Recruitment of a Foreign Quinone into the A1 Site of Photosystem I

Interruption of the menAor menB gene in Synechocystis sp. PCC 6803 results in the incorporation of a foreign quinone, termed Q, into the A1 site of photosystem I with a number of experimental indicators identifying Q as plastoquinone-9. A global multiexponential analysis of time-resolved optical spectra in the blue region shows the following three kinetic components: 1) a 3-ms lifetime in the absence of methyl viologen that represents charge recombination between P700+ and an FeS− cluster; 2) a 750-μs lifetime that represents electron donation from an FeS−cluster to methyl viologen; and 3) an ∼15-μs lifetime that represents an electrochromic shift of a carotenoid pigment. Room temperature direct detection transient EPR studies of forward electron transfer show a spectrum of P700+ Q− during the lifetime of the spin polarization and give no evidence of a significant population of P700+ FeS− fort ≤ 2–3 μs. The UV difference spectrum measured 5 μs after a flash shows a maximum at 315 nm, a crossover at 280 nm, and a minimum at 255 nm as well as a shoulder at 290–295 nm, all of which are characteristic of the plastoquinone-9 anion radical. Kinetic measurements that monitor Q at 315 nm show a major phase of forward electron transfer to the FeS clusters with a lifetime of ∼15 μs, which matches the electrochromic shift at 485 nm of the carotenoid, as well as an minor phase with a lifetime of ∼250 μs. Electrometric measurements show similar biphasic kinetics. The slower kinetic phase can be detected using time-resolved EPR spectroscopy and has a spectrum characteristic of a semiquinone anion radical. We estimate the redox potential of plastoquinone-9 in the A1site to be more oxidizing than phylloquinone so that electron transfer from Q− to F X is thermodynamically unfavorable in the menA and menB mutants.

Interruption of the menA or menB gene in Synechocystis sp. PCC 6803 results in the incorporation of a foreign quinone, termed Q, into the A 1 site of photosystem I with a number of experimental indicators identifying Q as plastoquinone-9. A global multiexponential analysis of time-resolved optical spectra in the blue region shows the following three kinetic components: 1) a 3-ms lifetime in the absence of methyl viologen that represents charge recombination between P700 ؉ and an FeS ؊ cluster; 2) a 750-s lifetime that represents electron donation from an FeS ؊ cluster to methyl viologen; and 3) an ϳ15-s lifetime that represents an electrochromic shift of a carotenoid pigment. Room temperature direct detection transient EPR studies of forward electron transfer show a spectrum of P700 ؉ Q ؊ during the lifetime of the spin polarization and give no evidence of a significant population of P700 ؉ FeS ؊ for t < 2-3 s. The UV difference spectrum measured 5 s after a flash shows a maximum at 315 nm, a crossover at 280 nm, and a minimum at 255 nm as well as a shoulder at 290 -295 nm, all of which are characteristic of the plastoquinone-9 anion radical. Kinetic measurements that monitor Q at 315 nm show a major phase of forward electron transfer to the FeS clusters with a lifetime of ϳ15 s, which matches the electrochromic shift at 485 nm of the carotenoid, as well as an minor phase with a lifetime of ϳ250 s. Electrometric measurements show similar biphasic kinetics. The slower kinetic phase can be detected using timeresolved EPR spectroscopy and has a spectrum characteristic of a semiquinone anion radical. We estimate the redox potential of plastoquinone-9 in the A 1 site to be more oxidizing than phylloquinone so that electron transfer from Q ؊ to F X is thermodynamically unfavorable in the menA and menB mutants.
The overall goal of these studies is to replace phylloquinone (vitamin K 1 ; 2-methyl-3-phytyl-1,4-naphthoquinone) in the A 1 site of photosystem I with a foreign quinone of different redox and/or kinetic properties but still capable of forward electron transfer to the FeS 1 clusters. In two previous papers, we described the construction and physiology of phylloquinone biosynthetic mutants in Synechocystis sp. PCC 6803 (1), and we reported EPR and electron nuclear double resonance studies that implied the presence of a foreign quinone, termed Q, in the A 1 site of PS I (2). To summarize briefly, the interruption of the menA and menB genes, which code for dihydroxynaphthoic acid synthase and phytyltransferase, respectively, results in the termination of phylloquinone biosynthesis as judged by high pressure liquid chromatography/mass spectroscopy (HPLC/ MS) and gas chromatography/mass spectroscopy of pigment extracts from isolated PS I complexes. However, the menA and menB mutant strains grow photoautotrophically under low light intensities, and isolated PS I complexes are capable of sustaining high rates of steady-state electron transfer from cytochrome c 6 to flavodoxin. HPLC and HPLC/MS studies show that quantitative amounts of plastoquinone-9 are present in PS I complexes isolated from the menA and menB mutants, whereas plastoquinone-9 is present only in vanishingly small amounts in PS I complexes isolated from the wild type. EPR studies indicate that Q Ϫ has a considerably larger g anisotropy than the native phylloquinone, consistent with the presence of a 1-ring benzoquinone rather than a 2-ring menaquinone. In addition, the prominent hyperfine splittings due to the 2-methyl group in phylloquinone are absent in the CW and transient EPR spectrum of Q Ϫ . Spin-echo and transient EPR studies show that Q Ϫ is located at the same distance from P700 ϩ , and with the same orientation with respect to the membrane, as phylloquinone in the wild type. Pulsed electron nuclear double resonance studies additionally show features that arise from nearly axially symmetric hyperfine couplings tentatively assigned to two methyl groups on Q. These results indicate that in the absence of phylloquinone PS I recruits plastoquinone-9 into the A 1 site of the menA and menB mutants and that plastoquinone-9 functions as an efficient elec-tron cofactor from A 0 Ϫ to the FeS clusters. The premise of the present study is that since plastoquinone-9 and phylloquinone have different one-electron reduction potentials in dimethylformamide, they likely possess different reduction potentials and hence different kinetic properties in the A 1 site. The kinetics of forward electron transfer from A 1 Ϫ to the FeS clusters in wild-type PS I have been well documented using transient EPR spectroscopy and time-resolved UV spectroscopy. Both methods are consistent in showing a forward electron transfer time of ϳ280 ns in spinach and cyanobacterial PS I complexes (3)(4)(5)(6). Photovoltage measurements of oriented cyanobacterial PS I complexes reveal an electrogenic phase with a similar forward electron transfer time that was ascribed the A 1 Ϫ to F X transition (7). Optical studies reveal an additional component with a forward electron transfer time of ϳ10 ns in detergent-treated samples but not in whole cells (5). In the absence of detergent, the fast phase represents only 30% of the total contribution, leading to the speculation that in the native membranes the forward transfer time is the same in spinach and cyanobacteria.
