainin HPLC system controlled by Rainin Dynamax Method Manager was used. Two to five milliliters of sample were injected at the start of a run. The flow rate was 0.25 ml/min for the first 30 min (100% start buffer, 0.025 M Tris/acetate) and then 0.5 ml/min for the remainder of the run (90 min) with 100% gradient buffer. The gradient buffer contained polybuffer PB 96 (Pharmacia) (final dilution, 1:10.5) and Pharmalyte S-10.5 (Pharmacia) (final dilution, 1:9) and was adjusted to pH 6.7 with acetic acid. A gradient from pH 8.7 to pH 6.7 was achieved as determined by measuring the pH of the eluted fractions. The column was run at room temperature, amd fractions (1 ml) were collected at 4°C. Dimethylglycine and homocysteine were added to all tubes immediately after each run to a concentration of 1 mM each to stabilize the enzyme. Fractions containing BHMT activity were pooled and stored at 4°C. Samples used for specific activity calculations were dialyzed to remove polybuffer prior to enzyme and protein assays.

Gel Electrophoresis Determination

and Molecular

Weight

The homogeneity of BHMT preparations was analyzed by SDS-PAGE at pH 8.30 using a 10% acrylamide running gel and 3% acrylamide stacking gel, and bands were detected using the silver stain technique

METHYLTRANSFERASE

79

(56). The molecular weight of the native BHMT was determined by gel filtration on a Sephadex G-150 column (1.6 X 70 cm). Calibration was done with the following protein standards: urease hexamer (545,000 Da) and trimer (272,000 Da), amylase (200,000 Da), and bovine albumin dimer (132,000 Da) and monomer (66,000 Da). Aliquots of post-Sephadex and post-MonoP were applied to the column which had been previously equilibrated with 0.025 M Tris/acetate buffer, pH 8.30, containing 1 mM dimethylglycine and 1 mM homocysteine.

HPLC

Assay Procedure

o-Phthalaldehyde reagent was prepared as follows: 50 mg of crystalline OPA was mixed with 1 ml HPLC-grade methanol, 0.1 ml Brij 35, 0.2 ml mercaptoethanol, and 48.7 ml potassium borate buffer (0.5 M, pH 10.4) which had been previously sparged with nitrogen. This OPA solution was divided into 2-ml aliquots and stored at 4°C in the dark. The assay solution contained the following: (a) 80 ~1 of a solution containing 40 mM potassium phosphate buffer, pH 7.40, and 2.5 mM betaine; (b) 7 11 of 0.10 M homocysteine; and (c) 13 ~1 HzO. Enzyme (5-50 ~1) and phosphate buffer were then added to give a total volume of 150 ~1. After the samples were incubated at 37°C for 6-24 h, the desired products were separated from protein using a micropartition system (MPS 1, Amicon). A sample of the filtrate was mixed with an equal volume of the OPA reagent and incubated 1 min, and 20 ~1 of the mixture was injected into a reverse phase C,-cyano column (4.6 mm X 25 cm, Rainin Microsorb) which had been preequilibrated with the mobile phase: 0.04 M potassium phosphate, pH 7.40/methanol/acetonitrile (SO:lO:lO, v/v). A standard curve was produced from stock solutions of methionine in the range 0.1 to 1 mM. A flow rate of 0.8 ml/min was used with a running time of 15 min. The retention times and peak areas were monitored at 340 nm.

Synthesis Reactions involving air-sensitive materials were run under argon. Dry solvents were freshly distilled prior to use. Tetrahydrofuran (THF) was predried for 24 h over CaCl, and distilled over sodium metal using benzophenone as an indicator for the absence of water. Diethyl ether (Et,O) was distilled over sodium metal with benzophenone indicator. Methylene chloride (CH,Cl,) was predried for 24 h over CaCl, and then distilled over phosphorous pentoxide. Acetonitrile was shaken with anhydrous potassium carbonate, filtered, and distilled over phosphorous pentoxide. Glassware was oven dried overnight, assembled hot, and flushed with argon. Liquids were transferred with syringes and needles through rubber septums. Solutions of trimethylamine and dimethylamine in THF were prepared by bubbling the gases into dry THF for 5-10 min in a septumcapped bottle with stirring. The concentration of the amine in each solution was determined as follows. Benzene (0.5 ml) was added to a dry 5-ml volumetric flask, and the alkylamine/THF solution was added to bring the total volume to 5 ml. The molarity of the amine in THF was determined from the relative ratios of the alkylamino protons to the benzene protons. ‘H NMR spectra were recorded at 60 MHz (Varian EM-360) or at 100 MHz (Varian T-100). Chemicals and NMR solvents were purchased from Aldrich Chemicals. Elemental analyses were done by Galbraith Laboratories (Knoxville, TN). (+)-Pinanediol was prePinanyl phenylthiomethaneboronate (2). pared from (+)-a-pinene by the method of Ray and Matteson (57). Phenylthiomethaneboronate (1) was prepared from thioanisole, n-butyl lithium, and trimethylborate in the presence of (1,4)-diazabicyclo[2.2.2]octane as described in the literature (58). A solution of phenylthiomethaneboronate (23 g, 0.13 mol) and (+)-pinanediol (20 g, 0.12 mol) in 250 ml dry diethyl ether was stirred for 1 h. The reaction mixture was extracted with saturated sodium bicarbonate solution (2 X 150 ml). The organic layer was dried over magnesium sulfate and concentrated under vacuum to yield 35.5 g (100%) of the ester (2) as a viscous liquid which was used in the next step without further purification.

