High-pressure-induced transitions in microsomal cytochrome P450 2B4 in solution: Evidence for...

ARCHIVES OF BIOCHEMISTRY AND BIOPHYSICS

Vol. 320, No. 2, July 10, pp. 330-344, 1995

High-Pressure-Induced Transitions in Microsomal Cytochrome P450 2B4 in Solution: Evidence for Conformational Inhomogeneity in the Oligomers

Dmitri R. Davydov,’ Eric Deprez, Gaston Hui Bon Hoa, Tatiana V. Knyushko,” Galina P. Kuznetsova,” Yakov M. Keen,* and Alexander I. Archakov” Znstitut de Biologie Physico-Chimique, ZNSERM U.310, Paris, France; and ‘“Znstitute of Biomedical Chemistry, Russian Academy of Medical Sciences, Moscow, Russia

Received October 24, 1994, and in revised form April 27, 1995

Pressure-induced changes in ferric P450 2B4 (LM2) were studied as a function of benzphetamine concentration (0.05 + 2 mu) and state of aggregation of the hemoprotein in solution. Application of factor analysis to the spectral changes in the Soret region allowed us to resolve two particular pressure-induced processes in 2B4 oligomers. The first process was identified as the conversion of the low-spin P450 into the P420 state. At 25°C it was followed by decay (bleaching) of about 50% of the newly formed P420. The second process was a pressure-induced high- to low-spin shift. Both transitions were reversible, except the hemoprotein bleaching. The amplitude of the P450 --t P420 transition accounted for 67 + 5% of the total hemoprotein content. Furthermore, the fraction of the hemoprotein exposed to spin equilibrium was not affected by the P450 -, P420 conversion and was estimated to be only about 31 + 5% of the total hemoprotein content. After the dissociation of the oligomers by 0.2% Triton N-101, the inhomogeneity vanished: 95% of the monomers were involved in the P450 + P420 transition (AV” = -86 ml/ mol) followed by intense bleaching of the hemoprotein. This agrees with our earlier observations on the reduced carbonyl complex of P450 2B4 and suggests some conformational difference between subunits in P450 LM2 oligomers. The parameters of the P450 + P420 conversion (AV” = -32 ml/mol, Plls = 1560 bar) show no dependency on the substrate concentration. Analysis of the pressure-induced spin shift versus benzphetamine concentration shows this transition to be caused mainly by changes in the spin equilibrium of

i Present address: Institute of Biological and Medical Chemistry, Russian Academy of Medical Science, 10 Pogodinskaya ul., 119832, Moscow, Russia.

‘To whom correspondence should be addressed. Fax: (33-l) 43-29- 80-88. E-mail: HUIBONHOQIBPC.FR.

330

both substrate-bound (AV” = -49 ml/mol) and substrate-free (AV” = -21 mYmo1) hemoprotein, whereas the substrate binding step itself has a very weak pressure dependency (AV” = -8 ml/mol). v 1995 Academic

Press. Inc.

In the past few years pressure-induced transitions in protein structure have been used successfully to study changes in the heme moiety of cytochromes P450,,,, and P45Oiin on substrate binding (l-6). Recently, we reported the first results of pressure-induced transitions in rabbit liver microsomal cytochrome P450 2B4 (7). This hemoprotein (also known as cytochrome P450 LM2) is denoted “2B4” in this paper.” Pressure-induced inactivation (P450 + P420 transition and disappearance of the hemoprotein-specific bands at high pressures) of the complex of the reduced hemoprotein with carbon monoxide was studied (7). This membrane hemoprotein pressure-related behavior was certainly different from that of the soluble bacterial cytochromes P45OCXn and P450ii,. This difference is caused partly by the oligomeric structure of the microsomal hemoprotein in solution. The molecular masses reported by several authors for microsomal cytochromes P450 suggest the size of the aggregates to be in the range of 6-10 protomer molecules (8-12). Electron microscopy studies of P450 2B4 (13) and P450 lA2 (14) oligomers show them to be hexamers. Within the membranes, microsomal P45Os were also proven to form oligomers, which are believed to be hexamers as well (15-18).

a Abbreviations used: 2B4, cytochrome P450 2B4 (LM2) from rabbit liver microsomes; P450 and P420, cytochrome P450 2B4 in the P450 and P420 states, respectively; P450,,, and P450,,, high-spin and low- spin states of the cytochrome; PCA, principal component analysis.

0003-9861195 $12.00 Copyright ~3 1995 by Academic Press, Inc.

All rights of reproduction in any form reserved.

PRESSURE-INDUCED TRANSITIONS IN CYTOCHROME P450 2B4 331

An application of high hydrostatic pressure was shown to induce conversion of 2B4 reduced carbonyl complex into an inactive P420 form. However, in contrast with monomeric soluble bacterial cytochromes P450 (4), only about 65% of the oligomeric 2B4 could be converted into P420 (7). The rest of the protein was stable up to 6 kbar. The same behavior was observed for 2B4 incorporated into the proteoliposomes. After monomerization of hemoprotein oligomers by 0.2% Tri- ton N-101 this inhomogeneity of 2B4 pool vanished. Monomers could be completely converted into P420 (7). Thus, in the monomeric state the pressure-related behavior of 2B4 reduced CO complex is analogous to that of the bacterial hemoproteins. These results were interpreted as evidence of the existence of the static splitting of the 2B4 pool in the oligomers between two conformers with different pressure sensitivities (7). Such splitting is supposed to exist in solution as well as in membranes.

To further test this hypothesis, to compare the ability of these two presumed conformers to bind substrate, and to determine the position of their spin equilibrium, we have undertaken the present investigation. It is devoted to studies of pressure-induced transitions in 2B4 in solution as a function of its aggregation state and of the substrate (benzphetamine) concentration.

MATERL4LS AND METHODS

1. Experimental Materials. Electrophoretically homogeneous cytochrome P450

2B4 was purified from the livers of phenobarbital-treated rabbits (19). Cytochrome P450 preparations were stored in 0.1 M potassium phosphate buffer containing 1 mM EDTA, 1 mM dithiothreitol, and 20% glycerol at -70°C. Neither a detectable peak nor a shoulder at 420 nm was detected in the differential spectrum of reduced carbonyl complex. However, the use of such a spectrum is inconvenient in the quantitative determination of the concentration of P420 in the sample since the trough of the band of CO-free reduced P450 overlaps here with the peak of the reduced P420 carbonyl complex. Measure- ments of absolute spectra of the reduced carbonyl complex by the method of Imai and Sato (20) and fitting of the absolute spectra of oxidized hemoprotein by 2B4 absorbance standards (see below) both show the presence of up to lo-15% P420 in the preparations.

Hepes, EDTA, and nL-dithiothreitol were obtained from Sigma Chemicals. Triton N-101 from Fluka and sodium dithionite from Merck were used.

High-pressure system. The optical pressure system, capable of generating pressures up to 6200 bar, has been described previously (21). The bomb was fitted directly into the sample compartment of a Cary-219 spectrophotometer. The high-pressure system was fully computer-controlled. Remote control of the system, data acquisition, and spectra decomposition analysis were done with an IBM PC/AT computer equipped with our data acquisition and fitting software package “SpectraLab,” which is described in the miniprint supplement of the present article.