This third paper in the menA/menB series focuses on the kinetics of electron transfer in PS I complexes from the menA and menB mutants of Synechocystis sp. PCC 6803. Since electron transfer rates are sensitive to changes in Gibbs free energy as well as to alterations in distances and reorganization energies among donor and acceptor pairs, the replacement of phylloquinone would be expected to translate to a change in the rate of electron transfer through A 1 . The analysis is simplified by the findings that the orientation of Q Ϫ and the distance between Q Ϫ and P700 ϩ in the mutants are the same as for phylloquinone in the wild type (2). Similarly, the reorganization energy is not expected to differ significantly given that the replacement quinones are structural analogs of phylloquinone. We find that the rate of forward electron transfer from Q Ϫ to F X is slowed by a factor of ϳ100 -1000 compared with the wild type. The forward electron transfer kinetics allows us to measure a light-induced difference spectrum of Qminus Q in the UV. Based on the behavior of plastoquinone-9 and phylloquinone in organic solvents, and based on rate versus free energy relationships derived from electron transfer theory, we estimate the redox potential of plastoquinone-9 in the A 1 site.

MATERIALS AND METHODS
Optical Kinetic Spectroscopy in the Near IR Region-Optical absorbance changes in the near IR were measured using a laboratory-built spectrophotometer (8). To ensure resolution of kinetics in the microseconds time domain, a high frequency roll-off amplifier described in the original specifications was not used, and the signal was fed directly into the plug-in (11A33 differential comparator, 100-MHz bandwidth) of the Tektronix DSA601 oscilloscope. The sample cuvette contained the PS I trimers isolated from the menA or menB mutants at 50 g/ml Chl in 25 mM Tris-HCl, pH 8.3, 10 mM sodium ascorbate, 4 M DCPIP, and 0.04% Optical Kinetic Spectroscopy in the Visible Region-Optical absorbance changes in the visible region were measured using a laboratorybuilt spectrometer consisting of a 400-watt tungsten-halogen source (Oriel), a 1 ⁄4-meter monochromator (Jarrell Ash model 82-410) prior to the sample cuvette (to select a measuring wavelength), a sample compartment for a 1 ϫ 1-cm fluorescence (4-sided clear) cuvette, a second 1 ⁄4-meter monochromator (Jarrell Ash model 82-410) after the sample cuvette (to reject the flash and fluorescence artifacts), and a PIN-10 photodiode detector (United Detector Technology). Suitable lenses were placed to focus the light on the monochromator slits and to provide a collimated beam through the sample cuvette. The photocurrent was converted to a voltage with a 30,000-ohm resistor; the voltage was amplified using a model 100 amplifier (EG & G) set to a DC bandwidth of 100 kHz, and the signal was digitized using a Model 4094A digital oscilloscope (Nicolet Instruments, Madison, WI) interfaced via an IEEE-488 bus using a GPIB-TNT board (National Instruments, Austin, TX) in a Power Macintosh 7100/88 computer. The data were transferred to the computer and stored as binary files using LabView 4.1 (National Instruments, Austin, TX) and further processed in Igor Pro 3.14 (Wavemetrics, Inc., Lake Oswego, OR). Actinic flashes were supplied using a frequency-doubled Nd-YAG laser (Spectra Physics) operating at 532 nm at a flash energy of 1.4 mJ. Each kinetic trace represents 16 averages within the digital oscilloscope. The sample cuvette (10 ϫ 10-mm) contained 3.0 ml of PS I complex at 10 g/ml Chl suspended in 25 mM Tris, pH 8.3 with 0.04% ␤-DM, 10 mM sodium ascorbate, and 4 M DCPIP.
Optical Kinetic Spectroscopy in the UV Region-Optical studies in the UV (240 -340 nm) were conducted using a pulse-probe spectrometer described in Ref. 9. The monochromator slit was fixed at 4 mm, equivalent to a bandwidth of 8 nm. Measurements in the UV were performed using a xenon flashlamp as the actinic source filtered by a Schott and a Kodak Wratten 34. The photodiodes were protected with Corion Solar Blind UV-transmitting filters. The optical path length of the sample cuvette was 1 cm. Each data point represents the average of eight measurements, taken with a flash spacing of 20 s to allow complete reduction of P700 ϩ by the external donor. A background measurement was taken similarly except that the sample was shielded from the detecting flash to allow for correction of the actinic flash artifact. The absorbance change shown represents the difference between the two measurements. The sample cuvette contained the PS I trimers isolated from the menB mutant at 10 g/ml Chl, 25 mM Tris-HCl, pH 8.3, 10 M sodium ascorbate, 4 M DCPIP, and 0.03% ␤-DM.