80

LEE

Pinanyl iodomethaneboronate (3). Pinanyl phenythiomethaneboronate (64 g, 0.21 mol), sodium iodide (47.7 g, 0.32 mol), and methyl iodide (74.2 ml, 1.2 mol) were dissolved in 500 ml dry acetonitrile. The solution was stirred for 96 h at room temperature. Unreacted methyl iodide was removed at ambient temperature on a rotary evaporator, and acetonitrile was then rotoevaporated at an elevated temperature (warm water bath) to yield a deep red oil. Anhydrous ether (500 ml) was added to precipitate the sodium iodide which was removed by vacuum filtration. The ether was subsequently removed on a rotary evaporator. The product was stored under argon and then applied to a silica gel column (4.5 X 40 cm, 40-140 mesh). The column was eluted under house vacuum with 15% Et,O/hexane, and the eluted sample was concentrated on a rotary evaporator to yield a light yellow oil (3) which was stored desiccated under argon at -20°C. The yield was 77 g (87%). Purity was judged to be approximately 83% by NMR analysis (the major impurity was thioanisole, approximately 17%). This preparation could be used in the next step without further purification. Identifiable peaks, 60 MHz ‘H NMR (CDCls): 6 0.87 (s, 3, pinanyl CH,), 1.32 (s, 3, pinanyl CH,), 1.42 (s, 3, pinanyl -0-C-CN,), 2.27 (s, 2, ICH,B), 4.38 (dd, 1, -0-CH-). Pinanyl N,N,N-trimethylaminomethaneboronate, iodide salt (4). To a three-necked round bottom flask containing 25 ml anhydrous THF was added via syringe 6.4 ml (4.3 mmol) of a trimethylamine/THF solution (0.67 M in THF). A 1.5-ml (4.3 mmol) aliquot of (3) in THF was added dropwise at a rate of about 1 ml/min, and the reaction was stirred for an additional hour. Anhydrous ether (200 ml) was added and the precipitate was filtered to yield a white powder. The white solid was dissolved in a minimal amount of acetonitrile and reprecipitated with ether. The yield of (4) was 1.3 g (81%), mp 220-222°C; identifiable peaks, 60 MHz ‘H NMR (CDCls): 6 0.084 (s, 3, pinanyl CH,), 1.34 (s, 3, pinanyl CHJ, 1.46 (s, 3, -O-C-C&), 3.61 (s, 9, (C&),-N-), 3.66 (s, 2, -N-CH,-B), 4.46 (dd, 1, -O-C&); 13C NMR (CDCls): 6 56.73 (CH,),N), 55.16 (N-CHP-B). Anal. Calcd for C,,Hz7N02BI: C, 44.3; H, 7.12; N, 3.69. Found: C, 44.27; H, 7.17; N, 3.73. N,N,N-Trimethylaminomethaneboronate, iodide salt (5). A 105-mg (277 mmol) sample of (4) and 3 ml of glass-distilled water were added to a 12-ml centrifuge tube. The sample was solubilized by adding several drops of 0.2 M HCl (to approximately pH 3) and vortexing. The aqueous solution was extracted with 5 X 1 ml ethyl ether, adjusted to pH 7.5, and evaporated. The residue was washed with acetone to yield 72 mg (91%), mp >300°C (dec). 100 MHz NMR (D,O): d 2.80 (s, 2, -NCH,B), 3.36 (s, 9, (C&),-N-). Although (5) was not isolated as an analytically pure sample, the pinanediol ester (4) could be regenerated in almost quantitative yield from (5) and pinanediol using the following procedure. Pinanediol (35 mg, 0.2 mmol) was added to a solution of (5) (25 mg, 0.1 mmol) in acetonitrile. The solution was stirred for 3 min and poured into 10 ml of dry diethyl ether. A white precipitate was obtained (35 mg, 95% yield) which had identical properties (mp, NMR, TLC) to the pinanediol ester (4) isolated originally. Pinanyl N,N-dimethylaminomethaneboronate, hydrochloride salt. A solution of dimethylamine (36.9 ml, 25.5 mmol) from a 0.69 M solution in THF was added to 100 ml dry CH2C!12. Pinanyl iodomethaneboronate (3) (3.5 ml, 8.5 mmol) was dissolved in CH#.& in a dropping funnel. This solution was added dropwise over 30 min and the combined solution stirred overnight. The solvents were removed by rotoevaporation and 150 ml ether was added to precipitate the dimethylamine hydroiodide salt. The product was isolated as the hydrochloride salt by bubbling anhydrous HCl through the filtrate and removing the precipitated salt by suction filtration. Four successive recrystallizations from CH,Cl.J pentane yielded 0.99 g (42%) of the pinanyl N,N-dimethylaminomethaneboronate, hydrochloride salt, mp 185-188’C; identifiable peaks, 60 MHz ‘H NMR (CDCls): d 0.85 (s, 3, pinanyl CH,), 1.30 (s, 3, pinanyl CH,), 1.46 (s, 3, -0-C-Cl&), 2.78, (d, 2, -NH-C&B), 2.94 (d, 6, (CH,),NH-), 5.4 (broad multiplet, 1, -NH-), 4.9 (dd, 1, -0-CH-). 13C NMR (CHCls): 6 44.53 (CH,),-N-), 43.87, (-N-CH,-B). Anal. Calcd for C13H25N02BC1: C, 57.05; H, 9.11; N, 5.12. Found: C, 56.96; H, 8.98; N, 5.06.

