Regioselectivity of Catechol 0-Methyltransferase - Semantic Scholar

THEJOURNALOF BIOLOGICAL CHEMISTRY

Vol. 261, No. 1. Issue of January 5,pp. 176184,1986 Printed in U.S.A.

Regioselectivity of Catechol 0-Methyltransferase THE EFFECT OF pH ON THE SITE OF 0-METHYLATION OF FLUORINATED NOREPINEPHRINES* (Received for publication, July 15, 1985)

Dhiren R. ThakkerS, CharlesBoehlertS, Kenneth L. Kirks, Rosa Antkowiak8, and CyrusR. Crevelingn From the $Laboratoryof Molecular Pharmacology, Divisionof Biochemistry and Biophysics, Centerfor Drugs and Biologics, Food and Drug Administration, Bethesda, Maryland20205 and the §Laboratory of Chemistry and Uuboratoryof Bioorganic Chemistry. National Institute of Arthritis. Diabetes. and Digestive and Kidney Diseases, National Institutesof Health, Bethesda, Maryland 20205

Selectivity of catechol 0-methyltransferase hasbeen examined for the three ring-fluorinated norepinephrines to elucidate the role of acidity of the phenolic groups intheir methylation. Substitution of fluorine at the 5-position of norepinephrine reverses theselectivity of catechol 0-methyltransferase so that pO-methylation predominates. The&fluor0substituent also causes the pKa of the p-hydroxyl group to decrease substantially. In contrast, 2- and6-fluoronorepinephrines are methylated predominantly at the m-hydroxyl group. These results suggest that acidity of a phenolic group canplay an important role in its ability to be methylated by catechol 0-methyltransferase. Percentages of p-0-methylation of norepinephrine and its fluorinated derivatives increase withpH. This relative increase in p-0-methylation appears to accompany ionization of a group with pKa of 8.6, 7.7, 7.9, and 8.4 for norepinephrine andits 2-, 5-, and 6-fluor0 derivative, respectively. These pKavalues are the same as or similar to thepKa values of a phenolic hydroxyl group of these substrates. 3,4-Dihydroxybenzyl alcohol and its 5-fluor0 derivative are 0-methylated by catechol 0-methyltransferaseto form p- andm-0methyl products in approximately 1:1 and 4: 1 ratios, respectively, at all pH values. Based on the above results, a catechol-binding site model for catechol 0methyltransferase is proposed in which the two phenolic hydroxyl groups of catechol substrates are postulated to be approximately equally spaced from the methyl group of the cosubstrateS-adenosylmethionine.

stances. The methyl group in these methyl-transfer reactions is donated by the cosubstrate S-adenosylmethionine (AdoMet) (Fig. 1).Evidence is rather strong that the methyl transfer proceeds via a direct nucleophilic attack by one of the phenolic hydroxyl groups of catechol substrates on the electron-deficient methyl group of AdoMet (5). Hence, NE and other catecholamines, with their two phenolic groups of somewhat similar nucleophilicity, should be metabolized by catechol 0-methyltransferase to yield approximately equal amounts of the m-0-methylated andp-0-methylatedproducts in the absence of other electronic and steric interactions at the active site. However, NE ispredominantly methylated by catechol 0-methyltransferase at the m-hydroxyl group (6, 7). This result is indicative of the role played by other stereoelectronic interactions between the substrate and enzyme at the active site in determining the site of methylation. Extensive investigation with a large number of substituted catechols led to the proposal that relative m uersus p-0methylation is dependent on the orientation in which the substrate binds to the enzyme (6). According to thisproposal, polar ionic groups on the side chain cause the substrate to bind predominantly in the orientation that brings only the m-hydroxyl group in the vicinity of the methyl group being transferred. It is believed’that this binding mode leads to the formation of predominantly the m-0-methylated products from catecholamines (6, 7). Thus, nucleophilicity of the phenolic hydroxyl groups was accorded a secondary role to the polarity and orientation of the functional groups on the side chains of the catechol substrates as a determinantof the site of methylation. However, subsequent studies with ring-fluorinated norepinephrines (8)and with 5-fluorodopa (9) provided evidence that methylation occurred predominantly at the Enzymatic 0-methylation represents an important metaphenolic hydroxylic group which ionized more readily. These bolic pathway that leads to extraneuronal inactivation of studies, thus, provided strong evidence that nucleophilicity of catecholamine neurotransmitters such as norepinephrine the phenolic hydroxyl group played a more important role (NE’) and dopamine (1).The enzyme catechol methyltransthan polarity of the groups on the side chain in determining ferase (EC 2.1.1.6) catalyzes methylation of catecholamines and their metabolites (l),catecholic steroids such as 2-hy- the site of methylation. In a separate study, almost a 4-fold droxyestradiol (2), and of a large number of other catechol increase in relative p-0-methylation of NE was observed as pH of the incubation medium was increased from 7 to 9 (10). substrates (see Refs. 1,3, and 4 and references therein) leading The increased p-0-methylation appeared to accompany ionito their metabolic inactivation or detoxication in most inzation of a phenolic hydroxyl group of NE (p% = 8.61, but * The costs of publication of this article were defrayed in part by the role of a functional group at theactive site of catechol 0the payment of page charges. This article must therefore be hereby methyltransferase could not be excluded. marked “aduertisement” in accordance with 18 U.S.C. Section 1734 It appears that themechanism of methyl transfer catalyzed solely to indicate this fact. by catechol 0-methyltransferase and the stereoelectronic in‘The abbreviations used are: NE, norepinephrine; AdoMet, S- teractions between the enzyme and catechol substrates that adenosylmethionine; 5-fluorodopa, 5-fluoro-3,4-dihydroxyphenylalanine; 2-FNE, 2-fluoronorepinephrine; 3,4-DHB, 3,4-dihydroxyben- lead to selective methylation of some substrates but not of zyl alcohol; 5F-3,4-DHB; 5-fluoro-3,4-dihydroxybenzylalcohol; the others are far from clear. In the present study, we have investigated the regioselectivity of methylation in the three HPLC, high performance liquid chromatography.