Time-resolved CW EPR Spectroscopy-EPR spectral changes shown in Fig. 1B were measured using a Bruker ESP 300E spectrometer equipped with a TM 110 cylindrical resonator (Bruker ER4103). A frequency-doubled Nd-YAG laser (Spectra Physics DCL) provided the excitation flash at a wavelength of 532 nm and at an energy of 14 mJ. The computer was configured to capture and average the data and to flash the laser at 10-s intervals. The data represent an average of 64 flashes. The flat cell (Wilmad WG-813-Q) contained the PS I trimers isolated from the menA or menB mutants at 400 g/ml Chl, 25 mM Tris-HCl, pH 8.3, 10 mM sodium ascorbate, 4 M DCPIP, and 0.03% ␤-DM.
Electron Spin Polarized Transient EPR Spectroscopy (Direct Detection)-Room temperature-transient EPR experiments were carried out using a setup described in detail elsewhere (10). The samples were pumped continuously through a flat cell mounted in a Varian rectangular resonator equipped with a rough-surfaced glass window that scatters the laser light to provide a more even distribution of light intensity in the cell. Time/magnetic field data sets were collected using direct detection by measuring light-induced transients at fixed magnetic field positions over an appropriate spectral region. Decay-associated spectra were then generated by fitting the transients with a kinetic function and plotting the amplitude(s) against the magnetic field as discussed (3,4).
Transient EPR Spectroscopy with Field Modulation Detection-The same setup and conditions were also used to collect time/field data sets shown in Fig. 7 but with field modulation and lock-in detection. By using direct detection, the decay of the spin polarization limits the accessible time range to times shorter than a few microseconds. By using field modulation, the spectrometer has a rise time of ϳ50 s but a much higher sensitivity so that slow forward electron transfer and charge recombination can be monitored.
Photovoltage Measurements-Measurements of transmembrane electric potential difference generation by PS I-containing proteoliposomes adsorbed onto the surface of azolectin-impregnated collodion film were done at room temperature as described elsewhere (11). The instrument rise-time was 200 ns. Saturating light flashes were provided by a frequency-doubled Quantel Nd:YAG laser ( ϭ 532 nm; pulse half-width, 15 ns; flash energy, 20 mJ).
Analysis of Kinetic Data-The multiexponential fits of optical and EPR kinetics were performed by the Marquardt algorithm in Igor Pro version 3.14 (Wavemetrics Inc., Lake Oswego, OR) on a G3/300 Macintosh computer. For global analysis of kinetics in the visible region ( Fig.  2), individual kinetics were analyzed first assuming the presence of either three components or two components and a base line. The results of these analyses were then used for fitting the whole set of data to global lifetimes, and the best solution was chosen based on the analysis of 2 , standard errors of the parameters, and the residuals of the fits.

RESULTS
P700 ϩ Recombination Kinetics, Optical and EPR Spectroscopy-In PS I trimers isolated from the wild type with ␤-DM, the reduction of P700 ϩ is multiphasic after a saturating flash (8). When measured by optical spectroscopy in the near-IR and in the absence of external electron acceptors, the majority of P700 ϩ is reduced with lifetimes of ϳ86 and 12 ms in an ϳ10:1 ratio (minor microsecond kinetic phases are present due to back reactions from earlier acceptors in damaged reaction centers). There also exists a long lived kinetic phase of P700 ϩ reduction due to direct reduction of P700 ϩ by reduced DCPIP that represents ϳ20% of the total absorbance change. In PS I trimers isolated from the menB mutant, the reduction of P700 ϩ is also multiphasic after a saturating flash (Fig. 1A). When measured in the absence of external electron acceptors, P700 ϩ is reduced with lifetimes of ϳ2.6 and 7 ms in a 0.63:0.37 ratio (minor microsecond kinetic phases are similarly present due to back reactions from earlier acceptors in damaged reaction centers). There also exists a minor long lived kinetic phase of P700 ϩ reduction due to direct reduction of P700 ϩ by reduced DCPIP that represents 7% of the total absorbance change. The menA mutant showed nearly identical kinetics (data not shown).
When measured in the same reaction medium using timeresolved EPR spectroscopy with field modulation detection (Fig. 1B), the majority of P700 ϩ in the menB mutant is reduced with a lifetime of 3.2 ms, a value in good agreement with the optical study. Since the signal-to-noise limits the precision of these measurements, the single kinetic phase found in the EPR measurement may correspond to a convolution of the 2.6-and 7-ms lifetimes found in the optical experiment. The EPR study also shows a long lived kinetic phase of P700 ϩ reduction that corresponds to ϳ13% of the total spin concentration. The menA mutant showed similar kinetics (data not shown). If we consider only the 3.2-ms kinetic phase, then the charge recombination of the electron acceptor(s) with P700 ϩ is ϳ25 times faster in the menA and menB mutants than in the wild type.
