Cytochrome b6 Arginine 214 of Synechococcus sp. PCC 7002, a Key Residue for Quinone-reductase Site Function and Turnover of the Cytochrome bf Complex*

Quinone-reductase (Qi) domains of cyanobacterial/chloroplast cytochrome bf and bacterial/mitochondrial bc complexes differ markedly, and the cytochrome bf Qi site mechanism remains largely enigmatic. To investigate the bf Qi domain, we constructed the mutation R214H, which substitutes histidine for a conserved arginine in the cytochrome b6 polypeptide of the cyanobacterium Synechococcus sp. SPCC 7002. At high light intensity, the R214H mutant grew ∼2.5-fold more slowly than the wild type. Slower growth arose from correspondingly slower overall turnover of the bf complex. Specifically, as shown in single flash turnover experiments of cytochrome b6 reduction and oxidation, the R214H mutation partially blocked electron transfer to the Qi site, mimicking the effect of the Qi site inhibitor 2-N-4-hydroxyquinoline-N-oxide. The kinetics of cytochrome b6 oxidation were largely unaffected by hydrogen-deuterium exchange in the mutant but were slowed considerably in the wild type. This suggests that although protonation events influenced the kinetics of cytochrome b6 oxidation at the Qi site in the wild type, electron flow limited this reaction in the R214H mutant. Redox titration of membranes revealed midpoint potentials (Em,7) of the two b hemes similar to those in the wild type. Our data define cytochrome b6 Arg214 as a key residue for Qi site catalysis and turnover of the cytochrome bf complex. In the recent cytochrome bf structures, Arg214 lies near the Qi pocket and the newly discovered ci or x heme. We propose a model for Qi site function and a role for Arg214 in plastoquinone binding.

The cytochrome bf complex of oxygenic photosynthesis transfers electrons between the photosystem II and I reaction cen-ters. The functional complex is a dimer comprising four major subunits (1,2): the Rieske iron-sulfur protein (ISP) 1 with a 2Fe-2S cluster; cytochrome f, which binds a c-type heme; and cytochrome b 6 , which binds high and low potential b hemes, b H and b L , and a "subunit IV" (3)(4)(5)(6). Four or five additional small subunits, PetG, PetL, PetM, PetN, and PetO (in chloroplasts) have also been evidenced (reviewed in Refs. 7 and 8). In addition, the recent x-ray crystal structures of cytochrome bf complexes have revealed a previously undetected and unique heme designated heme x in the cyanobacterium Mastigocladus laminonsus (9) or heme c i in the alga Chlamydomonas reinhardtii (10). This extra heme, a chlorophyll, and a carotenoid molecule/ monomer of the complex are features that distinguish cytochrome bf complexes from the related, bacterial/mitochondrial cytochrome bc complexes (Refs. 9 and 10; see also Refs. 11 and 12 for reviews).
Cytochrome bf complexes couple proton translocation across the membrane to electron transfer from a lipophilic, two-electron donor (plastoquinol) to a hydrophilic, one-electron acceptor protein (plastocyanin or a c-type cytochrome). The most widely accepted model to explain the mechanism of bc-bf complexes is the "Q cycle" hypothesis of Mitchell (13). This scheme, as modified by Crofts et al. (14) and others (for example see Ref. 15), postulates both an oxidation of plastoquinol and a reduction of plastoquinone at two distinct sites within the protein, the Q o and Q i sites, located on opposite, positive and negative sides of the membrane. Plastoquinol oxidation at the Q o site involves transfer of electrons to both a high potential chain (the Rieske ISP, cytochrome f, and plastocyanin or cytochrome c 6 ) and a low potential chain (hemes b L and b H and plastoquinone at the Q i site) as illustrated in Fig. 1. Plastoquinol oxidation releases two protons into the lumen, and plastoquinone reduction at the Q i site picks up two protons from the stromal space (reviewed in Refs. 5, 11, 12, and 16).
Structural elucidation of mitochondrial cytochrome bc complexes by x-ray crystallography has suggested a mechanism for plastoquinol oxidation at the Q o site (Refs. 17-20; see also Ref. 21). The extramembrane, cluster-binding domain of the Rieske protein assumes different conformations with respect to its transmembrane helix (18 -20, 22, 23). A movement of the qui-nol molecule has also been predicted, based on the location within the Q o site of residues responsible for inhibitor resistance (reviewed in Ref. 12). An interpretative model has been proposed (12), according to which the quinol and the oxidized Rieske ISP remain within hydrogen bonding distance at the Q o site until an electron is transferred from the quinol to the Rieske ISP cluster. These partners then move apart to reduce hemes b L and c 1 (equivalent to heme f in the bf complex), respectively. The recent three-dimensional structures of cytochrome bf complexes (9,10) are compatible with this model for Q o site function.
