Omeprazole, a Specific Inhibitor of Gastric (H+-K+)-ATPase, Is a H+-activated Oxidizing Agent of Sulfhydryl Groups*

Omeprazole (5-methoxy-2-[[(4-methoxy-3,5- dimethylpyridinyl)methyl]sulfinyl]-1H-benzimidazole) appeared to inhibit gastric (H+-K+)-ATPase by oxidizing its essential sulfhydryl groups, since the gastric ATPase inactivated by the drug in vivo or in vitro recovered its K+-dependent ATP hydrolyzing activity upon incubation with mercaptoethanol. Biological reducing agents like cysteine or glutathione, however, were unable to reverse the inhibitory effect of omeprazole. Moreover, acidic environments enhanced the potency of omeprazole. For example, in vivo pretreatment of rats with carbachol, a secretagogue, enhanced the activity of omeprazole to inhibit gastric (H+-K+)-ATPase, while pretreatment with cimetidine, an antisecretory agent, reduced its potency. In vitro, lowering pH of incubation media from 7.4 to 5.0 improved the ability of omeprazole to inhibit hog gastric (H+-K+)-ATPase almost 60-fold. The inhibitory effect of the drug was accompanied by a dose-dependently decreased amount of free sulfhydryl groups in the isolated hog gastric membranes. The chemical reactivity of omeprazole with mercaptans is also consistent with the biological action of omeprazole. The drug, only under acidic conditions, reacted with a stoichiometric amount of ethyl mercaptan (or beta-mercaptoethanol) to produce regio-isomers of N-sulfenylated omeprazole sulfide (5-methoxy-2[[(4-methoxy-3,5- dimethyl-2-pyridinyl)methyl]thio]-1- or 3-(ethylthio)benzimidazole). The N-sulfenylated compound reacted at neutral pH with another stoichiometric amount of ethyl mercaptan to produce omeprazole sulfide quantitatively. The gastric polypeptides of 100 kilodaltons representing (H+-K+)-ATPase in the rat gastric mucosa or isolated hog gastric membranes were covalently labeled with [14C]omeprazole. The radioactive label bound to the ATPase, however, could not be displaced by mercaptoethanol under the identical conditions where the ATPase activity was fully restored. These observations suggest that the essential sulfhydryl groups which reacted with omeprazole did not form a stable covalent bond with the drug, but rather that they further reacted with adjacent sulfhydryl groups to form disulfides which could be reduced by mercaptoethanol.

of the ATPase which represents the terminal step of the acid secretory process (3-8). The biochemical mechanism for the drug action, however, has not been reported (2). During preliminary experiments, we have observed that omeprazole is no longer inhibitory to the ATPase in the hog or rat gastric microsomal membranes which had been prepared in the presence of dithiothreitol (1 mM j. This observation prompted us to explore possible interactions of the drug with sulfhydryl groups. In this report, we will present evidence that omeprazole is a H+-activated oxidizing agent of sulfhydryl groups.

RESULTS
The effect of p H on the inhibitory action of omeprazole on gastric (H'-K')-ATPase was examined on isolated hog gastric membranes which were rendered permeable to ions (Fig. 1). Omeprazole, when incubated with the gastric membranes in a Tris acetate buffer, pH 5.0, for 30 min, dose-dependently blocked membrane (H+-K+)-ATPase activity with an apparent Ki value of 6 FM. When the pH of the buffer was raised to 7.4, the K; increased to 200 PM. Clearly acidic environments enhanced the inhibitory action of omeprazole. As further shown in Fig. 1, the ATPase inhibitory activity of omeprazole was accompanied by a dose-dependent reduction of free sulfhydryl groups in the gastric membranes. However, adding 2 mM dithiothreitol (or 4 mM mercaptoethanol) to the media completely nullified the inhibitory effect of omeprazole regardless of the pH of the media.
