Picosecond Detection of an Intermediate in the Photochemical Reaction of

Transcription

1 Proc. Nat. Acad. Sci. USA Vol. 72, No. 6, pp , June 1975 Picosecond Detection of an Intermediate in the Photochemical Reaction of Bacterial Photosynthesis (bacteriochlorophyll/photooxidation/picosecond spectroscopy) MARK G. ROCKLEY*, MAURICE W. WINDSOR*, RICHARD J. COGDELLt, AND WILLIAM W. RSONt * Department of Chemistry, Washington State University, Pullman, Wash ; and t Department of Biochemistry, University of Washington, Seattle, Wash Communicated by Hans Neurath, April 7, 1975 ABSTRACT Preparations of photosynthetic reaction centers from Rhodopseudomonas sphaeroides were excited with flashes lasting approximately 8 psec. Immediately after the excitation, there appeared a transient state which was characterized by new absorption bands near 5 and 68 nm, by a bleaching of bands near 54, 6, 76, and 87 nm, and by a blue shift of a band near 8 nm. The transient state decayed with an exponential decay time, r, of 246 i 16 psec after the flash. As the transient state decayed, the radical cation of the reaction center bacteriochlorophyll complex appeared. This indicates that the transient state is an intermediate in the photooxidation of the bacteriochlorophyll. The absorption spectrum of the transient state shows the state to be identical with a state (PF) which has been detected previously in reaction centers that are prevented from completing the photooxidation, because of chemical reduction of the electron acceptor. Analysis of the spectrum suggests that the formation of pf involves electron transfer from one bacteriochlorophyll molecule to another within the reaction center, or possibly from bacteriochlorophyll to the bacteriopheophytin of the complex. The initial absorbance changes after flash excitation also include a bleaching of an absorption band at 8 nm. The bleaching decays with T AZ 3 psec. The bleaching appears not to be a secondary effect, but rather to reveal another early step in the primary photochemical reaction. The primary photochemical reaction of bacterial photosynthesis is the transfer of an electron from a bacteriochlorophyll complex, P, to an acceptor, X, whose identity is uncertain (1-3). The electron transfer reaction occurs with a quantum yield of essentially 1% (4), and it occurs with great speed, even at temperatures below 4K. It can be studied in preparations of isolated "reaction centers," which contain three different polypeptides, four equivalents of bacteriochlorophyll, two of bacteriopheophytin, and one each of ubiquinone and nonheme iron (1-3). In a search for clues to the mechanism of the electron transfer reaction, Parson et al. (5) have studied reaction center preparations in which the photochemical reaction was blocked by the chemical reduction of X. When such preparations were excited with a 2 nsec flash, optical absorbance changes revealed the formation of two short-lived states. One of the transient states, called pf, appeared instantaneously and decayed with a half-time of approximately 6 nsec. State pf was formed in high quantum yield, suggesting that it might be an intermediate in the photochemical electron transfer reaction. The second transient state, PR, was of less interest, because of its slower formation and lower quantum yield. Abbreviation: IR, infrared. State pf is unlikely to be the lowest excited singlet state of the bacteriochlorophyll complex, P*, because measurements of the fluorescence yield indicate that the lifetime of P* is only 2 to 4 psec when X is in the reduced form (and less when X is not reduced) (6, 7). The proposal that the primary reaction involves an intermediate state such as pf, rather than proceeding directly from P*, would explain a longstanding anomaly in the quantitative relationship between the quantum yields of photochemistry and fluorescence (3-5). The possibility has remained, however, that pf is a side-product which forms only if the normal reaction is blocked. A decision on the role of state pf in photosynthesis, therefore, requires information on whether pf is formed under conditions that permit the electron transfer reaction to occur. We report here on an investigation of this point, using picosecond kinetic techniques. The results appear to establish pf as an intermediate in the electron transfer reaction. While this work was in progress, Kaufmann et al. (8) reached the same conclusion, using somewhat different techniques of psec spectroscopy. MATERIALS AND METHODS Reaction centers were obtained from cells of Rhodopseudomonas sphaeroides strain R-26 by the method of Clayton and Wang (9). The detergent lauryldimethylamine oxide was replaced by Triton X-1 by dialysis (5, 1). The apparatus used in the psec experiments was essentially that of Magde and Windsor (11). We shall summarize the approach only briefly here for clarity; ref. 11 provides additional details. A single pulse from a mode-locked Nd+3/glass laser was frequency-doubled to provide a 53 nm excitation flash