ET

AL.

RESULTS

AND

DISCUSSION

Purification of Betaine:Homocysteine Methyltransferase Studies concerning the mechanism of BHMT and the potency and interaction of enzyme inhibitors with BHMT require a readily purifiable source of enzyme. Previous studies have used partially purified (lo- to 25-fold) preparations from rat liver (2,3,10,15-18). We have developed a rapid, efficient procedure for the preparation of highly purified BHMT and have shown that the isolated enzyme remains in the native hexamer form. Betaine:homocysteine methyltransferase was purified approximately 25fold after heat treatment, centrifugation to remove particulate material, and Sephadex G-100 chromatography. Figure 1 shows the elution pattern of a typical Sephadex chromatographic run, in which the BHMT activity elutes with the first protein peak. Fractions containing active enzyme (approximately 30 ml) from three runs were pooled and concentrated prior to chromatofocusing. Active BHMT was obtained after chromatofocusing on a Pharmacia MonoP HR 5/20 FPLC column. Several different gradients and pH ranges were investigated, and good resolution was obtained with a Trig/acetate buffer and a gradient from pH 8.7 to pH 6.7. As indicated in Fig. 2, three active forms of BHMT were eluted from the MonoP column. Peak (2) consistently had the highest specific activity and had an isoelectric point in the pH range 7.6-7.8. Peak (1) exhibited a pI near 8.0, and peak (3) had a pI near 7.0. Although the enzyme activity of the peaks varied somewhat from one run to the next, the relative activities remained consistent. Peaks (1) and (3)

7000

6000

-

5000

-

4000

a C 2

E 8

: 3000

&

2000

ii

1000

0 20

40

60

Fraction

80

100

120

140

160

number

FIG. 1. Sephadex G-100 chromatography after heat denaturation and centrifugation. Absorbance at 280 nm (solid line), BHMT activity (dashed line) in total cpm per incubation period (30 ~1 enzyme). The solvent was 0.01 M potassium phosphate, pH 7.5, containing 1 mM dimethylglycine and 1 mM homocysteine. Column size was 2.6 X 80 cm, flow rate was 15 ml/h, and 3-ml fractions were collected.

RAT

LIVER

BETAINE:HOMOCYSTEINE

81

METHYLTRANSFERASE

-4500

60 - c5

-4000

50 e iz 8 s -2 s: 3 2 '2

40- 7.5 30'0.

20-

z

- 6.5

2

-3500

^M

-3000 . -2500

i

_ 2000

B B

-1500

FIG. 3. SDS-PAGE of fractions from the purification of rat liver BHMT. Lane 1, crude rat liver homogenate; Lane 2, heat-treated and centrifuged homogenate; Lane 3, Sephadex G-100; Lane 4, MonoP peak (1); Lane 5, MonoP peak (2); Lane 6, MonoP peak (3). Molecular weight markers are indicated and were obtained from a set of Bio-Rad low and high molecular weight proteins. Protein bands were visualized with silver stain. Details of the run are described in the Experimental Procedures section.

ij -1000

IO ,-so0 0 10

20

40

30

Fraction

-0 50

number

FIG. 2. FPLC chromatofocusing of pooled fractions from Sephadex G-100 chromatography. Relative full screen absorbance at 280 nm (solid line), BHMT activity in cpm/h incubation/mg protein (dashed line), and pH of effluent fractions (dotted line) are plotted against fraction number. One-milliliter fractions were collected. Active fractions eluted from left to right; peak (l), peak (2), and peak (3). Conditions of the FPLC run are described in the Elxperimental Procedures section.