Experimental procedure. All pressure experiments were made in 0.1 M Na Hepes buffer, pH 7.4, containing 1 mM EDTA and I mM dithiothreitol at 25°C. if not otherwise indicated. Control measure- ments showed that the pH of this buffer decreases by approximately 0.06 unit per kilobar. Solubilization of the oligomers was achieved

by the addition of a 10% solution of Triton N-101 in the same buffer to yield a final concentration of 0.2%. After addition of the detergent, the samples were incubated for 30 min at 4°C. Changes in the aggregation state of the hemoprotein were followed by measurement of the sedimentation velocity. Analytical ultracentrifuging experiments were performed by means of a Beckman XLA ultracentrifuge with an AN6OTi rotor and photoelectric scanning at 420 nm. Absorbance spectra were recorded using computer-controlled Gary-219 Warian, U.S.A.) or Uvikon-940 (Kontron, Switzerland) spectrophotometers.

2. Methods of Spectra Processing and Data Fitting Principal component analysis technique for separation ofparticular

processes in a multicomponent system. Pressure-induced changes in absorption spectra of ferric 2B4 in the Soret region were found to be very complex. They reveal several simultaneous individual processes taking place in the same pressure range. To resolve these particular processes we used the method of principal component analysis (PCA), which is also known as bilinear factor analysis (28) or “singular value decomposition” analysis (26). This method, first introduced in chemistry by Malinowski and Howery (22), is now widely used to interpret absorbance and fluorescence spectral changes in complex biochemical systems (23-28). The PCA method in a version close to that used by Durell and co-workers (28) was realized in the SPAN module of our SpectraLab software. A brief description of the PCA algorithm and some details of our implementation are described in the miniprint supplement at the end of this article. Some specific terms of the method necessary for our description are also explained under Results.

In the present study we usually used PCA on sets of absorbance spectra in the range of 340-600 nm digitized with a step of 2 nm if not otherwise indicated. Prior to the analysis, all spectra were corrected for water volume contraction using the known data on the compressibility of water (29). Spectral changes, caused by turbidity, which appears at pressures about 3 kbar due to denaturation (unfolding) of the protein, were resolved by using PCA and were selec- tively removed from the spectra by subtraction of the third-order polynomial approximation of these changes. However, such a correction was used only for qualitative interpretation of the changes. Quantitative analysis was performed only for pressures below 3 kbar.

The set of standards of the extinction spectra of pure states of 2B4 (high- and low-spin states of P450 form and P420 state of the hemoprotein) was used to analyze pressure-induced spectral changes. These standard spectra were obtained from the measure- ments of substrate- and temperature-induced spectral changes in 2B4 and also from the spectral changes accompanying the P450 --t P420 conversion caused by addition of urea and sodium iodide. PCA algorithm was used to delineate the standards from these data using those of bacterial cytochromes P450,, and P450ii, as initial estimates. The set of extinction standards used in this work is shown in Fig. 1. As one can see from Fig. 1, the standards of P450 2B4 high-spin and low-spin spectra are similar to those known (30). The spectrum of the oxidized P420 state of 2B4 is characterized by two maxima at 365 and 426 nm. It also corresponds well to that reported for the P420 state of cytochrome P450,,, where these maxima have been found near 360-370 and 420-425 nm (5, 31). The extinction coefficient for the P420 state of 2B4 at 426 nm was determined to be 0.128 PM-’ cm-‘.

Results of determination of P450 heme iron spin equilibrium by least-squares fitting of the spectra by linear combination of these standards were in good agreement with those obtained using the temperature-difference approach of Ristau et al. (30), whenever ap- plicable. Concentrations of P420 hemoprotein determined by our method were no more than 5% different from those determined from the absolute spectra of CO-saturated reduced hemoprotein using extinction coefficients given by Imai and Sato (20).

Analysis ofpressure-induced transitions. Fitting the pressure dependencies of the concentrations of 2B4 species was based on the

332 DAVYDOV ET AL.

E, umor’cm-’

0.04

420 500 580 660 Wavelength, nm

FIG. 1. Reconstituted extinction spectra of the pure species of cytochrome P450 2B4: low-spin P450 (P450,, , bold line), high-spin P450 (P45Di,,, thin line), and P420 (dashed line) obtained by iterative application of factor analysis on the series of 2B4 spectra measured as a function of temperature and benzphetamine concentration (P450,, and P450,,.) or as a function of NaI and urea concentrations (P420).

classical equation for the equilibrium constant of the first-order reaction A --t B (321,

[Al [II

where K&l and K,(O) are the equilibrium constants of the reaction at pressure p and at zero pressure, respectively; Pm is the pressure at which K, = 1 (“half pressure” of the conversion); and AV is the molar reaction volume, associated with the conversion at P = 0. To fit the concentration curves directly this equation was transformed to produce the relationship

where Co = [Al + [Bl is the sum of the concentrations of species A and B. Marquardt (331 and Nelder-Mead (341 optimization algorithms were used to fit the data to this equation treating Co, K,(O), and AV as the parameters. An adjustable offset (A,) was added to the right part of this equation to adapt the equation to the case where the initial position of equilibrium is unknown. It occurs, for instance, when the concentration measured reflects the sum of the concentration of the compound exposed to equilibrium and the constant concentration of another, spectrally similar component.

The fitting was performed using the SpectraLab software package described in the miniprint supplement.

RESULTS

1. Prekure-Induced Transitions in Cytochrome P450 2B4 Oligomers

Preliminary analysis of the transitions. Pressure- induced changes in absorbance of oligomeric 2B4 in

solution were studied as a function of substrate (benzphetamine) concentration in the range 0.05-2 mM. A series of spectra measured at increasing hydrostatic pressure and the corresponding pressure dependencies of the concentrations of the hemoprotein species (P450 high-spin, P450 low-spin, and P420 states) are shown in Fig. 2. It should be noted that at pressures above 3 kbar a noticeable turbidity component appears in the spectra, indicating that at these pressures a partial denaturation (unfolding) or aggregation of the protein takes place. In the present work we have analyzed the changes below these pressures.

As one can see from Fig. 2a (inset), a pressure increase leads to some decrease in concentrations of both high-spin and low-spin states of P450 and to the appearance of the inactivated P420 state of the hemoprotein. At 25°C a considerable decrease in total hemoprotein content is also detected. However, at 4°C such hemoprotein bleaching has a negligibly low amplitude (data not shown). Remarkably, significant reconversion of the pressure-induced P420 back to P450 is observed after decompression to 1 bar. A large part of pressure- generated P420 reconverts into P450 immediately after decompression; however, there is a slow phase of the process, which takes about 2 h.

The samples of 2B4 after pressure treatment were also analyzed by measuring the absolute spectra of the reduced carbonyl complex of the hemoprotein according to Imai and Sato (20). The difference between the results of such determination and P450 and P420 concentrations found from the spectra of oxidized 2B4 was within 5% limits.

It should be noted that the reversibility of pressure- induced P450 + P420 conversion takes place only in the presence of dithiothreitol. On the other hand, P450 + P420 conversion by means of high concentrations of salt, by detergents, or by SH modifiers is irreversible (20). This means that in these cases the P420 state- provoking changes in the heme moiety are fixed by some severe conformational transitions in the hemoprotein, while for pressure-generated P420 in the presence of dithiothreitol these changes are maintained only by pressure and disappear upon pressure release.

