Kinetics of Oxygen and Carbon Monoxide Binding to Synthetic Analogs of the Myoglobin and Hemoglobin Active Sites*

Proc. Nat. Acad. Sci. USA Vol. 72 No. 3, pp. 1166-1170 March 1975

Kinetics of Oxygen and Carbon Monoxide Binding to Synthetic Analogs of the Myoglobin and Hemoglobin Active Sites* (oxyheme/base effect/solvent effect/oxygenation kinetics) C. K. CHANG AND T. G. TRAYLORt Department of Chemistry, University of California, San Diego, La Jolla, Calif. 92037 Communicated by Nathan 0. Kaplan, January B, 1976

ABSTRACT Kinetics of reversible oxygenation and hemoglobin oxygenation varies over about 5-fold (lc). On the carbon monoxide complex formation of the simple heme compounds pyrroheme-N-13-(1-imidazolyl)propyllamide other hand, carbon monoxide affinity varies very little (ld). and pyrroheme4-(3-pyridyl)propyl ester have been mea- The effects suggested to contribute to these changes are sured in different solvent environments. The oxygen on and listed on Fig. 1 and enumerated below. off rates and equilibria ofthese compounds can be made to 1. Strain: Because the proximal imidazole is invariant in closely match those of myoglobin, of hemoglobin a chains, or of the various steps for hemoglobin by varying solvent oxygen carriers and because it seems to change position upon environment or the basicity of the proximal base. These oxygenation, the changes in oxygen affinity as hemoglobin is results suggest that the protein could alter oxygen on oxygenated have been attributed to movement of this rates by varying the basicity of the proximal base and the imidazole toward or away from the heme plane (le, 5). off rates by changing the polarity of the distal environment. 2. Proximal imidazole basicity: It has also been suggested that the hydrogen bonding of the proximal imidazole proton The ability of the single compound protoheme to either re- to a neighboring peptide increases the basicity of this imid- versibly bind oxygen (la), catalyze oxidation of organic sub- azole, thus increasing the oxygen affinity (6a, 7). strates (2a), decompose hydrogen peroxide (3b, 4) or transport 3. Hydrophobic pocket: The hydrocarbon-like residues electrons (2b), depending upon its environment, is remarkable pointing into the distal side of the heme pocket provide a (3b). Even among the first class of compounds, the oxygen "hydrophobic pocket" which is thought to both prevent carriers, there are large variations in the kinetics and equi- oxidation and favor oxygenation (8, 9). Examination of the

2. Proximal basicity

5. Electronic effects

Distal imidazole H bond Distal imidazole polarity

3a. Hydrophobic pocket 3b. Steric interference

FIG. 1. Heme protein site environment and proposed effects on oxygenation kinetics and equilibria. libria of oxygenation which have been the subject of extensive distances within this distal pocket also suggested that steric investigations. The oxygen affinity of various oxygen carriers interference between the entering ligand and certain side varies about 1000-fold (lb) and that isq the four steps of chains would affect binding (if). 4. Distal imidazole: The distal imidazole hydrogen could and retard dissociation The fact *Presented, in part, at the Priestley Bicentennial Confer- bond to bound oxygen (id). or ence, Pennsylvania. State University, August 1, 1974, and the that some hemoglobins having, e.g., arginine (10) iso- Annual Meeting of the American Chemical Society, Atlantic leucine (11) instead of this histidine, nevertheless bind oxygen, City, N.J., September 2, 1974. shows that this is not a requirement. t To whom requests for reprints should be addressed. 5. Electronic effects: Reconstituted hemoglobins and myo- 1166 Downloaded by guest on September 29, 2021 Proc. Nat. Acad. Sci. USA 72 (1975) Kinetics of Reversible Heme Oxygenation 1167

