Vol. 52 ERYTHROCYTE PERMEABILITY 511 Kolthoff, I. M. (1926). Biochem. Z. 176, 101. Parpart, A. K., Jacobs, M. E. & Dzienmian, A. J. (1937). Lohmann, K. & Jendrassik, L. (1926). Biochem. Z. 178, Biol. Bull. Woods Hole, 73, 381. 419. Peters, J. P. & Van Slyke, D. D. (1932). Quantitative Luckner, H. & Lo-Sing (1938). Pflug. Arch. gem. Phy8iol. Clinical Chemistry, Vol. 2. London: Bailliere, Tindall and 239, 278. Cox. Maizels, M. (1933). J. Physiol. 77, 22P. Ponder, E. (1948). Haemolysis and Related Phenomena. Maizels, M. (1934). Biochem. J. 28, 337. London: Churchill. Maizels, M. (1944). Quart. J. exp. Physiol. 32, 143. Warburg, E. J. (1922). Biochem. J. 16, 153.

The Reaction between Metmyoglobin and Hydrogen Peroxide

BY P. GEORGE AND D. H. IRVINE Department of Colloid Science, Univer8ity of Cambridge (Received 8 Augu8t 1951) Transient complexes with well defined spectro- ferrocyanide shows it to have 1 oxidizing equi- scopic characteristics play an important role in the valent, and magnetic measurements suggest that it reaction of the haemoprotein , catalase may be a compound containing quadrivalent . and peroxidase, with hydrogen peroxide. Two and A preliminary account of these experiments has sometimes three different complexes can be ob- been published (George & Irvine, 1951). served depending on the experimental conditions. For example, Keilin & Mann (1937) obtainedtwo red MATERIALS AND METHODS complexes with peroxidase and hydrogen peroxide, and later Theorell (1941) showed that a green com- Metmyoglobin. The metmyoglobin (MetMb) used in these plex preceded the formation ofthe first red complex. experiments was obtained from horses' hearts, the method of This system has been studied intensively by preparation following that of Theorell (1932). Briefly, the minced hearts are extracted with water at 0° and muscle Chance (1949 a, b) who extended the observation to globulins removed by addition of basic lead acetate to the systems of catalase and hydrogen peroxide or extract, the pH of which has been previously adjusted to alkyl hydroperoxides (1949c). There is, however, 7 0 with 1 N-NaOH. The excess Pb is removed as phosphate very little information from which structural with Na3PO4, and haemoglobin precipitated by adding differences accounting for the various complexes (NH4)2S04 to 75 % saturation. MetMb is precipitated when can be inferred. The transient nature of the com- the solution is saturated to 100% with (NH4)2SO4 and is plexes leads to experimental difficulties. taken up in the minimum quantity of 3M-phosphate buffer, These difficulties were absent in the case of the pH 6-8. Phosphate is removed by dialysis against distilled complex formed with metmyoglobin (MetMb) and water and MetMb again salted out. It is then taken up in 0 6M-phosphate buffer, pH 5 7, passed through a Seitz filter hydrogen peroxide. The existence of a well defined and stored at 00. red complex in this system and with haemoglobin Quantities of this stock solution were thoroughly di- too is well established. Keilin & Hartree (1935) alysed against distilled water before use and the concentra- found that the formation requires 1 molecule of tion determined following the method of de Duve (1948). hydrogen peroxide per haematin iron atom. With Hydrogen peroxide. Pure 97% (w/w) H202, free from reducing agents the complex is reduced back to inhibitors, was used. This was kindly supplied by Laporte metmyoglobin or methaemoglobin. As far as is Chemicals Ltd. Before each experiment a stock solution known, only this single complex is formed and (approx. 0.1 N) was standardized against KMnO4 and then because it shares diluted to the desired concentration. certain properties with the Buffer 8olutiona. In the pH range 58-7-0 phosphate secondary complexes of peroxidase and catalase, a buffers (Green, 1933) were used. For the pH range 78-10-0 detailed knowledge of its behaviour is valuable for borate buffers were used. Experiments at pH 11.1 were the further elucidation of reaction mech- carried out in a glycine buffer, while at higher pH values anisms. dilute alkali was used. The pH values of all buffer solutions In this paper experiments are described on the were measured on a Cambridge pH meter. formation of the metmyoglobin-hydrogen peroxide complex over a wide pH range which reveal that Method8 only between pH 8 and 9 is the formation un- Colorimetric or spectrophotometric determinations of attended by oxidative side reactions affecting the optical density were made in the majority of the experi- porphyrin ring. Reduction of the complex with ments. Because ofits easier and more rapid application, the 512 P. GEORGE AND D. H. IRVINE I952 former method was used wherever possible (pH range The complex is a good oxidizing agent and will 5.8-9.0). Yellow filters were found to give the most suitable oxidize ascorbic acid, quinol, p-phenylenediamine, wavelength region. In the more alkaline regions and also in leuco-malachite green, iodide ion (but not bromide) experiments requiring a complete analysis of spectra in the and ferrocyanide ion, itselfbeing reduced to MetMb. visible and Soret regions, spectrophotometric measure- The rate of oxidation is dependent on pH and is ments were made using a Unicam quartz spectrophoto- meter. faster in more acid solutions. Owing to the relative slowness of the formation and reduction ofthe H202-MetMb complex at the concentrations used and especially in the alkaline region, it was necessary 10 to follow the reactions kinetically in order to ascertain when full formation or reduction had taken place. 8- The percentage complex formation or reduction was cal- culated by interpolation using Lambert and Beer's law. The optical density of a solution ofMetMb and H202 was found to reach a maximum value when a slight excess of H02 was -4 present, e.g. a molar ratio between 1-5 and 2-5 to 1 de- 4 pending on the pH ofthe solutions. This optical density was taken to be that corresponding to 100% complex forma- 2 tion. The procedure is justified by the experiments of Keilin & Hartree (1950), who found that H202 con- 0 centrations in great excess (100 to 1) were required 650 600 550 500 before appreciable reaction of H202 and the complex itself Wavelength (m,u.) occurred. Fig. 2. Effect of pH on complex. Visible region 650- For the magnetic measurements a modification of the 500 0, complex at pH 8-4; x, complex at pH 10 0; Theorell (1942) microbalance was used. The calibration was mp. carried out with ferrous ammonium sulphate solutions. *, complex at pH 6-8; ®, complex at pH 5-8.

