Mendeleev Communications Electronic Version, Issue 4, 1998 (pp. 129–168) Oximes, amidoximes and hydroxamic acids as nitric oxide donors Leonid N. Koikov,*a Natalia V. Alexeeva,a Elena A. Lisitza,a Emmanuil S. Krichevsky,a Nikita B. Grigoryev,a Alexandr V. Danilov,a Irina S. Severina,b Natalia V. Pyatakovab and Vladimir G. Granika a Centre for Medicinal Chemistry, All-Russian Chemical-Pharmaceutical Institute, 119815 Moscow, Russian Federation.Fax: +7 095 246 6633 b Institute of Biomedical Chemistry, 119832 Moscow, Russian Federation. Fax: +7 095 245 0857 Quaternized pyridine aldoximes (2- and 4-PAM), some hydroxamic acids and amidoximes produce NO under mild oxidation and activate soluble guanylate cyclase. Recently we have reported that some oximes of 2-arylmethylene- and 2-arylmethylquinuclidin-3-one gave NO and the parent ketones under mild oxidation.The most active NO donors containing a 2-HO-phenyl group proximal to the oxime fragment activated soluble guanylate cyclase; 4-HO-isomers, 2-MeO- and 4-MeO-derivatives as well as the oxime of quinuclidin-3-one were much less active.1–2 The aim of this work is an investigation of the factors affecting the ability of the oxime group to serve as NO precursor, with special attention given to the influence of its conjugation with heteroatoms (O, N) and aromatic rings.For this purpose three series of compounds: oximes RCH=NOH 1 and 2, hydroxamic acids RC(=O)NOH 3 and amidoximes RC(NH2)=NOH 4 were chosen. Methyl derivatives 1a, 3a, 4a were used as the reference compounds and aromatic substituents in the above series were designed for an elucidation of the possible role of H-bonding (R: b, Ph; c, 2-HO; d, 2-MeO; e, 4-HO; f, 4-MeO-phenyl; h, 2-Py, i, 4-Py; j, 2,6-Py) through interaction at the enzyme site at physiological pH values.Quaternized pyridine aldoximes 2h,i served as strongly electron-deficient nuclei lacking H-bond acceptor properties.The required compounds were obtained according to the published procedures from hydroxylamine and the corresponding aldehydes for syn-aldoximes3 [1a: bp 113–114 °C (lit.,4 112– 114 °C); 1b: mp 29–31 °C (lit.,5 35 °C); 1c: mp 58–59 °C (lit.,6 56–57 °C); 1d: mp 90–92 °C (lit.,5 92 °C); 1e: mp 62–65 °C (lit.,7 68–70 °C); 1f: mp 52–54 °C (lit.,3 61–62 °C)], or the corresponding esters for hydroxamic acids [3d: mp 132–134 °C (acetone–diethyl ether, lit.,8 129–131 °C); 3e: mp 179 °C (H2O, lit.,9 168 °C); 3f: mp 159–160 °C (acetone–hexane, lit.,9 162 °C); 3h·0.5H2O: mp 136 °C (H2O, lit.,10 120 °C, anhydrous); 3i: mp 175 °C (H2O, lit.,10 161 °C); pyridine-2,6-dihydroxamic acid 3j·0.5H2O: mp 185 °C (H2O, lit.,10 217 °C, anhydrous)] or the corresponding nitriles for amidoximes [4h: mp 113–114 °C (H2O, lit.,11 115 °C); 4i: mp 205 °C (H2O, lit.,11 199 °C); pyridine-2,6-diamidoxime 4j: mp 235–237 °C (H2O, lit.,12 214 °C)].Commercial preparations of 1h (Merck), 1i (Lancaster), 2a (Aldrich) and 3b,c (Reakhim, Russia) were used. Quaternization of 1h,i with MeI was carried out in MeOH at room temperature. Amidoximes 4a,b,g and acetamidine hydrochloride were a gift from Dr S.M. Vinogradova. According to 1H NMR spectroscopy the purity of all samples was more than 98% and compounds 1e,f contained ca. 5% of anti-isomers. The results of oxidation of the above compounds with K3[Fe(CN)6] are given in Table 1. NO was trapped and determined as [Fe(CN)5NO]2– (nitroprusside) anion by pulsed differential polarography according to the previously described protocol.1–2 Under these conditions oximes 1 do not generate NO, which is in sharp contrast with our earlier finding that (E)-2-[(2'-hydroxyphenyl)methylene]- and 2-[(2'-hydroxyphenyl) methyl]quinuclidin-3-one gave NO in 17% and 10% yields, respectively.A possible explanation for this difference may lie in the