178

0-Methylation of Fluorinated Norepinephrines possible ring-fluorinated norepinephrines as well as in other fluorinated and nonfluorinatedcatechols as afunction of pH. The substitution by fluorine at selected positions of the substrates allows selective alteration in acidity of the phenolic hydroxyl groups without introducing stericperturbations in the molecule. Thesestudies were designed to elucidate the role of stereoelectronic factors in the regioselectivity of catechol 0-methyltransferase. MATERIALS ANDMETHODS

General-HPLC analyses were performed on a Gilson HPLC system equipped with a Hewlett-Packard Model 1040A diode array UV/ VIS detector and with a Gilson data master. Radioactivity was measured with a radiomatic flow-detector model HS-I. DuPont Zorbax ODS and Perkin-Elmer C18 columns were used as indicated. UV spectra were recorded in a flow cell with the Hewlett Packard model 1040A detector or with a Hewlett Packard model 8451A UV/VIS spectrophotometer in conventional cuvettes in the staticmode. NMR spectra were recorded with a varian XL-300 MHz spectrometer; chemical ionization mass spectra were measured with a Finnigan model 1015Dgas chromatograph/mass spectrometer. Chemicals-(-)-@)-NE (Calbiochem), (+)-normetanephrine (Calbiochem), (+)-norparanephrine (Sterling Winthrop Research Institute), 3-methoxy-4-hydroxybenzylalcohol, and 3-hydroxy-4-methoxybenzylalcohol (Aldrich), AdoMet (hydrogen sulfate salt, BoehringerMannheim), ["C-CH3]AdoMet(New England Nuclear, 59.9 mCi/mmol), dithiothreitol (Bethesda Research Laboratories) were obtained from the indicated sources. 3,4-Dihydroxybenzylalcohol (3,4-DHB) (11) was prepared by reduction of 3,4-dihydroxybenzaldehyde (Aldrich) with lithium aluminum hydride in ethyl ether. After standard work-up, the benzyl alcohol was purified by HPLC on a Waters Radial PAK column (C-18, 5 pm, 0.8 X 10 cm) eluted with

*M YR

FIG. 1. Enzymatic 0-methylation of NE by catechol O-methyltransferase (GOMT).

179

10% methanol in water (3 ml/min). The product eluting at 3 min was identified as 3,4-dihydroxybenzyl alcohol by its chemical ionization mass spectrum (NH,, m/z at 123, (M + 1)+- HzO) and by its UV spectrum in methanol (X,at 220 and 282 nm). Synthesis of Fluorine-substituted Norepinephrines, Normetanephrines, and Norparanephrines-The 2-fluoro-, 5-fluoro-, and 6fluoronorepinephrine (2-FNE, 5-FNE, and 6-FNE) substrates were synthesized by previously published procedures (12). Fluorine-substituted normetanephrines (Zb,e) and norparanephrines (2a-c) were prepared from the benzyl ethers of the corresponding ring-fluorinated vanillins and isovanillins bywayof the trimethylsilylcyanohydrins analogous to the procedure used in the synthesis of ring-fluorinated norepinephrines (12) (Fig. 2). The compounds served as synthetic standards for metabolites. The preparation of 5-fluoronormetanephrine (lb) illustrates the procedure. 5-Fluorovanillin (13) (340 mg, 2 mmol) was refluxed for 1 h in 3 ml of 95% ethanol with 300 mg of benzyl chloride, 300 mg of potassium carbonate, and 10 mg of sodium iodide to give, after work-up (12) and chromatography over silica gel (carbon tetrachloride/chloroform, l / l ) , 445mgof noncrystalline 4benzyloxy-5-fluoro-3-methoxybenzaldehyde (36). This was converted to the trimethylsilylcyanohydrin by stirring overnight with excess (0.3 ml) of trimethylsilylcyanide and a few crystals of anhydrous zinc iodide. Removal of excess reagent in vacuo gave the crude product which was reduced to the ethanolamine (lb) with lithium aluminum hydride in ether, using the isolation procedure developed by Fieser (13). Catalytic hydrogenolysis of the benzyl group (10% Pd/C, methanol) and addition of 1.3 eqof oxalic acid gave 5-fluoronormetanephrine as the neutral oxalate. 2-Fluoronorparanephrine (2a), 5fluoronorparanephrine (2b), 6-fluoronormetanephrine (IC), and 6fluoronorparanephrine (2c) were prepared by similar procedures. Identity and purity of all compounds were confirmed by a combination of TLC, HPLC, mass and NMR spectral data, and for crystalline compounds, by combustion analysis; yields and physical data are reported in Table I. 2-Fluoronormetanephrine was prepared in small quantities by enzymatic @methylation of 2-fluoronorepinephrine and was characterized by UV (Amax at 218 and 268 nm) and mass spectral analysis (CI, NH,, m/z at 184, (M + 1)+- HZO). It was distinguished from its isomer 2-fluoronorparanephrine by their different retention times (HPLC, cf Table 11). Of the required fluorinated vanillins and isovanillins, 5-fluorovanillin (36 was prepared as previously described (8). 5-Fluoroisovanillin (46) and 6-fluoroisovanillin (4c) were prepared in improved yield by regioselective methylation (methyl iodide, potassium carbonate, and acetone) of the more acidic 4-hydroxyl group of 5-fluoro- and 6fluoroprocatecheualdehyde, respectively. Similarly, selective benzylation of the 4-hydroxyl group of 6-fluoroprocatecheualdehyde folOH