We have also examined the effect of omeprazole on H' accumulation in isolated hog gastric membrane vesicles. Typically, the distribution ratio of [14C]aminopyrine, an index used for estimating the degree of intravesicular acidification, varied only from 1500 to 1200 over the incubation period from 5 to 30 min in the presence of ATP, KCl, and valinomycin. As shown in Fig. 2, the ability of omeprazole to dissipate a pH gradient across the gastric membranes intensified as a function of incubation time; apparent K; values were 15.0,3.0, 1.5, and 0.5 p~ at the incubation times of 5, 10, 20, and 30 min, respectively. These time-dependent changes in the po-tency of omeprazole could be due to either a progressive inhibition of (H+-K+)-ATPase or time-dependent interferences with aminopyrine accumulation by omeprazole as a competing weak base of PIC, 4.0 (19). The aminopyrine movement is not likely to be subject to interference by omeprazole, however, since omeprazole sulfide, a similar weak base but not a potent inhibitor of (H+-K+)-ATPase, had no effect on the intravesicular acidification at the micromolar doses. Furthermore, the low Ki values of omeprazole suggest that the compound interacted with the ATPase under acidic environments of the intravesicular space, where the lumenal side of the ATPase is accessible in these inside-out gastric membrane vesicles (20,21).
Omeprazole, when administered subcutaneously in the rat, dose-dependently reduced (H+-K+)-ATPase activity in the gastric mucosa (Fig. 3). The dose to inhibit 50% of the ATPase activity as obtained 3 h after dosing was 1.5 f 0.4 mg/kg. However, when rats were induced to a resting state of acid secretion with cimetidine treatment 30 min prior to the dosing of omeprazole, then the drug even at 10 mg/kg blocked only 20% of the ATPase activity in the gastric mucosa. With the rats stimulated to secrete acid by carbachol treatment, on the other hand, the dose of omeprazole to inhibit 50% of the ATPase activity was less than 1 mg/kg. These observations are consistent with the above studies in vitro showing acid activation of omeprazole.
We have studied the effect of sulfhydryl reducing agents on gastric (H+-K+)-ATPase inactivated by omeprazole in vivo or in vitro. For instance, the gastric microsomes which were prepared from the rats treated with omeprazole (5 mg/kg) lost 70% of their (H+-K+)-ATPase activity as compared to those from the untreated rats. Upon incubation with &mercaptoethanol, the omeprazole-treated microsomes recovered their ATPase activity as a function of incubation time (Fig.   4) and of concentration of the reducing agent (Fig. 5). With a saturating dose of mercaptoethanol (200 mM), it took about 45 min of incubation time to recover half of the lost ATPase activity. With a fixed incubation time of 3 h, half-maximal activation occurred at 25 mM mercaptoethanol. Dithiothreitol was as effective as mercaptoethanol in activating the omeprazole-inhibited ATPase activity, while cysteine and reduced glutathione were not effective. Similar studies were carried out with intact isolated hog gastric membrane vesicles interacted with omeprazole (50 PM) under the conditions developing a pH gradient; more than 70% of (H+-K+)-ATPase activity was lost during incubation for 30 min. Again, the (H+-K+)-ATPase activity was restored by mercaptoethanol ( Fig. 6) with similar kinetics as in uiuo. Cysteine and glutathione were not effective (data not shown).
[14C]Omeprazole was used in order to study labeling patterns of gastric (H+-K+)-ATPase in vivo and in vitro. The doses of the labeled drug, 1.5 mg/kg in the rat and 5 X loT5 M in intact isolated hog gastric membrane vesicles developing a pH gradient, were sufficient to produce 50-70% inhibition of (H+-K+)-ATPase activity. Table I shows the amounts of 14C radioactivity tightly bound to the hog and rat gastric membranes, that is, the radioactivity in the membranes after freeze-thawing several times and filtration over a Millipore filter (HAWP, 0.45 PM). The specific 14C activity associated with the hog membranes as labeled in vitro (35,800 cpm/mg of protein) was much higher than that of the rat membranes labeled in vivo (13,500 cpm/mg of protein). It would be expected, however, that the in vitro I4C labeling involved higher levels of nonspecific binding of the drugs. Table I also shows that the amounts of "C radioactivity associated with the hog and rat membranes were not significantly affected by incu-bation of the membranes with 200 mM mercaptoethanol for 3 h.
