Measurements of CO Geminate Recombination in Cytochromes P450 and P420
1995; Elsevier BV; Volume: 270; Issue: 15 Linguagem: Inglês
10.1074/jbc.270.15.8673
ISSN1083-351X
AutoresWeidong Tian, Andrew V. Wells, P. M. Champion, Carmelo Di Primo, Nancy Counts Gerber, Stephen G. Sligar,
Tópico(s)Pharmacogenetics and Drug Metabolism
ResumoThe kinetics of CO geminate recombination in cytochrome P450cam are studied at room temperature subsequent to laser photolysis. The geminate rebinding kinetics of P450 are strongly affected by the presence of the camphor substrate. We observe a ∼2% geminate yield for substrate-bound P450 and a 90% geminate yield when the substrate is absent. The drastic difference in the geminate kinetics suggests that the presence of camphor significantly alters the CO rebinding and escape rates by modifying the heme pocket environment. Two geminate phases and two bimolecular rebinding phases in the substrate free protein were observed, which could arise from slowly interconverting protein conformations. When the temperature or the viscosity of the solution is changed, the fast geminate rate remains the same, whereas the slow geminate rate and the two bimolecular rates change significantly. The geminate rebinding yield of substrate-free P420 is smaller than that of substrate free P450, but its geminate rebinding rate is faster. This demonstrates that in the absence of substrate, CO escapes from the pocket of P420 much more rapidly than from P450 and suggests that the distal pocket environment is altered in the P420 form. The kinetics of CO geminate recombination in cytochrome P450cam are studied at room temperature subsequent to laser photolysis. The geminate rebinding kinetics of P450 are strongly affected by the presence of the camphor substrate. We observe a ∼2% geminate yield for substrate-bound P450 and a 90% geminate yield when the substrate is absent. The drastic difference in the geminate kinetics suggests that the presence of camphor significantly alters the CO rebinding and escape rates by modifying the heme pocket environment. Two geminate phases and two bimolecular rebinding phases in the substrate free protein were observed, which could arise from slowly interconverting protein conformations. When the temperature or the viscosity of the solution is changed, the fast geminate rate remains the same, whereas the slow geminate rate and the two bimolecular rates change significantly. The geminate rebinding yield of substrate-free P420 is smaller than that of substrate free P450, but its geminate rebinding rate is faster. This demonstrates that in the absence of substrate, CO escapes from the pocket of P420 much more rapidly than from P450 and suggests that the distal pocket environment is altered in the P420 form. Cytochromes P450 are globular heme protein enzymes that catalyze the oxidation of various xenobiotics and endogenous compounds. Cytochrome P450cam, which catalyzes the hydroxylation of camphor, is obtained from Pseudomonas putida and has been widely studied as a model for P450 monoxygenases using many techniques (1Gunsalus I.C. Meeks J.R. Lipscomb J.D. Debrunner P. Münck E. Hayaishi O. Molecular Mechanisms of Oxygen Activation. Academic Press, Yew York1974: 559-613Google Scholar, 2Dawson J.H. Sono M. Chem. Rev. 1987; 87: 1255-1276Crossref Scopus (491) Google Scholar, 3Debrunner P.G. Gunsalus I.C. Sligar S.G. Wagner G.C. Siegal H. Metal Ions in Biological Systems. Vol. 7. Marcel Dekker, New York1978: 241-271Google Scholar, 4Champion P.M. Spiro T.G. Biological Applications of Raman Spectroscopy. Vol. 3. John Wiley & Sons, New York1988: 249-292Google Scholar). The three-dimensional structure of P450cam in both substrate-free and -bound forms has been determined by x-ray studies (5Poulos T.L. Finzel B.C. Gunsalus I.C. Wagner G.C. Krant J. J. Biol. Chem. 1985; 260: 16122-16130Abstract Full Text PDF PubMed Google Scholar, 6Poulos T.L. Finzel B.C. Howard A.J. Biochemistry. 1986; 25: 5314-5322Crossref PubMed Scopus (557) Google Scholar). One important result is that the heme active site of P450cam is deeply buried inside the protein, and there is no obvious access channel for substrate and diatomic ligand binding. This means that the protein must undergo structural fluctuations or interconversions, which allow the passage of small molecules to the active site. Recent photoacoustic calorimetry studies (7Di Primo C. Hui Bon Hoa G. Douzou P. Sligar S.G. J. Biol. Chem. 1990; 265: 5361-5363Abstract Full Text PDF PubMed Google Scholar) show that the presence or absence of the camphor substrate in P450cam significantly affects the dynamics of the protein. The camphor evidently leaves the heme pocket when CO is dissociated (7Di Primo C. Hui Bon Hoa G. Douzou P. Sligar S.G. J. Biol. Chem. 1990; 265: 5361-5363Abstract Full Text PDF PubMed Google Scholar, 8Wells A.V. Li P. Champion P.M. Martinis S.A. Sligar S.G. Biochemistry. 