In contrast to the P450 -+ P420 conversion, the pressure-induced decrease in the total hemoprotein concentration was irreversible. This was also proven by the Imai and Sato method. It should be noted that such disappearance (bleaching) of the hemoprotein absorbance bands under certain conditions is a well- known peculiarity of microsomal cytochromes P450. This bleaching was detected during P450-catalyzed hy- droxylation (351, upon phosphorylation of the hemoprotein (36), in the presence of sodium cholate (371, and under other conditions. Remarkably, the P420 state of the hemoprotein is much more sensitive than the na- tive P450 state to this process (37). Apparently, there


O.D. 0.2

0 I

340

I

380

I I I I ,

420 460 500 540 580

Wavelength, nm

A0.D. 0.04

A

340 380 420 460 500 540 580

Wavelength, nm

FIG. 2. Pressure-induced transitions in cytochrome P450 2B4 in solution in the presence of 2 mu benzphetamine followed by Soret-region spectroscopy. Conditions: 4 FM 2B4 in 100 mu Na Hepes buffer, pH 7.4, 1 mM EDTA, 1 mu dithiothreitol, 2 mu benzphetamine, 25°C. Optical path length, 5 mm. Spectra at 1 bar and 0.6, 1.0, 1.6, 2.5, 3.0 kbar are shown. Data were corrected for the pressure-induced contraction using water compressibility coefficients (19). Spectral changes caused by an increase in the turbidity at 3 kbar were suppressed by factor analysis-supported algebraic correction by third-order polynomial. (a) Absolute spectra. Inset: Changes in the concentrations of 2B4 species: P450,, (l), P450h, (2), P420 (3), and total hemoprotein concentration (4) calculated by the least-squares fitting of the spectra by linear combination of 2B4 spectral standards (Fig. 1). The lines correspond to the results of the fitting of these curves by the definitive model of the transitions (see below). (b) The same set of data shown as the differences between spectra at particular pressures and the spectrum of the initial state (at I bar). Inset: Difference between the spectrum measured 2 h after decompression and the spectrum of the initial state. Dashed line represents the fitting of the spectrum by the combination of 2B4 extinction standards. It corresponds to the disappearance of 1.3 pM P450,, and appearance of 0.3 pM P420.

is no single mechanism responsible for such bleaching. release of the heme from the pocket as well (34). In Depending on the conditions it could result from vari- our case we cannot exclude any of these possibilities. ous combinations of heme loss and oxidative destruc- However, studies of the exact mechanism of bleaching tion of the heme tetrapyrrole ring by oxygen species were beyond the aims of the present investigation. directly in the heme moiety of P450 (P420) or after Both P420 appearance and total hemoprotein disap-

334 DAVYDOV ET AL.

pearance could be fitted to Eq. [21 with a square coefficient of correlation (p2) higher than 0.995. Such fitting yields the same Pvz and AV’ values for both processes (about 1.6 kbar and 35 ml/mol, respectively). This suggests that one of the pressure-induced processes in 2B4 oligomers is the P450 + P420 transition accompanied or immediately followed by some bleaching of the hemoprotein.

The pressure-induced changes in the concentrations of the high- and low-spin states of P450 were quite different in shape from those of P420 and the total hemoprotein, indicating that after the P450 + P420 conversion some additional transition(s) took place. The decrease in the concentration of the high-spin state of P450 was much faster than that of the low-spin state. Moreover, at high concentrations of benzphetamine a slight increase in the low-spin content was detected during the initial compression steps. Thus, the second transition was assumed to be a high- to low-spin shift, by analogy with that shown for P450,,, (3, 4).

Application of PCA to pressure-induced spectral changes in 2B4 oligomers. To confirm our suggestions on the nature of pressure-induced processes and to separate individual transitions, we used the PCA procedure. The ultimate goal of PCA is to determine the minimal number of particular processes reflected in the overall spectral changes and to separate these changes into so-called principal components used to understand the nature of each process. PCA generates two series of quantities for each principal component. These quantities are basis spectra and vectors of loading factors, which can be thought of correspondingly as differential spectra and dynamics of apparent particular processes (‘basis processes”). The number of statistically significant principal components is an estimate of the minimal number of real transitions in the system. However, these components do not necessarily correspond to the real transitions. They represent only one of the possible ways of decomposition and should be considered a linear combination of the real transitions. To delineate such transitions some additional information is required, such as the extinction spectra of pure components of the system and some idea of the nature of the transitions. This information is then used to transform the initial results of PCA (to redistribute the spectral changes between the components) to obtain the exact stoichiometry and dynamics of the real processes.

Results of the application of PCA to the pressure- induced spectral changes in 2B4 oligomers are shown in Fig. 3. In the whole range of benzphetamine concentrations, two first principal components (Fig. 3a) were sufficient to fit 99.5% of the total spectral changes. Only these two components could be adequately approximated by linear combination of 2B4 absorbance standards. The surface of the square residuals of such an

approximation reveals no systematic behavior (Fig. 3~). Therefore we can deduce that there are only two independent pressure-induced transitions in 2B4 oligomers.

As mentioned above, principal components found by PCA do not correspond to the real processes in the system but rather represent some linear combination of them. Accordingly, there is no direct physical mean- ing to the apparent stoichiometry and dynamics of the principal components found here (Figs. 3a and 3b). To proceed to a real “physical” transition, we suggest that one of them is a stoichiometric high- to low-spin shift, as observed for P450,,, . The fit of the differential spectrum of the spin transition as calculated from 2B4 spectral standards using the first two basis spectra (Figs. 3a and 3b) was quite good, especially at high concentrations of benzphetamine. Such transformation yields a set of differential spectra of two transitions and also their pressure dependencies. This set is absolutely equivalent to the initial results of PCA in terms of approximation of the observed spectral changes in the system, but the spectral changes are now “redistrib- uted” between the components to obtain the best fit of the target proposition (stoichiometric high- to low-spin transition as one of the components).

This approach gives the same pair of differential spectra of apparent transitions in the whole range of benzphetamine concentrations studied (square of the coefficient of correlation between the differential spectra obtained at different benzphetamine concentrations was better than 0.96). The first transition, in accordance with the initial assumption, represents the pressure-induced spin shift (Fig. 4a). The averaged differential spectrum of the second transition deduced from the first two PCA basis spectra could also be fitted by combination of 2B4 spectral standards (Fig. 4b). Con- sistent with our preliminary analysis, such fitting reveals that this process corresponds to a conversion of low-spin P450 into the P420 state. This conversion is accompanied by a 50% disappearance (bleaching) of the hemoprotein. It should be noted, however, that both differential spectra of apparent pressure-induced transitions have some systematic deviations from their fitting by 2B4 absorbance standards (see insets to Figs. 4a and 4b). Such deviations could be caused by some pressure-induced changes in the spectra of 2B4 (red shifts of P450 low-spin and P420 bands, broadening of P450 high-spin band) or by absorbance of hemine or hemine-protein adducts as a result of heme loss.

As stated above, P450 + P420 and spin shift transitions were reversible on decompression, whereas the bleaching was not. Principal component analysis was applied also to the series of spectra registered at de- creasing pressure. The results were interpreted in the same manner, as described above. Two main transitions were found here: stoichiometric spin shift and 1:l


A0.D. A O.D.

400 500 600 Wavclcngth, nm

Loadinn factor Wavelength, nm

Square deviation, (0.D.)2 I

Pressure, kbar

1.5.10-6 d

400 500 Wavelength, nm

600

FIG. 3. PCA application to the data set shown in Fig. 2b. (a) First basis spectrum (96.5% of the total changes in terms of the square deviation from the spectrum of the initial state). Dashed line shows the results of its fitting by the linear combination of the standard spectra of 2B4 (-0.41 /IM P&O,., -0.59 pM P@&, +0.51 pM P420, p* = 0.997). (b) Second basis spectrum (3.3% of the total changes). Results of its fitting by the linear combination of the standard spectra of 2B4 (+1.5 pM P450,., -0.65 ~.LM P450,,., -0.35 pM P420, p* = 0.95) are shown by dashed line. (c) Loading factors for the first (1) and second (2) components versus pressure. The differential spectra for each pressure (Fig. 2b) could be approximated by the sum of the basis spectra multiplied by the corresponding loading factors. (d) Surface of the square deviations for such approximation (overall accuracy is 99.8%).