The "C-labeled membranes were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis under reducing conditions. Then autoradiographs of the gels were obtained (Fig. 7). With the hog membranes, we have detected the incorporation of 14C radioactivity into the polypeptides of 100 kilodaltons corresponding to (H+-K+)-ATPase. Addition of excess nonradioactive omeprazole during the in vitro incubation only marginally reduced the intensity of radiation associated with the ATPase. It should be noted, however, that a majority of the membrane-bound radioactivity was located ahead of the dye front, clearly representing nonspecific binding of the labeled omeprazole to the membranes. In the rat membranes labeled in vivo, we have barely detected the incorporation of 14C radioactivity into the polypeptides of 100 kilodaltons. No noticeable radioactivity was detected in the lower molecular mass regions of the gels.
The chemical reactivity of omeprazole with p-mercaptoethanol was examined in the presence or absence of 2 molar equivalents of HC1 (Fig. 8). Omeprazole only under acidic conditions diminished free sulfhydryl groups with a tH of 7 min at 22 "C. Analysis of the reaction mixtures by thin-layer chromatography (TLC) indicated conversion of omeprazole to a new compound (Compound I); TLC RF values (chloroform/methanol, 5:l) were 0.47 and 0.73 for Compound I and omeprazole, respectively. The amount of Compound I formed was inversely proportional to the level of free sulfhydryl groups remaining in the reaction mixture; at an incubation time of 40 min, omeprazole was almost quantitatively transformed to Compound I. The reaction occurred with ethyl and benzyl mercaptans as well to give related products. Timoprazole, the unsubstituted parent compound of omeprazole, also reacted in the same manner with the various mercaptans, although at somewhat slower rates. On the other hand, the N-methyl analog of timoprazole (l-methyl-2-[ (2-pyridinylmethyl)sulfinyl]benzimidazole) failed to react under the identical reaction conditions and could be recovered unchanged. Fig. 9 shows the 200-MHz 'H NMR spectrum of Compound IA. Compound IA differs from Compound I in that ethyl mercaptan was employed in place of p-mercaptoethanol. With the exception of a singlet which appeared at 6 4.90 (S-C&py), the remaining 'H resonances of Compound IA appeared as double lines in the 200-MHz spectrum. This notable feature in line positions suggested Compound IA was a mixture of two closely related regio-isomers. The appearance of a triplet at 6 1.02 and a quartet at 6 2.30 clearly indicated the incorporation of an ethyl mercaptyl group into the omeprazole structure. Interestingly, the methylene hydrogens at 6 4.90 were exchangeable as evident by the addition of DzO. Under the same conditions, the methylene hydrogens of omeprazole or the sulfide of omeprazole did not exhibit this behavior.
The infrared spectrum of Compound IA showed the absence of strong absorption bands at 1070-1030 cm", indicating the loss of the sulfoxide oxygen. The mass spectrum (fast atom bombardment) of Compound IA showed (M + H) at m/e 389, which was consistent with the incorporation of one ethyl mercaptyl group and the loss of one oxygen from the parent molecule. Finally, treatment of Compound IA with 1 eq of ethyl mercaptan gave the sulfide of omeprazole as the stable final product.
Three sites of the benzimidazole moiety of omeprazole can be considered for substitution with the mercaptyl group, N-1, C-2, and N-3. At the present time, we believe Compound IA is a mixture of N-1 and N-3 sulfenylated compounds. The exchangeability of the methylene hydrogens can be rational-spectrum of Compound IA in CDClS.

FIG. 9. Proton nuclear magnetic
The spectrum was obtained using the been determined, however, whether the pH effect was due to a change in the reactivity of the inhibitor or to the protein. In this study, we have presented several lines of biochemical evidence that omeprazole acted as an oxidizing agent of sulfhydryl groups under acidic conditions, for example, the reversibility of the omeprazole effect by mercaptoethanol, the facile oxidation of membrane-free sulhydryl groups, and concomitant inhibition of gastric (H+-K+)-ATPase by omeprazole under acidic conditions.