RESULTS Qualitative experiments indicated that a reddish violet complex was formed throughout the pH range 150 5 0-12 0 when H202 was added to MetMb. The

10

100 9 ,

8 w

7

50 6

5 I - I I I I~~.

4 I,,,,,, 1 ol I I - 450 440430420410400390380 450 430 410 370 3 Wavelength (m/l.) Wavelength (m,u.) 650 600 550 500 Fig. 3. Effect of pH on complex. Soret region; (a) smooth

Wavelength (mM.) curve represents complex at pH 8-4; ®, MetMb, pH 6-8; Fig. 1. Recovery of MetMb from complex. Azide spectra; 0, MetMb, pH 11-1. (b) CZ, complex at pH 8-4; V7, complex at pH 10 0; A, complex at pH 11 1; *, complex smooth curve represents MetMb + azide; 0, reduced at pH 6-8; ®, complex at pH 5-8. complex at pH 8-4 + azide; 0, reduced complex at pH 11-1 +azide. Experiments show that between pH 8-0 and 9 0 complex is comparatively stable. Between pH 8 the reaction between H202 and MetMb is straight- and 9 slight decomposition could be detected at the forward in the sense that the minimum amount of end of half an hour after formation and complete side reactions occur. Thus it is possible within this destruction was apparent after about 48 hr. region to recover 90-100% of MetMb from the standing at room temperature. The end product complex by the addition of mild reducing agents. was mainly MetMb. The rate of destruction was By comparison only about 65 % MetMb is recover- more rapid in acid solution. able at pH 11i1 (Fig. 1). Below pH 8-0 and especi- VoI. 52 METMYOGLOBIN AND HYDROGEN PEROXIDE 513 ally at pH values < 7 0 side reactions occur pro- addition to the bands at 542 and 573 m,u. of the ducing one or more compounds with strong ab- true haemochromogen, a band at 630 m,u. There is sorption in the red. Above pH 9*0 the products of also a considerable reduction of the bands at 542 side reactions have no such absorption. Thus in and 573 m,u. (see Fig. 4). At pH 6-8 and 5.8, Figs. 2 and 3, which show the spectra ofthe complex in the visible and Soret region, it will be seen that 16 the difference between the spectrum of the complex at pH 10*0 and at pH 8-4 is mainly a difference in 14 intensity. Table 1 gives the millimolar extinction coefficients of the complex at 549 and 423 m,u. at 12 pH values 8-3, 10-0 and 11 1. On the other hand, there is a definite shift of the 10 6 band in the Soret region when the complex is formed at a pH < 8-0; at pH 6-8 the band is at 420-423 m,. and at pH 5 8 it is at 413 m,u. (Fig. 3b). In the visible region the difference is borne out by the appearance 6- of a new band at 590 m,. (Fig. 2). 4- Table 1. Millimolar extinction coefficient (e x 10-3) of complex at 549 and 423 mbt. at different pH values 2- ex10-8 at Ex10-8 at pH 549 mg. 423 m,. 0 8.4 9 8 107.0 650 600 550 500 10.0 8-3 87*0 Wavelength (my.) 11.1 6-5 62.0 Fig. 4. Haemochromogen spectra, 650-500 m,., from complex formed at pH> 8-0; smooth curve is haemo- Table 2. Effect on extinction and position of peak of chromogen from MetMb; ®3, from complex formed at changing the pH after complexformation at a given pH 8-6; *, from complex formed at pH 8-0; A, from pH complex formed at pH 11 1. (a) Complex formed at pH 6-8 Position of peak pH (MP.)(m)x< 10-3 14 6-8I 420-423 9930 9.0 420-423 94~0 12 11.1 420-423 94~0 (b) Complex formed at pH 9.0 10 6 Position of peak pH (mf&.) c x :10-3 9-0 423 10'o2-0 6'8 423 100-0 11.1 423 1011.0 6-

An interesting observation on this effect of pH is 4- that once the complex is formed changing the pH seems to have little influence on the spectrum. 2 Table 2a, b shows some of the results obtained from measurements in the Soret region. This shows 0 that the influence of different hydrogen-ion con- 650 600 550 500 centration is exerted during the formation reactions Wavelength (mp.) and the effect is not due to a change in spectrum of Fig. 5. Haemochromogen spectra, 650500 me. from the complex as the pH is varied. complex formed at pH<8.0. Smooth curve is haemo- The nature ofthe side reactions is not fully under- chromogen from MetMb; ®D, from complex formed at stood. It would seem that the reduction of the pH 6*8; 0, from complex formed at pH 5-8. intensity, which is observed when the complex is formed at a pH > 9 0, is due to varying amounts of however, a different band appears at 615 mu., and cholemyoglobin formed. Thus the - there is very little reduction in the bands of the haemochromogen (obtained by reducing the com- true haemochromogen (Fig. 5). In addition, the plex at pH 11.1 with Na2SO4 in the presence of Soret spectrum of the haemochromogen obtained carbon monoxide and adding 20 % alkali, cf. from the complex formed at pH 5-8 is even more Lemberg, Legge & Lockwood (1941)) shows, in intense than that of the true haemochromogen Biochem. 1952, 52 33 514 P. GEORGE AND D. H. IRVINE I952 (Fig. 6). Substitution ofthe porphyrin ring without concentration and rules out the possibility of an ring fission seems a likely explanation. equilibrium of the type The following observations show that these side MetMb + H202 = complex. reactions occur during the actual formation of the complex. If MetMb is allowed to stand with the The results of the experiments are shown in Fig. 7. preformed complex for about 1 hr. at either pH 5.8 or 11-1, the haemochromogen spectrum obtained from the mixture is the same as that produced when 0-01~~~~~~~~~~~ C- 15 E 50 -, x E |, 0 0 00 05 1.0 1.5 Ratio: H202/MetMb Fig. 7. Effect of concentration on complex formation. i, MetMb 1.0 x 10-4M; FC, MetMb 4.9 x 10-51M; 0, MetMb 8-8 x 10-6M. Broken line is theoretical curve °x W based on a 1:1 ratio of H,02 to MetMb. 10j 100