1992; 31: 4384-4393Crossref PubMed Scopus (109) Google Scholar), which suggests that important conformational changes within the protein take place upon photolysis that may be linked to the opening of an access channel. Camphor-free P450 has a high CO affinity, but when camphor binds, the CO affinity is reduced by a factor of 10, and the association rate is slowed by about 2 orders of magnitudes (9Peterson J.A. Griffin B.W. Arch. Biochem. Biophys. 1972; 151: 427-433Crossref PubMed Scopus (56) Google Scholar). Infrared experiments (10O'Keefe D.H. Ebel R.E. Peterson J.A. Maxwell J.C. Caughey W.S. Biochemistry. 1978; 17: 5845-5852Crossref PubMed Scopus (71) Google Scholar, 11Jung C. Hui Bon Hoa G. Schröder K. Simon M. Doucet J.P. Biochemistry. 1992; 31: 12855-12862Crossref PubMed Scopus (99) Google Scholar) reveal that substrate-free P450cam-CO has a broad and slightly structured CO-stretching band. The multiple signals indicate that P450cam exists in a dynamic equilibrium involving several conformational substates. Binding of camphor or camphor analogues strongly influences this equilibrium (11Jung C. Hui Bon Hoa G. Schröder K. Simon M. Doucet J.P. Biochemistry. 1992; 31: 12855-12862Crossref PubMed Scopus (99) Google Scholar) and analogous resonance Raman experiments have demonstrated significant differences in the Fe-CO vibrational frequencies as a function of substrate (8Wells A.V. Li P. Champion P.M. Martinis S.A. Sligar S.G. Biochemistry. 1992; 31: 4384-4393Crossref PubMed Scopus (109) Google Scholar, 12Bangcharoenpaurpong O. Champion P.M. Martinis S.A. Sligar S.G. J. Chem. Phys. 1987; 87: 4273-4284Crossref Scopus (33) Google Scholar, 13Uno T. Nishimura Y. Makino R. Iizuka T. Ishimura Y. Tsuboi M. J. Biol. Chem. 1985; 260: 2023-2026Abstract Full Text PDF PubMed Google Scholar). Resonance Raman investigations (8Wells A.V. Li P. Champion P.M. Martinis S.A. Sligar S.G. Biochemistry. 1992; 31: 4384-4393Crossref PubMed Scopus (109) Google Scholar) have also shown that the presence of camphor substrate in P420 (the inactive form of P450) samples has little effect on the Raman spectra in the oxidized, reduced, or CO-bound states. The P420 heme appears to be in equilibrium between a high-spin, five-coordinate form and a low-spin six-coordinate form in both the ferric and ferrous oxidation states (8Wells A.V. Li P. Champion P.M. Martinis S.A. Sligar S.G. Biochemistry. 1992; 31: 4384-4393Crossref PubMed Scopus (109) Google Scholar). In the ferric state of P420, H2O remains as a heme ligand, whereas alterations of the protein tertiary structure lead to a significant reduction in affinity for Cys357 thiolate binding to the heme iron. In ferrous P420, H2O and histidine are the most likely axial ligands (8Wells A.V. Li P. Champion P.M. Martinis S.A. Sligar S.G. Biochemistry. 1992; 31: 4384-4393Crossref PubMed Scopus (109) Google Scholar). This evidence indicates that the heme environment and the camphor-binding site of P420 differ significantly from that of P450. It is likely that the altered tertiary protein structure leads to changes in key rate constants associated with diatomic ligand binding to the heme as well as substrate binding to the protein. The reaction cycle of P450cam basically includes four stable intermediates (1Gunsalus I.C. Meeks J.R. Lipscomb J.D. Debrunner P. Münck E. Hayaishi O. Molecular Mechanisms of Oxygen Activation. Academic Press, Yew York1974: 559-613Google Scholar, 14Sligar S.G. Biochemistry. 1976; 15: 5399-5406Crossref PubMed Scopus (353) Google Scholar). The initial state is the substrate-free cytochrome, P450mo(ferric iron, spin 1/2). Upon binding the camphor substrate, a heme-ligated H2O molecule is expelled to form the high-spin state, P450mos (ferric iron, spin 5/2). When the heme iron is reduced to P450mrs (ferrous iron, spin 2), dioxygen can bind to the heme forming P450mO2rs. Subsequent to the input of a second electron, the oxygenated intermediate decays with rapid product formation. To better understand the catalytic process, we have studied the kinetics of diatomic ligand binding to P450cam in the presence and absence of the camphor substrate. Such studies are sensitive to the heme environment as well as the protein structure and dynamics. In this investigation we explore the ligand binding kinetics of CO to P450. We utilize CO because it serves as a stable “reporter group” that allows easy comparison to other heme protein systems (MbCO, HbCO) that have been extensively studied (15Frauenfelder H. Sligar S.G. Wolynes P.G. Science. 