UNSATURATED LIPIDE OXIDATION CATALYZED BY HEMATIN COMPOUNDS*

Hematin catalysis of unsaturated lipide oxidation is probably a primary reaction in many types of pathological unsaturated fat oxidation in viva (l-3) and in the spoilage of frozen meat products by oxidative fat rancidity (4). Some of the properties of hematin catalysis have been described (4-8), but knowledge of the reaction mechanism and the reaction products is relatively meager. The catalytic activity of myoglobin is unknown. The structural requirements for the reactants have not been characterized. Hematin catalysis may involve only a peroxidative reaction @-lo), but the possibilities of a direct oxidative reaction between the hematin compound and the unsaturated lipide have not been eliminated. It is the purpose of this research to broaden our knowledge of unsaturated lipide oxidation catalyzed by hematin compounds and to define more precisely the mechanism of the reaction.

Hematin catalysis of unsaturated lipide oxidation is probably a primary reaction in many types of pathological unsaturated fat oxidation in viva (l-3) and in the spoilage of frozen meat products by oxidative fat rancidity (4).
Some of the properties of hematin catalysis have been described (4)(5)(6)(7)(8), but knowledge of the reaction mechanism and the reaction products is relatively meager.
The catalytic activity of myoglobin is unknown. The structural requirements for the reactants have not been characterized. Hematin catalysis may involve only a peroxidative reaction @-lo), but the possibilities of a direct oxidative reaction between the hematin compound and the unsaturated lipide have not been eliminated.
It is the purpose of this research to broaden our knowledge of unsaturated lipide oxidation catalyzed by hematin compounds and to define more precisely the mechanism of the reaction. EXPERIMENTAL The hemoglobin solutions were prepared from fresh blood by laking the red blood cells with distilled water after washing the cells four times with cold saline. Myoglobin was extracted from cattle heart by the methods of Theorell (11). Hemoglobin was removed from the myoglobin by a method of differential heat denaturation involving heating of the myoglobin solution to 60' for 10 minutes (12). Myoglobin was assayed by the spectrophotometric method of Poe1 (13) to determine when it was free of hemoglobin.
The concentrations of myoglobin and hemoglobin employed as oxidation catalysts were determined by the method of Drabkin and Austin (14).
The methods used in preparing colloidal linoleate (8) and emulsions of unsaturated lipides (2) have been described previously. Emulsions of other unsaturated compounds were prepared in the same manner. Conjugated linoleic acid was prepared by alkali isomerization (15)

Catalysis of Colloidal Linoleate Oxidation by Myoglobin and Other Hematin
Compounds-Typical oxygen absorption versus time plots for myoglobincatalyzed linoleate oxidation at 0" and 37" are given in Fig. 1. Myoglobin, like hemin but unlike hemoglobin (4), catalyzes a rapid initiation of linoleate oxidation, accounted for by the oxygen lability of myoglobin. Because myoglobin readily undergoes oxidation to metmyoglobin, it was oxidized during preparation, gassing, and temperature equilibration prior to the oxygen absorption measurements.
In contrast, hemoglobin was oxidized to hemiglobin (methemoglobin) (18) only after it was added to the linoleate.
The large decrease in the rate of linoleate oxidation after 1 hour is another effect of the oxygen lability of myoglobin, in contrast to the relative stability of hemin, cytochrome c, and hemoglobin.
The rate of linoleate oxidation is a linear function of the square root of the myoglobin concentration (Fig. 2). The rates for the catalytic oxidation of colloidal linoleate by myoglobin, hemin, hemoglobin, and cytochrome c are quantitatively similar. The differences in the catalytic activities of cytochrome c, hemin, and hemoglobin previously described (8) are not significant because of the effect on the reaction rate of linoleate oxidation by small concentrations of linoleate peroxide which unavoidably develop during the preparation for manometric studies. The linear correlation between the rate of linoleate oxidation and the square root of the myoglobin concentration shows that a simple mechanism involving chain termination by a recombination of free radicals similar to that previously described (8) would be sufficient to explain these data on myoglobin-catalyzed the and Similar studies with hemoglobin from the rat, hog, and turkey ( Fig. 3) indicate that class and species differences of the hemoglobin source cause no significant difference in the rate of linoleate oxidation or deviations from the linear correlation of the oxidation rate and [hemoglobin]*.