C e0 50 --

E so - 450 425 400 370 0 Wavelength (mix.) Fig .6. Haemochromogen spectra. Soret region; *, 0 5 1.0 1F5 from MetMb; E3, from complex: formed at pH 5-8. Ratio: H202/MetMb Fig. 8. Effect of pH on complex formation. Broken line is MetMb is present with the comaplex for a few theoretical curve based on a 1:1 ratio of H202 to MetMb. minutes. This suggests that an oxridizing entity 0, pH 8-0, MetMb 4.7 x 10-5M; E, pH 8-4, MetMb more powerful than the complex itself is present 5.0 x 10-5m; ®, pH 9-0, MetMb 4-8 x 10-5M. during the formation of the complex. Qualitative Fig. 8 shows the effect of pH on complex forma- experiments with luminol support this. If luminol tion which is small within the pH region 8-0-9 0. is added to the complex a dull luminescence is Although in the figure the maximum complex obtained, whereas if luminol and MetMb axe mixed formation occurs at about pH 8-0-8-4, the pH at and H202 then added (under conditions where the which maximum formation is observed appears to complex is formed in a few seconds) a bright flash is differ to a small extent for different samples of observed followed by a dull luminescence. MetMb. In none of the samples used, however, was this pH outside the range 8 0-9-0. Quantitative mea,surement8 on the formation of the complex Reduction of the complex The formation of the complex between pH 8-0 Experiments with ferrocyanide ion as the re- and 9 0 is characterized by the fact that a l: l ratio ducing agent show that 1 mole of ferrocyanide will of H202 to MetMb does not produce 100 %/ complex reduce 1 mole of complex completely to give formation even in a concentration ofMetMb as high MetMb, within the pH region 8&0-90. Typical a,s 10-4m. At this concentration 80-90 %/ ofcomplex results of these experiments are shown in Table is produced, the percentage varying between these 3a, b. limits for different samplesi of MetMb. If, instead of adding ferrocyanide to the pre- The effect of concentration on the formation is formed complex, the ferrocyanide is first added to only very slight over a tenfold change of absolute MetMb in varying amounts and hydrogen peroxide VoI. 52 METMYOGLOBIN AND HYDROGEN PEROXIDE 515 Table 3. Reduction of the metmyoglobin-hydrogen peroxide complex byferrocyanide (a) pH 8*0 Initial Observed Complex 4- complex Ferrocyanide complex left reduced Ratio Fe(CN) ('"M) (PM) (MM) (PM) Complex 24 10 14 10 1-00 24 15 10 14 1-07 24 24 1 23 1-04 40 30 9 31 0-97 (b) pH 8-4 Initial Observed Complex complex Ferrocyanide complex left reduced Ratio: Fe(CN)64- (MM) (MM) (MM) (PM) Complex 12 10 2 10 1.00 30 30 2 28 1-07

figures may be compared with those quoted by Table 4. Reaction of metmyoglobin with hydrogen Hartree (1946): reduced peroxidase Xm X 1o-6 c.g.s. peroxide in the pre8ence offerrocyanide at pH 8 units= 11 410; haemoglobin Xm X 10-6 c.g.s. units (In these experiments ferrocyanide was added before =12290, uB=5-43; xmxO-6c.g.s. hydrogen peroxide.) units= 12 400, MB= 5 46. (See footnote p. 516.)