OH

F

FIG. 2. The scheme for the synthesis of m- and p-0-methylated 2-, 5-, and 6-fluoronorepinephrines.

F

3 (b,c) (3.methyl ether)

5 (b,cl (3-methyl etherl 6 (a-cl (4-methyl etherl

4 k c ) Wmethyl ether)

1 1b.c) (3-methyl ether)

2 (a-c) 14-methyl ehterl

series a , 2-fluor0 6 , 5-fluor0 c , 6-fluoro

TABLE I Physical data andyields of fluorine-substituted rwrmetanephrines, norparanephrines, and their precursors Compounds Ib,c 2a-c, 4c, 5b,c, 6a-c (cf. Fig. 2) were analyzed for C, H, and N. Yield and mp for precursors of the compounds in this table are as follows: 3b (86%, ND), 3c (85%, 65-66 "C), 4a (69%, ND), 46 (71%, ND), 4c (65%, 88-90 " c ) , 56 (83%, 117.5-118.5 OC), 5c (69%, 103-105 "C), 6a (79%, 71-72 "C), 6b (75%, 104.5-105.5 "C), & (79%, 103-105 "C). ND, not determined because the compounds were noncrystalline or Dartiallv crvstalline. ~~

NMR

uv

m.p.

Yield Compound

protons Methyl protons Aromatic

%

202-227 6.95 42

Ib IC

6.90 202-204 47 187-190 7.0740

2a 2b 2c

6.81 188-191 47 28

E

"C

202-208 6.97

(Am, "1.

(m (ABX), H-2, 6) (d, J~.F= 6.8 H-2) HZ, 6.75 (d, J ~ . F = 11.2 HZ, H-5) (m 6)(ABX), H-5, (m (ABX), H-2,6) (d, Jz-F= 6.8H-2) HZ, 6.73 (d, J ~ .= F 11.7 HZ. H-5)

nm)

em-'

3.90 (S) 3.86 (S)

7600 (2281, 1100 (270) 6650 (222), 3500 (282)

3.89 (S) 3.89 (S) 3.86 (S)

7950 (2261, 1100 (270) 7800 (222), 1100 (272) 5950 (224), 3600 (282)

0-Methylation of Fluorinated Norepinephrines

180

TABLE I1 HPLC analysis conditions for metabolites of catechol substrates Retention times Substrate

Substrate

rn-OCHB p-OCH3

and Internal standard

Columna solvent systems (flowrate, ml/min)”

min

7.8 3.0 13.2 16.6 (metanephrine) A (1.0) Norepinephrine 1.8 5.9 8.1 4.8 (5-fluoronormetanephrine) B (0.8) 2-Fluoronorepinephrine 7.3 14.5 26.3 35.1 (3-methoxy-4-hydroxyhenethylamine) c (1.0) 5-Fluoronorepinephrine 6.8 10.5 20.5 25.2 (3-methoxy-4-hydroxyhenethylamine) D (2.0) 5-Fluoronorepinephrine* 6.0 3.2 5.0 9.4 (5-fluoronoremetanephrine) A (0.8) 6-Fluoronorepinephrine 1.9 3,4-Dihydroxybenzylalcohol 5.7 8.0 21.9 (benzoic acid) E (1.0) 1.6 4.1 5-Fluoro-3,4-dihydroxybenzyl alcohol 7.3 14.4 (benzoic acid) F (1.0) a Columns and solvent systems: A, DuPont Zorbax ODS (0.46 X 25 cm), 0.1 M potassium phosphate (pH 3); B, Perkin-Elmer C18 HS-3 (0.46 X 8.5 cm), 0.1 M potassium phosphate (pH 3); C, DuPont Zorbax ODS (0.62 X 25 cm), gradient of methanol in 0.1 M potassium phosphate (pH 3) (cf. Fig. 3); D, DuPont Zorbax ODS (0.94 X 25 em), 10%methanol in 0.1 M potassium phosphate (pH 3); E, Perkin-Elmer C18,3 X 3 (0.46 X 3 cm), 5% methanol in 0.1 M potassium phosphate (pH 3); F, Perkin-Elmer C18 3 X 3 (0.46 X 3 cm), 10%methanol in 0.1 M potassium phosphate (pH 3). These chromatographic conditions for 5-fluoronorepinephrine were used when the products were determined radiochemically.