globins have increased oxygen affinity as the heme side chains techniques, they were reduced in a flask attached to the side- are made less electron withdrawing. However, there are also arm. Aqueous solutions were reduced by careful titration with suggestions that this effect is steric rather than electronic (if). aqueous sodium dithionite, using the Soret band to monitor The dominant theme in these discussions is that the mag- reduction. Solutions in organic solvents were reduced with nitudes of equilibria are determined by specific steric inter- palladium under hydrogen gas (18). After reduction, the actions which vary the strain engendered by converting a solutions were decanted into the cuvette under argon and the five-coordinate iron to a six-coordinate iron (5). However, the sidearm flask was replaced with a bulb of known volume. almost invariant carbon monoxide affinity in these heme The apparatus was then evacuated and placed in the photo- proteins (lb) seems to argue against purely stereochemical flash apparatus and flashed with an approximately 40 joule effects and to suggest that a variable polarity effect in the light of 100 jssec duration and 40 psec decay time. There was distal pocket (12, 13) might be important. The increasing no change in absorption at 407 nm. Aliquots of carbon mon- oxygen affinity of an isolated heme compound with increasing oxide and then oxygen were added with gas-tight syringes polarity of solvent also suggests a polarity effect (13, 14). and the solution was shaken vigorously to equilibrate the The uncertainty in these suggestions could be reduced gases with the solutions. The solutions were flashed between greatly by a detailed kinetic study of oxygenation of an each aliquot. Cleanly first-order decays were obtained in each isolated heme protein active site in different environments. case, exactly like those obtained with myoglobin on the same We have made two discoveries which now make such kinetic apparatus. Decay constants were calculated from photo- studies possible. First, by attaching 3-aminopropyl-1-imid- graphs of the oscilloscope traces of transmission at 407 nm azole to pyrroheme (15, 16), we produced a single compound versus time. Alternatively the transmission output voltage which not only shows the same spectra as deuteromyoglobin was electronically processed through a log amplifier and the in its oxy, deoxy, carbon monoxy, and hemin forms (15, 16), resulting absorbance was passed through a second log ampli- but more importantly, behaves kinetically (kinetic order, etc.) fier to give a linear trace whose slope is the decay constant. like heme proteins such as myoglobin. Secondly, we found that Rate constants obtained after successive gas addition agreed small amounts of carbon monoxide stabilize simple hemes to +5%. The flash photolysis apparatus, constructed from a toward oxidation much more effectively than had been Zeiss PMQ-II spectrometer, a Braun F022 flashgun, and a thought (6b), thus making kinetic studies of oxygenation of Tektronix 564B Oscilloscope, will be described elsewhere. isolated heme compounds almost as routine as those on heme In some cases oxygen off rates were obtained by stopped- proteins (17). flow technique using the oxygen pulse method of Gibson (19), We report here the kinetics of the reversible reaction of and carbon monoxide on rates were also determined by the two heme compounds shown below with carbon monoxide and usual rapid mixing techniques. The results are in agreement with oxygen in several solvent systems. with those reported here. RESULTS The reactions under investigation are: it Heme + CO = Heme-CO [1] hiv Heme + 02 = Heme-02 [2] k R Heme-02-- Heme-CO [3] where the heme is either 1 or 2. With carbon monoxide present I' is measured directly as a first-order decay, l'(CO) = Ptobs, which includes concentration of CO (Fig. 2a). However, when oxygen is added two decays are observed (Fig. 2b). The fast decay to a certain absorbance AHO, (where H represents the 1 = 1 Base hemes or 2) is k'(02) k'ob., followed by a decay R from -NH(CHA-N-3N this absorbance to AHCO. Under our conditions the Eq. 4 of Gibson (la, 20) is valid. 2 Base= -O(CH2?3 1 c'02 = _ [4] MATERIALS AND METHODS R k kl'CO Pyrroheme-N-[3-(1-imidazolyl)propyl]amide (1) (15) and If k is small then k' can be obtained directly from the first pyrroheme-N-13-(3-pyridyl)propyl]ester (2) (13) were pre- decay, k'obs. However, when k is large (>1000) then k'0b5 can- pared by the acid chloride method previously described (15). not be measured directly because large oxygen concentrations Solutions, approximately 6 ml of 2 AM, were prepared in are necessary to get appreciable concentrations of heme-02, purified solvents or solvent mixtures (by weight), in a 1 cm X and this makes k'ob. too fast to measure. In these cases k' 1 cm cuvette which was attached through 20 cm of 1 cm tubing is obtained from the slope of Eq. 4. In all cases, k is obtained to a sidearm and a vacuum T stopcock. This tonometer was from the intercept of a plot of 1/R versus O2/CO as in Eq. 4. equipped with a silicon rubber septum through which gases For k, only gas pressures are required, but to obtain 1' and k' could be quantitatively admitted. Total tonometer volume the concentrations and thus the solubilities of CO and 02 are was 100 ml. After the solutions were degassed by freeze-thaw needed. The results are tabulated in Table 1. Downloaded by guest on September 29, 2021 1168 Chemistry: Chang and Traylor Proc. Nat. Acad. Sci. USA 72 (1975) TABLE 1. Kinetic constants for reactions of hemes with carbon monoxide and oxygen at 22°*