fact that the rigid structure of the quinuclidine derivatives allows for intramolecular interaction of the 2'-OH group and the oxime fragment, which is impossible for syn-oxime 1c.Even so, the absence of any difference within the series 1, comprising both the strong acceptor pyridine and strong donor hydroxyphenyl rings, as well as the nonconjugated methyl group, is rather surprising. Probably, the competition between the donor–acceptor properties of R and the acidities of oxime and phenolic OH groups, responsible for the proportion of oxime anions, accounts for the phenomenon. The difference is more pronounced for hydroxamic acids 3, which are completely ionized at the pH employed13–15 (Table 2).Among them the most active NO donor proved to be acetohydroxamic acid 3a. The pyridine ring essentially inhibited the activity 3h–j, though 3j shows some residual activity.The absence of any visible trend in the aromatic sub-series 3b–f allows us to suggest the operation of several opposing factors. In any case, the involvement of a formal oxime group in conjugation with the oxygen lone pair (hydroxamic acids exists as 3 and not 11) in 3a–f leads to a marked increase in activity compared to the corresponding oximes 1a–f, 3a being half as active as NH2OH·HCl.The simple amidoximes 4a,b essentially mimic the behaviour of the corresponding hydroxamic acids 3a,b. At the same time, the activity of non-conjugated phenylacetic derivative 4g, is only half of that of benzamidoxime 4b instead of the expected value between 4a and 4b. In contrast to hydroxamic acids, the substitution of a benzene ring for pyridine leads to a sudden leap in amidoxime activity. The values for compounds 4h,i are close to the value for 4a.The activity of more conjugated 4i is 1.5 times higher than that of 4h and pyridine-2,6-diamidoxime 4j is twice as active as its 2-substituted prototype 4h. The pathways leading to NO generation from oximes, amidoximes and hydroxamic acids are to be investigated separately but the observed facts may be rationalized as N H R OH N Me I N H OH R NH O OH R N NH2 OH N 1 2 3 4 h R = 2-Py i R = 4-Py j R = Scheme 1 aNa2[Fe(CN)5NO] 100%, NH2OH·HCl 55.8%.b2,6-Disubstituted pyridyl. Table 1 Yield of NO (%) from oximes, amidoximes and hydroxamic acids: oxidation (c = 2×10–4 M, 4% EtOH–H2O) by K3[Fe(CN)6] at pH 12.a R Compound 1 2 3 4 Me a 0.0 — 22.9 25.0 Ph b 0.5 — 9.0 10.0 2-HOC6H4 c 0.0 — 14.0 — 2-MeOC6H4 d 0.0 — 5.1 — 4-HOC6H4 e 0.0 — 6.5 — 4-MeOC6H4 f 0.0 — 16.4 — 4-HOC6H4CH2 g — — — 5.5 2-Py h 0.0 6.4 0.0 20.6 4-Py i 7.4 62.2 0.8 29.5 2,6-Pyb j — — 3.1 40.0Mendeleev Communications Electronic Version, Issue 4, 1998 (pp. 129-168) follows. At pH 12 all the oximes (except 1a) and hydroxamic acids studied are ionized > 90% (Table 2) and the first step in the oxidation is undoubtedly the loss of an electron from their anion (Scheme 2).Within the series of oximes 1a–i, 2h–j only acceptor pyridine rings provide the favourable combination of an increase in content of 5 and in radical 6 (presented in two resonance structures 6a and 6b) the stabilisation responsible for the subsequent NO generation.The absence of a correlation between 1h,i, 2h,i acidity and the yield of NO shows the importance of the stabilisation. The latter is higher for 4-Py than for the 2-Py substituent, 1-methyl-4-pyridinealdoxime 2i being the most active of all the compounds tested. Nor can the adverse contribution of ortho-effects in 1h, 2h, especially steric shielding by N–Me in 2h, be excluded.A control experiment demonstrated that 4h,i are stable at pH 12, i.e. NO formation is not caused by hydroxylamine release due to hydrolysis or other processes. The behaviour of hydroxamic acids is essentially the same (Scheme 3). Both predominant oxo-form 3 and minor