Oxidation by Molecular Oxygen. II. The Oxygen Atom Transfer

Elena I Karasevich , Vera S Kulikova , Aleksandr E Shilov , Al'bert A Shteinman ... Jack R. Giacin , Thomas M. Hellman , Maurice E. Snook , Jonathan W...
0 downloads 0 Views 281KB Size
Aug. 20, 1964

if H202(approximately equivalent to the amount which might be formed if ascorbic acid reacted with 0 2 to give H 2 0 2and dehydroascorbic acid) is added with no O2 present then again no products are formed (4). The results of these experiments thus clearly indicate t h a t H202is not an intermediate in the oxidation with molecular oxygen. Line 5 of Table I shows the amounts of products formed when the hydroxyl radical (generated by the Fenton reaction) I 3 is the oxidizing agent. This reaction did give some of the same products observed with the Udenfriend reactiqn. However, gas chromatographic analysis revealed a t least five other products in the reaction of H O . with cyclohexane and two others with cyclohexene. Under the Udenfriend conditions these products were not observed. Consequently i t appears that the Udenfriend reaction on cyclohexane or cyclohexene does not involve the hydroxyl radical as the oxidizing agent. Further evidence for this is obtained if cyclohexene oxide is added a t the beginning of the r u n ; i t is not significantly reacted under the Udenfriend conditions but under the Fenton conditions i t is largely decomposed. If cyclohexanol is added under the Udenfriend conditions then a small amount of i t is oxidized to cyclohexanone, and presumably some, if not all, of the cyclohexanone observed under these conditions arises by oxidation of the cyclohexanol formed in the reaction. Cyclohexanone added a t the beginning of the run does not appear to be affected under either conditions. Recently, KaufmanI4 has identified the cofactor for phenylalanine hydroxylase as a tetrahydropteridine of general structure I. Compound I contains a partial structure analogous to the enediol structure of ascorbic acid (11) which is a necessary reagent in the Uden-

n

OH

339 1

COMMUNICATIONS TO THE EDITOR

H' CHOH

OH I11

I

pyrimidine (111). The hydroxylated products obtained are compared in Table 11. T h e similarity of the products formed with I1 and I11 is suggestive t h a t the two reactions proceed by related mechanisms. Furthermore, the obvious close similarity of I11 to the natural cofactor I suggests t h a t the mechanisms of the enzymatic and model hydroxylations are related. A mechanism for the model and enzymatic hydroxylations is considered in the following communication. l5 (15) G . A Hamilton, J. A m . Chem. Soc., 86, 3391 (1964)

FRICKCHEMICAL LABORATORY PRIFCETON UNIVERSITY PRINCETON, S E W JERSEY

GORDON A . HAMILTON ROBERT J . WORKMAN LAURA Woo

RECEIVED APRIL4, 1964

Oxidation by Molecular Oxygen. 11. The Oxygen Atom Transfer Mechanism for Mixed-Function Oxidases and the Model for MixedFunction Oxidases1t2

Sir: The Udenfriend oxidation3 by molecular oxygen has been considered a model for aromatic hydroxylases and other mixed-function oxidases4 It has now been shown t h a t H20zand the hydroxyl radical are not intermediates in the model oxidation, and t h a t the model system oxidizes aromatic compounds to phenols, saturated aliphatic compounds to alcohols, and olefins to epoxide^.^-^ Such reactions are reminiscent of carbene reactionsIl8 and the implication is that an oxygen species with six electrons and similar to a carbene or carbenoid species is responsible for the oxidations. Other information relevant to a consideration of the mechanism of the model and enzymatic mixed function oxidases is the following : the tetrahydropteridine cofactor ( I ) , required by phenylalanine h y d r o x y l a ~ e , ~ has structural features similar to ascorbic acid (11) and 2,4,5-triamino-6-hydroxypyrimidine (111) which n

CHzOH I1 friend model system. If the enzymatic and model H' 'CHOH OH systems react by similar mechanisms then compounds I U I I CHIOH more closely related to I should also react in the model I1 system. Consequently, the hydroxylation of anisole by the model system has been investigated with the are necessary for the model oxidations4; both the ascorbic acid replaced by 2,4,5-triamino-G-hydroxy-model and enzymatic mixed-function oxidases appear I

TABLE I1 THEHYDROXYLATIOS OF ANISOLEB Y MOLECULAR OXYGEN 5% Conditionsa

yield of methoxyphenolsh

Isomer distribution

ortho

meta

pava

UdenFriend system 8 43 18 39 with I1 Udenfriend system 4 49 13 38 with I11 a The reactions were carried out in 61.5 ml. of water saturated with anisole, buffered with acetate a t p H 4.5, and contained 0.04 mmole of Fe(CIOa)a and 1 mmole of ascorbate or pyrimidine. The reactions were shaken overnight under an atmosphere of air. * Yield based on the initial amount of ascorbate or pyrimidine. The products were analyzed by gas chromatography. (13) For a review see iV Uri. C h e m Rea' , 6 0 , 375 (1952). (14) S K a u f m a n , Pvoc S a t l . Acad. Sci. L- S . , 60, 1085 (19631, J C h e m . , 239, 332 (1964). i b d , 237, PC2712 (1962).