Global Multiexponential Analysis of Optical Spectra in the Blue Region- Fig. 2A depicts spectra obtained in the blue region by global multiexponential analysis of kinetics in the absence of methyl viologen in PS I complexes from the menB mutant. Three discrete kinetic components are present. The component fitted to the slowest kinetic phase approximated by a base line (checked boxes) corresponds to the spectrum (P700 ϩ minus P700) and represents a population of P700 ϩ reduced by the external electron donor, DCPIP. In these reaction centers, the electron has already escaped from FeS Ϫ to an external acceptor, probably molecular oxygen. The 3.2-ms kinetic phase (solid squares) corresponds to the spectrum of (P700 ϩ FeS Ϫ minus P700 FeS) and represents a close match to the kinetics of P700 ϩ relaxation measured optically and by EPR (Fig. 1, A and B). The fastest kinetic phase is the 13-s kinetic event (solid circles), which will be discussed in detail later. Fig. 2B depicts spectra obtained by global multiexponential analysis of kinetics in the presence of methyl viologen. Three discrete kinetic components are also present. The component fitted to the slowest kinetic phase (open squares) corresponds to the spectrum of (P700 ϩ minus P700), and it represents the entire population of P700 ϩ reduced by the external electron donor, DCPIP. The spectrum of the 18-s component (open circles) resembles that of the 13-s component in the absence of methyl viologen (this component will be discussed later). An additional third component, with a lifetime of 744 s (solid diamonds), is present with a broad bleaching from 400 to 500 nm characteristic of an S 3 Fe charge-transfer transition and corresponds to the spectrum of FeS Ϫ minus FeS. It is not possible to identify unambiguously the FeS cluster giving rise to this absorbance change because the oxidized minus reduced difference spectra of F X , F B , and F A are nearly identical. However, the electron is likely in equilibrium between F A and F B , and the kinetics likely represents the forward electron donation from the terminal electron acceptor F B Ϫ to methyl viologen. The larger absorbance change of the 3.2-ms phase in the absence of methyl viologen ( Fig. 2A) compared with the slow phase in the presence of methyl viologen (Fig. 2B) is derived from the additional contribution of a reduced electron acceptor. Assuming that the spectrum of the 3.2-ms component in the absence of methyl viologen corresponds to the absorbance changes brought about by P700 ϩ plus a reduced acceptor, whereas the spectrum of the slow kinetic component in the presence of methyl viologen corresponds to the absorbance changes of P700 ϩ alone, the difference will correspond to the spectrum of the reduced acceptor. The resulting spectrum, presented in Fig. 2C, is similar to that of the 744-s kinetic phase in the presence of methyl viologen (Fig. 2B, solid diamonds), showing a broad bleaching from 400 to 500 nm characteristic of an S 3 Fe charge-transfer transition. (The amplitude of the absorbance change is lower than in Fig. 2B (solid diamonds) because a fraction of the reaction centers has already lost the electron from the FeS Ϫ clusters.) The electron that back-reacts with P700 ϩ with a 3.2-ms lifetime is therefore located in a FeS cluster, although for reasons mentioned above, it is not possible to differentiate between F B and F A . The menA mutant showed similar results (data not shown).
Forward Electron Transfer, Decay of the Electron Spin Po- larized EPR Signal from P700 ϩ A 1 Ϫ -An attempt was made to measure the forward electron transfer from Q Ϫ to the FeS clusters in the menA and menB mutants by accumulating time/ field transient EPR data sets at 9 GHz (X-band) in PS I complexes isolated with ␤-DM. Fig. 3 shows spectra extracted from the data sets by fitting the individual transients as described in Ref. 4. Fig. 4 shows selected transients corresponding to the field position marked by the arrow in Fig. 3. For the wild-type sample (Fig. 3, top, and Fig. 4, top), the onset of the laser flash results in the appearance of an E/A/E polarization pattern within the rise time of the spectrometer due to the generation of the spin-polarized radical pair P700 ϩ A 1 Ϫ (3,4). At the field position chosen for the transients in Fig. 3, this radical pair gives the absorptive contribution to the top trace. This spectrum decays with a time constant of ϳ280 ns to the emissive spectrum due to P700 ϩ FeS Ϫ shown in Fig. 3, top. In the corresponding transient in Fig. 4 (top), the electron transfer results in a transition from an absorptive contribution at early time to an emissive contribution at later times. We have shown that in samples devoid of F A and F B , the late signal is retained indicating that the electron transfer proceeds via F X (4). However, it is not possible to identify unambiguously the FeS cluster involved in the radical pair giving the emissive spectrum in intact samples because its contribution is spread over a large spectral range and because of fast relaxation at room temperature (see Ref. 12 for discussion). The subsequent decay of the emissive spectrum with a time constant of 1.5 s represents the relaxation of the spin-polarized signal. This represents the limit on the time resolution of the spin polarization method.
For the menA (Figs. 3, middle, and 4, middle) and menB (Figs. 3, bottom, and 4, bottom) mutants, the onset of the laser flash results in the appearance of a spectrum that we assign to the state P700 ϩ Q Ϫ , where Q is tentatively identified as plastoquinone-9 in the A 1 site. This spectrum does not show any indication of singlet-triplet mixing in the precursor state, and its rise time is governed by the response time of the spectrometer. Thus, we conclude that electron transfer to Q occurs on a time scale of less than ϳ0.5 ns. Compared with the wild type, the spectrum of P700 ϩ Q Ϫ decays more slowly than that of P700 ϩ A 1 Ϫ , and there is no indication of P700 ϩ FeS Ϫ during the 1.5-s decay of the spin polarization pattern. Thus, we can place a conservative lower limit on the lifetime of P700 ϩ Q Ϫ at 2-3 s. This result indicates that electron transfer from A 1 Ϫ to F X is slowed by a factor of at least 10, from ϳ280 ns in the wild type to Ն 2-3 s in the menA and menB mutants; one kinetic phase will be shown to be ϳ300 s when measured directly in  3. Electron spin polarized EPR spectra of the wild-type, menA, and menB mutants at room temperature. Top trace, native PS I. The two sequential spin polarized spectra have been extracted from the complete time/field data set as described in detail (4). The solid curve corresponds to the radical pair P700 ϩ A 1 Ϫ , whereas the emissive spectrum is due to P700 ϩ FeS Ϫ . Middle and lower traces, menA and menB mutants. The curves are decay-associated spectra extracted from the complete time/field data sets. In the mutants only one kinetic component is observed which we assign to P700 ϩ Q Ϫ . The data were collected using direct detection as described under "Experimental Procedures" and are plotted with absorption (A) in the positive direction and emission (E) in the negative direction. the field modulation transient EPR experiment described later.