lasting about 8 psec. This illuminated only a small band across the center of the sample. A measuring flash of white light, which was generated by self phase modulation of residual 16 nm light from the original pulse, traversed an optical delay path and interrogated the sample at an adjustable time after the actinic flash. The relatively weak measuring pulse passed through both the center of the sample and the unexcited regions above and below the center, allowing a measurement of the absorbance differences caused by the excitation. After passing through the sample, the measuring pulse proceeded to a spectrograph which was equipped with a silicon vidicon detector coupled to an optical multichannel analyzer (SSR Instruments, Inc., model 125 A). To correct for spatial inhomogeneities in the measuring beam, we first measured the apparent absorbance of the central region of the sample in the absence of the excitation flash. This was then 2251

2 2252 Biophysics: Rockley et al. Proc. Nat. Acad. Sci. USA 72 (1975) ~ v A 4 9O 6 7 AA.2 B.1 -vaoi WAVELENGTH (nm) Absorbance changes caused by excitation of reaction FIG. 1. centers at moderate redox potential, as measured at 2 psec () or 24 psee () after flashes lasting approximately 8 psec. See the text and ref. 11 for experimental details. To facilitate comparison with other spectra, the ordinate is scaled to show absorbance changes for 1 1sM reaction centers and a 1 cm light path; the actual measurements employed 59 IMM reaction centers and a light path of either 1 or 2 mm. The spectrum measured at 24 psec does not show the full conversion of pf to P+, because the conversion is only about 63% complete at this time. In the region around 8 nm, the spectrum measured at 2 psec includes a partial bleaching, as well as a blue shift of the 8 nm absorption band of the reaction centers. At 2 psec, the bleaching has decayed to about 5% of its initial level. Scattered excitation light interferes with the measurements between 51 and 55 nm. Each point is an average of at least two measurements. compared with a measurement that included the excitation flash. For work in the near infrared (IR), a Schott OG 55 filter was placed in front of the spectrograph to remove light of shorter wavelengths from the measuring continuum. The excitation pulse and the measuring pulse both were planepolarized. To check whether induced dichroism caused a distortion of the measurements, we repeated measurements of the flash-induced absorbance changes at six key wavelengths, after rotating the polarization of the measuring beam by 9; within experimental error, the rotation had no significant effect on the results. The samples used in the psec experiments contained 59 um reaction centers, 5 mm Trisp HCl (ph 7.5), and.5% (v/v) I oo WAVELEGTH (fun) FIG. 2. (A) Absorbance changes accompanying the formation of state pf, as measured after excitation of reaction centers at low redox potential with flashes lasting approximately 2 nsec. See the text and ref. 5 for experimental details. The ordinate is scaled for 1 MAM reaction centers and a 1 cm light path. Some of the data for wavelengths below 4 nm are taken from ref. 5. (B) Sum of absorbance changes accompanying the one-electron oxidation of bacteriochlorophyll to the cationic free radical in CH2Cl2 and those accompanying the one-electron reduction to the anionic free radical in dimethylformamide. Data for calculation of the spectrum were taken from the work of Fajer et al. (13, 14), who prepared the two radicals by electrochemical oxidation and reduction. The ordinate is scaled for the formation of 1 MuM of each of the radicals and for a 1 cm light path. In the composite spectrum, the bleaching at 38 nm and the new band at 65 nm are due mainly to the formation of the anion, whereas the bleaching at 6 nm and the new band at 42 nm are due mainly to the formation of the cation. Triton X-1. The light path generally was 2 mm, but it was reduced to 1 mm for some of the measurements in the IR, where the reaction centers absorb strongly. The nsec absorbance measurements were performed as described in ref. 5 with a 2 nsec 834 nm actinic flash from a ruby-pumped dye laser. The samples contained 9 AM reaction centers in the same Tris-Triton buffer that was used in the psec work. Working with a higher concentration of reaction centers than Parson et al. (5) used improved the accuracy of the measurements in the 5-6 nm region con- 7