P450 + P420 conversion. Thus, it appears that revers- cesses. The newly formed P420 state seems to be unsta- ible P450 -+ P420 and irreversible hemoprotein bleach- ble and loses the heme at high pressures. We are ing are not simultaneous, but are consecutive pro- unable to resolve this process from the P450 + P420

336 DAVYDOV ET AL.

a

AO.0. 0.1

!

Residual error. &O.D.

-0.02 , 400 500 Wavelength. nm

a.05 I I I 4 350 400 450 500 550

Wovelength, nm

0.02

0

-0.04

-0.06

A0

/

.D.

b

Residual error. n0.D.

0.005 1

Oj,p

-0.005

I 400 SOD

Wavelength. nm

350 400 450 500

Wovelength, nm

550

FIG. 4. Basis spectra of two independent transitions resolved by the target-oriented transformation of the results of PCA of pressure- induced spectral changes in 2B4 oligomers. The spectra represent the averages of the results of 11 independent experiments at various concentrations of benzphetamine (O-2 mu). Residual errors of the fitting of these spectra by 2B4 spectral standards are shown in the insets. (a) Pressure-induced spin shift. Dashed line corresponds to the differential spectrum of 1:l P450,,. -+ P450,. transition calculated from the 2B4 spectral standards (p2 = 0.96). (b) P450 + P420 transition. Spectrum shown by dashed line was calculated from 2B4 spectral standards (p* = 0.985) and corresponds to the disappearance of 1 pM P4501. and appearance of 0.53 pM of P420, therefore suggesting some disappearance (bleaching) of the chromophore.

transition because of the complexity of the system and, presumably, because of a small reaction volume of the hemoprotein bleaching process.

Analysis of pressure-induced P450 inactivation process. The pressure profile of the P450 + P420 transition did not depend on the benzphetamine concentration (Fig. 5). The whole set of pressure dependencies of the molar fraction of initial 2B4 concentration, which underwent the transition, could be fitted at once by Eq. [21 with a single set of parameters (AV” = -32 -+ 5 ml/ mol, Py2 = 1560 + 90 bar, p2 = 0.985, Fig. 5). Remark- ably, fitting gives the total amplitude of the transition (Co in Eq. [21), which is only about 67 + 5% of the initial content of P450 in the system. This means that only about two-thirds of the hemoprotein in oligomers undergoes pressure-induced inactivation. The rest of the hemoprotein was stable at pressures below 3 kbar. This finding is in good agreement with our previous observation that only 65% of the hemoprotein is exposed to the P450 + P420 conversion in the reduced carbonyl complex of the oligomeric 2B4 as well (7).

Analysis of pressure-induced spin shift. As we have shown above, the P450 + P420 transition is characterized by the differential spectrum and concentration profile at any concentration of the substrate. The basis

spectrum suggests that only low-spin hemoprotein is exposed to the inactivation, independent of the position of the apparent overall spin equilibrium of the hemoprotein. This finding appears to be very important. It suggests, that the pool of P450 is not homogeneous here. The hemoprotein subjected to pressure-induced inactivation is represented by the low-spin state only and is not exposed to the substrate-dependent spin equilibrium.

In agreement with this, all the changes in the concentration of the high-spin P450 resulted only from spin shift transition. The pressure profile of the loading factor of the spin shift component and the profile of the high-spin P450 concentration (Fig. 2a, inset) have the same shape and could be equally used to follow spin shift. These profiles strictly obey Eq. [2] (p2 > 0.998). This clearly requires that Co (total concentration of the compounds subjected to the equilibrium) in Eq. [2] be constant and not affected by P450 inactivation. Thus, the former conclusion on the inhomogeneity of 2B4 pool is inevitable: the hemoprotein subjected to spin equilibrium is not exposed to pressure-induced inactivation and vice versa.

Fitting the pressure dependencies of the spin state gives an estimate of the amount of hemoprotein ex-


P-450 converted

5oxz

0 0.5 I 1.5 2 2.5 Pressure, kbar

FIG. 5. Pressure profile of P450 + P420 transition (corresponding basis spectrum is shown in Fig. 4b). Percentage of the hemoprotein that underwent the transition was calculated from the corresponding loading factor normalized by the initial concentration of 2B4. The whole set of curves obtained at 0 CO), 0.0.05 (+I, 0.1 (x), 0.2 (*), 0.3 (01, 0.8 (O), and 2.0 ([XI) mu benzphetamine was fitted to Eq. 121. Solid line corresponds to the results of fitting.

posed to the equilibrium (parameter Co in Eq. [2]) as 31 -+ 5% of the total initial content of P450. Therefore only about one-third of the cytochrome P450 2B4 in oligomers is exposed to the spin equilibrium. This fraction does not undergo any P450 + P420 or bleaching processes at pressures below 2.7 kbar. The remaining two-thirds of the total hemoprotein content is always low-spin under conditions of the experiment. This fraction undergoes only pressure-induced P450 + P420 transition and subsequent partial bleaching and is probably unable to bind the substrate.

Further analysis of spin transitions in 2B4 was undertaken to test whether the pressure-induced spin shift is a result of substrate complex dissociation or reflects only the changes in the spin state of substrate- bound and substrate-free hemoprotein. This analysis was based on the following scheme of substrate binding and spin equilibrium:

Kl PI’ ,. - Ph

K, - [pl’ Kf _ by1 bhl ’ [pii1 ’

Kd = 1,K = ([PI] + bhl)’ [sl a Ml + bf;l ’

where P and P” are cytochrome P450 and its complex with substrate, respectively. Indexes “1” and “h” desig- nate, respectively, low- and high-spin states of the hemoprotein. It should be noted that this simplified scheme assumes no difference in substrate binding equilibria for high- and low-spin hemoprotein. Since substrate- free hemoprotein exists predominantly (more than 90%) in the low-spin form, such simplification appears to be acceptable here. Analysis of this model yields the equation for the fraction of high-spin state of hemoprotein (Fh),

Fh = bhl + [&I

[PI 1 K&31 1

= 1 + K&3]’ 1 + Kf + ( 1 l+K, ’ [41

where [pl is the total concentration of the hemoprotein exposed to equilibrium system 3. Combining this relationship 141 with Eq. 111, it is possible to deduce the equation for the pressure dependence of Fh . This equation includes six independent parameters: equilibrium constants of the elementary transitions at normal pressure (Kz , KY, KY) and the corresponding reaction volumes. To evaluate these parameters we have fitted the whole set of the pressure dependencies of high-spin content obtained at various concentrations of benzphetamine (0.05, 0.1, 0.2, 0.3, 0.4, 0.8, and 2 mu). Such fitting was made by the optimization program, based on the Nelder-Mead algorithm of minimization of the sum of square deviations (34).

Since the whole pool of P450 is exposed to both spin equilibrium and P450 + P420 transition, Fh is expected to be the ratio of concentration of high-spin P450 (IPhI + [Psh]) to the current total concentration of both high- and low-spin hemoprotein (which decreases with pressure increase). However, using this definition of Fh our data do not fit the above equations satisfactorily, especially at high concentrations of substrate and high pressures. If we use the above conclusion that only one- third of the hemoprotein, which does not undergo the inactivation process, is exposed to spin equilibrium, Fh is expressed as the ratio of the current concentration of high-spin P450 to one-third of the initial concentration of the hemoprotein. In this case the whole set of data fits the model (4) with a square correlation coefficient of about 0.992 without any systematic deviations (Fig. 6).