exhibited approximately 85 and 60% of the activity of peak (2). A 170-fold purification was achieved from purified/dialyzed peak (2) with a specific activity of 1120 units/mg of protein (Table I). A 7.4% overall yield resulted from the combined activit,y of all three active peaks. The most active form (peak 2) of BHMT was dialyzed to remove the polybuffer and wed for the kinetic studies. Determination

of Purity

and Molecular

Weight

The major form of purified BHMT from human liver has a native molecular weight of 270,000 and a subunit molecular weight of 45,000, consistent with a hexameric enzyme (18). A minor form has been reported (59) which has a native molecular weight identical to the major form, but a subunit molecular weight of 32,000. The pH dependence, stability, and amino acid composition of the two

forms are similar. It was suggested that these results could be accounted for by limited proteolysis of subunits. Pig liver BHMT also has a native molecular weight of 270,000 (60). To assess the subunit molecular weight and to measure the purity of the preparations, SDS-PAGE was run. Figure 3 illustrates the SDS-PAGE patterns of the different enzyme fractions. A major band near 45,000 Da is evident among the many bands in the Sephadex G-100 sample. After chromatofocusing, each of the three different active forms of BHMT exhibits a major doublet band near 45,000 Da, with only a few faint minor bands in evidence at approximately 66,000,55,000, and 33,000 Da. These results indicate that BHMT is substantially purified by this efficient two-column procedure. In addition, all three purified active forms are composed of subunits having a molecular weight virtually identical to that of the major form of human liver BHMT (18). Sephadex G-150 chromatography was used to determine the molecular weight of the purified native BHMT and also to ascertain whether native BHMT dissociates under the conditions of chromatofocusing, where homocysteine and dimethylglycine are absent during the col-

TABLE P’urification

of Betaine:Homocysteine Total

Purification

step

Rat liver homogenate Heat treatment Gel filtration (Sephadex G-100) (peak 2) FPLC chromatofocusing

protein bd

4050 640 67 0.83

I Methyltransferase

Specific activity (units/mg) 6.8 21 170 1120

Note. For experimental details see the Experimental Procedures section. needed to catalyze the methylation of 1 nmol of homocysteine in 1 h. The were done after dialysis to remove the polybuffer. a Total percentage recovery from all three peaks was 7.4%.

from Total

activity (units)

2.8 1.7 1.1 9.3 One unit of enzyme enzyme and protein

X x x x

lo4 104 104 lo2

Rat

Liver Purification factor

Percentage recovery

1.0 4.0 25 170

100 61 40 3.4”

activity is defined as that amount assays on the FPLC chromatofocusing

of enzyme sample

82

LEE

Fraction

ET

AL.

270,000. No active monomer peak appears to be present in the MonoP-purified sample, indicating that native BHMT remains intact in the polybuffer gradient in the absence of homocysteine or dimethylglycine. These results are consistent with native BHMT hexamers composed of 45,000 Da subunits for all three active forms of the rat liver enzyme separated by chromatofocusing. The fact that the pZ’s of these forms differ can be explained by limited proteolytic degradation which could generate subunit forms with similar molecular weights (a doublet near 45,000 is observed) but different net charges. Different combinations of these 45,000-Da subunits could yield multiple active hexameric forms possessingdifferent pl values.

number

Stability

20

‘lo

60 Fraction

so 100 number

120

140

FIG. 4. (bottom) Sephadex G-150 chromatography of a BHMT-active fraction from the Sephadex G-100 column, and (top) Sephadex G-150 chromatography of a BHMT-active fraction from the chromatofocusing column. Absorbance at 280 nm (solid line) and BHMT activity (cpm per incubation period, 30 al enzyme) (dotted line). Column size was 1.6 X 70 cm, flow rate was 12 ml/h and l-ml fractions were collected. Eluant buffer was 0.025 M Tris/acetate, pH 8.30, containing 1 mM dimethylglycine and 1 mM homocysteine.

umn run. Protein standards were used to generate a calibration curve of log molecular weight versus elution volume as described in the Experimental Procedures section. Active fractions from the Sephadex G-100 column and from the MonoP column were run separately. Figure 4 (bottom) shows the elution pattern of a post-sephadexG-100 sample. The major peak of activity (>90%) from the Sephadex column corresponds to a protein with a molecular weight between 260,000 and 270,000, whereas a smaller peak corresponds to a protein with a molecular weight of 45,000. This result suggeststhat a small amount of active monomer may be generated during the first stagesof purification even in the presence of homocysteine and dimethylglycine. Figure 4 (top) shows the Sephadex G-150 chromatogram of the FPLC chromatofocusing sample (peak 2 shown here). The enzyme activity occurs in a major protein peak which elutes from the G-150 column at a volume consistent with a molecular weight near

of Purified

BHMT

Purified human liver BHMT is moderately labile as shown by substantial loss of activity and the observance of lower molecular weight active products during its purification (18). The inclusion of homocysteine and dimethylglycine produced a substantial stabilization of activity and a decrease of lower molecular weight active products. Human BHMT was stable for months at -20°C in glycerol solutions as well. To examine the stabilizing effects of products and reactants on purified rat liver BHMT, samples of peak (2) from the MonoP column were stored under several different conditions, and enzyme activity was measured at intervals up to 10 weeks. Table II summarizes the results of these studies. Samples 1, 3, and 4 were taken from an active peak (2) fraction and stored at -20°C with the additions indicated. Sample 2 was concentrated fivefold and the relative activity has been adjusted accordingly. Samples 1 and 2, which lack homocysteine and dimethylglycine, lose considerable activity within 2 weeks time, although activity in the more concentrated sample persists longer. The addition of homocvsteine and dimethvlslvcine was narticularlv effective

TABLE

Stability

II

of Betaine:Homocysteine

Methyltransferase Relative % activity remaining

Sample 1 2 3 4

1 + 1

mM mM

homocysteine dimethylglycine

10% glycerol

-

-

+ +

+

2 weeks

4 weeks

10 weeks

65 60 105 80

20 50 60 40

0 15 30 0

Note. Samples were taken directly from peak (2) of the chromatofocusing column eluate and stored at -20°C. Sample 2 was concentrated fivefold before storage. Enzyme activity was determined as described under Experimental Procedures.