Less information is available on the oxidation mechanism of the b 6 hemes and reduction of plastoquinone at the Q i site (for examples see Refs. 16 and 24 -26). According to the Q cycle model, this occurs via a two-step reduction of a quinone molecule at the Q i site. Thus, under oxidizing conditions, plastoquinol at the Q o site reduces cytochrome b L , which in turn reduces cytochrome b H . A second turnover places both b L and b H in a reduced state causing the reduction of a plastoquinone to plastoquinol at the Q i site and proton uptake from the stromal space. This model has been confirmed experimentally for cytochrome bc complexes, mainly from results obtained with the inhibitor antimycin (26 -28), which binds selectively to the Q i site (18 -20). However, these results cannot be transposed directly to cytochrome bf complexes. Inhibitors of the cytochrome bf Q i site have been identified, but none are completely effective. The inhibitor NQNO (29), which seems to operate by a mechanism similar to that of antimycin, only partially blocks electron flow (29,30). Moreover, the redox midpoint potentials of the b hemes differ in bc and bf complexes (reviewed in Ref. 6). Thus, the equilibrium between heme b H and the Q i site quinone lies in favor of the latter in bc complexes and the former in cytochrome bf complexes (16,25). This suggests that although two electrons can be injected consecutively onto the quinone molecule in bc complexes, the transfer must occur essentially in a concerted manner in cytochrome bf complexes (16,25). Finally, the additional heme near the Q i domain of bf complexes suggests that the mechanism of quinone reduction may be very different from that in cytochrome bc complexes. Based on its ligation properties, Stroebel et al. (10) have suggested that the new heme belongs to the family of the penta-coordinate high spin c hemes. Both the algal (10) and cyanobacterial (9) structures suggest that the c i (or x) heme may be involved in electron flow from heme b H to the Q i plastoquinone and cyclic flow from photosystem I.
No functional evaluation of the cytochrome bf Q i site mechanism has yet been obtained because mutations of this site have not been available from oxygenic, photosynthetic organisms. Prior to elucidation of the cytochrome bf crystal structures, we have initiated a molecular genetic study to probe the mechanism of quinone reduction in the cytochrome bf complex. Although loss-of-function mutations are not allowed in cyanobacterial bf complexes (reviewed in Ref. 5), the feasibility of generating stable mutations within this complex in cyanobacteria is now well established (31,32). To dissect the catalytic mechanism of the Q i site, we generated mutation R214H in the PetB, cytochrome b 6 , polypeptide of Synechococcus sp. PCC 7002. This changes a conserved arginine in the cytochrome bf complex to a histidine that is the conserved, corresponding residue near the Q i pocket of cytochrome bc complexes. The converse H217R mutation in the bc complex of Rhodobacter capsulatus increases the binding affinity of the Q i semiquinone, increases the amplitude of b heme reduction, and slows electron transfer through the complex (33). We report here that the PetB R214H mutation in Synechococcus decreases the overall growth and cytochrome bf turnover rates, slows electron transfer across the membrane between the two b hemes, and thus defines a key residue for quinone-reductase (Q i site) function in cytochrome bf complexes. These data are interpreted in light of the bc and bf structures, and a model is proposed for Q i site catalysis in the cytochrome bf complex.

MATERIALS AND METHODS
Cyanobacterial Cultures-Stock cultures of Synechococcus sp. PCC 7002 were grown under cool white fluorescent lamps (20 -30 mol photons m Ϫ2 s Ϫ1 ) in medium A as described previously (31). R214H mutant cultures were supplemented with streptomycin (Sm) and spectinomycin (Sp) at 50 g/ml. For growth rate determinations, mid-log phase cells were inoculated into medium A (bubbled with 1% CO 2 in air) and incubated at 39°C and 250 mol photons m Ϫ2 s Ϫ1 light intensity. Cell density was determined by direct turbidity measurements on 18mm-diameter culture tubes in a Spectronic 20D spectrophotometer (Milton Roy, Rochester, NY). Cultures for assays were grown at 200 -250 mol photons m Ϫ2 s Ϫ1 light intensity and harvested at mid-log phase. Chlorophyll concentrations were determined as described in Ref. 34. Synechococcus strain BstUI (31) is wild type for petB but carries the same Sm/Sp resistance cassette upstream of petB as in the R214H mutant. The BstUI strain was used as a control in some experiments. Its salient properties are indistinguishable from the wild type.