(a) Initial MetMb and H202 concentrations 1-04 x 10-4M and 1.01 x 10-4M respectively DISCUSSION Oxidizing Complex equivs. of H202 The first significant fact which emerges is the effect Fe(CN)- formed accounted for of pH on the reaction between hydrogen peroxide (1MM) (MM) (,uM)* and metmyoglobin. The reaction can be said to be 0 91 91 45-0 'normal' within the pH range 8 0-9-0. Outside this 120 72 192 95-0 140 56 196 97-0 region there is the peculiarity on the alkaline side of 160 38 198 98-0 a reduction of the intensity of the complex, and on * the acid side the production of a different spectrum On the basis 1 mole complex= mole Fe(CN)64- (Figs. 2 and 3). The = 1 equiv. H202. 202 equivs. present initially. haemochromogen spectrum obtained from the complex formed at pH 11 1 (b) Initial MetMb and H202 concentrations 5-2 x 10-5M (Fig. 4) shows a peak at 630 m,. and a reduction of and 5-0 x 10-5M respectively the intensities of the peaks at 542 and 573 m,. of Oxidizing the true Complex equivs. haemochromogen. This is typical of the Fe(CN) 4- formed accounted for behaviour which Lemberg et al. (1941) associate (MM) (PM) (,uM)* (%O) with choleglobin formation. Moreover, the forma- 0 37 37 37-0 tion of choleglobin, more specifically in this case of 60 29 89 89-0 cholemyoglobin, would account for the absence of 70 23 93 93-0 any marked in the visible 80 14 94 94*0 absorption and Soret 90 5 95 95.0 regions other than that attributed to the complex. On the acid side, however, this explanation will * On the 1 = 4- basis mole complex mole Fe(CN) not hold. The spectra = 1 equiv. H202. 100 equivs. present initially. haemochromogen obtained from the complex at pH 6-8 and 5 8 show a peak at then introduced, complex formation is observed 615 m,u., not at 630 mn., and there is little reduction of the peaks at 542 and 573 mit. (Fig. 5). This fact, even when the amount of ferrocyanide present is in together with the observation that in the excess of the amount ofcomplex expected. Table 4a, Soret b gives the results of two typical experiments of region there is not only no decrease, but an actual this nature at different absolute concentrations. rise in the intensity of the haemochromogen spectrum obtained from the complex at pH 5*8 above the true haemochromogen (as shown in Magnetic properties of the complex Fig. 6), suggests that a different type of oxidative Preliminary measurements ofthe molar magnetic attack on the porphyrin ring has occurred. This type susceptibility ofthe complex at pH 8 6 indicate that ofattack presumably does not affect the conjugated it is about 11 000-11 800 x 10-6 c.g.s. units at room double-bond system to the same extent that temperature. This corresponds to a magnetic cholemyoglobin formation does, for the intensity of moment of 5-1-5-3 Bohr magnetons (MB) and sug- light absorption in the Soret region is believed to be gests the presence of four unpaired electrons. These a characteristic of the conjugated system. 33-2 516 56P. GEORGE AND D. H. IRVINE I952 The absence of any effect of pH on the spectrum On the basis of this last feature, the equation for of the complex once it has been formed and the the formation of the complex can be written in its fact that the complex does not apparently react simplest form either with itself or added metmyoglobin to give the MetMb + -+ complex + X, side reactions, indicate clearly that these side H202 reactions occur during the formation ofthe complex. 2 oxidizing equivs. The experiments with luminol suggest that a -* 1 oxidizing equiv. +1 oxidizing equiv. transient oxidizing agent, more powerful than the A tentative mechanism accounting for the com- complex, is produced during the formation reaction. plex and X each with 1 oxidizing equivalent is the The data of the other experiments reveal three formation of an OH radical and ferryl myoglobin in important features: a reaction of the type (1) The complex is not formed in any set of Fe+(H2O)+ H202 - FeO + OH + H2O+ H+, equilibrium reactions involving metmyoglobin and (complex) (X) hydrogen peroxide, either in their ionized or un- ionized forms, e.g. where Fe+ represents the haematin iron atom of metmyoglobin. Ferrylmyoglobin by analogy with MetMb + H202 = complex. the hypothetical ferryl ion FeO2+ and the well known vanadyl ion VO2+ would be a compound of In this respect its formation is unlike that proposed quadrivalent iron and would thus have 1 oxidizing by Chance (1949b, 1951) for the peroxidase and equivalent when reduced to a compound in the catalase complexes, i.e. ferric state. The identification of the complex with ferrylmyoglobin would be in accordance with ...... FeOOH FeOH+ H202 + H20. magnetic measurements which show the presence of This suggests that neither H202 nor HO2- is present fourunpaired electrons.