TIME (mid

FIG. 3. A, HPLC separation of 5-FNE, its m- andp-0-methylated derivatives (synthetic), and the internal standard 3-methoxy-4-hydroxyphenethylamine. The inset shows UV spectra of m- and p - 0 methyl derivatives of 5-FNE, and of AdoMet and its demethylated derivative S-adenosylhomocysteine (AdoHcy). The arrows indicate that at 225 nm (the wavelength chosen to monitor the column effluent) interference due to AdoMet or AdoHcy is minimum. B, HPLC separation of 5-FNE and its metabolites formed by catechol 0-methyltransferase. The chromatogram also shows internal standard, AdoMet, AdoHcy, and two peaks for dithiothreitol (DTT).The late eluting less polar peak for DTT is presumably due to thedisulfide produced by oxidation of DTT. Thechromatographic conditions are given in Table 11. The dashed line indicates solvent composition.

Pd/C, methanol) of the benzyl ether gave 3,4-dihydroxy-5-fluorobenzyl alcohol (5F-3,4-DHB) as a glass. Compounds in this series were purified by TLC andidentified by chemical ionization mass spectrometry. Incubations-Incubation mixtures contained catechol substrate (1 mM), magnesium chloride (2.4 mM), AdoMet (0.5 mM), dithiothreitol (4 mM), Tris buffer, pH 6.5 to 10 (50 mM), and partially purified catechol 0-methyltransferase (calcium phosphate eluate step) (specific activity, 65 nmol of product/mg of protein/min, 3,4-dihydroxybenzoic acid) (15, 16) (2.4mg) in a total volume of 0.2 ml. Reactions were initiated by addition of catechol, and after incubation at 37 “C for 15 min were terminated by addition of 0.1ml of 1 N HCl containing appropriate internal standard.Appropriate blanks were run by omission of (i) catechol substrate and (ii) catechol 0-methyltransferase. After centrifugation, aliquots of the reaction mixtures were analyzed by HPLC. (-)-@-NE (K,,, = 0.15 mM (p-0-methylation) and 0.17 mM (m-0-methylation)) and5-FNE (K, = 0.06 mM (p-0-methylation) and 0.12 mM (m-0-methylation)) were used a t saturating concentration (1mM); this concentration was presumed to be saturating for other catechol substrates. Incubations with [“C-CHa]AdoMet (0.8 mCi/mmol) were carried out similarly in experiments to determine the kinetic constants for (-)-@)-NE and 5-FNE. For these experiments, incubation time was reduced to 10 min and theamount of enzyme was reducedto 1.6 mg. Analysis of Metabolites by HPLC-The quantitation of m- and p0-methylated products of catechol substrates was achieved by separation on reverse phase columns and integration of the UV signal by Gilson data master orof the radioactivity eluting in metabolite peaks by the radiochemical detector. The HPLC conditions used for separation of these metabolites from several catechol substrates investigated are shown in Table 11.

RESULTS

Enzymatic 0-Methylation of Fluorinated Norepinephrines-Relative amounts of m- and p-0-methylated products formed by catechol 0-methyltransferase from the fluorinated norepinephrines were determined by their separation on HPLC systems (cf. Table 11) and integrationof the individual lowed by methylation of the 3-hydroxyl group provided a convenient metabolite peaks (UV). A representative chromatogram for 5synthesis of 3c, prepared previously in low yield by the photochemical FNE is shown in Fig. 3. The m- and p-0-methylated metabschiemann reaction (8).2-Fluoroisovanillin, which could not be pre- olites of 5-FNE were identified by their cochromatography pared by selective monoalkylation of 2-fluoroprocatecheualdehyde, with the corresponding synthetic standards. The identity of was prepared in good yield bythe methionine-mediated regioselective the metabolites was further confirmed by their UV spectra cleavage of the 3-methyl ether of 2-fluoroveratraldehyde in methane (inset to Fig. 3) which were identical to those of the corresulfonic acid according to theprocedure of Fujii (14). sponding authentic standards. Furthermore, several normalSynthesis of 5-FluorobenzylAlcohols-5-Fluoro-4-hydroxy-3-methoxybenzyl alcohol and 5-fluoro-3-hydroxy-4-methoxybenzyl alcohol ized UV spectra were obtained as each metabolite peak eluted were prepared by reduction of 5-fluorovanillin and 5-fluoroisovanillin, through the diode array detector and were found to be idenrespectively, with sodium borohydride (ethanol). Reduction of 3,4- tical for the individual metabolite peak. This assured the dibenzyloxy-5-fluorobenzaldehydefollowed by hydrogenolysis (10% purity of the individual metabolite peaks. Integration of the


181

FIG. 4. Percentages of p-0-methylation as a function of pH for 2FNE ( A ) ,5-FNE ( B ) ,and 6-FNE (C). For comparison, data for NE (D)(10) are also included. The arrows indicate apparent pK. values of the groups being ionized, leading to the increased p - 0 methylation of the norepinephrines. Points represent mean of 2-3 determinations; bars represent the range. Incubation conditions are described under “Materials and Methods.”

t ?