Site-directed Mutagenesis of the Synechococcus petB Gene-Escherichia coli strains were grown in LB medium supplemented as needed with 150 g/ml of ampicillin, 15 g/ml tetracycline, or 25 g/ml each of Sm and Sp as described previously (31). Standard molecular genetic manipulations were performed as in (35). Plasmid pA6.2 carrying the Synechococcus petBD genes was used for site-directed mutagenesis of petB as described in Ref. 31. The petB R214H mutation was created by means of the mutagenic oligonucleotide 5Ј-CCACTTCCTCATG/AT-tCaTAAGCAAGGTATTTC-3Ј. The lowercase letters identify the modified bases (wild type ATC CGT Ͼ ATt CaT in the R214H mutant), and the underlining indicates the newly created HinfI restriction site tag. The R214H mutation in plasmid pA6.2 was verified by plasmid isolation from E. coli transformants, PCR amplification from the primers TK-1 (5Ј-GACAGAGCAAGCTGTGTTAC-3Ј) or TL-F (5Ј-ACGATCAC-CGTTTCCTTC-3Ј) and TK-2 (5Ј-AGCCATCGCCACCGGACGAC-3Ј) (locations shown in Fig. 2), and restriction tests for the HinfI tag.
Plasmid pA6.2-R214H carrying the petB R214H mutation was introduced into Synechococcus by transformation (31). Because this plasmid cannot replicate in Synechococcus, selection for Sm/Sp resistance forces integration into the genome of the Sm/Sp resistance cassette and flanking petBD sequences as illustrated in Fig. 2. Integration of the petB-R214H mutation and segregational loss of the wild type petB allele were confirmed by: (i) PCR amplification of petB from Synechococcus transformants and restriction tests for the HinfI tag as described above for E. coli, (ii) allele-specific PCR detection of the petB-R214H allele according to Ref. 31 as shown in Fig. 2B, and (iii) DNA sequencing of petB PCR products from the TL-F and TK-2 primers. Forward primers for allele-specific PCR were 5Ј-CCACTTCCTCATGATCCG-3Ј (for detection of wild type petB) or 5Ј-CCCACTTCCTCATGATtCa-3Ј (for detection of Spectroscopic Measurements-Turnover of the cytochrome bf complex was monitored by measuring the rate of reduction of cytochrome f/c 6 following a saturating 40-ms light pulse as described previously (31,34). Time-resolved absorbance changes were measured in the ␣-band region (530 -570 nm) of the cytochromes. Because the spectra of cytochromes f and c 6 are largely superposed, and transfer between them is rapid, no effort was made to monitor these carriers separately (31). To improve the signal to noise ratio, the measurements were averaged with an interval between flashes of 7-10 s. The cells were suspended at 1-5 M chlorophyll in reaction medium containing 5 mM Hepes, pH 7.5, 10 mM NaCl, and 10 mM NaHCO 3 , at 39°C. Prior to recording data, samples were light adapted by exposure to a series of saturating light flashes. The data shown are averages from three samples.
Single turnover experiments were performed at room temperature, using a "Joliot-type" spectrophotometer (36). Actinic flashes were provided by a xenon lamp (3 s at half-height) filtered through a Schott filter (RG 695). Flashes were fired at a frequency of 0.15 Hz and were nonsaturating (hitting ϳ 20% of the reaction centers) unless otherwise indicated. Cyanobacteria were kept in the dark under an argon atmosphere in a large reservoir, connected to the measuring cuvette, to ensure dark reduction of the plastoquinone pool. Cytochrome b 6 redox changes were evaluated as the difference between absorption at 563 nm and a base line drawn between 545 and 573 nm. At intervals following an actinic flash, the transmission of weak, monochromatic light passing through the sample was measured relative to a reference sample shielded from actinic light. Data were expressed as the difference (⌬I) in light transmission between the sample and reference cuvettes, normalized to light transmission (I) by the reference cuvette as in (36,37). Cells were collected in the exponential phase of growth and resuspended in 20 mM Hepes buffer, pH 7.2, containing 20% Ficoll (w/v) to prevent cell sedimentation. 3-(3Ј,4Ј-Di-chloroprenyl)-1-1-dimethylurea and hydroxylamine were added at concentrations of 10 M and 1 mM, respectively, to block photosystem II activity. Alternatively, the cells were harvested and resuspended in the same buffer containing D 2 O (99.8% D atom) with a pellet/buffer volume ratio of 1/10 and stirred for 1 h. After a second centrifugation, the cell pellet was resuspended in the same buffer, with a pellet/buffer volume ratio of 1/20. No increase in the kinetic isotope effect was observed when the duration of the incubation in D 2 O was increased.