* The electronic structures of in the metmyoglobin complex. the Fe2+, Fe3+ and Fe4+ ions are given in Fig. 9. (2) Full formation ofthe complex is not obtained 3d 4s 4p Unpaired for a 1:1 ratio of hydrogen peroxide to metmyo- electrons . In view ofthe fact that the system is not an Fei+X 0 0 0 0 o 000 4 equilibrium system this result must mean that hydrogen peroxide or metmyoglobin or complex (or more than one of'these) is disappearing during the Fe3+ 2)00000 0 000 5 reaction. The disappearance of metmyoglobin can be ruled out since increasing the peroxide ratio increases complex formation. Furthermore, within FeW 00000 0 000 4 the pH region 8 0-9 0 nearly all the metmyoglobin Fig. 9. Electronic structures of Fe2+, Fe3+, Fe4+. can be even in this 100 recovered, though region % It can be 2 formation is not observed for 1: I-ratio. If complex seen that both Fe and Fe4+ ionic is the destruction must take compounds would have four unpaired electrons. being destroyed, place But in the case of the this while formation is occurring, since once the complex MetMb-H202 complex, has been formed no destruction is detectable for clearly cannot be a ferrous compound for it is a strong oxidizing agent which on reduction yields a about half an hour (within the pH region 8.0-9-0). ferric In the absence of an the destruction of compound. equilibrium The identification of X with an OH radical fits in could be as complex only interpreted producing with the observations that X is an some new entity which would be both colourless (1) oxidizing and reducible to to for agent. (2) X is transient; only during the actual metmyoglobin, account, first, formation of the complex is its the lack of absorption and, secondly, for the com- oxidizing ability of at certain observed. (3) During the formation of the complex plete recovery metmyoglobin pH a flash is observed if is No flash is values. This does not seem likely. luminol present. (3) The oxidizing equivalent of the complex is * Note added in proof: Theorell & Ehrenburg (1952) unity (Table 3). Hence the full oxidizing equivalent have reported a susceptibility value of 3800 x 10-6 c.g.s. ofthe peroxide can only be accounted for by assum- units: our more recent experiments show that the high ing that some other oxidizing entity also with value of about 11000 x 10-6 C.g.s. units quoted above is in 1 oxidizing equivalent is produced the re- error on account of decomposition of the complex during during the measurements. We find the correct value to be action of hydrogen peroxide and metmyoglobin. < 8400 x 10-6 c.g.s. units. A covalent instead of an ionic Evidence supporting this is obtained by adding a structure containing quadrivalent iron would be in accord reducing agent to the metmyoglobin before intro- with these new values, for d28ps covalent bond formation ducing hydrogen peroxide when the full oxidizing of Fe4+ results in two unpaired electrons (Xm= 3390 x 10-6 capacity of the peroxide is realized (see Table 4). c.g.s. units) as can be seen from Fig. 9. Vol. 52 METMYOGLOBIN AND HYDROGEN PEROXIDE 517 obtained if luminol is added after the complex has but arise by some other mechanism during the for- been formed. mation of the complex. The side reactions could be accounted for be- 3. The system MetMb+H,H0, is not an equi- cause of the high oxidizing activity of the OH librium system. The effect of concentration on the radical. formation of the complex is only very small over a SUMMARY tenfold change of absolute concentration. 4. The oxidizing equivalent of the complex is 1. Metmyoglobin forms a complex with hydrogen unity. By adding a reducing agent to the metmyo- peroxide throughout the pH range 5-0-12-0. Only globin before peroxide it is possible to account for within the pH range 8 0-9 0 is this reaction un- the two oxidizing equivalents of the peroxide. This attended by oxidative attack on the porphyrin ring. indicates the reaction to be of the type 2. On the alkaline side the attack on the ring seems to give cholemyoglobin, for a regular decrease MetMb + H202 -* complex + X. in the intensity of the absorption bands is observed both in the visible and Soret regions ofthe spectrum. 5. A tentative mechanism is proposed identi- In solutions more acid than pH 8-0 a different type fying the complex with ferrylmyoglobin, a com- ofoxidative attack occurs, for no regular decrease in pound of quadrivalent iron, and X with the OH intensity is observed, but there is a shift ofthe band radical. The high oxidizing activity of the OH maxima. These side reactions are not due to the radical could account for the oxidative side re- reaction of the complex itself with metmyoglobin, actions.