Non-Spec$icity of Hematin Calalysis-The common property of the wide variety of unsaturated reactants listed in Table I is the formation of peroxides during oxidation (19). The intimate mechanism of hematin catalysis involves the reaction of hematin iron with a peroxide and subsequent decomposition of this activated complex into free radicals capable of initiating a chain reaction. All of the unsaturated reactants listed in Table I  3. The relation between the rate of colloidal linoleate oxidation at 0" and the concentration of hemoglobin. 0, 0, and c7, hog, rat, and turkey hemoglobin, respectively. be expected to form cyclic peroxides or polymeric peroxides (20). All the others form hydroperoxides, including squalene which forms a diperoxide consisting of one cyclic peroxide and one hydroperoxide (21). The stabil-ity of sorbic acid peroxide may prevent its reaction with and decomposition by hematin. The pronounced inhibition by small concentrations of hematin suggests that it may have been involved in chain termination of sorbic acid oxidation.
The oxidation rate at 60" was increased by hematin catalysis above the autocatalytic rate in the order cyclohexene < squalene < cumene < pcymene < tetralin.
Cumene, p-cymene, and tetralin have similar reactive structures consisting of activated methylene groups adjacent to a phenyl group.
The ratios of the oxidation rate catalyzed by 5 X lO+ M hemin to the oxidation rate without the catalyst are 158, 117, and 14 for tetralin, p-cymene, and cumene, respectively.
This effect can be partially explained by assuming that the peroxides of tetralin and p-cymene are The hemin atmosphere more stable than cumene peroxide and that all three aromatics form peroxide-hematin complexes of similar stability.
Similar to the general observation that hematin-catalyzed oxidation of free fatty acids is faster than that of esterified fatty acids was the finding that hematin catalyzed the oxidation of linseed alcohol more than that of linseed oil glycerides (2).
Linoleate Peroxide Decomposition and Concurrent Reactions-The use of emulsion systems of linoleate peroxide as reactants facilitated the direct determination of peroxides by the iodometric method. They were rapidly decomposed in the presence of hematin (Fig. 4). The decrease in the rate of hematin-catalyzed peroxide decomposition after 10 hours was due to a rapid destruction of hematin and the large decrease in the peroxide concentration. At 40" the non-catalyzed decomposition of peroxides in emulsion systems was much greater than that found in single phase lipide systems (19,22). Changes in carbonyl content during the initial part of this reaction were measured by the method of Henick et al. (23) At 40" in the presence of light, peroxide decomposition and concurrent hematin destruction were very rapid (Table II).
The differences between hematin-catalyzed and non-catalyzed peroxide destruction are best observed at 0". The ratio of moles of peroxide decomposed per mole of hematin destroyed was 1 X lo3 at 0" and 3 X lo2 at 40". The studies of Dubouloz et al. (24) likewise show concurrent destruction of lipide peroxides and hematin compounds, but the ratio of peroxide to hematin destroyed was 50 under their conditions. The observation that hematin destruction was a linear function of peroxide decomposition suggests that hematin is probably destroyed at random by a reaction with the free radical products from lipide peroxide decomposition which, during the early stages, approximated a first order reaction. To determine whether peroxide decomposition involves a scission of the carbon chain at the double bonds, simultaneous measurements of peroxide and diene conjugation were made. The decrease in conjugated dienes corresponded to approximately half of the peroxide decomposition (Table  III).
The increase in chromophores that absorb at 278 rnp was greater than that found for hemin-catalyzed linoleate peroxide decomposition and is a measure of linoleate decomposition products, principally carbonyl compounds, which absorb in this spectral region.

Spectral Studies of Hemoglobin-Catalyzed
Linoleate Oxidation-The spectral curves (Fig. 5) for linoleate oxidation products of hog hemoglobin catalysis show two maxima at 232 and 280 rnp, similar to those previously  The inset shows the relation between the spectral absorption of linoleate oxidation products and oxygen absorption.
The symbols 0, a, and q represent spectral absorption at the 278 rnp maximum when linoleate oxidation was catalyzed by rat, turkey, and hog hemoglobin, respectively.