3 Proc. Nat. Acad. Sci. USA 72 (1975) Intermediate Detected in Bacterial Photosynthesis ran 6At 61 nm.2.1 K I I Ị I TiME AMTER EXCITATION (pmec) In(AAt-&A, nws) -2. I I I I I I TWE AFTER EXCITATION (psec) FIG. 3. Kinetics of decay of state pf and formation of the radical cation, P+, as measured after excitation of reaction centers at moderate redox potentials with 8 psec flashes. See the text for experimental details. The upper figure shows the formation and decay of the absorbance band at 68 nm due to PF (e) and the development of the absorbance decrease at 61 due to the formation of P+ (). The lower figure shows semilog plots of the absorbance changes observed at a given time (t) minus those observed at a long time (1 nsec) after the flash. The straight lines in these plots were obtained by linear regression; the correlation coefficient was.975 for the 68 nm data and.986 for the 61 nm data. The T values are 246 i 16 psec at 68 (-) and 22 ± 14 psec at 61 nm (). Each point is an average of between two and four measurements. The uncertainties quoted indicate the standard deviations of the least-squares fit. siderably, and allowed an extension of the measurements to 7 nm. To block the electron transfer reaction, we reduced the reaction centers with Na2S24 (5). RESULTS Fig. 1 shows spectra of the absorbance changes that result from the excitation of reaction centers with flashes lasting about 8 psec. The filled circles represent measurements that were made at 2 psec after the excitation; the open circles, measurements that were made at 24 psec. The absorbance changes measured at 24 psec after the flash show all of the features that are characteristic of the photooxidation of P to the radical cation, P+. These are well known from measurements on slower time scales (1-3). They include a bleaching of absorption bands at 6 and 87 nm and a blue shift of a band at 8 nm. One expects the photooxidation to occur under the conditions of the present experiments, because the redox potential was high enough to hold X in the oxidized form before the flash. The spectrum measured at 2 psec after the flash is similar in many respects to that measured at 24 psec, but it also includes a prominent absorption band at 68 nm, and a bleaching of bands near 55 and 75 nm. These effects have largely disappeared in the spectrum measured at 24 psec, and the bleaching near 6 nm has become much more pronounced than it is at 2 psec. Fig. 2A shows the difference spectrum for the formation of state pf, as measured after 2-nsec flashes under conditions that block the electron transfer reaction. Over the spectral regions where the two types of measurements overlap, the spectrum is identical with that seen at 2 psec after the flash in the psec experiments. The difference spectrum for the formation of state pr is not shown here, but it is quite unlike these spectra in the region around 8 nm (5). It also differs from them in having an absorption band that is essentially flat between 62 and 7 nm and that is only about half as strong as the 68 nm band in the pf spectrum (our unpublished observations). We therefore interpret the initial absorbance changes in the psec experiments as being due to the transient formation of state pf. Fig. 3 shows the kinetics of the formation of the radical cation P+ after an 8 psec flash. The rate of formation of P+ can be measured from the rate of the absorbance decrease at 61 nm, because the conversion of P to pf does not cause an absorbance change at this wavelength (Fig. 2A). The rise kinetics are first-order, with an exponential decay time, T, of 22 ± 14 psec. Fig. 3 also shows the kinetics of decay of state pf as measured from the rate of decay of the 68 nm absorption band. The decay kinetics are first-order, with r = 246 i 16 psec. The finding that the kinetics of the disappearance of pf are essentially identical with the kinetics of formation of P+ supports the conclusion that P+ forms directly from P'.

4 2254 Biophysics: Rockley et al. Proc. Nat. Acad. SCi. USA 72 (1976).6, 79nm.4 I I I I I 68rnm TIME AFTER EXaTATION B7rnm FIG. 4. Kinetics of decay of the bleaching of the 8 nm absorption band of reaction centers, after excitation at moderate redox potentials with 8 psec flashes. The increase in the absorbance at 79 nm is plotted as a function of time after the excitation. At this wavelength, the decay of the bleaching reveals an absorbance increase due to the blue shift of the 8 nm absorption band. Each point is an average of at least two measurements. In the spectral region between 78 and 82 nm, the initial absorbance changes appear to contain two distinct components. The 8 nm absorption band of the reaction center complex shifts to shorter wavelengths, resulting in absorbance increases on the blue side of the band and absorbance decreases on the red side (Fig. 1). The blue shift persists as pf decays to P+. Superimposed on the blue shift, there also is a bleaching of the 8 nm absorption band. The bleaching reverses rapidly, relative to the rate of conversion of pf to P+. Fig. 4 shows the recovery kinetics, which are first-order with = 3 psec. It seemed possible that the bleaching of the 8 nm band was a secondary effect which resulted from the absorption of a second quantum, after the conversion of P to pf. To test this possibility, we measured the dependence of the bleaching on the strength of the excitation flash (Fig. 5C). Within experimental error, the bleaching had the same dependence on the flash intensity as did the absorbance changes reflecting the formation of pf (Fig. 5A and B). The bleaching therefore appears