Forward Electron Transfer, Optical Absorbance Changes in the Near-UV-We found it possible to measure accurately forward electron transfer from Q Ϫ to the FeS clusters in the UV using a flash-detection, pump-probe spectrophotometer (13). Absorption changes were determined in the menA mutant from 245 to 330 nm at time points 5 s, 100 s, and 5 ms after an actinic flash. The light-minus-dark difference spectrum recorded 5 s after the flash shows a maximum at 310 nm, a cross-over at 280 nm, and a minimum of 255 nm, as well as a shoulder at 290 -295 nm (Fig. 5A, circles). The difference spectra at 100 s (Fig. 5A, squares) and 5 ms (Fig. 5A, triangles) after the flash resemble the spectrum at 5 s, implying that all kinetic phases are derived from the same species. This spectrum is strikingly similar to the plastosemiquinone-9 anion radical minus plastoquinone-9 difference spectrum in methanol (Fig. 5B, checked circles), which shows a maximum at 315 nm, a crossover at 280 nm, and a minimum of 255 nm as well as a shoulder at 280 -300 nm (14). The reader should note that the absolute value of the absorbance change of the trough on the short wavelength (blue) side is larger than the absolute value of the absorbance change of the peak on the long wavelength side in both the Q Ϫ /Q (Fig. 5A, circles) and the plastoquinone anion radical minus plastoquinone-9 (Fig. 5B, checked circles) difference spectra. The spectrum of Q Ϫ /Q is also strikingly similar to the Q A minus Q A difference spectrum in deoxycholate-isolated PS II particles (15), except that the latter is shifted to the red about 15 nm (Fig. 5B, checked squares). It is noteworthy that the spectrum in Fig. 5A does not resemble that of wild-type PS I, in which the flash-induced difference spectrum (with phylloquinone in the A 1 site) shows a peak centered at 380 nm and a crossover of ϳ325 nm (5). It is also interesting that the Q Ϫ /Q difference spectrum differs from the PQH/PQ difference spectrum, in which the peak occurs at 295 nm, and in which the absolute value of the extinction coefficient at 250 nm is over twice that at 300 nm (14). This indicates that the plastoquinone radical in the A 1 site remains unprotonated during its measured lifetime. The light-minus-dark difference spectrum of Q Ϫ /Q therefore supports the assignment of Q as plastoquinone-9 in the A 1 site (1, 2).
The kinetics of the absorbance change at the 310 nm maxi-mum of the menA mutant are depicted in Fig. 6A. The decay kinetics are (at least) biphasic, and the lifetime of the fast phase is estimated to be 17.6 s. The slow phase is difficult to fit precisely due to the limited number of data points. Nevertheless, the extrapolated absorbance at the onset of the flash indicates that ϳ60% of Q Ϫ decays within 100 s. Assuming that 100 chlorophyll molecules are associated with each PS I monomer, then at a chlorophyll concentration of 10 g/ml, the P700 concentration in this sample would be 112 nM. The flashinduced absorption change of Q Ϫ /Q at the peak maximum of 315 nm (Fig. 5A, circles) would correspond to 0.94 mOD measured 5 s after the flash. The differential extinction coefficient of PQ Ϫ /PQ at the peak in the UV is reported as 13,000 M Ϫ1 cm Ϫ1 in solution (14), and this value has also been used for Q A in PS II (15). By using this value and the absorption difference at the onset of the flash gives a total of 72 nM of Q undergoing light-induced reduction. Assuming an equimolar concentration of plastoquinone-9 in the A 1 site with P700, this would correspond to 64% of the redox-active Q. However, the difference spectrum is recorded 5 s after the flash, and given that the lifetime of the fast kinetic phase is 17.6 s, the absorption change of Q Ϫ /Q at the onset of the flash can be estimated as 1.04 mOD. Hence 80 nM, or 71% of the redox-active Q, is associated with forward electron transfer. This estimation suffers from uncertainty in the extinction coefficient of plastoqui- none-9 in the A 1 site and in the estimate of the absorbance at the onset of the flash. Nevertheless, the calculation shows that the majority of electrons that are transferred from A 0 Ϫ to the FeS clusters are mediated by plastoquinone-9. This rules out a significant electron bypass of the quinone at room temperature in the menA and menB mutants.
Forward Electron Transfer Kinetics, Electrochromic Bandshift at 490 nm-The global multiexponential analysis of the flash-induced changes in the absorption spectra of PS I complexes from the menB mutant revealed a 13-s kinetic event in the absence of methyl viologen with a derivative-shaped spectrum centered at 460 nm ( Fig. 2A, solid circles connected with dotted line). This spectrum is characteristic of an electrochromic shift of a pigment that occurs in response to electron transfer. The same electrochromic shift can be seen in the presence of methyl viologen, except that the extracted lifetime is 18 s and the amplitude is higher (Fig. 2B, open circles with dotted line). The longer lifetime may be instrument-related; to obtain a reasonable signal-to-noise, the rise time (1/e) of the amplifier was limited to 10 s, which would tend to underestimate the initial amplitude in Fig. 2A. The electrochromic shift also shows a slower kinetic phase in the absence (Fig. 6B) and presence (data not shown) of methyl viologen. The kinetics in the near-millisecond time domain are complicated by absorbance changes at 490 nm due to the decay of P700 ϩ , and no further attempt was made to separate their relative contributions. Similar kinetic phases were measured in the menA mutant (data not shown).
In wild-type PS I, the light-minus-dark difference spectrum shows a positive-going absorption band from 440 to 500 nm with a shape that resembles the red shift of a pigment centered at about 470 nm (5). A similar set of absorption changes were recently measured in mutant cells of Chlorella sorokiniana that lack PS II (16). These features were attributed in the wild type to an electrochromic red shift of an absorption band of a carotenoid that is induced by A 1 Ϫ . Similarly, the spectrum of the ϳ18-s component measured in menB mutant most probably represents an electrochromic bandshift due induced by Q Ϫ on a nearby carotenoid. The kinetics of the flash-induced carotenoid bandshift therefore represent an indirect, but reliable, method to measure the oxidation kinetics of the plastosemiquinone anion Q Ϫ in the visible region.