The corresponding filled symbols are for the 232 rnp maximum. Colloidal linoleate, 0.02 N, buffered at pH 7, was used. The hemoglobin concentration ranged from 1 X 10-6 to 40 X 10-S M.

LIPIDE
OXIDATION BY HEMATIN COMPOUNDS found for hemin catalysis (8). One of the most interesting and unusual features of hemoglobin catalysis is the large absorbance of the linoleate oxidation products in the region of 280 mp. Under the same experimental conditions, the linoleate oxidation products of autoxidation, lipoxidase catalysis (8), and copper-protein catalysis (25) have low absorbance at 280 rnp. The decrease in the extinction at 232 rnp of Curves 2 and 3 indicates a decreased level of diene conjugation on the basis of the oxygen absorbed.
A decrease in the ratio of diene conjugation to oxygen absorption and an increase in the extinction at 280 rnp are the typical effects of higher hemoglobin concentrations.
The increase in extinction at 280 rnp serves as a measure of the increased rate of production of peroxide decomposition products, mainly carbonyl compounds.
The inset in Fig. 5 shows spectral absorption at the two maxima for the linoleate oxidation products of turkey, rat, and hog hemoglobin catalysis as a function of oxygen absorption.
Because three different hemoglobins were used at various concentrations, the peroxidative decomposition of conjugated diene peroxides was a variable, causing wide scattering of the data.
In general, these data show that the relatively high absorption in the region 278 rnp and the relatively low conjugated diene absorption at 232 rnp are characteristic for hemoglobin catalysis, irrespective of source. Based on the increase in spectral absorption at 232 or 278 rnp as a function of time, a rapid and direct spectrophotometric determination of the rate of linoleate oxidation catalyzed by hematin compounds has been developed.
The use of this method is shown later in this paper. Denatured Globin Hemichromes-Attempts to observe the intermediate complex of hemiglobin or metmyoglobin and linoleate peroxide were made. Mixing various concentrations of hemiglobin or metmyoglobin and salts of linoleate peroxide resulted in the formation of denatured globin hemichromes (18) whose spectral characteristics are so similar to those anticipated for the intermediate complex (26) that its formation could not be detected.
The very rapid formation of denatured globin hemichrome by mixing hemiglobin or metmyoglobin with sodium or potassium salts of linoleic, oleic, stearic, or lauric acid (Fig. 6) can be ascribed to a nonspecific denaturation of the globin by these surface-active compounds (18).

Absence of Valence Change during Hematin Catalysis-The
effect of carbon monoxide, aside, and cyanide on hemoglobin catalysis was determined by use of a direct spectrophotometric measure of the formation of the 278 rnp chromophore as a function of time.
The results (Fig. 7) show the inhibition of cyanide. Non-inhibition by carbon monoxide is strong evidence against a valence shift in iron during hematin compound catalysis of linoleate oxidation (18), since it combines strongly with ferrous ion in hematin compounds.
To circumvent the possibility of light dissociation 729 of carboxyhemoglobin, linoleate oxidation catalyzed by hemin and hemoglobin was carried out in the dark in Thunberg tubes containing gas mix- and hemiglobin azide. The reaction at 22" was fol-Iowed by the direct spectrophotometric method. The reactant was 3 ml. of 0.02 M colloidal linoleate, pH 9, saturated with oxygen.
The concentration of hematin compounds was 3.4 X lOmE M. The concentrations of cyanide and azide were 4 X 10-4 and 2 X 10e4 M, respectively. tures of 0.5 atmosphere of CO pressure + 0.5 atmosphere of O2 and 0.5 atmosphere of Nz + 0.5 atmosphere of OZ. Measurement of linoleate oxidation by spectral absorption at 232 and 278 rnp after 1 hour showed that the reaction rate was the same for both gas mixtures. Thus, there was no inhibition by carbon monoxide. If hematin compound catalysis of linoleate oxidation involved a valence change, the activity of the catalyst should be a function of its oxidationreduction potential (27). Conversion of hemin to hemichromes by reaction with nitrogenous bases varied the oxidation-reduction potential over a wide range. Ammonia and pilocarpine hemichromes accelerated the reaction to 175 and 160 per cent of the control, whereas cyanide, nicotine, and histidine inhibited to 39, 52, and 66 per cent of the control. The pyridine, a-picoline, and piperidine hemichromes gave reactions equal to hemin.