I' V (M-1 (M-1 k'/k sec1 see-' (M-1 X X k X Heme Solvent 106) 10-) (sec-) 10-) 1 CH2C12 10 90 600 0.15 HCON(CH3s), 8 90 160 0.5 HCON(CHs)2/H20 (2/1) 10 85 90 1.0 HI0, pH = 7.3, sodium phosphate- buffer, 2% cetyl- trimethylammo- nium bromide 30 60 35 1.7 2 Toluene 7.5 35 2500 0.001 HCON(CH3)2 14 9.6 450 0.02 HCON(CH,)2/H20 (2/1) 15 20 400 0.05 HIO, pH = 7.3 sodium phosphate buffer, 2% cetyl- trimethylammo- nium bromide 20 14 150 0.09 * Solubilities of CO and 02 in various solvents were obtained from ref. 21. With solvent mixtures, it was assumed that solubilities are linear functions of the composition.

be excluded by the observations that the on rates are different for 1 and 2. We, therefore, conclude that both hemes are reacting in the configuration shown in the drawing of 1 and 2. The first important finding is that oxygenation of both 1 and 2 is enhanced by polar solvents, whereas carbon monoxide FIG. 2. (a) Oscilloscope trace of voltage versus time for the equilibria are rather insensitive to solvent polarity (22, 23; recombination of 1 and CO after flash photolysis at 220 in 2% C. K. Chang, unpublished results). This agrees with solvent cetyltrimethylammonium bromide, pH 7.3 sodium phosphate effects on oxygenation equilibria previously determined by buffer. [CO] = 8.3 torr (1.1 kPa) = 10 JAM; sweeptime = 2 titration with oxygen (13). The fact that these solvent msec/division; wavelength = 407 nm. (b) The recombination of changes have their predominant effect upon dissociation 1 with 02 and CO at 407 nm. [CO] = 10 [02] = 16.6 torr ;&M, rates is strong confirmation of the ability of polar solvents (2.2 kPa) = 29 IAM. Upper trace: sweeptime = 0.5 msec/division; lower trace: sweeptime = 0.1 sec/division. In both (a) and (b) to stabilize the FeOO bond. This and the proximal base vertical change = 2.5% transmission/division. effect (22) have been interpreted as evidence for highly di- polar character of this bound (13) DISCUSSION 8+ Fe-O Before discussing solvent or structural effects on oxygenation \a- 0 the state of the deoxy compound must be established. Hemes 1 and 2 are presumably pentacoordinate, but could have water The effect of increasing the basicity of the fifth ligand upon or solvent in the sixth position. Although solutions of 1 or 2 oxygen binding, which was previously determined by titration in dimethylformamide or H20 have broad absorptions at (22), is corroborated by the kinetic data. However, it appears about 550 nm at room temperature which split into usual a that the "proximal basicity" affects both the oxygen on and off and bands at lower temperatures, exactly the same behavior rates with, in some cases, the effect being mainly on the on is found in anhydrous toluene or polystyrene (15, 16). How- rates, k'. This introduces the possibility of controlling on and ever, addition of imidazole retards both 1' and k', the effect off rates separately to obtain any desired oxygenation equi- increasing with increasing base concentration. But, especially librium. for 1, both 1' and k' are rather insensitive to solvent and the Effect of on more basic solvents dimethylformamide and H20 do not protein pocket oxygenation retard 1' or k'. Thus, there does not seem to be a kinetically The idea that the "hydrophobic pocket" affords oxyheme significant specific iron solvation of the heme in these solu- some special stability towards not only oxidation but dis- tions. sociation has been repeatedly expressed (8, 9). The simplest The possibility that the hemes 1 and 2 are reacting with way to test this idea is to remove all the protein from the carbon monoxide and with oxygen as four-liganded hemes can heme (except for the proximal imidazole) and to see what Downloaded by guest on September 29, 2021 Proc. Nat. Acad. Sci. USA 72 (1975) Kinetics of Reversible Heme Oxygenation 1169 TABLE 2. Kinetics of oxygenation of heme proteins TABLE 3. Hemoglobin* versus model compounds and their "active sites" (kinetics at 20 or 220)