1991; 254: 1598-1603Crossref PubMed Scopus (2614) Google Scholar, 16Gibson Q.H. J. Biol. Chem. 1989; 264: 20155-20158Abstract Full Text PDF PubMed Google Scholar, 17Olson J.S. Mathews A.J. Rohlfs R.J. Springer B.A. Egeberg K.D. Sligar S.G. Tame J. Renaud J. Nagai K. Nature. 1988; 336: 265-266Crossref PubMed Scopus (226) Google Scholar, 18Henry E.R. Sommer J.H. Hofrichter J. Eaton W.A. J. Mol. Biol. 1983; 166: 443-451Crossref PubMed Scopus (241) Google Scholar, 19Tian W.D. Sage J.T. rajer V. Champion P.M. Phys. Rev. Lett. 1992; 68: 408-411Crossref PubMed Scopus (123) Google Scholar). Although the bimolecular dissociation and association kinetics of CO to P450 have been measured earlier (9Peterson J.A. Griffin B.W. Arch. Biochem. Biophys. 1972; 151: 427-433Crossref PubMed Scopus (56) Google Scholar, 20Debey P. Balny C. Douzou P. FEBS Lett. 1973; 35: 86-90Crossref PubMed Scopus (16) Google Scholar, 21Gray R.D. J. Biol. Chem. 1982; 257: 1086-1094Abstract Full Text PDF PubMed Google Scholar, 22Shimada H. Iizuka T. Ueno R. Ishimura Y. FEBS Lett. 1979; 98: 290-294Crossref PubMed Scopus (18) Google Scholar, 23Oertle M. Richter C. Winterhalter K.H. Di Iorio E.E. Proc. Natl. Acad. Sci. U. S. A. 1985; 82: 4900-4904Crossref PubMed Scopus (12) Google Scholar, 24Shimizu T. Ito O. Hatano M. Fujii-Kuriyama Y. Biochemistry. 1991; 30: 4659-4662Crossref PubMed Scopus (31) Google Scholar, 25Miura Y. Kawato S. Iwase T. Ohta S. Hirobe M. Biochemistry. 1991; 30: 3395-3400Crossref PubMed Scopus (3) Google Scholar, 26Koley A.P. Robinson R.C. Markowitz A. Friedman F.K. Biochemistry. 1994; 33: 2484-2489Crossref PubMed Scopus (15) Google Scholar, 27Shiro Y. Kato M. Iizuka T. Nakahara K. Shoun H. Biochemistry. 1994; 33: 8673-8677Crossref PubMed Scopus (28) Google Scholar), the nanosecond geminate recombination of CO to P450 has not been reported previously. Such observations are crucial for a determination of the fundamental kinetic rate constants of heme proteins (18Henry E.R. Sommer J.H. Hofrichter J. Eaton W.A. J. Mol. Biol. 1983; 166: 443-451Crossref PubMed Scopus (241) Google Scholar) (e.g. ligand entry to the active site, kin; ligand escape to the solvent, kout; and ligand binding to the heme, kBA) as shown below for the three-state model. P450CO⇄kBAkLP450:CO⇄kinkoutP450 + COMODEL 1 The fundamental rate constants can be calculated from the experimentally determined geminate rebinding rate (kg), geminate amplitude (Ig), and bimolecular rate (ks), kBA=Igkg,kin=ks/Ig,kout=kg(1−Ig)(Eq. 1) under the conditions that kout, kBA ≫ kin, kin = k′in [CO] with [CO] ≫ [P450], and the photolysis/off rate, kL ∼ 0 when t > 0. In this work, we present the CO geminate rebinding kinetics of P450 and P420, both in the presence and absence of substrate and in different glycerol/buffer solutions at different temperatures. P450mos was generated and purified from P. putida as described previously (28Gunsalus I.C. Wagner G.C. Methods Enzymol. 1978; 52: 166-188Crossref PubMed Scopus (302) Google Scholar). High purity P420mos was prepared by exposure of P450mos to a pressure of 2 kbar at room temperature for 1-2 h (29Hui Bon Hoa G. Di Primo C. Dondaine I. Sligar S.G. Gunsalus I.C. Douzou P. Biochemistry. 1989; 28: 651-656Crossref PubMed Scopus (61) Google Scholar, 30Martinis, S. A., 1990, Ph.D. thesis, University of Illinois, Urbana-Champaign, IL.Google Scholar). Substrate-bound samples contained 0.4-1.0 m M camphor and 100 m M KCl. Concentrated frozen samples were thawed and diluted in 0.1 M sodium phosphate buffer at pH 7.0-7.2. The substrate-free protein P450mowas obtained by passing a concentrated P450mos sample through a small column of Sephadex G-25 which contained the same buffer. A cold centrifuge was used to reconcentrate the sample. The purity of the P450mopreparation was determined by checking the absorbance ratio at 416.5 nm (the Soret peak of P450mo) relative to 390 nm (the Soret peak of P450mos). In our experiments, the ratio was ≥ 2.7 which indicates the purification of P450mowas better than 97%. P450mowas reduced to P450mrby anaerobically adding a small amount of concentrated sodium dithionite solution (∼20:1 with respect to protein concentration). P450mCOr was obtained by bubbling CO gas slowly through the P450mrsample for 20 min. When preparing glycerol/buffer mixed samples, a longer gas equilibration time (∼1 h) was used. All the preparation procedures were carried out at 0−2°C, in order to maintain the sample (8Wells A.V. Li P. Champion P.M. Martinis S.A. Sligar S.G. Biochemistry. 