RAT

LIVER

BETAINE:HOMOCYSTEINE

in that virtually no activity was lost during the first 2 weeks of storage, and 60% activity remained after 4 weeks. Addition of glycerol as well as homocysteine and dimethylglycine did little to potentiate stability. The results indicate that the presence of dimethylglycine and homocysteine is necessary to stabilize the rat liver enzyme and suggest that the inclusion of these compounds in solutions during the purification is beneficial. Development

of an HPLC

of a Boronic

83

12

Assay for BHMT

Amino acids and their derivatives can be detected at low concentrations when reacted with o-phthalaldehyde and mercaptoethanol or other thiol compounds (61-64). These OPA derivatives absorb maximally near 340 nm. In addition, since they a:re fluorescent (excitation, 338 nm; emission, 425 nm), the potential for an even more sensitive assay exists. The formation of product from the BHMT reaction was monitored using a reverse phase HPLC column and precolumn OPA derivatization. OPA reacts quickly with many amino acids, and this was found to be the case with meth:ionine under the conditions of our assay. After a l-min incubation the reaction of OPA with methionine appeared to be complete, and longer incubation times gave no increase in the OPA-methionine peak. Homocysteine and dimethylglycine appeared to be derivatized more slowly. As a quaternary amine, betaine is not expected to form a derivative with OPA and did not under the conditions of this study. An isocratic solvent system was found to be sufficient for the separation of the three OPA derivativeis on a Rainin Microsorb 5-mm column (4.6 X 25 cm), and retention times for each were determined for homocysteine (4.0 min), methionine (9.9 min), and dimethylglycine (11.6 min). Tris, present as a buffer in some solutions, xwas detected as the OPA derivative with a retention time of 6.6 min. The rapid and quantitative derivatization of methionine by OPA was used as the basis for this, assay. A calibration curve for methionine over a lo-fold concentration range was linear, and as indicated in Fig. 5 nanomolar amounts of product were detected (A& and quantitated. The rate of formation of methionine (nanomoles per incubation time) could be measured automatically using an internal standard curve (see Experimental Procedures section) or by extrapolation of the area value from a given run on the calibration curve. Figure 6 illustrates that the formation of methionine was proportional to the amount of enzyme present over the range of 1.65 to 16.5 pmol BHMT. This sensitive and rapid assay procedure offers an alternative to the radiochemical assay which may be prohibitively expensive due to the requirement for custom-synthesized methyl[14C]betaine. Synthesis

METHYLTRANSFERASE

Acid Analog of Betaine

Matteson and Majumdar (65) synthesized several alkylaminoboronate esters from the dibutyl or pinacol esters

FIG. 5. Calibration curve for the quantitation of methionine via HPLC. Methionine samples of differing concentrations were mixed with an equal volume of OPA solution and incubated 1 min prior to injection of 20 ~1 into a Cs-cyano reverse phase HPLC column. The area of each OPA-methionine derivative peak from a given run is plotted versus total nanomoles methionine injected. HPLC run conditions are described in the Experimental Procedures section.

of iodomethaneboronate and secondary (piperidine, morpholine) or tertiary (triethyl, tributyl, dimethylaniline) amines, respectively. Pinacol and pinanyl esters have also been used in the synthesis of aminoboronates (66) and peptide boronates (28-30, 32, 33). These ester moieties hydrolyze readily in aqueous buffer solutions to yield the free acids (28, 34, 35), and thus the purified esters of desired aminoboronates can be used in kinetic studies after an appropriate incubation period. We have synthesized the boronic acid esters of betaine and dimethylglycine from pinanyl iodomethaneboronate and trimethylamine and dimethylamine, respectively. Reaction of phenylthiomethaneboronate (1) with pinanediol gave pinanyl phenylthiomethaneboronate (2) in essentially quantitative yield. Pinanyl iodomethaneboronate (3) was then synthesized from (2) in the presence of sodium iodide, methyl iodide, and acetonitrile. Treatment of (3) with trimethyl amine yielded the desired pinany1 N,N,N-trimethylaminomethaneboronate as the iodide salt (4) as shown in Scheme 1. This compound proved to be quite stable, and a sample stored at -70°C under argon for several years gave essentially the same NMR peaks as the original sample, indicating very little decomposition. Free boronic acids are typically converted to their anhydrides by drying, or they may crystallize out with HCl present (66). Although an analytically pure sample of the free acid (5) was not isolated, the NMR spectrum remained unchanged for at least 2 months, and the pinanediol ester (4) could be regenerated from aqueous acidic preparations of (5) in better than 95% yield. Since preincubation in aqueous buffers is sufficient to hydrolyze pinanediol esters to the free acids (28), the purified pinanediol ester (4) was used in all kinetic studies. The boronate analog of dimethylglycine, pinanyl NJ-dimethylaminomethaneboronate, was also prepared, and the

84

LEE

ET

AL.