Parameters of Eqs. 121 and 141 evaluated by this fitting are shown in Table I. One can see that dissociation of the substrate complex itself is characterized by a small negative reaction volume. Therefore, an increase in the hydrostatic pressure provokes only a slight shift of substrate binding toward dissociation of the complex.

338 DAVYDOV ET AL.

1, Fh

0 05 1 15 2 25 Pressure, kbor

025

Oi I

0 05 1 15 2 Benzphetomine, mM

FIG. 6. Pressure-induced spin shift in cytochrome P450 2B4 in solution at various concentrations of benzphetamine (corresponding basis spectrum is shown in Fig. 4a). In accordance with the proposed model, the high-spin content in the P450 fraction exposed to the spin equilibrium (F,,) was calculated from the concentration of high-spin state normalized by one-third of the total initial concentration of the hemoprotein (see text). Solid lines show the results ofthe data fitting in accordance with the suggested model of the substrate binding and spin equilibria (scheme [3]). Resulting parameters of the fitting are shown in Table I. The same changes in F,, are shown in a as a function of pressure at various benzphetamine concentrations (0 CO), 0.05, 0.1, 0.15, 0.2, 0.3, 0.8, and 2 (*) mu) and in b as a function of benzphetamine concentration at various pressures (1 (01, 300, 600, 900, 1200, 1500, 2000 and 2500 (. ) bar).

The spin equilibrium constants at zero pressure for substrate-free and benzphetamine-bound hemoprotein differ by a factor of about 30, indicating a considerable shift of the spin equilibrium toward the high-spin state after substrate binding. The reaction volumes of the high- to low-spin shift for substrate-free and substrate- bound species are much higher (-21 and -49 ml/mol). Thus, the pressure-induced high- to low-spin shift is caused in 2B4 not by the dissociation of the substrate complex, as has been proposed for P450,,, (2-4), but mainly by the shift of the spin equilibrium of both substrate-bound and substrate-free hemoprotein.

2. Pressure-Induced Transitions in Cytochrome P450 2B4 Monomers

To dissociate 2B4 oligomers, the nonionic detergent Triton N-101 was used. Sedimentation experiments were used to control the dissociation. It was found that addition of 0.270 of this detergent results in a decrease in the sedimentation coefficient from 7.0s to 3.9s. Thus, in the presence of 0.2% Triton N-101, an exten- sive dissociation of 2B4 oligomers was observed. This observation is in good agreement with the results of sedimentation equilibrium experiments made by Mya- soeclova and Bern& (38), who reported that 2B4 hexamers dissociate completely into monomers under these conditions.

Pressure-induced spectral changes in 2B4 monomers are shown in Fig. 7. There was a small amount of high- spin P450 at normal pressure which simplifies the analysis. One can see that the application of pressure leads to a considerable irreversible decrease in the amplitudes of 2B4-specific absorbance bands. Corresponcl- ingly, changes in the concentrations of 2B4 species indi- cate a fast decay of the hemoprotein accompanied by some appearance of the P420 state. The whole pool of 2B4 is exposed to inactivation: at 2500 kbar only about 5% of the initial P450 concentration was found in the system.

Application of the factor analysis technique here reveals two pressure-induced transitions that represent more than 99.9% of the total spectral changes in the system (Fig. 8). Target-oriented transformation of these principal components based on the suggestion that one of the processes is a stoichiometric P450 + P420 conversion shows the second one as P420 disappearance (bleaching). Some systematic deviation of the basis spectrum of this latter process from its fitting by P420 spectral standard takes place at about 370-380 and 440-450 nm. Such deviation probably indicates the appearance of free hemin species absorbing in these regions. These results agree with our previous hypothe-

TABLE I

Equilibrium Constants and Reaction Volumes of Benzphetamine Binding and Spin Shift

Transitions of Cytochrome P450 2B4

Equilibrium constant .lv”” Transition” at zero pressure (ml/mol)

P450, --t P450h K, = 8.31 -49.1 (P450-S), --t (P450-S),, @ = 0.32 -21.2 P450 + s - P450-s K,, = 0.132 rTIM -8.1

o The designations used here correspond to those specified foi scheme 131.

’ The sign of the reaction volume corresponds to the high- to- low- spin and substrate complex dissociation transitions.


O.D.

3

Pressure, kbar

I I

450 500 Wavelength, nm

A0.D.

Wavelength. nm 1 I

350 400 450 500 Wavelength, nm

550

FIG. 7. Pressure-induced transitions in cytochrome P450 2B4 in solution in the presence of 0.2% Triton N-101. Other conditions were similar to those indicated for Fig. 2. Spectra measured at 1 bar and 1.0, 1.2, 1.5, 1.7, 2.8 kbar are shown. (a) Absolute spectra. Inset: Corresponding changes in the concentrations of 2B4 species: P4501, (1). P450+,, (21, P420 (3) and total hemoprotein concentration (4). The lines correspond to the results of data fitting. (b) Differential spectra as referred to the spectrum of the initial state (at 1 bar). Inset: Difference between the spectrum measured 2 h after decompression and the spectrum of the initial state (at 1 bar before pressure increase). Dashed line represents the fitting of this spectrum by the combination of the extinction standards of the pure 2B4 species and corresponds to the disappearance of 4.0 pM P450 and appearance of 1.0 pM P420.

sis that the P450 -+ P420 transition is followed by P420 transition accounted for 95 t 1% of the total hemopro- hemoprotein decay (bleaching) at high pressures. Pres- tein content. Thus, on monomerization an apparent in- sure dependencies of both processes strictly obey Eq. homogeneity of 2B4 pool disappears. 121 (Fig. 9). Monomerization leads to a large increase in the absolute value of the reaction volume of the P450 + P420 conversion (from 35 ml/m01 in oligomers to 86 ml/

DISCUSSION

mol in monomers), which is presumably caused by the The data on the pressure-induced changes in ferric increase in protein mobility and/or water accessibility cytochrome P450 2B4 are in good agreement with those on monomerization. The amplitude of the P450 -+ P420 reported earlier (7) for the 2B4(Fe2+)-CO complex. To-

340 DAVYDOV ET AL,.

a0.D. A O.D.

Wavelength, nm

Loading factor

Wavelength, nm

Pressure, kbar

Square deviation, (0.D.)2

1.0.10-6

500 Wavelength, nm

600

FIG. 8. PCA of the data set shown in Fig. 7b. (a) First basis spectrum (99% of the total changes in terms of the square deviation from the spectrum of the initial state). Dashed line shows the results of its fitting by the linear combination of the standard spectra of 2B4 (-0.97 /LM P450,,, -0.03 PM P45Oh., +0.29 pM P420, p* = 0.993). (b) Second basis spectrum (0.9% of the total changes). Results of its fitting by the linear combination of the standard spectra of 2B4 (-1.0 FM P450,., +0.09 pM P450h., + 1.0 P420, p2 = 0.94) are shown by the dashed line. (c) Loading factors for the first (1) and second (2) components versus pressure. (d) Surface of the square deviations for the approximation of the data by two first principal components (overall accuracy is 99.9%).

gether they show that the pool of 2B4 in the oligomers is divided into two fractions with different pressure-

sure-related behavior. Sixty-seven percent of the hemoprotein undergoes pressure-induced P450 + P420 tran-

related behavior and, therefore, are represented by two sition in oxidized and reduced CO-bound (7) states. The conformers of the hemoprotein. Each has distinct pres- rest of the hemoprotein is stable up to 3 kbar in the


P-450 converted 100% ,

0 0.5 I I .5 2 2.5 3

Pressure, kbor

FIG. 9. Pressure course of the two suggested particular pressure- induced transitions in P450 monomers: 1:l P450 + P420 transition (1) and P420 bleaching (2). The lines correspond to the results of data fitting by Eq. 121.

oxidized state and up to 6 kbar in the reduced CO- bound state (7). No evidence for an interconnection between these two conformers was found and the constant ratio of these forms in 2B4 oligomers (hexamers) was evaluated to be about 2:l.