Biol

(1) Presented a t t h e 146th S a t i o n a l Meeting of t h e American Chemical Society, Denver, Colo., J a n . , 1964; Abstracts of Papers, Division of Biological Chemistry, p . 13A. (2) This investigation was supported b y P H S research grant G M 09586 from t h e Division of General Medical Sciences, Public Health Service (3) S. Udenfriend, C . T . Clark, J Axelrod, and B B . Brodie, J . Biol C h e m , 208, 731 (1954) (4) G A . Hamilton, R . J . Workman, and L. Woo, J . A m C h e m . Soc., 86, 3390 (1964), and references therein ( 5 ) R . 0 C . A-orman and G K . R a d d a , Pvoc Chem. Soc., 138 (1962) ( 6 ) G . A. Hamilton and J P. Friedman, J A m . Chem. S o c . , 86, 1008 (1963) T Y , (7) ( a ) J Hine, "Divalent Carbon," Ronald Press, New York, i 1964; (bj E Chinopores, C h e m Reu., 63, 235 (1963). (8) T h e reaction of a carbene with anisole h a s been studied by H Meerwein, H . DisselnkBtter, F Rappen, H . U . Kintelen, and H Van de Vloed [Ann., 604, 1.51 (1957)], and i t was observed t h a t t h e isomer distribution of methylanisoles was o v l h o . m e l a : p a v a , 34:35:31, which is not unlike t h e isomer distribution of phenols formed on hydroxylation of anisole b y t h e model system? (ovlho: mela p a v a , 43: 1 8 : 3 9 ) . T h e striking result is t h a t a large amount of meta isomer is formed in both cases. (9) S Kaufman, Proc. S a t ! Acad. Sci C . S.,60, 1085 (1963), J. B i d C h e m . , 239, 332 (1964); i b i d . , 137, PC2712 (1962).

COMMUNICATIOSS TO THE EDITOR

3392

to require a metal ion, usually F e + f the stoichiometry of the enzymatic oxidations is 0 2 S AH2 --+ SO A H20 where S, the substrate, is oxidized by two electrons to SO, and a biological reducing agent, AH2 (eventually reduced diphosphopyridine nucleotide or triphosphopyridine nucleotide in most cases), is oxidized by two electrons to A I o ; for both the enzymatic and model hydroxylations the oxygen in SO comes from molecular oxygen l o ; and the phenylalanine hydroxylase cofactor (I) is oxidized to the level of a dihydropteridine in the same step t h a t phenylalanine is hydroxylated to tyrosine.g All this evidence is consistent with the following mechanism for the model and enzymatic reactions.

+ +

+ +

\

,H

alcohols

"\/

The enediol structure, F e + + ,and O2could be in equilibrium with a complex such as V. This complex is proposed as the actual oxidizing agent. By a shift of electrons as shown an oxygen atom with six electrons could be transferred from V to an organic substrate if a t the same time the enediol complex is oxidized to the dicarbonyl complex V I . I t is not proposed that c o m plex V would generate a free oxygen atom b u t in the presence of a substrate molecule complex V could react as an "oxenoid" species (named in analogy to the carbenoid species which can transfer a divalent carbon during reaction). The formation of phenols, alcohols, and epoxides by the reaction of V with appropriate substrates should occur in a way analogous to the formation of methyl compounds and cyclopropanes by the reaction of carbenes or carbenoid species with similar substrates. The transformation of V to VI is an overall four-electron oxidation and requires that the enediol be oxidized a t the same time as the substrate is oxidized. As mentioned earlier, this is observed for phenylalanine h y d r o x y l a ~ e . ~Complex VI would be expected to add water to give V I I , which should be in equilibrium with F e + + , OH-, and the dicarbonyl compound VIII. In the model system the F e + + could recycle b u t during the oxidation the ascorbic acid would be oxidized to dehydroascorbic acid and could not recycle. l 2 However, in the enzymatic reaction the oxidized cofactor (10) ( a ) H S Mason, Adzman E n z y m o l . , 1 9 , 128 (1957); ( b ) for a recent review see "Oxygenase?." 0 Hayaishi. E d . , Academic Press, Inc , New 'iork, S Y , 1962 ( I I ) T h e mechanism is depicted for t h e ascorbate reaction b u t one of the oxygens in the enediol of ascorbate could he replaced by a n S-R a s in t h e phenylalanine hydr,,xylase cofactor ( 1 ) or t h e pyrimidine (111) with little change in mechanism expected (12) I)eh>-droascorbic acid is formed under t h e conditions of t h e oxidation h u t this is not good evidence for t h e proposed mechanism since ascorbic acid is oxidized by O? t o dehydroascorbic acid even in t h e absence of other substratec When 2 . 4 . 3 - t r i a m i n o - 6 ~ h y d r o x y ~ y ~ i m i d i n(111) e was used instead of ascixbic acid, t h e ultraviolet spectrum of t h e products indicated t h a t t h e same pyrimidopteridines were formed a s when t h e triaminohydroxy~ pyrimidine is uxidizeri in air in t h e absence of other substrates l 3 This