1992; 31: 4384-4393Crossref PubMed Scopus (109) Google Scholar). CO solubility changes as a function of temperature (31Wleth V.L. Seidell A. Solubilities of Inorganic and Metal Organic Compounds. Vol. 1. D. Van Nostrand Co., New York1940: 217-219Google Scholar) and glycerol concentration (32McKinnie R.E. Olson J.S. J. Biol. Chem. 1981; 256: 8928-8932Abstract Full Text PDF PubMed Google Scholar). The changes of the concentration of CO as a function of the temperature in 75% glycerol were treated as proportional to the change in water. The CO concentrations used for the samples at different temperatures and solutions are listed in Table I. These concentrations are used to find the bimolecular CO association rates, kon.Table I:The kinetics of CO rebinding to P450 and P420All the data are least squares fit using Equation 2. Since the time resolution is limited, the kinetics faster than 10 ns are not observed. This leads to 10–20– “missing” amplitude as determined by the ▵A(O) obtained from the equilibrium CO bound and unbound species. We estimate that the cumulative effect of all sources of uncertainty in the geminate rebinding give an error in the range of 10–20– for amplitudes and 20–50– for rates. All the data are least squares fit using Equation 2. Since the time resolution is limited, the kinetics faster than 10 ns are not observed. This leads to 10–20– “missing” amplitude as determined by the ▵A(O) obtained from the equilibrium CO bound and unbound species. We estimate that the cumulative effect of all sources of uncertainty in the geminate rebinding give an error in the range of 10–20– for amplitudes and 20–50– for rates. The details of the flash photolysis experiment have been presented elsewhere (19Tian W.D. Sage J.T. rajer V. Champion P.M. Phys. Rev. Lett. 1992; 68: 408-411Crossref PubMed Scopus (123) Google Scholar). The sample was photolyzed by the 532-nm line of a frequency-doubled Nd:YAG laser (10-ns pulse width) and the kinetics of the ligand rebinding was probed by an argon-pumped cw dye laser. The absorption signal was monitored by a highly linear photomultiplier circuit and averaged about 103times using a 350-MHz digital oscilloscope. The voltage signal was subsequently converted to a differential absorbance between the bound and unbound states, using Δ A(t) = log(V(t)/ V(0)), where V(0) and V(t) are proportional to sample transmittance before and after photolysis. The quantity Δ A is renormalized to give the surviving unbound population, N(t) = Δ A(t)/Δ A(0), where Δ A(0) is the absorbance difference at t = 0 as determined from equilibrium measurements on bound and unbound samples (or from their relative extinction coefficients). Fitting functions are allowed to have a variable normalization, N0, and the deviation of N0from unity, which is a measure of the goodness of fit at t = 0, is considered in the assessment of the fitting function's validity. If N0lies outside the range 1.0 ± 0.2, the fit is considered unreliable, since non-equilibrium relaxation processes are unlikely to alter the t = 0 absorbance changes by more than ± 20% (19Tian W.D. Sage J.T. rajer V. Champion P.M. Phys. Rev. Lett. 1992; 68: 408-411Crossref PubMed Scopus (123) Google Scholar). Since the P450mrsample is less stable at high temperature, most experiments were carried out at 275 K. Special anaerobic cells were used to maintain CO gas pressure, and the sample temperature was stabilized with a circulating bath (±1 K). The optical spectra of the samples were taken before and after addition of CO and at the conclusion of the experiment to check the quality of the sample and to determine the absolute difference between the P450mCOr and P450mrabsorbance (Δ A(0)). Photoexcitation of each heme in the sample volume is assured by use of a polarization scrambler, counter propagating beam geometry, and sufficient pumping energy (10 mJ/pulse). Due to the limited time resolution of the instruments, kinetics faster than 10 ns are not observed. This leads to 10-20%“missing” amplitude, as determined by the expected equilibrium absorption change (Δ A(0)) between P450mCOr and P450mrat the probe wavelength. It should be noted that this methodology does not account for the possibility of spectral evolution caused by nonequilibrium protein relaxation (19Tian W.D. Sage J.T. rajer V. Champion P.M. Phys. Rev. Lett. 1992; 68: 408-411Crossref PubMed Scopus (123) Google