not to be a secondary effect, but rather to reveal another early step in the primary photochemical reaction. DISCUSSION The results that are presented above appear to establish state pf as an intermediate in the photooxidation of P to P+. When reaction centers are excited with an 8 psec flash, pf forms immediately, and P+ forms as pf decays. The conversion of PF to P+ evidently does not involve other intermediate states with significant lifetimes, and there is no reason to believe that this process requires more than a single step. Our results agree well with those of Kaufmann et al. (8), although our value of x = 246 A 16 psec for the decay of pf is somewhat greater than theirs. Kaufmann et al. (8) report a x of 12 psec, based on measurements at 54 and 64 nm. Because the formation of P+ causes absorbance changes at these wavelengths, the absorbance changes that initially follow a flash decay toward an asymptote that differs from zero. Kaufmann et al. (8) calculated the decay kinetics on the assumption that the asymptote was zero, and this may have led them to a slight underestimate of the half-time. 82 nm FLASH INTENSITY (relative unit) FIG. 5. Dependence of the initial absorbance changes at 68 nm (A), 87 nm (B), and 82 nm (C) on the intensity of the excitation flash. The absorbance changes were measured as in Fig. 1, at 2 psec after the excitation. The sample had a light path of 1 mm in all cases. The flash intensity was varied with neutral density filters; timing adjustments were made to correct for the effects of the filters on the optical path length. The wavelength used in the experiment of part C (82 nm) is isosbestic for the blue shift of the 8 nm absorption band; the absorbance changes shown here reflect the bleaching of the absorption band, and no absorbance change remains at this wavelength after the decay of the bleaching. The ordinate scale indicates the actual measurements, with no correction for the fact that the bleaching had decayed to about 5% of its initial level by the time of the measurement. The strongest flashes used in these experiments had an energy of approximately 3 mj, and all of the data presented elsewhere in this report were similar to these. Each point is an average of between two and four measurements. Error bars indicate standard deviations. The measurements at 68 nm involved relatively large errors, because of the small size of the absorbance changes. In the experiments of Figs. 1 and 3, the errors were less severe, because the sample had a 2 mm path and the flashes were all of maximum strength. The present results also provide some insight as to the possible identity of state pf. From the bleaching of the absorption bands at 6 and 87 nm, it is clear that pf involves the bacteriochlorophyll of the reaction center. The blue shift of the 8 nm absorption band is particularly significant, because it is similar to the band shift that accompanies the oxidation of P to P+. Although the blue shift is partly masked by the initial bleaching of the 8 nm band, comparison of Figs. 3 and 4 shows that the shift occurs well before the conversion of pf to P+. The blue shift of the 8 nm band suggests that the formation of pf involves electron transfer, rather than simply the promotion of the reaction center to an excited singlet or triplet state. In support of this conclusion, no such shift accompanies the conversion of P to state pr, which is most likely the lowest excited triplet state of the reaction center (5). The suggestion that the formation

5 Proc. Nat. Acad. Sci. USA 72 (1975) of pf requires electron transfer is consistent with other arguments presented in refs. 3 and 5 that pf is not the lowest excited single state, P*, and that the transfer of an electron to X involves an intermediate state which forms very rapidly from P* İf the formation of pf requires electron transfer, the next question is the identity of the electron donor and acceptor. There are two major possibilities. First, the state could involve reduction of the bacteriopheophytin of the reaction center. This would account for the bleaching of the absorption bands at 54 and 76 nm (Figs. 1 and 2A), because these bands are generally attributed to the bacteriopheophytin. Arguing against this interpretation, however, is the small magnitude of the bleaching at 76 nm, relative to the bleaching at 87 nm (Fig. 1). The amplitude of the 76 nm absorption band of reaction centers is approximately the same as that of the 87 nm band, and one would expect that the reduction of one of the bacteriopheophytin molecules could cause at least half as large an absorbance decrease at 76 as one sees at 87. Unfortunately, this argument must remain tenuous, in the absence of solid information of the spectral properties of the radical anion of bacteriopheophytin. The second possibility is that pf involves the transfer of an electron from one of the four bacteriochlorophyll molecules to another. This interpretation could explain the unusual spin polarization of the triplet