Photovoltage Measurements on PS I Complexes in Proteoliposomes-Excitation of oriented PS I complexes with a single turnover flash leads to the generation of a transmembrane electric potential difference from which the forward electron transfer rates and dielectrically weighted transmembrane distances can be measured (7,11,17). PS I complexes were incorporated into proteoliposomes, and the flash-induced response corresponded to the negative charging of the proteoliposome interior. For wild-type PS I in the absence of an external electron donor to P700 ϩ , the photoelectric response due to charge separation between P700 and the terminal electron acceptors, F A /F B , occurs within the ϳ0.2-s rise time of the instrument (8). Thus, the 20-and 200-ns kinetic phases of forward electron transfer between A 1 and F X in wild-type PS I are not resolved using our instrumentation. However, for PS I complexes isolated from the menB mutant (Fig. 6C, inset) and menA mutant (data not shown), the photoelectric response shows an instrument-limited rise followed by a slower rise in the submillisecond time range. Decomposition of these kinetics, presented in Fig. 6C, reveals components with lifetimes of ϳ11.4 and 306 s, and equal relative contributions (ϳ10%) to the overall photoelectric response. The decomposition also includes an offset that represents a longer lived component in the low millisecond time range. The lifetime of the fast kinetic phase is in reasonable agreement with the measurements of Q Ϫ oxidation at 315 nm and the carotenoid bandshift at 485 nm. The slow kinetic phase is also probably related to the oxidation of Q Ϫ , although the optical data at 315 and 485 nm are compromised by the presence of millisecond lifetime P700 ϩ / P700 changes at these wavelengths. We assign these two components to vectorial electron transfer from Q Ϫ forward to the FeS clusters in the menA and menB mutants. The data points contributing to the 11.4-s phase are overwhelmed by the large spike at early time, which is probably due to dielectric relaxation of the sample following the charge separation (8). Elimination of this artifact is difficult, and as a result there is a reasonably large error associated with the lifetime and amplitude of the faster of the two phases.
Forward Electron Transfer, Field Modulation Transient EPR-If the slow kinetic phase that extends into the hundreds of microseconds time range represents the oxidation of Q Ϫ , then the semiquinone anion radical should be detectable at these times using conventional time-resolved EPR spectroscopy (i.e. using field modulation detection). Fig. 7 shows the results of an analysis of a room temperature field modulation transient EPR experiment. Fig. 7, top, shows boxcar spectra of PS I complexes from the menA mutant at two times compared with the decay-associated spectrum of the 32-ms major kinetic phase in the wild type. At late times of ϳ5 ms after the laser flash, the menA mutant and the wild type give the same spectrum because the electron is on one of the FeS clusters (see Fig.  2B). At early times of ϳ100 -200 s after the laser flash, the zero crossing in the spectrum of the mutants is clearly shifted to lower field because of the contribution from Q Ϫ . The g anisotropy and crossover of this radical are similar to that obtained in dark-adapted whole cells of the menA mutant after exposure to white light and can be attributed to a semiquinone anion radical (2) . Fig. 7, bottom, shows a fit at the field position marked with an arrow in Fig. 7, top. From this fit, an electron transfer lifetime of ϳ300 s and a recombination lifetime of ϳ5 ms can be extracted. The identification of a slow kinetic phase with the decay of Q Ϫ is consistent with the presence of the Q Ϫ /Q difference spectrum measured in the UV 100 s after the flash (Fig. 5A, squares). It is, of course, possible that the slow kinetic phase represents a distribution of rate constants that extend into the low millisecond time range (see Fig. 5A, triangles); this will require further study. DISCUSSION The results from time-resolved optical, electrometric, and EPR techniques support the identity of Q in the menA and menB mutants as plastoquinone-9 and show that despite the altered forward and back electron transfer kinetics, plastoquinone-9 functions as an efficient electron cofactor in PS I. To accomplish forward electron transfer, it is necessary that the midpoint potential of in the A 1 site be sufficiently reducing to transfer electrons to the acceptor flavodoxin and at a rate that outcompetes the inherent back reaction of reduced plastoquinone-9 with P700 ϩ . We estimate the midpoint potential of Q Ϫ /Q using two approaches as follows: a comparison of the redox properties of plastoquinone-9 and phylloquinone in organic solvents, and a consideration of rate versus free energy relationships from electron transfer theory.