Scholar); however, such effects are expected to be relatively small (5%). Another source of uncertainty in the geminate kinetics arises from radiofrequency noise from the Nd:YAG laser. This effect has been limited by shielding and grounding the instruments and by subtracting the background response, with the probe light blocked, from the signal. The remaining noise still affects the accuracy of the measurement at the level of ≤ 1% near 10−7 s. The difference between the absorption spectra of the CO-bound and unbound states of P450 (P420) can be utilized in flash photolysis experiments to optically study the kinetics. The left panel of Fig. 1 shows the absorption spectra of camphor-free P450 and P420 in both the CO-bound and unbound states. In the flash photolysis experiments, CO is dissociated from the heme during the 10-ns photolysis pulse. Some CO molecules remain in the heme pocket and rebind to the heme, which gives rise to the geminate rebinding kinetics. The remaining CO molecules diffuse out of the heme pocket into the solvent. On longer time scales CO rebinds from the solvent giving rise to the bimolecular rebinding kinetics. The time course of CO binding measured at 447 nm was best resolved using two geminate (CO concentration-independent) exponential phases and two bimolecular (CO concentration-dependent) phases. In the right panel of Fig. 1, we present the kinetics of P450mCOr in 50% glycerol solution at 275 K. The solid line is a fit with two exponential geminate phases, the dash-dot line is a fit with one exponential geminate phase, and the dashed line is a fit with a “stretched” exponential geminate phase (e-(kt)β), where β is held fixed at 0.5. 1When β is reduced to 0.25 (dotted line) the fit improves between 10−7 to 10−6 s, but the absorbance normalization factor, N0, deviated significantly from unity (N0= 6.69). This signals that the fitting function is overshooting the zero time point determined by Δ A(0) and N(0) = 1. Such effects lead to a gross overestimate of the geminate amplitude at the expense of the bimolecular amplitude. The chi-square for the β = 0.25 fit is a factor of 3 better than for β = 0.5, but still a factor of 5 worse than the bi-exponential fits. The bi-exponential also gives a superior fit at the zero time point (N0= 1.18) compared with the β = 0.5 case (N0= 1.7). Since the two exponentials give a better description of the geminate phase, the CO rebinding kinetics were fit using, N(t)/N0=Ig1e−kg1t+Ig2e−kg2t+Is1e−ks1t+Is2e−ks2t(Eq. 2) where N(t) is the normalized (▵A(t)/▵A(O)) surviving population of unbound P450. Ig1,2 and kg1,2 represent the geminate rebinding amplitudes and rates, and Is1,2 and ks1,2 refer to the bimolecular amplitudes and rates. N0 is a scaling factor that allows comparison of the fit with the expected equilibrium absorbance change at time 0 (N0 = 1 ifthere is perfect agreement between the fit extrapolated to t = 0 and the expected equilibrium absorbance change). Earlier work on P450 CO rebinding kinetics evidently did not have sufficient signal to noise to reveal the two bimolecular phases. One simple explanation for the two geminate phases and two bimolecular rebinding rates observed for P450mCOr is that there are two conformations having different CO affinity. 2The possibility that the slow bimolecular phase of P450mCOr is due to camphor-bound P450mCOrs contamination in the sample cannot be completely ruled out, because its slow bimolecular rate is dominant and similar to the value of k2observed for P450mCOrs. The level of P450mCOrs contamination would need to scale roughly with the amplitude of the slow bimolecular phase of P450mCOr in order to account for this effect. However, contamination by P450mCOrs cannot explain the two phases associated with the P450mCOr geminate kinetics, since its small geminate amplitude would preclude its observation. The two geminate phases could also arise from protein relaxation processes, and further studies, involving isobestic monitoring and double pulse protocols, will be needed to assess this possibility. The differences in affinity could be due to a variety of conformationally controlled factors, ranging from distal pocket steric effects to heme geometry and iron spin state. Analogous kinetic (19Tian W.D. Sage J.T. rajer V. Champion P.M. Phys. Rev. Lett. 1992; 68: 408-411Crossref PubMed Scopus (123) Google Scholar) and resonance Raman investigations (33Morikis D. Champion P.M. Springer B.A. Sligar S.G. Biochemistry. 