These effects are not pronounced, and the catalytic activity does not correlate with the known oxidation-reduction potentials (18). E$ect of Hydroxyl Ion Concentration--The bonding of hydroxyl ion to the sixth bond position of the iron in hematin compounds at pH values above 7 (18) often changes their peroxidative activity (28). Since such changes furnish valuable information about the ability of the peroxides to replace the bound hydroxyl groups, the effect of increased hydroxyl ion concentration on hemoglobin-catalyzed linoleate oxidation was studied. Between pH 7.8 and pH 9.5 the oxidation rate was increased over 2-fold. This indicates that the formation of the lipide peroxide-hemiglobin complex is not suppressed by the increased binding of hydroxyl ion. This reaction is markedly different from that of hydrogen peroxide with alkaline peroxidase or alkaline hemiglobin, which is suppressed by increased hydroxyl ion concentrations (28). DISCUSSION The critical part of the mechanism of hematin compound catalysis of lipide oxidation is the initiation reaction in which free radicals are formed. Initiation involves the reaction of a hematin compound with a lipide peroxide to form an intermediate complex and the subsequent decomposition of the intermediate complex into free radicals. Reaction of hematin compound and peroxides is a very general phenomenon.
The peroxidative reactions of hemoglobin, myoglobin, catalase, and peroxidase with hydrogen peroxide and short chain alkyl peroxides are well known (18,26). This research and others (4,8) have shown that hematin, cytochrome c, metmyoglobin, hemiglobin, and many of their derivatives are powerful catalysts for the oxidation of a wide variety of unsaturated lipides and other peroxide-forming organic compounds. The iron porphyrin and the peroxide group are the essential reactants.
The probable reaction between the hematin nucleus and linoleate peroxide is represented as in the accompanying arrangement.
The reaction between linoleate peroxide and hematin to form the intermediary complex is probably reversible, as are reactions of simpler peroxides with hematin compounds (28 The destruction of the hematin catalysts which occurs during linoleate peroxide decomposition (Table II) and during the subsequent linoleate oxidation is probably caused by a random reaction of the hematin compound with a free radical.
This reaction may be similar to the concurrent oxidation of carotenoids and antioxidants in the hematin-catalyzed reaction (2,4). Cooxidation of carotenoids and antioxidants most probably involves abstraction of hydrogens by linoleate peroxide radicals.
Because of the high reactivity of the free radical intermediates, many non-specific cooxidations can be expected to occur during hematin-catalyzed lipide oxidation.
The evidence presented in this research against a valence change in the iron strongly suggests that the iron porphyrin-lipide peroxide formation and subsequent decomposition involve no compulsory valence changes. Information relative to the bonding of lipide peroxides to hematin compounds can be obtained from inhibition studies. Two types of inhibition have been observed.
Cyanide forms a very stable hemichrome and probably inhibits because the lipide peroxide cannot displace the cyanide ion from the iron atom in the heme nucleus.
Other nitrogenous compounds like ammonium ion form weak hemichromes and cause increased catalysis.
Increasing the hydroxide ion concentration has the same effect of increasing the catalytic reaction.
Ammonium ions and hydroxyl ions may be readily replaced by peroxides and may function in this manner to increase hematin catalysis.
Other nitrogenous compounds like histidine and tryptophan and dyes like methylene blue and thionine are not known to form hemichromes of great stability, yet are good inhibitors (2). These relatively large molecules may function as inhibitors primarily through steric effects. A large molecule attached to the iron in the heme nucleus could block the lipide peroxide from a close approach to the iron atom. SUMMARY Hematin compounds are non-specific catalysts for the decomposition of peroxides.
Myoglobin has catalytic activity quantitatively similar to that of hemin, cytochrome c, and hemoglobin.
A wide range of unsaturated lipides and peroxide-forming organic compounds is catalytically oxidized by hemin.
During the decomposition of linoleate peroxide there is a loss of double bonds and formation of carbonyl compounds.
Also, hematin is concurrently destroyed. The spectral absorption of linoleate oxidation