(M-1 k'/k (M-1 sec1- k K02 sec-1 k (M-1 Heme X 10-7) (sec') (M-1 X 10-) Heme X 10-7) (sec-1) X 10d) Ref. 2 (Toluene) 0.3 2500 0.001 1 (in H20, pH = Hb 0.2 1080 0.009 7.3) 6.0 35 1.7 This work Hb(02) 0.4 244 0.017 Isolated a chains 5.5 31 1.8 24 2 (120) 1.4 150 0.09 Myoglobin (sperm Hb(02)2 0.35 28 0.12 whale) 1.9 10 1.9 24 Hb(02)3 4.0 48 0.8 Myoglobin 1 (H20) 6 35 1.7 (Aplysia) 1.5 70 0.22 25 * See ref. 26.

effect this has upon oxygenation. The result depends some- tively, hydrogen bonding of the hydrogen on this imidazole what upon the heme protein chosen for comparison, as shown could change its basicity, as suggested by Caughey (7). in Table 2. This table shows clearly that, in water, the Because we can duplicate most of the oxygenation kinetics of isolated site binds oxygen with the same constant as does heme protein oxygen carriers with simple hemes in solution, myoglobin and shows kinetics almost identical to those of we conclude that no special structural environment such as isolated a chains. While it is possible that there is a coinci- direct steric interaction of substituents in the pocket of the dental cancellation of effects which causes myoglobin to bind heme protein is necessary for this binding. Furthermore, it oxygen as does the isolated site in a very polar environment, seems possible that polarity effects could be the dominant it is simpler to search for some similarity which brings this factor in differentiating oxygen affinities in heme proteins. about. How can the "hydrophobic pocket" be polar? One However, our models invariably react faster with carbon possibility is a dipole-dipole interaction -between the monoxide than does myoglobin (1' = 5 X 105 M-1 sec1 ref. or Fe--O--O dipole and the distal imidazole. lg) or hemoglobin (lh, 28). Because proximal basicity solvent changes do not appear to reduce 1' substantially, some

H 1 other factor seems to be responsible for the slow reaction of some heme proteins with carbon monoxide. Steric interference (le) could account for the reduced rate, but such interference should also destabilize the carbon monoxide complex, which I a+ Fe- it apparently does not do. An alternative possibility relates to the geometry and spin state of the heme (29, 30). Because the singlet CO molecule -0/ must combine with a high spin -Fe(II) to pair these spins giving a singlet product, the necessity for spin inversion could N retard the rate of reaction. But triplet oxygen, by spin inter- / action at the transition state, might not be retarded by slow H spin inversion to the same extent. Perhaps triplet oxygen can induce the S=0 state of iron even before the heme becomes The center of the imidazole is only a little over 3 X from the planar. This effect could explain the faster on rates of both center of the Fe-00- (le). Furthermore, although Aplysia oxygen and nitric oxide compared to carbon monoxide and the myoglobin has a much greater oxygen dissociation rate con- differences between the isolated sites and heme proteins. stant than does sperm whale myoglobin (Table 2), they have Indeed, activation volumes indicate that the transition states similar carbon monoxide kinetics and equilibria (25). We for the interaction of these two ligands with myoglobin are suggest that conformational and other changes could alter quite different (28). But the evidence for such a spin effect on the relative positions and directions of these two dipoles, rates is not very compelling and the reasons for the differences both of which are in a nonpolar environment, and this change in on rates for these rather similar molecules remains obscure. would greatly affect k and thus the oxygenation equilibria. Conclusions Hemoglobin oxygenation The kinetics and equilibria of oxygenation steps of oxygen- The four steps in the oxygenation of hemoglobin show large carrying heme proteins can be approximately duplicated by a differences in both on and off rates (26, 27). We show in synthetic "myoglobin active site" in which proximal basicity Table 3 how, with a combination of proximal base strength and solvent environment are varied. We conclude from these and solvent effects, some of these steps can be duplicated. results that the protein serves more to protect the hemes from The comparison in this table suggests that a combination of oxidation than to bring about binding of oxygen. Variations distal side polarity adjusted by moving the distal imidazole in oxygen affinities among heme proteins are interpreted as and proximal basicity could bring about all the changes in resulting more from polarity of the heme environment than binding displayed by hemoglobin. from stereochemical factors. The proximal basicity could be changed by the movement We are grateful to Dr. Charles Perrin and Dr. John Wright for of the proximal imidazole as suggested by Perutz (5). Alterna- helpful advice, to David Epstein and Richard Threlkel for as- Downloaded by guest on September 29, 2021 1170 Chemistry: Chang and Traylor Proc. Nat. Acad. Sci. USA 72 (1975)