Chemically, omeprazole undergoes two different types of reactions with mercaptans depending on the state of protonation of the drug. A pathway shown in Scheme 2 is proposed for the reaction of the free base of omeprazole in the presence of a 20-fold excess of ethyl mercaptan. Isolation and identification of Compounds I1 and I11 support the scheme. This reaction pathway is only noticeable in the presence of excess sulfhydryl groups under neutral conditions. This route, therefore, is not likely to represent the major mode of reaction of omeprazole with essential sulfhydryl groups of gastric (H'-K+)-ATPase. However, this reaction is probably responsible for abolishing the inhibitory effect of omeprazole by excess dithiothreitol (or mercaptoethanol) when incubated  (24,25), N-sulfenylated benzimidazoles have been prepared previously by reaction of 2benzylthiobenzimidazoles with electron-deficient reagents, e.g. CClsSH and CFaSH. It should be noted, however, that Compound IA cannot be formed from omeprazole sulfide under our experimental conditions. The HCl salt of Compound IA was very stable in aqueous media. Its free base, on the other hand, was stable in CHC13, but decomposed fairly rapidly to unknown products in aqueous environments.
Apparently, essential sulfhydryl groups of gastric (H+-K+)-ATPase which react with omeprazole are not likely exposed directly to aqueous environments, since polar reducing agents like cysteine and glutathione could not interact with them. Nevertheless, the probable N-sulfenylated complex between omeprazole and essential " S H groups of the ATPase did not appear to survive neutral conditions employed during isolation of gastric microsomes, since we could not detect any displacement of the I4C label of omeprazole bound to the ATPase in uiuo by mercaptoethanol under the conditions where the ATP hydrolyzing activity of the enzyme was fully restored. Certainly, the specific activity of ['4C]omeprazole (35 mCi/mmol) appeared to be sufficiently high enough to detect any such changes. For example, if we assume that (H+-K+)-ATPase represents at least 2.5% of the microsomal proteins, a considerable underestimation judging from the intensity of 100-kDa protein in the sodium dodecyl sulfategel pattern (26), one molecule of ['4C]omeprazole bound to a 100-kDa unit of (H+-K+)-ATPase would be equivalent to 14,000 cpm/mg of the microsomal protein. Two possibilities may be cited 1) the 14C label bound to the ATPase rather represents the interaction of omeprazole with other than essential sulfhydryl groups via mercaptoethanol-insensitive W. B. Im Labelinq Gastric IH--K')-ATPasewmth I~4CI-omeprazale.

Chem~~alReacti~nrofOmepra~olew~thSulfhydrylGroulw.
A merhanlrm for the formatoon of Compound I A IS presented in Scheme 1. Attachment of themet~ap~lgroup~tC-2tog~vethe~wbenz~midazoler~ngsynem IBnsanotherp~wble Aadtc conditlonr and mercaptans alw drartcally affected the U V. spectrum of omcprazole (kg. 10). The free bare of the drug showed two abwrptmn mamma at 300 and 275 nm, whtch were rhtfted to 357 and 271 nm m the presence of 0 1 N HCI Addntmn of an equmolar concentratlon of Fmercaptoethanol mrtantaneourly brought about the disappearance of the abrorptlon maxmum at 357 nm and the formatoon of a new, broad peak at 285 nm. Analpr of the reaction munure on TLC tndocated a complete convewon of omeprazole to Compound I However. in the presence of an ences amount of mercaptans. #.e, 20 molar equwalents of As shown rn Flg. 8. the free base of omeprazole was not very reactwe wnth mercaptans ethylmercaptan, omeprazole was convened to two compounKwhlch were identhed as Compounds I1 and m. on the barir of thew NMR, mfrared and mass spectra. The structure of Compound II was flrmly ertabhrhed by cornpawon to an authentic sample prepared by lndependentrhemlcalrynth~ir N-l and N-3 rulfenylated compounds.