iminoform 11 give the better conjugated (compared to 9) imino-anion 10, converting into tautomeric radicals 12,13, of which the latter is involved in reactions similar to that of 6.TLC examination of the reaction mixtures originating from 3b,c,f reveals only the presence of the corresponding carboxylic acids. No detectable amounts of the products of ring hydroxylation or radical dimerisation have been found. Amidoximes are not so straightforward. Though their acidity is low and their basicity is close to pyridine (pKa: 4a 5.95, 4b 5.10, Py 5.1018), they form salts in water not only with acids but also with alkalis.19 So, in this case both the anion of 4 can be oxidized into 18 and the neutral amidoximes can give cation radical 17 followed by deprotonation to 18 (Scheme 4).The reactive 18 can behave similarly to the above mentioned key-radicals 6 and 13 (Schemes 2 and 3), forming NO via N,O-intermediate 19, but this nevertheless seems not to be the principal route, if at all.Despite the fact that 4b gives benzamide and some benzoic acid as the end-products (TLC), the first detectable reaction product is PhCN (GC-MS). According to the literature the latter is formed from 4b and strong oxidizers, while under conditions similar to ours a low yield of 5-amino-4,5-dihydro- 3,5-diphenyl-1,2,4-oxadiazole was obtained.20 On mixing of 0.11 mmol of 4b, 0.3 mmol of K3[Fe(CN)6] and 0.9 mmol of KOH in 0.5 ml of water at 80 °C the initially clear solution almost immediately becomes opaque (emulsion of hydrophobic PhCN) and a strong smell of bitter almond can be observed.After 2–3 min the solution becomes clear again and contains no 4b.A parallel experiment without K3[Fe(CN)6] showed that after 5 min only about half of 4b was hydrolyzed to benzamide. We have no exact kinetic data but the consumption of 4b in the presence of K3[Fe(CN)6] is much faster than the alkaline hydrolysis leading to benzamide and hydroxylamine, i.e. elimination of PhCN from some radical intermediate (probably 18 ® 20 ® 21) but not simple hydrolysis is the main source of NO from amidoximes. The radical 21 or its anion have been reported as NO precursors under these conditions.21 General considerations do not allow us to exclude oxidation of the amidoxime NH2 group instead of NOH as the first step, but the much more basic and more electron-rich MeC(=NH)NH2 does not give NO under the conditions used, which supports the suggested scheme.The two-fold drop in activity upon substitution of Me for Ph in 3a/3b and 4a/4b can be attributed to a decrease in reactivity of the radicals stabilized by the aromatic ring. Further manifestation of this effect aggravated by a reduction in the electron density on anion 10, might be responsible for the low activity of pyridyl hydroxamic acids 3h–j. Since the nature of R has very little influence on the acidity of the =NOH group in amidoximes,18 the acceptor properties of pyridyl substituents are not as important in steps 4 ® 17 ® 18, but they certainly play a role in the subsequent stages, which may account for the lack of a completely parallel route in the series 3 and 4.Preliminary tests22 show that about a two-fold activation of soluble guanylate cyclase from human platelets is achieved by 3a,c and 4h at a concentration 10–6 mol dm–3. 3e (4-OH isomer of 3c) is less active (×2.5 at 10–5 mol dm–3) and OMe-derivatives of OH-acids 3d,f are not active at all. 3b, 4a,b give a ca. 1.5-fold activation at 10–4–10–5 mol dm–3, while R CH N OH OH– R CH N O – e– R CH N O R CH N O OH or H2O R CH NO OH R CH NO O – e– RCHO + NO Scheme 2 1,2 5 6a 6b 7 8 OH– R C NHOH O OH– R C NHO O R C NOH OH 3 11 9 R C NOH O 10 OH– – e– R C NOH O 12 R C NO OH 13a R C NO OH 13b OH or H2O R C NO OH OH 14 R C NO OH O 15 R C NO OH O 16 RCOOH + NO Scheme 3 aFor Me2C=NOH.