Our analysis of pressure-induced spin shift shows that only one-third of the 2B4 pool in oligomers is exposed to spin equilibrium. This fraction corresponds to the conformer of the hemoprotein, which is not exposed to any pressure-induced inactivation. This finding is in agreement with the fact that in the microsomal cytochromes P450 from family II (including P450 2B4) the maximum percentage of the high-spin state, which is obtainable at the saturating concentrations of sub- strates, always seems restricted to some limit lower than 100% (32-34). For P450 oligomers in solution at 37°C this limit appears to be between 30 and 50% of the high-spin state (39-41).

It was found that monomerization of 2B4 oligomers by Triton N-101 leads to the disappearance of the apparent inhomogeneity of the 2B4 pool both in the reduced 2B4-CO complex (7) and in the oxidized state of the hemoprotein. Therefore we propose that the reported 2:l conformational splitting in 2B4 oligomers is a result of some peculiarities of the supramolecular organization of 2B4 oligomers. In other words, we propose that the oligomers that 2B4 forms in solution (8- 14) and, apparently, in the membrane (15-18) exhibit a static 2:l distribution of the conformers with different pressure sensitivity. This ratio is in agreement with the electron microscopy studies of Tsuprun and Mya- soedova et al,, who have reported that P450 2B4 hexamers in solution are organized as dimers or trimers (13). If we accept that this distribution is caused by

some inequivalence of the locations in the hexamer, no interconversion between these conformers without their dissociation-reassociation is expected.

It should be noted that we have previously proposed such inhomogeneity of the 2B4 pool in the oligomers to interpret the biphasic kinetics of cytochrome P450 2B4 dithionite-dependent reduction (42). It was found that the ratio of the amplitudes of the fast and slow phases was always about 2:l in the microsomes, in the proteoliposomes, and in solutions of P450 2B4 oligomers. No change in this ratio with temperature and substrate concentration was detected. As for P450 inhomogeneity in pressure stability, the biphasity of dithionite-dependent reduction also vanished after dissociation of the oligomers by Triton N-101. These biphasic reduction kinetics were attributed to a difference in the heme accessibility by dithionite anion monomer in two hypo- thetical states of P450 in the oligomers (42). Our present data support this hypothesis.

Our finding of the conformational inequivalence of P450 2B4 subunits in hexamers provides a possible key with which to solve the problem of the biphasic kinetics of NADPH-dependent reduction of P450 (43-45). The observation that two-thirds of 2B4 in oligomers is “fro- zen” in the low-spin state can be considered a possible reason for the well-known discrepancy between ex- tremely fast rates of spin transitions in microsomal P450 heme iron (46) and the apparent “freezing” of this equilibrium during NADPH-dependent reduction of the hemoprotein (44, 45). The fact that in the monomeric reconstituted system, consisting of purified microsomal electron carriers and detergent, these kinetics become monophasic agrees with the above model. How- ever, to expand our model to the case of membrane- bound P450 it is necessary to study pressure-related behavior of microsomal cytochromes P450 in the proteoliposomes and microsomes. The results of these studies are to be reported in a forthcoming paper.

Other important deductions were made by analyzing the pressure-induced spin shifts in 2B4. As reported above, apparently no changes in the volume of 2B4 molecule result from the binding of benzphetamine. This result is in contrast with that proposed earlier for P450,,, (2-4). To confirm the difference between these two hemoproteins it is necessary to perform systematic studies of pressure-induced changes in P450,,, at various concentrations of substrate (camphor).

Comparing pressure-related behavior of 2B4 monomers and oligomers, we can deduce that oligomeriza- tion results in a considerable stabilization of the hemoprotein. A very high absolute value of the reaction volume (-86 ml/mol) for the P450 + P420 transition reveals a high flexibility of the 2B4 molecule in the monomeric state compared to that of oligomers, for which all pressure-induced transitions have reaction volumes less than 50 ml/m01 by the absolute value.

342 DAVYDOV ET AL.

ACKNOWLEDGMENTS 19.

The authors are grateful to Dr. Michael C. Marden and Professor Peter Rand for critical reading of the paper and useful comments.

20.

They are also very much obliged to Dr. Gerard Batelier (Institut 21.

d’Enzymologie, CNRS, Gif sur Yvette, France) for his help in ultra- centrifugation experiments. The authors gratefully acknowledge that 22.

the present investigation became possible by a long-term fellowship given by INSERM to Dr. D. R. Davydov. The part of this work per- 23. formed at the Institute of Biomedical Chemistry (Moscow) was supported by The Russian Foundation of Fundamental Investigations (RFFI), Grant 94-06-13764-a. 24.

REFERENCES 25.

1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

Hui Bon Hoa, G., and Marden, M. (1982) Eur. J. Biochem. 124, 311-315.

Fisher, M. T., Sligar, S. G., and Scarlata, S. F. (1985) Arch. Biochem. Biophys. 240, 456-463.

Marden, M. C., and Hui Bon Hoa, G. (1987) Arch. B&hem. Biophys. 253, 100-107. Hui Bon Hoa, G., Di Prima, C., Dondaine, I., Sligar, S. G., Gunsa- Ins, I. C., and Douzou, P. (1989) Biochemistry 28, 651-656.

Hui Bon Hoa, G., Di Primo, C., Geze, M., Douzou, P., Kornblatt, J. A., and Sligar, S. G. (1990) Biochemistry 29, 6810-6815.

Di Primo, C., Hui Bon Hoa, G., and Sligar, S. (1990) Eur. J. Biochem. 193, 383-386. Davydov, D. R., Knyushko, T. V., and Hui Bon Hoa, G. (1992) Biochem. Biophys. Res. Commun. 188, 216-221.

Myasoedova, K. N., and Berndt, P. (1990) FEBS Lett. 270,177- 180.

Guengerich, F. P., and Holladay, L. A. (1979) Biochemistry 18, 5442-5449.

French, J. S., Guengerich, F. P., and Coon, M. J. (1980) J. Biol. Chem. 255.4122-4119.

Bachmanova, G. I., Scotselyas, E. D., Kanaeva, I. P., Kuznetsova, G. P., Gordeev, S. A., Korneva, E. N., Karyakin, A. V., and Archa- kov, A. I. (19861 Biochim. Biophys. Res. Commun. 139,883-888.

Kanaeva, I. P., Dedinskii, I. R., Skotselyas, E. D., Krainev, A. G., Guleva, I. V., Sevryukova, I. F., Koen, Ya. M., Kusnetsova, G. P., Bachmanova, G. I., and Archakov, A. I. (1992) Arch. Bio- them. Biophys. 298.395-402.

Tsupnm, V. L., Myasoedova, K. N., Berndt, P., Sograf, 0. N., Orlova, E. V., Chemyak, V. Ya., Archakov, A. I., and Skulatchev, V. P. (1986) FEBS Z&t. 205,35-40.

Myasoedova, K. N., and Tsuprun, V. L. (1993) FEBS Lett. 325, 251-254.