Electrochemical Redox Titrations-Membranes of the Synechococcus R214H mutant were prepared and redox titrations performed in an electrochemical cell as described previously for wild type Synechococcus (38). The membrane pellet was resuspended in 50 mM MOPS, pH 7.0, 50 mM KCl buffer containing the redox mediators anthraquinone 2-sulfonate (Ϫ225 mV), anthraquinone 1,5-disulfonate (Ϫ170 mV), 2-hydroxy-1,4-naphthoquinone (Ϫ145 mV), anthraquinone (Ϫ100 mV), 2,5-dihydroxy-2-benzoquinone (Ϫ60 mV), menadione (0 mV), and 1,4naphthoquinone (ϩ60 mV), each at a concentration of 100 M. Cytochrome b 6 absorbance spectra in the 540 -580 nm range were recorded at ϳ25-mV intervals from samples poised at electrical potentials of Ϫ200 mV to ϩ35 mV. The cytochrome b 6 peak absorbance values at 564 nm were plotted as a function of potential. The midpoint potentials of the b 6 hemes were calculated by fitting these data to a sum of two Nernst equations (each n ϭ 1) as in previous analysis of wild type Synechococccus (38). Alternatively, a global fit analysis was performed as in Ref. 38, where the entire spectra were deconvoluted and fitted to a sum of Nernst curves. Both procedures yielded essentially identical results.

Construction of the PetB R214H Mutant and
Characterization of Growth-As stated, both photosynthetic and respiratory electron transport in cyanobacteria require the cytochrome bf complex, and therefore essential subunits cannot be inactivated (reviewed by Ref. 5). Accordingly, the PetB R214H mutation was constructed by allele replacement (31). This method leaves a copy of the antibiotic resistance cassette upstream of the targeted gene in the Synechococcus genome as illustrated in Fig. 2A. The Sm/Sp cassette does not detectably alter the growth or electron transfer properties of control cells (31). Synechococcus R214H mutant segregants were confirmed as shown in Fig. 2B and detailed under "Materials and Methods." Synechococcus mutant cells carrying the R214H mutation in the presumptive quinone-reductase site were initially characterized for growth properties and possible sensitivity to the cytochrome bc Q i site inhibitor antimycin A. At 39°C, 250 mol photons m Ϫ2 s Ϫ1 , and 1% CO 2 , the R214H mutant grew ϳ2.5fold more slowly than the wild type. Fig. 3 5, 7, 9, and 11). Each pair of lanes (lanes 2 and 3; lanes 4 and 5; lanes 6 and 7; lanes 8 and 9; and lanes 10 and 11) contains one of five different DNA templates. The PCR product at 510 bp in lanes 3 and 11 (marked by arrows) and its absence in lanes 2 and 10 indicate that two of five transformants tested were segregants in which the petB R214H allele had replaced the wild type allele.
FIG. 3. Growth of the Synechococcus R214H mutant. The petB R214H mutant and control strain were grown at 250 mol photons m Ϫ2 s Ϫ1 and 39°C in A medium bubbled with 1% CO 2 in air. E, BstUI control strain; f, the R214H mutant. A linear curve fit was used for the exponential phase of growth. The BstUI strain is wild type for petB (cytochrome b 6 ). cells, respectively. Under slower growth conditions (32°C, 80 mol photons m Ϫ2 s Ϫ1 , and atmospheric CO 2 ), mutant and control cells grew at similar rates. Mutant and control cultures were insensitive to antimycin (30 M) under all conditions (not shown). These data indicate that the R214H mutation did not alter the antimycin insensitivity of the cytochrome bf Q i site but did suggest a partial blockage of electron flow to this site resulting in slower growth under conditions of high electron flux.