0P

PH metabolite peaks and of the internal standard(3-methoxy-4hydroxyphenethylamine) peak allowed determination of the quantities of the two methylated products formed. The 0methylated products of 2-FNE and 6-FNE were similarly identified and quantitated by HPLC (cf. Table I1 for chromatographic conditions). The percentages of p-0-methylated products of 2-FNE, 5FNE, and 6-FNEformed by catechol 0-methyltransferase are shown in Fig. 4. For comparison, similar data for NE (10) are also shown in Fig. 4.’ Of particular interest was the fact that substitution of fluorine at the 5-position of NE resulted in complete reversal of the site selectivity of catechol O-methyltransferase. The p-0-methylatedproduct accounted for 7088% of the total metabolites of 5-FNE; whereas p-0-methylation is aminor pathway for NE, contributing to only 9-26% of total metabolites (Fig. 4, B and D). Substitution of fluorine at the5-position of NE substantially decreases the pKa of the hydroxyl group para to the ethanolamine moiety3(12, 17). Therefore, the above results (Fig. 4, B and D)indicate that the pK. of the hydroxyl groups of catechol substrates has a pronounced influence on methylation of these hydroxyl groups by catechol 0-methyltransferase. Similar effects of pKa on the site of methylation has been previously demonstrated for norepinephrines (8) and for 3,4-dihydroxyphenylalanine derivatives (9). As expected, 2-FNE, with decreased pKa of the 3-hydroxyl group (pKa = 7.8), was predominantly methylated at the 3-position (m-0-methylation). The per*The data in Ref. 10 were obtained with (-)-(It)-NE. However, (+)-NE gives almost identical distribution of products a t all pH values. The hydroxyl group para to the ethanolamine moiety is the one that is nearest to the fluorine substituent and, hence, should be the one most readily ionized (17).

centages of the m-0-methylated product in this case were even higher (88-96%of total metabolites) than those with NE (74-91% of total metabolites). The small decrease in relative amounts of p-0-methylation of 6-FNE compared to NE (Fig. 4, C and D) was unexpected since fluorine at the6position is expected to cause a small decrease in pK, if any, of the hydroxyl group at the4-position. Percentage ofp-0-methylation increased for all three fluorinated norepinephrines, as it did for NE, with increase in pH of the incubation medium from 6.5 to 10 (Fig. 4). Sigmoidal relationships between percentages of p-0-methylated products and pH for the fluorinated norepinephrines (Fig. 4), like the one for NE (lo), suggest that ionization of a group is responsible for the relative increase in p-0-methylation of each of these substrates.The titrationcurves in Fig. 4 further suggest that thepKa of this group is 8.6 in NE, 7.7 in 2-FNE, 7.9 in 5-FNE, and 8.4 in 6-FNE. It is significant that these values are the same as or close to thepKa values for the first ionization of the phenolic groups of NE (8.7) (18), 2-FNE (7.8), 5-FNE (7.9), and 6-FNE(8.5) (12). Therefore, it appears that first ionization of a phenolic group of norepinephrines is accompanied by an increase in p-0-methylation relative to m-0-methylation of these substrates.Furthermore, when rates of p - and rn-0-methylation of these substrates are expressed as a function of pH (Fig. 5), it appears that the increase in percentages of p-0-methylation of NE (IO), 2FNE, and 6-FNE is a direct consequence of an increase in rates of p-0-methylation. Rates of p-0-methylation of these substrates aregoing up at increasing pH values even when m0-methylation is decreasing (Fig. 5). A bell-shaped relationship between the rates of overall 0-methylation and pH, like the ones obtained for m-0-methylation of NE (lo), 2-FNE, and 6-FNE, has been observed for several substrates of ca-

182

0-Methylation of Fluorinated Norepinephrines 0.08 0.06

H

0.04 0.02

.W L

g

. . -

1.0 0.8

&

I

I

0

pz

0.4

e

pH

3

I;

t

0.09

r

8 0.06 0.03

FIG. 5. Rates of p and rn-0-methylation of 2-FNE ( A ) ,5FNE ( B ) , and 6-FNE (0 by catechol 0-methyltransferase. Points represent mean of 2-3 determinations; burs represent the range. Incubation conditions are described under “Materials and Methods.”