% of anion = 100/[1 + 10(pKa – pH)]. Table 2 Ionization of oximes and hydroxamic acids in water at pH 12 and room temperature. Compound 1 2 3 pKa % of anion pKa % of anion pKa % of anion a 12.4a 28.47 — — 9.4613 99.71 b 10.685 95.43 — — 8.7515 99.94 c 9.316 99.80 — — 7.4313 99.997 d 10.885 92.95 — — — — f 10.925 92.32 — — 9.014 99.90 h 10.1417 98.64 8.017 99.99 8.714 99.95 i 9.9917 99.03 8.5717 99.96 7.814 99.994 R C N NH2 – e– Scheme 4 OH R C N NH2 OH OH– R C N NH2 O R C N NH2 O 4 17 18a 18b OH or H2O R C NO NH2 OH 19 R C N NH OH 20 NHOH + R CN .. . 21 OH– NO RCONH2 + RCOOH . . . – e– – NOMendeleev Communications Electronic Version, Issue 4, 1998 (pp. 129–168) pyridine aldoximes 1h, 2i,h give a ‘bell-shaped’ curve with a maximum at ca. 10–5 mol dm–3 (Table 3). In general, enzyme activation data for the hydroxamic acids do not contradict their behaviour under chemical oxidation (Table 1), but the influence of an aromatic substitution pattern in 3b–f is strikingly similar to that observed for oximes of 2-arylmethylenequinuclidin- 3-ones1,2 (see above).Consequently, at a physiological pH compatible with the existence of intramolecular hydrogen bonds 22, the 2-OH-phenyl group provides either superior stabilisation of the intermediates (or the released NO) or higher complementarity to the active site, i.e.we can see the pronounced ortho-effect in 3c. The data for pyridine aldoximes also suggest the involvement of some ortho-interaction or site complementarity, since 2-pyridyl-substituted oximes 1h, 2h activate soluble guanylate cyclase better than 4-substituted 2i and, at the lower concentrations, better than hydroxylamine (Table 3).Since the preparation of soluble guanylate cyclase used is not free from non-specific oxidative enzymes also capable of producing NO, we cannot tell whether these ortho-effects are connected with the complementarity to soluble guanylate cyclase itself, to non-specific oxidative enzymes (P-450 etc.) or to an increase in the lifetime of NO released by non-specific enzymes, which allows NO to reach the soluble guanylate cyclase more efficiently and to cause its activation.Such stabilisation of NO by N-hydroxyguanidine (23, R = H), its derivatives and L-N-hydroxyarginine (L-NOHA) 24 is well documented.23,24 It should be noted that anions of oximes do not interact with NO.25 Thus, although oxidation of =N–OH-containing substrates (or their tautomers 3) at pH 12 cannot be considered as truly biomimetic, this very simple method allows us to carry out a fast preliminary screening of potential NO precursors. As a result, we were able to reveal the unknown NO generating potential of the well-known oximes, amidoximes and hydroxamic acids (there is only one brief mention of NO formation from N-hydroxybenzenesulfonamide: Piloty acid, the rate of which correlates with the rate of hydroxysulfonamide hydrolysis26) and some correlation between the chemical oxidation and soluble guanylate cyclase activation for these hydroxylamine derivatives.Though the better NO yields and values of soluble guanylate cyclase activation achieved in vitro are comparable to that of hydroxylamine itself, the use of the three classes of readily available new NO donors as templates for design of NO releasing ‘pro-drugs’ with desired in vivo properties looks rather promising.One more argument in favour of such a suggestion lies in the similarity of the observed chemistry and NO biosynthesis by nitric oxide synthase from L-arginine 23 via L-NOHA 24 followed by formation of NO and L-citrulline 2527,28 (oxidation of simpler N-hydroxyguanidines is quite complicated29).In conclusion, we cannot avoid mentioning that hydroxylamine, hydroxamic acids and especially quaternary pyridine aldoximes 2h (2-PAM) and 2i (4-PAM) were the first