Phylloquinone has a reported E1 ⁄2 of Ϫ465 mV (versus NHE) in dimethylformamide (DMF) (21), whereas plastoquinone-9 has a reported E1 ⁄2 of Ϫ369 mV (versus NHE) in DMF (22). If the A 1 site has a polarity similar to that of DMF, then plastoquinone-9 would be ϩ96 mV more oxidizing than native phylloquinone in the A 1 site. However, this is only a crude estimate, and it should be possible to refine this value using the concept of "acceptor number." Jaworski and colleagues (18) showed that the redox potential of a quinone undergoing the first electron reduction in organic solvent is related to the electrophilic properties of the solvent. This work was based on a formulation by Gutman (19) of an acceptor number, a dimensionless number that expresses the acceptor properties of a given solvent relative to that of SbCl 5 . The central idea is that E1 ⁄2 values of different quinones show smaller differences to one another in solvents with low acceptor numbers and higher differences to one another in solvents with high acceptor numbers. This solvent effect is quantitatively described by the empirical Equation 1: where a is the slope (i.e. the sensitivity to the solvent effect based on Lewis acidity); AN is the acceptor number (which ranges from 0 in hexane to 100 in SbCl 5 , the reference solvents; benzene is 8.2, DMF is 16, and water is 54.8), and E1 ⁄2 0 is the intercept (the value of E1 ⁄2 corresponding to a solvent with AN ϭ 0). An important point is that the semiquinone radical is destabilized in solvents with low acceptor numbers, which leads to a lower redox potential for the first electron reduction. Itoh and co-workers (20) applied Gutman's ideas to a study of replacement quinone head groups (lacking the phytyl tail) in PS I and estimated the acceptor number for the A 1 site to be 4.0, which is similar to the acceptor number for diethyl ether of 3.9. By assuming the redox potential of a given quinone in organic solvent is strictly linearly related to the E1 ⁄2 value in the A 1 site, the following Equation 2 could be derived (20): where E1 ⁄2 is the redox potential in DMF and E m is the redox potential in the A 1 site. Given that phylloquinone has an E1 ⁄2 of Ϫ465 mV (versus NHE) in DMF (21), the E m of phylloquinone in the A 1 site would be Ϫ754 mV. Given that plastoquinone-9 has an E1 ⁄2 of Ϫ369 mV (versus NHE) in DMF (22), the E m of plastoquinone-9 in the A 1 site would be Ϫ687 mV. Accordingly, the redox potential of plastoquinone-9 in the A 1 site would be 67 mV more oxidizing than phylloquinone. Equation 2 was derived assuming that the added quinones are capable of passing the electrons forward to the FeS clusters. Since there remains a lingering uncertainty over whether added quinones without phytyl tails are properly oriented so as to accommodate forward electron transfer from A 1 Ϫ to the FeS clusters (23-26), we sought an independent approach to the determination of the midpoint potential of plastoquinone-9 in the A 1 site.
According to Moser and co-workers (27,28), the rate of intraprotein electron transfer between two electron carriers with distance R can be described by the following Equation 3: where R is the edge-to-edge distance in Å; ⌬G 0 is the standard reaction free energy in eV, and is the reorganization energy in eV. Any change in the reorganization energy, the distance (including the orientation), or the redox potential of the quinone in the A 1 site will have an effect on the rate constant of electron transfer. For the purpose of this argument, we will assume that there will not be any change in the reorganization energy by substitution of similar molecules in the A 1 site, i.e. a dimethylbenzoquinone for a naphthoquinone. We showed earlier (2) that the distance between the plastoquinone-9 anion radical and P700 ϩ in the menA and menB mutants was nearly identical to the distance between the phylloquinone anion radical and P700 ϩ in the wild type. Hence, any alteration in the rate constant of electron transfer will be the governed primarily by the difference in the redox potential between phylloquinone and plastoquinone-9. The rate relationship in Equation 4 is equivalent to Equation 3. k exergonic ϭ 10 ͑15Ϫ0.6RϪ3.1͑⌬G 0 ϩ͒ 2 /͒ (Eq. 4) The equilibrium constant K eq between Q and F X is defined in Equation 5 as follows: The equilibrium constant K eq is related to Gibbs free energy according to Equation 6.
Hence, the reverse electron transfer rate, k reverse , can be expressed in Equation 7 as follows: We employ Equation 4 to specify the rate of forward electron transfer in an exergonic reaction and Equation 7 to specify the rate of forward electron transfer in an endergonic reaction. The averaged edge-to-edge distance between Q K (the two identified quinones in the x-ray crystal structure) and F X is reported to be 11.3 Å (29). A reorganization energy of 0.7 eV is commonly used for photosynthetic electron transfer reactions (27), although a value of 1.0 eV was recently estimated for A 1 and F X from the temperature dependence of electron transfer (30). The value of k exergonic will therefore depend on the difference in Gibbs free energy between the quinone and the F X iron-sulfur cluster. This calculation requires knowledge of accurate redox potentials for both cofactors. The E1 ⁄2 (Q/Q . ) of phylloquinone in wildtype PS I has not been measured directly; however, a value of Ϫ810 mV versus NHE has been derived from calculations based on the kinetics of electric field-induced electron transfer rates (31), and a value of less than or equal to Ϫ800 mV versus NHE has been calculated based on the P700 triplet yield (32). As mentioned earlier, the considerably higher value of Ϫ754 mV has been deduced from the measured E1 ⁄2 values of phylloquinone and plastoquinone-9 in DMF and an application of Gutman's acceptor number of 4.0 for the A 1 site. The E1 ⁄2 (F X /F x Ϫ ) of phylloquinone in wild-type PS I has been measured directly; an E1 ⁄2 of Ϫ705 mV was found for F x Ϫ /F x in PS I complexes by electrochemical poising and EPR measurement (33). However, this approach suffers from an uncertainty that the midpoint potential of F X was determined in the presence of a reduced F A and F B and may be overestimated due to an electrostatic effect of nearly reduced acceptors. A considerably higher E1 ⁄2 of Ϫ670 mV has been measured in P700-F X cores that lack PsaC and, hence, any electrostatic influence from the F A and F B clusters. Recently, an equilibrium constant of 73.5 was determined between F X and F A in P700-F X cores by analysis of the back reaction kinetics, which indicates that F X may be 110 mV more electronegative than F A (34). Since the E1 ⁄2 of F A has been measured to be Ϫ520 to Ϫ540 mV, the E1 ⁄2 of F X would be Ϫ630 to Ϫ650 mV. However, these approaches suffer from an uncertainty whether there is any effect on the midpoint potential of F X from the removal of PsaC, due either to structural changes or to alterations in solvent accessibility of the iron-sulfur cluster. Table I shows a matrix of the predicted values for the forward rate constant, k exergonic , for the three published values of the midpoint potential of A 1 Ϫ /A 1 , the three published values of the midpoint potential of F x Ϫ /F x , and the two published values of the reorganization energy. Assuming a reorganization energy of 0.7 eV, the averaged rate constant of optimum forward electron transfer would be 223 Ϯ 99 ns. This value is similar to the ϳ280-ns lifetime of A 1 Ϫ measured in wild-type PS I complexes (3)(4)(5)(6)(7). If a reorganization energy of 1.0 eV is used instead, the averaged rate of optimum forward electron transfer is 1.83 s, a value considerably slower than that observed experimentally. We next use the measured lifetime of Q Ϫ in the menA and menB mutants to back-calculate the redox potential for Q Ϫ /Q. Given that the major kinetic phase of forward electron transfer from Q Ϫ in the menA and menB mutants has a lifetime of ϳ15 s (Fig. 6, A-C), a careful consideration of Equations 4 and 7 shows that electron transfer must be endergonic (according to Equation 7) between A 1 and F X to accom- /F x , and the two values of the reorganization energy. The results imply that regardless of whether the reorganization energy is 0.7 or 1.0 eV, electron transfer between A 1 and F X is endergonic by 12-95 mV. If we accept a value of Ϫ705 mV for the F x Ϫ /F x redox couple and a reorganization energy of 0.7, then plastoquinone-9 in the A 1 site of the menA and menB mutants will have a midpoint potential of Ϫ610 mV. A value of Ϫ670 mV for the F x Ϫ /F x redox couple leads to a midpoint potential of Ϫ575 mV for plastoquinone-9 in the A 1 site, a value which is nearly isopotential with the F B ironsulfur center.