(+) pinanediol

a aq HCI 3

0

I

2

Methionine

3

FIG. 6. The rate of formation of methionine at several different BHMT concentrations. From 5 to 50 al of enzyme solution was used in the reaction solutions (150 hl total volume) containing 0.7 pmol homocysteine, 0.2 pmol betaine, and 20 mM potassium phosphate buffer, pH 7.40. The solutions were incubated at 37°C for an appropriate time period, and the reaction was then stopped by centrifugation through an MPS micropartition system. The methionine was quantitated as the methionine-OPA derivative on a C,-cyano reverse phase HPLC using a standardized curve as described in the text. Error bars of f15% were estimated from several runs at the same enzyme concentration.

Kinetic

Inhibition

OH OH

1

d

(nmoks)

kinetic studies of this and related compounds vestigated.

I

5

4

SCHEME

+ W,),;-CH,B\

will be in-

Studies

Boronic acids have proven to be potent enzyme inhibitors by virtue of their ability to form covalent adducts with active site nucleophiles. Peptide boronic acids form complexes with proteinases which mimic the tetrahedral reaction intermediates, and thus they may bind orders of magnitude more tightly than the substrates for which they are analogs (28-30, 32-36). In these instances active site nucleophilic groups such as serine or histidine form covalent adducts with the boron atom. It is reasonable to expect that a boronic acid analog with the correct binding specificity might interact with enzyme active site nucleophilic groups other than those which may be involved in a covalent catalytic mechanism (e.g., proteinases). Indeed such seems to be the case for the recently reported inhibition of alanine racemase and D-AlanineDAlanine ligase by the boronic acid analog of alanine (40). Although the exact nature of the interactions at the active sites are not known for either enzyme-inhibitor complex, nucleophilic attack of a pyridoxal phosphate oxygen to form a boronophosphate adduct has been proposed. Alternatively, a boron hydroxyl may attack the electrophilic carbon of pyridoxal phosphate. In the absence of specific structural or crystallographic information concerning the active site of BHMT, the boronic acid analog of betaine has been synthesized as a potential inhibitor and probe for active site nucleophilic groups. In studies using partially purified (lo- to 25fold) BHMT from rat liver, a Km value of 0.048 mM has been

reported for betaine (15). Using the MonoP-purified enzyme a Km of 0.12 mM was obtained, a value which is similar to that previously reported for the purified human liver enzyme, K,,, = 0.10 mM (18). In the presence of high concentrations of both substrates (7 mM homocysteine, 2 mM betaine), TAMB causes significant inhibition (30%) at a concentration as low as 0.01 mM (Fig. 7). The doublereciprocal plot obtained when the concentration of betaine was varied from 50 to 500 pM (homocysteine held constant at 7 mM) is shown in Fig. 8. The inhibition appears competitive with a calculated Ki of 45 PM. A boronic acid substrate analog which binds covalently to an enzyme would be expected to display a much tighter binding constant than the corresponding substrate. Although TAMB binds more tightly to BHMT than does betaine, the magnitude of this difference suggests that this inhibitor is acting as a substrate analog and that any significant covalent interaction is absent. A complete loss of the OPA-methionine peak occurred when TAMB was substituted for betaine in the HPLC enzyme assay, indicating that TAMB does not act as a methyl donor to homocysteine. The analog may bind in a different mode than betaine or the boronic acid moiety may distort catalytic site groups enough to disrupt methyl group transfer.

FIG. 7. Kinetic inhibition assay of BHMT in the presence of N,N,Ntrimethylaminomethaneboronate (TAMB). The enzyme assay was performed as described in the Experimental Procedures section with substrate concentrations of homocysteine (7 mM) and betaine (2 mM).

RAT

LIVER

BETAINE:HOMOCYSTEINE

METHYLTRANSFERASE

85

10. Finkelstein, J. D., and Mudd, S. H. (1967) J. Biol. Chem. 242,873880. 11. Stern, P., and Hoffman, R. M. (1984) In Vitro 20, 663-670. 12. Tisdale, 13. Tisdale,

M. J. (1980) Biochim. Biophys. Actu 609, M. (1980) Br. J. Cancer 42, 121-128.

14. Finkelstein, J. D., Martin, Arch. Biochem. Biophys.

J. J., Harris,

296-305.

B. J., and Kyle,

W. E. (1982)

218, 169-173.

15. Finkelstein, J. D., Harris, B. J., and Kyle, W. E. (1972) Arch. Biochem. Biophys. 153, 320-324. 16. Finkelstein, J. D., and Martin, J. J. (1981) Anal. Biochem. 111,7276.

FIG.

8.