1989; 28: 4791-4800Crossref PubMed Scopus (179) Google Scholar), as well as double-pulse flash photolysis experiments (34Tian W.D. Sage J.T. Champion P.M. J. Mol. Biol. 1993; 233: 155-166Crossref PubMed Scopus (104) Google Scholar), have revealed the presence of pH-dependent “open” and “closed” distal pocket protein conformations in myoglobin. The specific conformational changes have recently been confirmed by x-ray crystallography studies (35Quillin M. Brantley R. Johnson K. Olson J. Phillips G. Biophys. J. 1992; 61: A466Google Scholar). The double-pulse protocol (34Tian W.D. Sage J.T. Champion P.M. J. Mol. Biol. 1993; 233: 155-166Crossref PubMed Scopus (104) Google Scholar) kinetically selects the most rapidly rebinding (open pocket) fraction of the ensemble and determines the time scale for averaging between the open and closed states. The interconversion time scale (1-10 μs) for myoglobin is fast compared with the rate of ligand migration from the solution to the heme pocket (∼10−4 s) so that a time-averaged, single exponential population analysis, rather than a superposition of the open and closed states describes the ligand association and dissociation kinetics of myoglobin. The ligand association kinetics of P450mCOr may also involve different protein conformations (to acknowledge that a variety of factors might affect the kinetics, we refer to them as “fast” and “slow” rebinding conformations), but the interconversion rate may be slower than the CO entry and exit rates. Under this circumstance, two bimolecular rates should be observed. As mentioned above, recent infrared (11Jung C. Hui Bon Hoa G. Schröder K. Simon M. Doucet J.P. Biochemistry. 1992; 31: 12855-12862Crossref PubMed Scopus (99) Google Scholar) and Raman (12Bangcharoenpaurpong O. Champion P.M. Martinis S.A. Sligar S.G. J. Chem. Phys. 1987; 87: 4273-4284Crossref Scopus (33) Google Scholar, 13Uno T. Nishimura Y. Makino R. Iizuka T. Ishimura Y. Tsuboi M. J. Biol. Chem. 1985; 260: 2023-2026Abstract Full Text PDF PubMed Google Scholar) experiments report multiple FeCO-stretching bands in substrate-free P450mCOr, supporting the view that several conformational substates exist. In Fig. 2 we present the CO rebinding kinetics of P450mCOrs, P450mCOr, and P420mCOr in pH 7 aqueous solution at 275 K. In the left panel we show the kinetics of the substrate-bound form, P450mCOrs and the substrate-free form P450mCOr monitored at 447 nm. The P450mCOrs data (trace A) show only a very small geminate yield (Ig= 2.5%, kg= 3.6 × 107s−1) with 97% of the amplitude associated with the two bimolecular phases (Is1 = 11%, ks1= 240 s−1, Is2= 86%, ks2= 26 s−1). On the other hand, the P450mCOr data (trace B) reveal a significant geminate yield (Ig1 = 91%, kg1 = 4.8 × 107s−1; Ig2 = 3.6%, kg2 = 6.9 × 106s−1) along with two bimolecular phases (see Table I). The bimolecular rates of P450mCOrs and P450mCOr agree reasonably well with other literature values (9Peterson J.A. Griffin B.W. Arch. Biochem. Biophys. 1972; 151: 427-433Crossref PubMed Scopus (56) Google Scholar, 27Shiro Y. Kato M. Iizuka T. Nakahara K. Shoun H. Biochemistry. 1994; 33: 8673-8677Crossref PubMed Scopus (28) Google Scholar). The significant difference in the geminate kinetics between P450mCOrs and P450mCOr indicates that the presence or absence of substrate not only affects the coordination and spin state of the ferric heme iron, but also significantly changes the heme environment in the reduced state. The observation of significant geminate rebinding in the absence of substrate explains the early work of Shimada et al. (22Shimada H. Iizuka T. Ueno R. Ishimura Y. FEBS Lett. 1979; 98: 290-294Crossref PubMed Scopus (18) Google Scholar) which found that the quantum yield of CO photodissociation was only ∼6% for P450mCOr and ∼100% for P450mCOrs. It also supports the observation that the CO affinity of camphor-bound P450mCOrs is 10 times weaker than camphor-free P450mCOr (9Peterson J.A. Griffin B.W. Arch. Biochem. Biophys. 1972; 151: 427-433Crossref PubMed Scopus (56) Google Scholar), primarily because of a decreased heme binding rate (∼factor 7), kBA, and an increased CO escape rate (∼factor 20) in P450mCOrs that reduces the geminate amplitude (Ig= kBA/(kBA+ kout)) and, thus, the probability of ligand binding when CO enters the heme pocket. The right panel of Fig. 2 shows the kinetics of P420mCOr (the pressure inactivated form of P450), which was monitored at both 420 nm (C) and 447 nm (D). At 420 nm (the peak absorption of P420mCOr), the kinetics shows a geminate yield of about 60% (Ig1 = 53%, Ig2 = 11%) with geminate rates that are significantly faster than for P450mCOr (kg1 = 1.5 × 108s−1, kg2 = 2.7 × 107s−1). This indicates that CO escapes from the pocket of P420 3The absorption and resonance Raman spectra do not show a difference between the camphor-bound and camphor-free P420 (8Wells A.V. Li P. Champion P.M. Martinis S.A. Sligar S.G. Biochemistry. 