Although the value obtained using electron transfer theory provides an interesting exercise, this approach cannot be used for precise estimations because the uncertainties approach an order of magnitude for the energy and the rate constant. Given the additional uncertainties in the midpoint potentials of the components, in the precise value of the reorganization energies when phylloquinone and plastoquinone-9 occupy the A 1 site, and in the exact edge-to-edge distance between the quinone and the iron-sulfur cluster, the estimated values must be used with caution. Nevertheless, these considerations support the appraisal based on the redox behavior of plastoquinone-9 and phylloquinone in organic solvents in suggesting that the midpoint potential of Q Ϫ /Q is more oxidizing than the iron-sulfur cluster, F X (Fig. 8). If so, then we must ask how a thermodynamically unfavorable reaction can be reconciled with both the high quantum yield of electron transfer observed in the steadystate measurements (1) and the efficient reduction of F A /F B observed by optical and EPR spectroscopy (2). The important factor here is that even with an endergonic electron transfer step between A 1 and F X , overall net electron transfer is exergonic between A 1 Ϫ and F A /F B (Fig. 8). A drop in the quantum yield and an inefficiency in F A /F B reduction would be expected only if electron transfer between A 1 Ϫ and F X were sufficiently slow that it would be outcompeted by the inherent back reaction between A 1 Ϫ and P700 ϩ . We see no evidence for a back reaction between Q Ϫ and P700 ϩ when the terminal iron-sulfur clusters F A and F B are oxidized and available for electron transfer (data not shown). A similar instance of an unfavorable electron transfer with an overall negative change in the free Gibbs energy has been postulated for the electron transfer stage from F X via F A and F B to ferredoxin (35,36,38).
We next compare the calculated values in Table II with the experimental data. The issue is whether the reduction of F A /F B on a single turnover flash (Fig. 2) corresponds to the amount expected on the basis of chlorophyll concentration. In the following analysis, the term FeS refers to the iron-sulfur cluster that accepts the electron on a single-turnover flash without specifying its identity as F A or F B . Assuming that 100 chlorophyll molecules are associated with each PS I monomer, then at a chlorophyll concentration of 10 g/ml, the concentration of P700, F A , and F B in this sample is 112 nM. The flash-induced absorption change due to FeS Ϫ in this sample, determined by global analysis of the kinetics in the blue (Fig. 2B, solid diamonds), corresponds to 1.4 mOD. Given the extinction coefficient for P430 at 430 nm to be 13,000 M Ϫ1 cm Ϫ1 (37), a total of 108 nM of FeS undergoes light-induced reduction. The flashinduced absorption change due to FeS Ϫ in this sample, determined by difference in the absence and presence of methyl viologen (Fig. 2C), corresponds to 1.1 mOD (however, this is a known underestimate; see "Results"). Here, a total of 84.6 nM of FeS undergoes light-induced reduction. Taking the midpoint potential of F A to be Ϫ530 mV and the amount of F A reduced to be between 76 and 96%, a straightforward application of the Nernst equation indicates that the midpoint potential of Q Ϫ /Q must be more reducing than Ϫ559 to Ϫ611 mV to account for the high quantum yield of FeS reduction on a single turnover flash (for simplification, we ignore equilibrium between F A and F B ). These values are in reasonable agreement with the those determined by applying electron transfer theory (Table II).
In conclusion, multiple approaches to the problem show that when plastoquinone-9 occupies the A 1 site, electron transfer between Q Ϫ and F x is likely to be endergonic. However, efficient forward electron transfer occurs in PS I because of the large, favorable free energy change from A 0 Ϫ to flavodoxin. A more rigorous assessment of the midpoint potential of Q Ϫ /Q based solely on kinetic arguments will be published elsewhere. 2  menA/B mutants (B). The back reaction rates in the wild-type refer to conditions where the succeeding electron acceptor has been removed, either biochemically or genetically. The values in bold represent the dominant kinetic phase. The majority of the forward and back electron transfer times in the menA/B mutant have not been characterized. The back reaction pathways are depicted as direct to P700 for the sake of clarity; the actual pathway likely proceeds, at least in part, back through the electron acceptor chain.