sistance, and to the National Institutes of Health for support 14. Stynes, H. C. & Ibers, J. A. (1972) J. Amer. Chem. Soc. 94, (Grant HL 13581). 5125-5127. 15. Chang, C. K. & Traylor, T. G. (1973) Proc. Nat. Acad. Sci. 1. Antonini, E. & Brunori, USA 70, 2647-2650. M. (1971) in Hemoglobin and Myo- 16. Chang, C. K. & Traylor, T. G. (1973) J. Amer. Chem. Soc. globin in Their Reactions with Ligands (North Holland 95, 5810-5811. Publishing Co., Amsterdam), (a) p. 153; (b) p. 351; (c) p. 17. Traylor, T. G., Chang, C. K. & Geibel, J. (1974) Proc. An- 256; (d) p. 349; (e) p. 85; (f) p. 355; (g) p. 226; (h) p. 266. nual Meeting of the American Chem. Soc., Atlantic City, 2. Lemberg, R. & Barrett, J. (1973) Cytochromes (Academic N.J. INOR-30. Press, New York), (a) p. 217; (b) p. 122. 18. Brinigar, W. S. & Chang, C. K. (1974) J. Amer. Chem. Soc. 3. Hayaishi, 0. (1974) Molecular Mechanism of Oxygen 96, 5595-5597. Activation (Academic Press, New York), (a) p. 10; (b) p. 19. Gibson, Q. H. (1973) Proc. Nat. Acad. Sci. USA 70, 1-4. 548. 20. Noble, R. W. & Gibson, Q. H. (1969) J. Biol. Chem. 244, 4. Winfield, M. E. (1961) in Haematic Enzymes, eds. Falk, J. 3905-3908. E., Lemberg, R. & Morton, R. K. (Pergamon Press, New 21. Linke, W. F. & Seidell, A. (1958) Solubilities of Inorganic York), p. 245. and Metal-Organic Compounds (Van Nostrand, Princeton, 5. Perutz, M. F. (1970) Nature 228, 726-734. N.J.), 4th Ed. 6. Wang, J. H. (1962) in Oxygenases, ed. Hayaishi, 0. (Aca- 22. Chang, C. K. & Traylor, T. G. (1973) J. Amer. Chem. Soc. demic Press, New York), (a) p. 499; (b) p. 503. 95, 8477-8479. 7. Caughey, W. S. (1966) in Hemezs and Hemoproteins, eds. 23. Chang, C. K. & Traylor, T. G. (1973) J. Amer. Chem. Soc. Estabrook, R. E. & Yonetani, T. (Academic Press, New 95, 8475-8477. York), p. 285. 24. Brunori, M. & Schuster, T. M. (1969) J. Biol. Chem. 244, 8. Wang, J. H., Nakahara, A. & Fleischer, E. B. (1958) J. 4046-4053. Amer. Chem. Soc. 80, 1109-1113. 25. Wittenberg, B. A., Brunori, M., Antonini, E., Wittenberg, J. B. & Wyman, J. (1965) Arch. Biochem. Biophys. 111, 576- 9. Wang, J. H. (1958) J. Amer. Chem. Soc. 80, 3168-3169. 579. 10. Winterhalter, K. H., Anderson, N. M., Amiconi, G., An- 26. Ilgenfritz, G. & Schuster, T. M. (1974) J. Biol. Chem. 249, tonini, E. & Brunori, M. (1969) Eur. J. Biochem. 11, 435- 2959-2973. 440. 27. Antonini, E. & Gibson, Q. H. (1960) Biochem. J. 76, 534- 11. Humber, R., Epp, 0. & Formanek, H. (1970) J. Mol. Biol. 538. 52, 349-354. 28. Hasinoff, B. H. (1974) Biochemistry 13, 3111-3117. 12. Cole, S. J., Curthoys, G. C. & Magnusson, E. A. (1971) J. 29. Hoard, J. L. (1971) Science 174, 1295-1302. Amer. Chem. Soc. 93, 2153-2158. 30. Gouterman, M. & Zerner, M. (1966) in Hemes and Hemopro- 13. Brinigar, W. S., Chang, C. K., Geibel, J. & Traylor, T. G. teins, eds. Estabrook, R. E. & Yonetani, T. (Academic (1974) J. Amer. Chem. Soc. 96, 5597-5599. Press, New York), p. 589. Downloaded by guest on September 29, 2021