1992; 31: 4384-4393Crossref PubMed Scopus (109) Google Scholar), so we might expect that the kinetics of these two forms should be similar. In fact, we found that the addition of substrate does not affect the kinetics of CO rebinding to P420 (not shown). This indicates that if cytochrome P420 binds substrate, it has little affect on heme ligation state and heme environment. As a result, CO escape from P420 is effectively independent of the presence of substrate. much more rapidly than from P450 (see Table II) and suggests that the distal pocket environment is altered in the P420 form. The bimolecular phase of P420mCOr could not be fit with only two exponentials, so a stretched exponential (Ise(kst)β) was used. The fitting result, with β fixed at 0.5, is Is= 36%, ks = 1.8 × 103s−1. The nonexponential behavior may be due to the fact that “P420” is actually composed of a heterogeneous mixture of slowly interconverting conformational states. Raman spectra support this suggestion and indicate that the P420 heme is in equilibrium between a high-spin five-coordinate form and low-spin six-coordinate form (8Wells A.V. Li P. Champion P.M. Martinis S.A. Sligar S.G. Biochemistry. 1992; 31: 4384-4393Crossref PubMed Scopus (109) Google Scholar).Table II:The fundamental rate constants of P450 and P420All the fundamental rates are calculated from Equation 1. Uncertainties in the fundamental rates can be estimated as described in Table I. All the fundamental rates are calculated from Equation 1. Uncertainties in the fundamental rates can be estimated as described in Table I. In order to compare the kinetics of P420mCOr to P450mCOr, trace D shows the N(t) data for P420mCOr after scaling by their relative extinction coefficient changes at 447 nm (see Fig. 1). In addition to having a relatively small absorption change, the signal also undergoes a change in sign for t < 10−3 s. This latter observation suggests that protein relaxation takes place in P420 following photolysis (i.e. in analogy to myoglobin (19Tian W.D. Sage J.T. rajer V. Champion P.M. Phys. Rev. Lett. 1992; 68: 408-411Crossref PubMed Scopus (123) Google Scholar, 36Lambright D.G. Balasubramanian S. Boxer S.G. Chem. Phys. 1991; 158: 249-260Crossref Scopus (69) Google Scholar), at short times following photolysis, the Soret band of P420mris red shifted or broadened relative to its equilibrium value.) More importantly, for a sample that is a mixture of P450mCOr (35%) and P420mCOr (65%), the kinetics (trace E) can be seen to be a combination of the kinetic traces (traces B and D), with proper weighting. These results demonstrate that the two geminate phases and two bimolecular phases of P450mCOr monitored at 447 nm (trace B) are not the result of P420mCOr contamination of the sample. In Fig. 3, we present the kinetics of P450mCOr in pH 7.2 (75% glycerol) solutions as a function of temperature. When the temperature decreases from 293 to 263 K, the amplitude of the fast geminate phase increases by a factor of 1.23, but the rate remains the same (see Table I). On the other hand, the corresponding changes of the slow geminate rate kg2 and two bimolecular rates (ks1 and ks2) are much larger over the same temperature range (Table I). All the least squares fitting parameters from Equation 2 are listed in Table I. The fundamental rates for ligand entry, exit, and binding to the heme can be calculated using Equation 1. The rates associated with the fast bimolecular and geminate rebinding kinetics are used to calculate the fundamental rates associated with the fast protein structure, whereas the slower pair of geminate and bimolecular parameters are used to find the rates associated with the slow structure. Table II presents the fundamental rates calculated for both fast (f) and slow (s) states of P450mCOr in 75% glycerol at different temperatures. 