FIG. 6. Percentages of pO-methylation as a function of pH Points ).repfor 3,4-DHB (0”Q) and 5F-3,4-DHB (M resent average of two determinations; burs represent the range. Incubation conditions are described under “Materials and Methods.”

nant role in determining the site of 0-methylation. The percentages of the two products of both benzyl alcohols, in contrast to those of the two products obtained from norepinephrines, were unaffected by pH. Thus, the selective increase in p-0-methylation as a function of pH must be related to interactions of the CHzNH2moiety with the catechol-binding site of catechol 0-methyltransferase. The overall rates exhibited broad optima centered at pH 7.8 and 9.5 for 5F-3,4-DHB and 3,4-DHB, respectively, which reflect both the dependence of the intrinsic activity of catechol 0-methyltransferase on pH and the pK, values of the two substrates (data notshown).

techol 0-methyltransferase (19-21). The enhancement in rate DISCUSSION of p-0-methylation caused by ionization of phenolic hydroxyl group(s), thus, appears to overcome the decrease in intrinsic Methylation of catechol substrates by catechol O-methylcatalytic activity of catechol 0-methyltransferase at higher transferase appears to result from direct nucleophilic attack pH values. For 5-FNE, the rate of p-0-methylation shows a of a phenolic hydroxyl group on the methyl group of AdoMet bell-shaped curve. This may be the consequence of the low (5). Hence, nucleophilicity as well as acidity (pK,) of the pK, of the p-hydroxyl group in this substrate. phenolic groups of catechols should play a significant role in Effect of Side C h i n on the Site of 0-Methylation of Catechol determining which of the two phenolic groups is predomiSubstrates-The phenolic groups of NE areexpected to have nantly methylated by catechol 0-methyltransferase. We have similar pK, values, yet its m-hydroxyl group is selectively examined the specificity of catechol 0-methyltransferase for methylated by catechol 0-methyltransferase. Specific inter- m- and p-0-methylation of the three ring-fluorinated norepactions between the catechol substrate and binding site of inephrines and have studied the change of the selectivity as a catechol 0-methyltransferase must be invoked to explain the function of pH. These studies have allowed us to elucidate high selectivity of the enzyme for m-0-methylation of NE. the relationship between the acidity (pK,) of the phenolic We have examined the methylation of 3,4-DHB and its 5- hydroxyl groups and theirability to be methylated by catechol fluoro derivative (5F-3,4-DHB) as a function of pH to eluci- 0-methyltransferase. Ring fluorinated norepinephrines are date the relative role of pK, of phenolic hydroxyl groups and particularly suited for this purpose, since pK, values of the of the ethanolamine side chain in the regioselectivity of m- or p-hydroxyl group can be selectively altered by substicatechol 0-methyltransferase. The percentages of p-0-meth- tution of fluorine at the appropriate position without introylated products of 3,4-DHB and of 5F-3,4-DHB formed are ducing much steric perturbation in the molecules. plottedasfunctions of pHin Fig. 6. For 3,4-DHB, p-0The selectivity of catechol 0-methyltransferase for m-0methylation accounts for 40-45% of total metabolites com- methylation of NE and othercatecholamines has been attribpared to 9-26% in NE. These results demonstrate that most uted to the repulsive interactions of polar (charged) amino of the selectivity of catechol 0-methyltransferase for m-0- group on the side chain and a proposed hydrophobic pocket methylation of NE is lost when CHJVHzfunctionality on the at or near the active site of catechol 0-methyltransferase (6, ethanolamine side chain of NE is replaced by hydrogen in 7). It has been postulated that repulsion between the side 3,4-DHB. Incontrast,the high selectivity of catechol 0- chains of these substrates andthe hydrophobic pocket forces methyltransferase for p-0-methylation of 5-FNE was essen- the substrates to bind in such a way as to bring only the mtially unchanged when its CHzNHzfunctionality was replaced hydroxyl group in thevicinity of the methyl group of AdoMet; by hydrogen in 5F-3,4-DHB (Figs. 4 and 6). The contribution the p-hydroxyl group in this binding mode is proposed to be of p-0-methylated products of both 5-FNE and5F-3,4-DHB too far for a nucleophilic attack on the methyl group. Hence, was in excess of 80% of total metabolites. Thus, theCHzNHz the selectivity for the m-hydroxyl group is believed to be due moiety of NE plays an important role in the selectivity of to its correct juxtaposition for nucleophilic attack on the catechol 0-methyltransferase for m-0-methylation,but in methyl group rather than its higher nucleophilicity. Small substrates with two phenolic hydroxyl groups of unequal pK,, percentages of the p-0-methylated products are believed to the acidity uf the hydroxyl group appears to play a predomi- be formed by binding of a small population of substrate


183

FIG. 7. A proposed catechol-binding sitemodel for catechol O-methyltransferase. A and B show how predominant m- and p-0-methylatedproducts can be formed from NE and 5-FNE although both the substrates are in the same preferred binding mode. C and D show how 3,4-DHB can be converted to m- andp-0-methylated products by binding in two different modes.