Kawato, S., Gut, J., Cherry, R. J., Winterhalter, K. H., and Rich- ter, C. (1982) J. Biol. Chem. 257, 7023-7029.

AIston, K., Robinson, R., Park, S. S., Gelboin, H. V., and Fried- man, F. K. (1991) J. Biol. Chem. 266, 735-739.

Greinert, R., Finch, S. A. E., and Stier, A. (1982)Xenobiotica 12, 717-726.

Schwartz, D., Pirrwitz, J., Meyer, H. W., Coon, M. J., and Ruck- Paul, K. (1990) Biochem. Biophys. Res. Commun. 171, 175-181.

26.

27.

28.

29.

30.

31.

32.

33. 34. 35.

36.

37.

38.

39.

40.

41. 42.

43.

44.

45.

46.

Imai, Y. (1987) in Cytochrome P450 (Sate, R., and Omura, T., Eds.), pp. 37-46, Kodansha, Tokyo. Imai, Y., and Sato, R. (1967) Eur. J. Biochem. 1, 419-426. Hui Bon Hoa, G., Douzou, P., Dahan, N., and Balny, C. (1982) Anal. Biochem. 120, 125- 145. Malinowski, E. R., and Howery, D. G. (1980) Factor Analysis in Chemistry, Willey-Interscience, New York. Ramos, L. S., Beebe, K. R., Karey, W. P., Sanches, M. E., Erick- son, B. C., Wilson, B. E., Wangen, B. E., and Kowalski, B. R. (1986) Anal. Chem. 58, 294R-515R. Appellof, C. J., and Davidson, E. R. (1981)AnaL Chem. 53,2053- 2056. Shrager, R. I., and Hendler, R. W. (1982)Anal Chem. 54,1147- 1152.

Hofrichter, J., Sommer, J. H., Herr-y, E. R., and Eaton, W. A. (1983) Proc. Natl. Acad. Sci. USA 80, 2235-2239. Halaka, F. G., Babcock, G. T., and Dye, J. L. (1985) Biophys. J. 48,209-219. Durell, S. R., Lee, C.-H., Ross, R. T., and Gross, E. L. (1990) Arch. Biochem. Biophys. 278, 148-160. Bridgman, P. W. (1931) The Physics of High Pressure, Macmillan Co., New York. Ristau, O., Rein, H., Janig, G., and Ruckpaul, K. (1978) Biochim. Biophys. Acta 536, 226-234. Wells, A. V., Li, P., Champion, P. M., Martinis, S. A., and Sligar, S. G. (19921 Biochemistry 31,4384-4393. Weber, G. (1992) Protein Interactions, Chapman and Hall, New York. Marquardt, D. W. (1963) J. Sot. Ind. Appl. Math 11,431-440. Nelder, J. A., and Mead, R. (1965) Comput. J. 7, 308-313. Karuzina, I. P., and Archakov, A. I. (1994) Free Radicals in Biol- ogy and Medicine, Vol. 16, pp. 73-97, Pergamon Press Ltd. Jansson, I., Curti, M., Epstein, P. M., Peterson, J. A., and Schenkman, J. B. (1990)Arch. Biochem. Biophys. 283,283-292. Ingelman-Sundberg, M., and Paul, K. G. (1986) Acta Chim. Stand. 40, 233-234. Myasoedova, K. N., and Berndt, P. (1990) FEBS Lett. 275,235- 238. Gibson, G. G., and Tamburini, P. P. (1980) in Biochemistry, Bio- physics and Regulation of Cytochrome P450 (Gustaffson, J. A., Ed. ), pp. 133- 136, Elsevier, Amsterdam. Cinty, D. L., Sligar, S. G., Gibson, G. G., and Schenckman, J. B. (1979) Biochemistry 18, 36-42. Guengerich, F. P. (1983) Biochemistry 22, 2811-2820. Davydov, D. R., Karyakin, A. V., Binas, B., Kurganov, B. I., and Archakov, A. I. ( 1985) Eur. J. Biochem. 150, 155- 159. Backes, W. L., Tamburini, P. P., Jansson, I., Gibson, G. G., Sli- gar, S. G., and Schenckman, J. B. (1985) Biochemistry 24,5130- 5136. Backes, W. L., and Eyer, C. S. (1989) J. Biol. Chem. 264,6252- 6259. Karyakin, A. V., and Davydov, D. R. (1988) Vestn. AMN SSSR 1, 53-62. Ziegler, M., Blanck, J., and Ruckpaul, K. (1982) FEBS Lett. 150, 219-222.

PRESSURE-INDUCED TRANSITIONS IN CYTOCHROME P450 2B4

MINIPRINT SUPPLEMENT

343

SpectraLab software package for data acquisition and analysis ia enzymology: Curve fitting using principal component analysis

algorithms

General characteristics of SpectraLob package.

Numerical methods of curve fitting are now in common use in biochemical an biophysical laboratories. At present several software packages for data fitting in biochemistry are available (Duggleby, 1984; Beynon, 1987; Knack and Rohm, 1981; Letherbarrow, 1990)‘. They are based on the same principles and have similar characteristics. Being very useful in most routine biochemical applications, these packages, however, are often inefficient in fitting complex kinetic models, such as multiexponential kinetics, complex inhibition mechanisms, ligand saturation curves at comparable concentrations of enzyme and ligand, etc. They do not permit the global fitting of three-dimensional data sets (e.g., reaction velocity vs. concentrations of substrate and inhibitor). Furthermore, these packages usually have a limited data manipulation and preprocessing toolkit.

Many approaches used to study the kinetic and physico-chemical properties of enzyme systems are based on spectroscopic analysis, which requires numerical techniques to resolve individual compounds and particular transitions in complex system with multiple chromophores (22 - 28). It is convenient to combine in a single package, these methods of data analysis and a real time data acquisition and handling routines having minimal instrumentation dependency. In an attempt to satisfy this need for a universal tool for data acquisition and fitting in spectroscopic and kinetic analysis in enzymology we have developed our software package “SpectraLab”. It consists of two main programs: a general purpose data acquisition and curve fitting program (AQUID) and the SPAN module for the analysis of absorbance and fluorescence spectra based on the PCA (principal component analysis) approach.

The package was developed for IBM PC XT/AT compatible computers and runs under MS-DOS (version 3.3 and higher). Size of the prop package is less than I MByte. It was written using Turbo Pascal 6.0 (about 95% of the code). Assembler language was used to write some instrument drivers. The package could be easily adapted for almost any analytical instrument (such as spectrophotometers, spectrofluorimeters, pH-states, etc.) having a digital interface or at least an analog output for an external plotter.

The architecture of SpectraLab resembles a general purpose spreadsheet program. The functions, such as tile input/output, data edit and display, data acquisition procedures, data fitting and preprocessing (smoothing, calculation of derivative and area, arithmetic manipulations, etc.) are available in three ways: using pop-up menus, functional keys or command line (command file). In the former case a kind of macro- definition language is used. This feature is useful for preparing macro- definitions for automatic data acquisition and processing.

The program operates simultaneously with up to 64 curves (series X/Y-data pairs) of unlimited length. Each curve is associated with an optional “third-coordinate” (Z) value that is useful in systematic studies of the response of biochemical system versus some physico-chemical factor (temperature, pressure, ligand concentration, etc.). It allows fitting of three-dimensional data sets (series of data curves) to models including two independent variables (pressure and concentration of the substrate, time and inhibitor concentration, etc.).