REFERENCES Chance, B. (1949a). Arch. Biochem. 21, 416. Keilin, D. & Hartree, E. F. (1935). Proc. Roy. Soc. B, 117, 1. Chance, B. (1949b). Arch. Biochem. 22, 224. Keilin, D. & Hartree, E. F. (1950). Nature, Lond., 166,513. Chance, B. (1949c). J. biol. Chem. 179, 1331. Keilin, D. & Mann, T. (1937). Proc. Roy. Soc. B, 122, 119. Chance, B. (1951). In The Enzyme8, 2, pt. 1, ed. by Sumner, Lemberg, R., Legge, J. W. & Lockwood, W. H. (1941). J. B. & Myrback, K. New York: Academic Press. Biochem. J. 85, 339. Duve, C. de (1948). Acta chem. 8cand. 2, 264. Theorell, H. (1932). Biochem. Z. 252, 1. George, P. & Irvine, D. H. (1951). Nature, Ld., 168, Theorell, H. (1941). Enzymologia, 10, 250. 164. Theorell, H. (1942). Ark. Kemi Min. Geol. 16A, no. 1. Green, A. (1933). J. Amer. chem. Soc. 55, 2331. Theorell, H. & Ehrenburg, A. (1952). 11th Int. Congr. Hartree, E. F. (1946). Rep. Progr. Chem. 43, 287. Biochem. Abetr. p. 285.