efficient antidotes against nerve gas poisoning.8,10,14,15,17 The commonly accepted ‘nucleophilic’ mechanism of acetylcholine esterase reactivation could not explain all the phenomena observed in vivo, which may account for the potential NO contribution.The last hypothesis is far beyond the scope of this work and must be investigated separately. This work was supported by the Russian Foundation for Basic Research (grant no. 96-04-48325). References 1 L. N. Koikov, N. V. Alexeeva, N. B. Grigoryev, V. I. Levina, K. F. Turchin, T. Ya. Filipenko, I.S. Severina, I. K. Ryaposova and V. G. Granik, Mendeleev Commun., 1996, 94. 2 L. N. Koikov, N. V. Alexeeva, N. B. Grigoryev, V. I. Levina, K. F. Turchin, T. Ya. Filipenko, M. D. Mashkovsky, M. E. Kaminka, V. B. Nikitin, G. N. Engalycheva, M. A. Kalinkina, I. S. Severina, I. K. Ryaposova and V. G. Granik, Khim.-Farm. Zh., 1997, 5, 28 (Chem. Abstr., 1997, 127, 272489). 3 E. Beckmann, Ber., 1890, 23, 1680. 4 H. Wieland, Ber., 1907, 40, 1677. 5 O. L. Brady and R. G. Goldstein, J. Chem. Soc., 1926, 1918. 6 Merck Index, 1996, 12, 8479. 7 M. Sekiya, N. Yanaihara and T. Masui, Chem. Pharm. Bull., 1961, 9, 945. 8 M. A. Stolberg, W. A. Mosher and T. Wagner-Jauregg, J. Am. Chem. Soc., 1957, 79, 2615. 9 J. Hase, K. Kobayashi, N. Kawaguchi and K. Sakamoto, Chem. Pharm. Bull., 1971, 19, 363. 10 B. E. Hackley Jr., R. Plapinger, M. Stolberg and T. Wagner-Jauregg, J. Am. Chem. Soc., 1955, 77, 3651. 11 R. Delaby, P. Reynaud and T. Tupin, Bull. Soc. Chim. France, 1957, 714. 12 G. A. Pearse Jr. and J. Wisowaty Jr., J. Heterocycl. Chem., 1973, 10, 647. 13 Y. K. Agrawal, V. P. Khare and A. S. Kapoor, J. Electroanalyt. Chem., 1974, 54, 433. 14 T. Wagner-Jauregg, Arzneimittel-Forshung, 1956, 6, 194. 15 A.L. Green, G. L. Sainsbury, B. Saville and M. Stansfield, J. Chem. Soc., 1958, 1583. 16 I. Hayashi, K. Ogihara and K. Shimitzu, Bull. Chem. Soc. Jpn., 1983, 56, 2432. 17 S. F. Mason, J. Chem. Soc., 1960, 22. 18 J. Barrans, Ann. Fac. Sci. Univ. Toulouse Sci. Math. Sci. Phys., 1964, 25, 7 (Chem. Abstr., 1964, 60, 12004). 19 F. Tiemann and P. Krieger, Ber., 1884, 17, 1685. 20 J. Stieglitz, Ber., 1889, 17, 3148. 21 V. I. Levina, A. V. Danilov and N. B. Grigoryev, Khim.-Farm. Zh., 1998, 53 (in Russian). 22 I. S. Severina, I. K. Ryaposova, L. B. Volodarsky, D. C. Mozhuchin, A. Ya. Tichonov, G. Ya. Schwartz, V. G. Granik, D. A. Grigoryev and N. B. Grigoryev, Biochem. Molec. Biology Intern., 1995, 36, 913. 23 A. Zembovicz, T. A. Swierkosz, G.J. Southan, M. Hecker and J. R. Vane, Brit. J. Pharmacol., 1992, 107, 1001. 24 J. G. Southan, S. S. Gross, H. Hecker, A. I. Marlett, H. G. Parkers, E. E. Anggrad and J. R. Vane, Portland Press Proc. (Biology of Nitric Oxide 4), 1994, 8, 255. 25 G. A. Russel, R. K. Norris and A. R. Metcalfe, J. Am. Chem. Soc., 1972, 94, 4959. 26 A. Gzzesiok, H. Weber, R. Z. Pino and M. Feelish, Portland Press Proc. (Biology of Nitric Oxide 4), 1994, 8, 238. 27 J. F. Kerwin, J. R. Lancester and P. L. Feldman, J. Med. Chem., 1995, 38, 4343. Table 3 Activation of soluble guanylate cyclase (2 series) relative to control = 1. Compound Concentrations tested/mol dm–3 10–6 10–5 10–4 10–3 Series 1 1h 1.3 2.0 1.6 — 2h 1.4 2.1 1.1 — 2i 1.1 1.6 1.5 — NH2OH·HCl 1.1 1.3 2.1 — Series 2 4a — 1.4 1.3 1.3 3b — 1.4 1.7 1.5 4b — 1.4 1.3 1.2 NH2OH·HCl — 2.5 2.2 3.6 NH O O H OH NH RHN NH2 N RHN NH2 OH O RHN NH2 + NO 22 O O 23 24 25 COOH NH2 R = LScheme 5Mendeleev Communications Electronic Version, Issue 4, 1998 (pp. 129-168) 28 V. G. Granik, S. Yu. Ryabova and N. B. Grigoriev, Usp. Khim. 1997, 792 (Russ. Chem. Rev., 1997, 66, 717). 29 J. M. Fukuto, G. C. Wallage, R. Hszieh and G. Chaudhuri, Biochem. Pharmacol., 1992, 43, 607. Received: Moscow, 28th October 1997 Cambridge, 15th December 1997; Com. 7/07978H