molecules in the unfavorable mode, with the charged amino a hydrophobic pocket (6). The presence of the hydrophobic group projecting in the hydrophobic pocket. The studies based pocket can contribute to the high selectivity of catechol 0 on product inhibition have supported the concept of two methyltransferase for rn-0-methylation of NE and other catemodes of catechol binding leading to rn- and p-0-methylated cholamines. Unfavorable interactions between the polar products (21). Although this binding-site model has success- (charged) ethanolamine moiety of NE and the hydrophobic fully explained the product ratios of several catechol sub- pocket are expected to force a large population of the substrate strates, the results obtained with fluoronorepinephrines and to bind predominantly in one mode4 (Fig. 7A). However, the fluorodihydroxybenzylalcohol in thepresent studyrequire contrary to what had been postulated previously (6, 21), we that alternate possibilities be considered. The above model is propose that both of the hydroxyl groups of the catechol unable to explain (i) preponderance of the p-0-methylated substratesare equally spaced from the methyl group of products from 5-FNE or 5F-3,4-DHB, and (ii) the titration AdoMet so that either of the hydroxyl groups can mount a curves for p-0-methylation with different pK, values for NE nucleophilic attack on the methyl group inboth binding and its several fluorinated derivatives unless additional as- modes. The selectivity for rn-0-methylation of NE may be a sumptions are made. For example, an assumption that the consequence of catalysis by a general base situated such that catechol substrates in anionic form (phenolate anion) bind it only assists in rn-0-methylation (Fig. 7 A ) . The presence of differently than the undissociated substrate molecules would a nucleophile, critical for catalytic activity, has been previexplain the above observations. However, such an assumption ously demonstrated at or near the active site of catechol 0would imply that unfavorable interactions between the pro- methyltransferase (3, 22, 23). General base catalysis of disposed hydrophobic pocket and the polar ethanolamine func- placement reactions a t a saturated carbon, although not very common (24,25), hasbeen demonstrated (26-30). (In all these tionality are overcome by yet undetermined factors. We propose an alternate catechol binding-site model so studies, the characteristic isotope effect was substantially that the above resultsare accommodated along with the 'A favorable interaction between a positively charged amino group previously observed selectivity of catechol O-methyltransferand a negatively charged group near the active site can also lead to ase (6, 7) toward catechol substrates. In this model (Fig. 7), the selectivity. However, substrates with negatively charged side as in the previous model (6, 7), the catechol substrates are chain such as 3,4-dihydroxy mandelic acid are also predominantly presumed to have two different potential modes of binding at methylated at the rn-hydroxyl group (>92% at all pH values, data the active site such that the side chain projects in different not shown, also cf. Ref. 7). This would argue against the above regions (Fig. 7, C and D),one of which appears to consist of possibility.

184


lower ( k & ~ = 1.13 to 1.37) than that observed in general base-catalyzed acyl-transfer reactions (kH/kD= 1.5-2.5). However, the studies by Coward et al. (27), Knipe and Coward (28), and by Irie and Tanida (29, 30) have suggested that a general base-catalyzed enzymatic transmethylation is a viable mechanism. Furthermore, Wong and Showen (25) have concluded that theisotope effect of 1.3-1.6 observed for catechol 0-methyltransferase-catalyzedmethylation of 3,4-dihydroxyacetophenone is consistent with such a general base-catalyzed methyl transfer by catechol 0-methyltransferase). At higher pH values, progressive ionization of the p-hydroxyl group allows it to compete more effectively with the “activ a t e d m-hydroxyl group for the methyl group. This would explain essentially superimposable pH curves for percentage p-0-methylation of norepinephrines and ionization of their respective phenolic groups. The observed dependence of p - 0 methylation on pH for NE and its fluorinated derivatives is consistent with theinherent assumption in the proposed model that the p-0-methylation requires ionization of the phenolic hydroxyl group, whereas methylation of the m-hydroxyl group is assisted by a base and hence can occur selectively at pH values considerably lower than its pKa in solution. In the 3,4-dihydroxybenzyl alcohols, absence of CHzNH2 functionality would allow the substrates to bind in both the modes with equal ease, leading to equal activation of the mand p-hydroxyl groups in binding modes C and D, respectively. Formation of relatively equal amouns of m- and p - 0 methylated products from 3,4-DHB at all pH values is, thus, entirely consistentwith the predictions of the model proposed here. The higher percentages of p-0-methylated products from 5F-3,4-DHB (79-84% of total metabolites) at all pH values appear to result from selective methylation of the p hydroxyl group in both the binding modes due to itslow pK,. In summary, the studies with the fluorinated norepinephrines have shown that the acidity of the phenolic hydroxyl groups is a major determinant of the regioselectivity of catechol 0-methyltransferase for substrates with phenolic groups of unequal pKa values.Furthermore, these studies have provided new insight into the juxtaposition of the methyl group of AdoMet and the catechol functionality. It appears that, unlike the previously held belief (6, 7, 21), the methyl group of AdoMet is equally accessible for nucleophilic attack from either of the two phenolic hydroxyl groups of the catechol substrate. The SNz reaction mechanism for transmethylation (5) would also require that themethionine functionality of AdoMet be positioned perpendicular to theplane of the ring (cf Fig.7) so that a linear transition state is possible for both m- andp-0-methylation. The present study supports the previous postulate that a hydrophobic pocket exists near the active site which repels the polar charged functional groups on the side chain of catechol substrates (6, 7). Finally, a role for general base catalysis in activation of one of the two phenolic hydroxyl groups is proposed to accommodate the observed regioselectivity of catechol O-methyltransferase. Activation by a general base of both phenolic groups had been previously proposed (21). An alternative hypothesis (31), which involves a dipole-dipole interaction between a positive center in the enzyme and the 6-position (or 1-position) of the aromatic ring of catechol substrates, can equally explain activation of the m-hydroxyl group (or the p hydroxyl group). These two mechanisms for the activation of the phenolic hydroxyl groups of catechol substrates are not

mutually exclusive, and more studies arein progress to determine the extent towhich each may contribute. Acknowledgments-We gratefully acknowledge the help of Noel Whittaker in obtaining the mass spectra and of Ellen Kirshhaum in the preparation of the manuscript.