4Under the assumption that the slow bimolecular phase reflects P450mCOrs contamination, the kinetics of P450mCOr contain two geminate and one bimolecular phase. The results for the fast state in Table II still hold approximately, because I2, associated with the slow geminate phase of P450mCOr, is much smaller than I1. The inset of Fig. 3 shows the Arrhenius fit to the fundamental rates for the fast rebinding conformation. The ratio of the rates in the fast and slow states are kBAf/kBAs ∼ 6-30, koutf/kouts ∼ 6-8, kinf/kins ∼ 16-20 over the temperature range 293-263 K. Surprisingly, the value of kBAf for the fast state does not appear to change significantly with temperature. This suggests a very small barrier for ligand binding at the heme in comparison with the slow state. Compared with the weak temperature dependence of kBAf, koutf, and kinf are much more sensitive to temperature. The Arrhenius barriers are found to be: HBAf ∼ 0 kJ/mol, Houtf ∼ 27kJ/mol, and Hinf ∼ 33kJ/mol. We note that, due to experimental error and the small amplitudes associated with the kinetic response of the slow state, Table II only contains estimates for the fast state barriers. The significant increase of the rates in the fast state indicates the possibility of protein conformational control of diatomic ligand binding in cytochrome P450. The time course of CO rebinding to P450mCOr in aqueous solution at different temperatures is shown in Fig. 4. The inset shows the expanded bimolecular process. When the temperature increases from 273 to 293 K, the amplitude of the fast geminate phase decreases ∼4%, the rate increases ∼5%, and the slow geminate rate increases ∼20% (see Table I). On the other hand, the bimolecular rates (ks1and ks2) increase significantly (factors of 2.5 and 3) over the same temperature range. All the least squares fitting parameters from Equation 2 are listed in Table I. The fundamental rates, analyzed using Equation 1, are listed in Table II and follow a trend that is similar to the kinetics of P450mCOr in 75% glycerol solution (i.e. kBAf is nearly independent of temperature, whereas koutf and kinf change significantly). A systematic difference in the rebinding kinetics of P450mCOr in aqueous solution and glycerol/buffer mixtures is observed by comparing Figs. 3 and 4. As the solution changes from aqueous to 75% glycerol at 273 K, the fast geminate amplitudes decrease from 92 to 77%, whereas the slow geminate amplitude increases from 3 to 10%. 5The slow geminate amplitude increases continuously as the percentage of glycerol in the solution increases from 0, 25, 50, and 75% (not shown). On the other hand, the geminate rates do not change significantly and the bimolecular rates change by factors ∼2 (see Table I). The glycerol-dependent results show that the kinetics are not dramatically sensitive to solvent environment. This indicates that, unlike the camphor substrate, glycerol has no major effect on the structural features of the heme pocket that determine the geminate amplitude and rebinding kinetics. In summary, the CO geminate rebinding kinetics of P450 and P420 were observed for the first time and experiments as a function of temperature and glycerol/water buffer have been carried out. P450mCOrs and P450mCOr have drastically different geminate rebinding kinetics, which indicates that substrate plays an very important role in the structure/function relationships that govern the dynamics of diatomic ligand binding and release. Such effects may occur via alterations of the heme pocket that affect the stability of CO within the pocket (this might involve displacement of water molecules, or direct steric interactions, for example). The two geminate and two bimolecular CO kinetic phases of P450mCOr suggest two conformational states of the protein, each one with a different CO affinity. The specific structural changes associated with the fast and slow states are not yet clear. One possibility is that they correspond to conformational interconversions of the protein that are linked to (camphor) substrate binding, which also affect the diatomic ligand entry and exit rates in addition to the rate of binding at the heme. In analogy to myoglobin (34Tian W.D. Sage J.T. Champion P.M. J. Mol. Biol. 1993; 233: 155-166Crossref PubMed Scopus (104) Google Scholar), these states may correspond to slowly interconverting open and closed substrate binding pockets. The open state has significantly faster fundamental rates (kBA, kin, kout) than the closed state, which indicates that the protein conformation can control ligand binding to the heme as well as its motion between the solvent and the heme pocket. The geminate kinetics of P450mCOr were affected by changes in the temperature and viscosity, but major structural changes were not apparent. This indicates that the structure of the P450mCOr heme pocket is stable to such perturbations. In comparison with P450, the P420mCOr data show faster geminate rebinding rates and a nonexponential bimolecular rate, suggesting that multiple conformations and an altered heme pocket are associated with the P420 form of the enzyme.
Referência(s)