state, which probably forms from pf if the transfer of the electron on to X is blocked (3, 12). To consider whether an anion-cation biradical of two bacteriochlorophyll molecules could account for the spectral properties of pf, we calculated the difference spectrum that is shown in Fig. 2B. The spectrum consists of the sum of the absorbance changes that accompany the conversion of free bacteriochlorophyll to its cation radical (13) and those for the conversion to the anion radical (14). Such a spectrum can give at best a rough approximation of the difference spectrum for the formation of a biradical, because it neglects the effect of interaction between the two molecules in the biradical. In addition, some of the absorption bands of the bacteriochlorophyll in reaction centers are shifted considerably from the positions of the corresponding bands of bacteriochlorophyll in solution. Nonetheless, it is clear that the calculated spectrum agrees well with the observed difference spectrum for the formation of pf (Fig. 2A). The simulation predicts the bleaching of the absorption bands near 38 and 6 nm; it predicts the development of new bands near 42, 5, and 65 nm, and it correctly predicts the relative intensities of these bands. It also predicts the bleaching of the major IR absorption band. This is not included in the figure, because the IR band lies at a considerably longer wavelength (87 nm) in reaction centers than it does in free bacteriochlorophyll (77 nm). The model does not predict the bleaching of the bacteriopheophytin absorption bands, and one would have to ascribe this bleaching to an indirect effect. Finally, we need to consider the bleaching of the 8 nm absorption band, which decays with T = 3 psec after the flash. One possibility is that the bleaching reflects P*, the excited singlet state which we assume gives rise to state pf. Because we did not detect absorbance changes with similar Intermediate Detected in Bacterial Photosynthesis 2255 decay kinetics at other wavelengths, this interpretation would require that the absorption spectrum of P* be very similar to that of pf, except near 8 nm. It also would require a revision of the 5-1 psec lifetime which has been calculated for P* from measurements of the fluorescence yield (6, 7). it would not be surprising if such a revision were to prove necessary, because of uncertainties in the natural radiative lifetime on which the calculations are based (3, 8). A second possibility is that the bleaching occurs (or remains) as P* is converted to pf, and that the decay of the bleaching reflects a relaxation of the interactions among the four bacteriochlorophyll molecules of the reaction center. Studies of the circular dichroism spectrum of reaction centers have suggested that the bacteriochlorophyll molecules initially interact strongly, so that the absorption bands at 8 and 865 nm both are properties of the complex as a whole (15, 16). If this is correct, the conversion of the complex to P* or pf could cause a bleaching of both of these bands. Subsequent molecular reorientation could result in the reappearance of the 8 nm band, at a somewhat shifted position. It is hoped that direct measurements of the lifetime of fluorescence from P*, and psec absorbance measurements over a wider spectral range, will distinguish between these two possibilities, and will resolve some of the many additional questions that the present work leaves unanswered. We are grateful to D)rs. K. J. Kaufmann, P. L. Dutton, T. L. Netzel, J. S. Leigh, and P. M. Rentzepis for sending us a copy of their manuscript (ref. 8) prior to its publication. We also thank Robin Anderson for assistance with some of the measurements. This work was supported in part by the Office of Naval Research, and by National Science Foundation Grant GB 3732X. 1. Clayton, R. K. (1973) Anau. Rev. Biophys. Bioeng. 2, Parson, W. W. (1974) Anntu. Rev. Microbiol. 28, Parson, W. W. & Cogdell, R. J. (1975) Biochim. Biophys. Acta 416, Wraight, C. A. & Clayton, R. K. (1974) Biochim. Riophys. Acta 33, Parson, W. W., Clayton, R. K. & Cogdell, IR. J. (1975) Biochim. Biophys. Acta, in press. 6. Zankel, K. L., Reed, D. W. & Clayton, R. K. (1968) Proc. Nat. Acad. Sci. USA 61, Slooten, L. (1972) Biochim, Biophys. Acta 256, Kaufmann, K. J., Dutton, P. L., Netzel, T. L., Leigh, J. S. & Rentzepis, P. M. (1975) Science, in press. 9. Clayton, R. K. & Wang, R. T. (1971) in Methods in Enzymology, ed. San Pietro, A. (Academic Press, New York), Vol. 23, pp Dutton, P. L., Leigh, J. S. & Wraight, C. A. (1973) FEBS Lett. 26, Magde, D. & Windsor, M. W. (1974) Chem. Phys. Lett. 27, Uphaus, R. A., Norris, J. R. & Katz, J. J. (1974) Biochem. Biophys. Res. Commun. 61, Fajer, J., Borg, D. C., Forman, A., Felton, R. H., Dolphin, D. & Vegh, H. (1974) Proc. Nat. Acad. Sci. USA 71, Fajer, J., Borg, D. C., Forman, A., I)olphin, D. & Felton, R. H. (1973) J. Am. Chem. Soc. 95, Sauer, K., Dratz, E. A. & Coyne, L. (1968) Proc. NVat. Acad. Sci. USA 61, Reed, D. W. & Ke, B. (1973) J. Biol. Chem. 248,