Inhibition of betairmhomocysteine methyltransferase by TAMB: no inhibitor (closed squares), 100 pM inhibitor (open circles). The reactions were run in 35 rnhl potassium phosphate buffer, pH 7.20, 7 mM homocysteine, and 0.054 to 0.54 mM total betaine concentration. The ratio of betaine/[‘%]betaine was kept constant in all runs by addition of 5 to 50 ~1 of [“Clbetaine. The reactions were initiated by the addition of 100 pl of a purified ELHMT sample, and the solutions (total volume, 1.0 ml) were incubated at 37°C and assayed using the standard procedure. The lines were obtained from a least-squares analysis; R2 = 0.99 (inhibited) and 0.96 (uninhibited).

Studies with the bisubstrate analog S-(d-carboxybutyl)DL-homocysteine (19) suggest that a direct methyl transfer does, in fact, occur between betaine and homocysteine. Although this methyl transfer is distant from the carboxylate of betaine, it is possible that one or more positively charged enzyme moieties might be present to facilitate the binding of the betaine carboxylate. The apparent lack of covalent interaction of TAMB with betaine:homocysteine methyl transferase suggests that if such groups exist, they are not available for nucleophilic attack on a boronic acid im the carboxylate position. Further studies on boronic acid analogs of dimethylglycine and non-nitrogen-containing analogs may elucidate the potential for the formation of more potent enzyme-inhibitor adducts.

17. Finkelstein, Commun. 18. Skiba, Awad,

J. D., and Martin,

J. J. (1984)

Biochem.

Biophys.

Res.

118,14-19.

W. E., Taylor, M. P., Wells, M. S., Mangum, W. M. (1982) J. Biol. Chem. 257, 14,944-14,948.

J. H., and

19. Awad, W. M., Whitney, P. L., Skiba, W. E., Mangum, J. H., and Wells, M. S. (1983) J. Biol. Chem. 258, 12,790-12,792. 20. Matteson, D. S., Sadhu, K. M., and Lienhard, G. E. (1981) J. Am. Chem.Soc. 103,5241-5242. 21. Baker, J. O., and Prescott, 130, 1154-1160. 22. Koehler, K., and Lienhard, 2483. 23. Abouakil,

N.,

and

Lombardo,

J. M. Biochem.

Biophys.

G. E. (1971)

Biochemistry

D. (1989)

Res. Commun.

Biochim.

10, 2477Biophys.

Acta

1004,215-220. 24. Garner, C. (1980) J. Biol. Chem. 255,5064-5068. 25. Amiri, P., Lindquist, R. N., Matteson, D. S., and (1984) Arch. Biochem. Biophys. 234, 531-536. 26. Sutton, L. D., Stout, D. M. (1986) Biockm.

J. S., Hosie, L., Spencer, Biophys. Res. Commun.

Sadhu,

K. M.

P. S., and Quinn,

134, 386-392. 13,5345-5350.

27. Koehler, K., and Hess, G. (1974) Biochemistry 28. Kettner, C. A., and Shenvi, A. B. (1984) J. Biol. Chem. 259,15,10615,114. 29. Shenvi, A. B. (1986) Biochemistry 25, 1286-1291. 30. Kettner, C. A., Bone, R., Agard, D. A., and Bachovchin, W. W. (1988) Biochemistry 27, 7682-7688. 31. Kinder, D. H., and Katzenellenbogen, J. A. (1985) J. Med. Chem. 28,1917-1925. 32. Bachovchin, W. W., Plaut, A. G., Flenkte, G. R., Lynch, M., and Kettner, C. A. (1990) J. Biol. Chem. 265, 3738-3743.

2. Finkelstein, J. D., Harris, B. J., Martin, J. J., and Kyle, W. E. (1982) Biochem. Biophys. Res. Commun. 108, 344-348. 3. Finkelstein, J. D., Kyle, W. E., and Harris, B. J. (1974) Arch. Biochem. Biobhys. 165,77.4-779.

33. Bone, R., Shenvi, A., Kettner, C. A., and Agard, D. A. (1987) Biochemistry 26, 7609-7614. 34. Bone, R., Frank, D., Kettner, C. A., and Agard, D. A. (1989) Biochemistry 28, 7600-7609. 35. Bachovchin, W. W., Wong, W. Y., Farr-Jones, S., Kettner, C. A., and Shenvi, A. B. (1988) Biochemistry 27, 7689-7697. 36. Farr-Jones, S., Smith, S. O., Kettner, C. A., Griffin, R. G., and Bachovchin, W. W. (1989) Proc. N&l. Acad. Sci. USA 86, 69226924.

4. Finkelstein, J. D., Kyle, Biochem. Biophys. 146,

37. Tulinsky, 7743.

REFERENCES 1. Finkelstein, 1582-1587.

5. Barak, (1987)

J. D., and Martin,

A. J., Beckenhauer, Biochem. Cell Biol.

J. J. (1986)

W. E., and

J. Biol.

Harris,

261,

Chem.

B. J. (1971)

Arch.

84.-92. H. C., Tuma, 65, 230-233.