Algorithms.

w Curve fitting is performed by direct optimization maximizing the square of correlation coefficient of the fitting. Methods of optimization implemented in the package include Marquardt (33) and Nelder-Mead (34) algorithms, Marquardt algorithm is realized in the p’rototypal form. In our implementation of Nelder-Mead algorithm we applied a normalization procedure on the optimizing parameters prior to optimization. Such normalization uses initial estimates of the parameters as

scaling factors. Therefore, the initial estimate for each parameter corresponds to the value of 1 .O in the normalized scales.

These two algorithms may be used either alone or in any sequential combination. Such combination is often very useful because of different features of these two algorithms. While the Marquardt algorithm is faster and more precise in the proximity of the minimum, the Nelder-Mead method is less local, and provides faster minimization if the initial estimates lie too far from the minimum. This method allows to find the minimum in some cases where the Marquardt algorithm is inefficient.

To provide the best fitting results, the selection of the method and the set of optimization parameters is done automatically depending on the fitting equation and data noise level. However, these parameters may be changed by the user, if necessary.

The program uses a fixed set of fitting equations, which includes many commonly used models (initial velocity calculation, fitting by polynomial, multiexponential equation, Michaelis-Menten equation, reversible and irreversible second order reaction kinetics, Arrhenius equation, Hill equation, ligand saturation equation at comparable concentrations of enzyme and ligand, etc.) as well as some more special models such as pressure dependence of equilibrium of first- and second order transitions, kinetics of modification of oligomeric enzyme attended with conformational transitions of subunits (Khailova, e.a., 1987; Kurganov, e.a., 1992) and some others. Initial estimates are mostly evaluated using appropriate linearization coordinates and a linear least squares technique as have been described by Knack and Rohm for (Knack and Rohm, I98 I ).

This method, being also known as bilinear factor analysis (28) or “singular value decomposition” (SVD) analysis (26), was first introduced in chemistry by Malinowski (22). It is widely used nowadays to interpret absorbance and fluorescence spectral changes in complex biochemical systems (23 - 28). The version of PCA algorithm realized in the SPAN module of the Spectralab package is close to that described by Durell and co-workers (28).

Briefly. this approach is based on the analysis of a covariance matrix built from a series of spectra reflecting the course of chemical transitions in the system induced by some variable z, such as temperature, pressure., etc. In our version of the method, we deal with the differential spectra referred to the spectrum of the initial state (at z=z,$. If the data are

collected at I wavelength and at J+l values of the independent variable z (including z=z,$, the resulting matrix D of the absorbance changes is of

order IxJ. If the change in the z-variable induces Q independent chemical transitions accompanied by changes in absorbance, according to the Beer’s law the element Dij of this matrix is represented as

where ci is the molar&y of the chromophore(s) that underwent the

transition q at “zj, and fiq is the change in the molar absorptivity (times

the pathlength of the cell) of the initial chromophore(s) on passing through the transition. This equation written in matrix notation becomes

D=FCT The F matrix describes the spectral shapes of the absorbance changes resulting from the individual transitions, while the C matrix contains information on the molar flux through each of the transitions induced by the changes in variable z. The vector fq, called the basis spectrum of the

component q, corresponds to the differential spectrum characterizing the q transition. The vector c

q’ called the vector of the loading factors for the

component q, represents the molar changes associated with the q transition versus variable z.

I References cited by number correspond to those from the list of references of the main paper. Additional references listed at the end of this miniprint are cried by author’s name and the year of publication.

344 DAVYDOV ET AL.

The ultimate goal of applying PCA is the determination of the number of independent components Q, as well as the columns of both F and C (i.e. basis spectra and vectors of loading factors): In our implementation of PCA the second moment matrix of the differential spectra, M, is analyzed. The matrix M is defined by

M=(l/J)DDT The rank of M gives an estimate of spectrally detectable independent processes in the system. Having constructed M, the number of components (rank) can be determined by diagonalizing the matrix. In principal, the number of non-zero diagonal elements (eigenvalues) is equal to the number of independent components (processes) in the system. However, due to experimental error, all the diagonal elements are usually non-zero.

To determine Q, Gr) is calculated by using 1,2,...r components and compared with the experimental matrix. In the presence of noise, one may suggest that the number of independent processes, Q, is the same as r when the difference between two matrices becomes comparable to the experimental error. These two components account for more than 99% of the total spectral changes.

When the minimal number of the independent processes taking place in the system, Q, is found. PCA also provides a factorisation of the matrix D into F (IxQ) and C (JxQ) matrices. In such a factorisation the most important independent part of the changes is represented in the tint component, the next one - in the second, etc. Unfortunately, fq and cq vectors found by PCA do not reveal the mechanisms behind the changes immediately. These vectors do not correspond, in general, to the individual “physical” processes, but should be considered as their linear combinations. To delineate the “physical” processes from the results of factor analysis some additional information is required, such as the extinction spectra of the pure components of the system and some suggestion as to the nature of the transitions. With such information available, so called “target testing” (22) can be applied to transform the factorisation of the D matrix (i.e., to redistribute the spectral changes between the components) into the best tit of the “targets.” Targets could be formulated from the propositions concerning the spectral shape of the absorbance changes resulting from some individual transition(s); and/or from some expectations of particular dependencies of the loading factors on the z parameter, for example Michaelis-Menten or Arrhenius equations, etc. (22.28). Such transformation allows one to get to the basic spectra and loading factors that represent the apparent set of the real “physical” processes taking place in the system.

Conclusion.

The software package “SpectraLab” designed to cover the needs of data acquisition and analysis in spectroscopic and kinetic studies in biochemistry. It was developed in an effort to achieve versatility in application, simplicity of usage and flexibility in the fitting algorithms. The combination of spectral analysis (PCA) and curve fitting algorithms tits many practical needs in physico-chemical analysis of complex enzyme systems. A peculiarity specific to the program is its ability to fit three- dimensional data sets (i.e., consisting of several curves for various pressure, substrate or inhibitor concentration, etc.). This facility is very useful in fitting complex models, such as the pressure dependence of the substrate binding (this paper), kinetics of modification of oligomeric enzyme accompanied by conformational changes of subunits (Khailova, e.a., 1987). etc.

SpectraLab may be easily adapted to a wide range of the analytical instruments. It was initially developed for the Aminco DW2a spectrophotometer with an Aminco-Morrow stop-flow. Later we developed the universal instrument driver, which is easily adaptable to almost any spectrophotometer or spectrofluorimeter with digital or analog output (using an ADC-board). For a couple of years we have been using the SpectraLab package in the automatised high-pressure spectroscopy system based on a Guy-219 spectrophotometer. Using this package with several various instruments, we have found that it allows the user to be almost independent on the instrumentation details.

Additional information concerning the program is available from the authors on request, which can be sent either by postal mail or by E-Mail to D.R.Davydov at [email protected].

References.

Beynon R.J. (1987) Curvefit software package. IRL Press Software. Duggleby, R.G. (1984) Comput. Biol. Med., l&447-55. Khailova, L.S., Kurganov, B.I., Nemerya, N.S. and D.R. Davydov (1987)

Mol. Biol. (Rus.), 2.L, 636-644. Knack, I., and Rohm, K.H. (198 I ) Hoppe-Seylers Z. Phys. Chem.,B,

1119-l 130. Kurganov, B.I., Khailova, L.S., and D.R. Davydov (1992) J. Chem.

Biochem. Kinetics 1992, I, 321 - 330 Letberbarrow R.J. (1990) Trends B&hem. Sci., l2,455-457.

High-pressure-induced transitions in microsomal cytochrome P450 2B4 in solution: Evidence for...

Documents

Transcript of High-pressure-induced transitions in microsomal cytochrome P450 2B4 in solution: Evidence for...