Turnover of Cytochrome bf Complexes in Wild Type and R214H Mutant Cells-Plastoquinol oxidation in the cytochrome bf complex operates via a branched pathway, where one electron is transferred to the high potential chain and the second electron is transferred to the low potential chain (Fig.  1). The overall rate of electron flow through the complex can therefore be evaluated by measuring the rate of cytochrome f/c 6 rereduction (the spectra of the two redox partners are largely superposed) after a long flash (31). The rationale for this measurement is that light induces the oxidation of the redox cofactors (the P700 reaction center of photosystem I, cytochromes f and c 6 , and the Rieske ISP 2Fe-2S cluster) that are functionally located downstream of the rate-limiting, plastoquinol oxidation step. When the light is turned off, plastoquinol oxidation drives the reduction of the ISP cluster, cytochromes f and c 6 , and P700. This process is proportional to the rate of plastoquinol oxidation but is also indicative of the overall turnover of the bf complex. If the Q cycle is continuously engaged during cytochrome bf turnover, as proposed (39), then the cytochrome f/c 6 rereduction rate also depends on the Q i site reduction of plastoquinone. The complex has to perform several turnovers to generate all of the reducing equivalents required to reduce the pool of electron carriers in the high potential chain (31,38).
We employed this technique as a first approach to gain information on the turnover efficiency of the cytochrome complex in the R214H mutant cells. Fig. 4A presents the absorbance changes from 540 to 570 nm measured in intact wild type and mutant cells following a 40-ms flash. Here reduction and oxidation processes are shown as increases or decreases in absorbance, respectively. The spectra show that the absorbance change following illumination is made up of contributions from cytochrome f (␣-band maximum 556 nm in Synechococcus sp. PCC 7002, Ref. 38) and cytochrome c 6 (␣-band maximum 553 nm in Synechococcus, Ref. 38) in both strains. Fig. 4B presents a typical experiment where the kinetics of electron flow in cytochrome f/c 6 were estimated in both the wild type and mutants strains. In the absence of inhibitors, the half-times of cytochrome f/c 6 reduction were 18 Ϯ 1 ms in the wild type, in substantial agreement with previous estimates (31), and 46 Ϯ 2 ms in the R214H mutant, indicating a significant decrease in cytochrome bf turnover efficiency. These values did not change upon the addition of antimycin A (data not shown) indicating, as in the growth experiments, that the simple replacement of cytochrome b 6 Arg 214 with histidine did not confer specific antimycin binding and inhibitor sensitivity.
Single Turnover Flash-induced Cytochrome b 6 Redox Changes-Having demonstrated that the overall turnover of the cytochrome bf complex was slowed in the R214H mutant, we tried to gain more direct information on the specific reaction step affected. To this end, we studied flash-induced, singleturnover absorption changes associated with redox events in the b 6 hemes. These reactions are indicative of electron transfer processes that occur both at the Q o (cytochrome b 6 reduction) and Q i (cytochrome b 6 oxidation) sites. The experiments were performed with intact cells, as described previously (38). The results of such measurements are presented in Fig. 5. Fig. 5 (A and B) shows flash-induced absorbance changes  (38). The difference between absorbance at 563 nm and a base line drawn between 545 and 573 nm was subsequently monitored to track the b 6 heme reduction and oxidation kinetics shown in Fig. 5, C and D. In wild type cells (Fig. 5C), illumination resulted in the appearance of a small signal increase (reflecting cytochrome b L reduction), followed by a large signal decrease. The latter is attributable to oxidation of the b H heme at the Q i site. Indeed, no redox signal changes are expected during electron transfer between the b L and b H hemes, because their spectra in vivo are almost identical (Refs. 11, 38, and 40 and references therein). Upon the addition of the Q i site inhibitor, NQNO (29), the cytochrome b 6 oxidation kinetics were slowed considerably, whereas the reduction process remained largely unaffected (compare the initial positive slopes). Consequently, the amplitude of the reduction phase increased at the expense of the oxidation signal, and the rate of electron injection into the low potential (cytochrome b L and b H ) chain could be correctly estimated. From these single-turnover data, we estimated a t1 ⁄2 for cytochrome b 6 reduction in the 2-ms range in agreement with a previous estimation based on the deconvolution of time-resolved spectra (38). Note that because of the bifurcated reaction mechanism at the Q o site, electron transfers to the high (cytochrome f/c 6 ) and low (b 6 heme) potential branches are expected to occur at similar rates as confirmed here (not shown) and demonstrated previously for Synechococcus 7002 (38). The almost 1-order of magnitude difference between our single-flash estimation of Q o site turnover based on b 6 heme reduction and that obtained from cytochrome f/c 6 reduction after long flash Intact BstUI control (WT) and R214H mutant cells were suspended as in A. The flash-induced absorbance changes at 557 minus 540 nm reflect the oxidation and reduction of cytochromes f and c 6 . The spectra were normalized to equivalent maximum absorbance changes. illumination (Fig. 4) can be explained by the occurrence of multiple turnovers in the latter as stated above.