The Ionization of Acidic Metmyoglobin

BY P. GEORGE A G. HANANIA Department of Colloid Science, University of Cambridge (Received 16 Augu8t 1951) The problem of establishing thermodynamic con- Some equilibrium data on the ionization of acidic stants for the ferric complexes of haemoproteins is methaemoglobin in water, which is equivalent to important in at least two respects. In the first place the formation of the hydroxide complex, have bcen such constants offer a workable criterion where a published by several workers but no values were parallelism is sought between the behaviour of obtained for AH or AS. Furthermore, there appears simple ferric iron and that of the iron atom in the to be an interesting anomaly in the effect ofneutral porphyrin ring of haemoproteins. Secondly, a salts on this ionization constant. An inspection of knowledge of such data would offer additional the results of Austin & Drabkin (1935) which were evidence for elucidating hitherto uncertain based on spectrophotometric studies of canine structures, for instance those of haemoprotein- methaemoglobin, and those of Coryell, Stitt & peroxide complexes, by comparing approximate Pauling (1937), derived from magnetic measure- thermodynamic constants for the formation of these ments on bovine methaemoglobin, shows that this complexes with corresponding data forthe formation ionization constant falls with increasing ionic of, say, the hydroxide, fluoride or cyanide complexes strength, i.e. pK moves away from the isoelectric where the structure is better known. point of the , whereas the effect of neutral