Vol. 86

would have to be reduced back to the tetrahydropteridine level so t h a t i t too can recycle. This would explain the requirement for a reducing agent in oxidations by mixed-function oxidases. Several variations on the proposed mechanism can be envisaged. One good possibility is that the enediol, F e + + , and oxygen form a complex such as IX which, A

by a shift of electrons and a proton as shown, could presumably transfer an oxygen atom to an organic substrate (S) and form complex X. Such a reaction would be subject to general acid and general base catalysis which could partially explain why the enzyme reaction proceeds more rapidly than the model reactions. A41so,if such a mechanism were true, then i t would not be necessary for both groups to complex with the metal ion and vinylogous enediols might also be reactants. The main features of the proposed mechanism are that a metal ion must complex with molecular oxygen and with some system capable of oxidation by two electrons. The latter is satisfied by the cofactor for phenylalanine hydroxylase and the reactants, I1 and 111, in the model system. The mechanism predicts that cofactors for other mixed-function oxidases should be capable of similar oxidation. The function of the metal ion in the proposed scheme is presumably to form an electronic link between the enediol and the oxygen.' In addition. i t probably allows conversion of the triplet oxygen molecule to a singlet species so that radical or diradical intermediates would not be necessary in the reaction of 1' with the substrate.1i A h o t h e rreaction of carbenoid species with aromatic compounds is the expansion of a benzene ring to a cycloheptatriene An analogous oxygen species would be expected to react with a benzene ring to give an oxepine and since such systems are known to cleave readily to open-chain compounds in aqueous solution l5 then a similar mechanism may be involved in the cleavage of aromatic rings in biological systems. Some such model reactions are being investigated. l 6 dimer is presumably formed from t h e d e h y d r o - 2 , 4 . . j - t r i a m i n ~ j ~ 6 - h y d r ~ j x ~ pyrimidine which the mechanism requires as the product in t h e hgdroxylation mechanism (13) E C Taylor. H . 11. Loux, l< .4 Falco, and G H . Hutchings, .I A m Chem 70, , 77, 2243 (19%); H 51 Imux, P h 1) Thesis, University uf Illinois, l9a6, p 92 (14) To describe t h e mechanism proposed here t h e author originally suggested the term "oxene mechanism' in order t u emphasize the analogy t u carhene reactions, b u t a referee has objected on t h e grounds t h a t a free "oxene," o r atomic oxygen, is not formed. This is a valid objection a n d t h e term "oxenoid," o r "direct oxygen a t o m transfer" is a better d e s c r i p t i ~ n (15) >I J Jorgenson, J 0 , ~< ' h e m , 27, 3234 (1962) (16) T h e concept of a direct oxygen a t o m insertion may be helpful in F o r example. in t h e oxidation of understanding other oxidation reactions hydrocarbons to alcohols by chromate a n d permanganate. retention of configuration is o f t e n observed 17 Ijirect carbon-hydrogen insertion would be expected t o give retention (17) K B lf'iberg and G Foster, J A m . Chem Soc , 83, 423 (19611. K H . Eastman a n d I< H Quinn. t b i d , 82, 4249 (19601, K . B U'iberg and A S F o x , i b i d , 86, 3487 (1463)

FRICKCHEMICAL LABORATORY

GORDON A . HAMIIJOY

PRINCETOS L:SIVERSITY PRINCETOS, S E \ V JERSEY

RECEIVEDAPRIL 4, 1964