REFERENCES 1. Axelrod, J. (1966) Phurmacol. Reu. 18, 95-112 2. Ball, P., Knuppen, R., Haupt, M., and Breuer, H. (1972) J. Clin. Endocriml. Metab. 3 4 , 736-746 3. Axelrod, J., and Tomchick, R. (1958) J. Biol. Chem. 2 3 3 , 702705 4. Borchardt, R. T. (1980) in EnzymaticBasis of Detoxification (Jakoby, W. B., ed) Vol. 11, pp. 43-62, Academic Press, New York 5. Woodard, R. W., Tsai, M.-D., Floss, H. G., Crooks, P. A., and Coward, J. K. (1980) J. Biol. Chem. 256,9124-9127 6. Creveling, C. R., Delgard, N., Schimizu, H., and Daly, J. W. (1970) Mol. Phurmucol. 6 , 691-696 7. Creveling, C. R., Morris, N., Ong, H. H., and Daly, J. W. (1972) Mol. Phurmacol. 8 , 398-409 8. Creveling, C.R., McNeal, E. T., Cantacuzene, D., and Kirk, K. L. (1981) J. Med. Chem. 2 4 , 1395-1399 9. Firnau, G., Sood, S., Pantel, R., and Garnett, S. (1980) Mol. Pharmacal. 19,130-133 10. Thakker, D.R., Kirk, K. L., and Creveling, C.R. (1982) in Biochemistry of S-Ademsylmethionine and Related Compounds (Usdin, E., Borchardt, R. T., and Creveling, C., eds) pp. 473477, Macmillan Press, London 11. Reimann, V. E. (1971) Liebigs Ann. Chem. 750,109-127 12. Kirk, K. L., Cantacuzene, D., Nimitkitpaisan, Y., McCulloh, D., Padgett, W. L., Daly,J. W., and Creveling, C. R. (1979) J . Med. Chem. 22,1493-1497 13. Fieser, L. F., and Fieser, M. (1967) in Reagents for Organic Synthesis, Vol. I, p. 584, Wiley, New York 14. Fujii, N., Erie, H., and Yajama, H. (1977) J. Chem. SOC.Perkin Trans. I2288-2289 15. Nikodejevic, B., Senoh, S., Daly, J. W., and Creveling, C.R. (1970) J. Phurmacol. Ezp. Ther. 174,89-93 16. Grossman, M. H., Creveling, C. R., Rybczynski, R., Braverman, M., Esersky, C., and Breakfield, X. (1985) J. Neurochem. 4 4 , 421-432 17. Vlasov, V. M., and Jacobson, G. G . (1974) Russ. Chem. Reu. 4 3 , 781-794 18. Martin, R. B. (1971) J. Phys. Chem. 7 5 , 2657-2661 19. Borchardt, R. T.(1973) J. Med. Chem. 16,377-382 20. Flohe, L., and Schwabe, K.-P. (1970) Biochim.Biophys.Acta 220,469-476 21. Coward, J. K., Slisz, E. P., and Wu, F. Y.-H. (1973) Biochemistry 12,2291-2297 22. Lutz, W. B., Creveling, C.R., Daly, J. W., and Witkop, B.W. (1972) J. Med. Chem. 1 5 , 795-802 23. Borchardt, R. T., and Thakker,D. R. (1977) in Methods Enzymol. 46,554-561 24. Jencks, W. P. (1969) in Catalysis in Chemistry and Enzymology, p. 169, McGraw Hill, New York 25. Wong, 0. S.-L., and Schowen, R. L. (1982) in Biochemistry of SAdenosylmethionine and Related Compounds (Usdin, E., Borchardt, R. T., and Creveling, C., eds) pp. 533-537, MacMillan Press, London 26. Swain, C.G., Kuhn, D. A., and Schowen, R. L. (1965) J. Am. Chem. SOC.87, 1553-1561 27. Coward, J. K., Lok, R., and Takagi, 0.(1976) J. Am. Chem. SOC. 98, 1057-1059 28. Knipe, J. O., and Coward, J. K. (1979) J. Am. Chem. SOC.1 0 1 , 4339-4348 29. hie, T., and Tanida, H. (1979) J. Org. Chem. 4 4 , 325-330 30. hie, T., and Tanida, H. (1980) J. Org. Chem. 4 5 , 1795-1800 31. Kirk, K. L., and Creveling, C.R. (1984) in Medieinal Research Reuiews, (de Stevens, G., ed) Vol. 4, pp. 189-220, John Wiley, New York