D. J., and Badakhsh,

6. Grzelakowska-Sztabert, M., Manteuffel-Cymborowska, Chmurzunska, W., and Sikora, E. (1986) Cancer Lett. 7. Finkelstein, J. D. (1978) in Transmethylation R. T., and Creveling, C. R.., Eds.), pp. 49-58, 8. Finkelstein, 9513.

J. D., Martin,

J. J. (1984)

9. Finkelstein, Chem. 263,

J. D., Martin, 11,750-11,754.

J. J., and Harris,

S. M.,

32,207-217.

(Usdin, E., Borchardt, Elsevier, New York.

J. Biol.

Chem.

259,

B. J. (1988)

9508J. Biol.

A., and Blevins,

R. A. (1987)

J. Biol.

Chem.

38. Matthews, D. A., Alden, R. A., Birktoft, J. J., Freer, Kraut, J. (1975) J. Biol. Chem. 250, 7120-7126.

262, S. T.,

7737and

39. Takahashi, L. H., Radhakrishnan, R., Rosenfield, R. E., and Meyer, E. F. (1989) Biochemistry 28, 7610-7616. 40. Duncan, K., Faraci, W. S., Matteson, D. S., and Walsh, C. T. (1989) Biochemistry 28, 3541-3549. 41. Crompton, I. E., Cuthbert, B. K., Lowe, G., and Waley, S. G. (1988) Biochem. J. 251.453-459. 42. Spielvogel, B. F., Wojnowich, L., Das, M. K., McPhail, grave, K. D. (1976) J. Am. Chem. Sot. 98, 5702-5703.

A., and Har-

86

LEE

43. Spielvogel, B. F., Das, M. K., McPhail, A. T., and Onan, K. D. (1980) J. Am. Chem. Sot. 102,6343-6344. 44. Spielvogel, B. F., Harchelroad, F., and Wisian-Neilson, P. (1979) J. Znorg. Nucl. Chem. 141, 1223-1227. 45. Wisian-Neilson, P., Das, M. K., and Spielvogel, B. F. (1978) Znorg. Chem. 17,2327-2329. 46. Spielvogel, B. F., Ahmed, F. U., Morse, K. W., and McPhail, A. T. (1984) Znorg. Chem. 23, 1766-1777. 47. Hall, I. H., Spielvogel, B. F., and McPhail, A. T. (1984) J. Pharm. Sci. 73, 222-225. 48. Norwood, V. M., and Morse, K. W. (1988) Znorg. Chem. 27, 302305. 49. Spielvogel, B. F., Ahmed, F. U., McPhail, A. T. (1986) J. Am. C&m.

Sot. 108,3824-3825. 50. Hall, I. H., Starnes, C. O., Spielvogel, B. F., Wisian-Neilson, P., Das, M. K., Wojnowich, L. (1979) J. Pharm. Sci. 68, 685-691. 51. Hall, I. H., Starnes, C. O., McPhail, A. T., Wisian-Neilson, P., Das, M. K., Harchelroad, F., and Spielvogel, B. G. (1980) J. Pharm. Sci. 69, 1025-1029. 52. McPhail, A. T., and Spielvogel, B. F. (1981) J. Pharm. Sci. 70,339341. 53. Hall, I. H., Williams, W., Gilbert, C. J., McPhail, A. T., and SpielVogel, B. F. (1984) J. Pharm. Sci. 73, 973-977. 54. Bradford, M. (1976) Anal. Biochem. 72, 248-253.

ET

AL. 55. Mudd, S. H., Finkelstein, J. D., Irreverre, J. Biol. Chem. 240,4382-4392. 56. Wray, W., Boulikas, T., Wray, Biochem. 118, 197-202. 57. Ray, R., and Matteson, 58. Matteson, 1325-1326.

V. P., and Hancock,

D. S. (1980)

D. S., and Ame,

F., and Laster,

Tetrahedron

D. J. (1978)

L. (1965)

R. (1981) Lett.

J. Am.

Anal.

21,449-450.

Chem.

Sot.

100,

59. Skiba, W. E., Wells, M. S., Mangum, J. H., and Awad, W. M., in Methods in Enzymology (Jacoby, W. B., and Griffith, 0. W., Eds.), Vol. 143, pp. 384-388, Academic Press, San Diego. 60. Ericson, 61. Buck, 265.

L. E. (1960)

62. Chow, J., Orenberg, 386, 243-249. 63. Jones, 482.

Actu Chem.

R. H., and Krummen,

Scund.

K. (1987)

J., and Nugent,

B. N., and Gilligan,

14, 2127-2134. J. Chromutogr.

K. D. (1987)

J. P. (1983)

387,

255-

J. Chromatogr.

J. Chromatogr.

266,

471-

64. Florance, J., Galdes, A., Konteatis, Z., Kosarych, Z., Langer, and Martucci, C. (1987) J. Chromutogr. 414, 313-322. 65. Matteson,

D. S., and Majumdar,

D. (1979)

J. Orgunomet.

K., Chem.

170,259-264. 66. Matteson, 618.

D. S., and Sadhu,

K. M. (1984)

Orgunometullics

3, 614-