The R214H mutant cells displayed rather different kinetics as shown in Fig. 5D. Here, the traces recorded in the absence and presence of NQNO were largely superposed. In the mutant, the rate of cytochrome b 6 reduction (which represents the oxidation of the plastoquinol pool at the Q o site) was substantially unaffected, but the oxidation of the b 6 hemes was greatly slowed (t1 ⁄2 ϭ ϳ80 ms relative to ϳ8 ms in the wild type). In sharp contrast to the wild type, the addition of the Q i site inhibitor did not further slow b heme oxidation in the R214H mutant. This clearly suggests that the turnover of the Q i site was specifically altered by the mutation, in agreement with our expectations.
Isotopic Effect on Cytochrome b 6 Oxidation Kinetics-In green algae, the substitution of D 2 O for H 2 O influences the kinetics of cytochrome bf electron transfer reactions. In particular, this effect mostly concerns the reactions that follow plastoquinol oxidation at the Q o site (41) and plastoquinone reduction (or cytochrome b 6 oxidation) at the Q i site (37). The interpretation for this isotopic effect is that deprotonation and protonation of the quinones kinetically limits the reactions occurring in both the Q o and Q i sites.
To investigate whether the slow-down of cytochrome b 6 oxidation in the R214H mutant resulted from a decreased efficiency of electron transfer or a modification of the protonation process, we measured b 6 heme kinetics in cyanobacteria following H 2 O-D 2 O exchange according to procedures developed for C. reinhardtii (see "Materials and Methods"). With the wild type, we obtained results closely resembling those previously reported for C. reinhardtii (37). Both the reduction and oxidation rates of the cytochrome b 6 hemes decreased substantially in D 2 O-enriched medium (compare the solid and open squares in Fig. 6A), consistent with the direct involvement of proton transfer events in these reactions. The isotopic effect on cytochrome b 6 oxidation was largely decreased by pretreatment with the Q i site inhibitor NQNO, suggesting that under these conditions, processes other than protonation of the Q i plastoquinone (for example, its binding to the Q i site or its reduction by the b H heme) limited the turnover of the Q i site.
In the R214H mutant, only a very small isotopic effect was observed on the cytochrome b 6 oxidation kinetics, independent of the presence of NQNO (Fig. 6B). This indicates that electron transfer reactions limited quinone reduction in the mutant and further supports the conclusion that the R214H mutation mimics the effect of NQNO.
Evaluation of the Cytochrome b 6 Redox Potentials in the R214H Mutant-Our data suggested that the kinetic consequences of the R214H mutation could be attributed either to a change in the redox potential of the Q i hemes (hemes b H and/or heme c i (x)) or to a decreased affinity of plastoquinone for its Q i -binding site. To determine whether the redox potential of the b H heme was modified in the mutant, we performed an equilibrium redox titration of membranes extracted from the R214H mutant. In the Ϫ200 to ϩ35 mV range, two components were clearly identified that had equivalent absorption spectra characteristic of cytochromes b 6 (Fig. 7). Their redox potentials were indistinguishable within experimental error from those previously determined for the wild type (38). We ruled out, therefore, that a modification of the b H heme redox potential was responsible for the kinetic effects observed in the R214H mutant.
In the spectral region analyzed, no evidence was found for a third redox component. This indicates that the c i (x) heme of the Q i domain was silent in the ␣ band region, consistent with its expected high spin nature (9,10). Because of strong chlorophyll absorbance, we were unable to perform redox titrations of membranes in the Soret region and therefore could not assess the redox properties of the c i (x) heme. DISCUSSION Functional Phenotype of the R214H Mutant-We have shown here that replacement of the conserved Arg 214 with histidine in cytochrome b 6 dramatically impairs Q i site function, turnover of the cytochrome bf complex, and the growth of Synechococccus cells. As mentioned, the cytochrome bf Q i site and transmembrane electron transfer between the b 6 hemes have been enigmatic and difficult to study. Principally (i) there is no highly efficient inhibitor such as antimycin A to block electron transfer at the cytochrome bf Q i site (16), (ii) no stable Q i -semiquinone and accompanying EPR signal have been detected (16), and (iii) the spectral properties of the b H and b L hemes are virtually identical within intact cells (38,40,42). All of these factors have made it exceedingly challenging to track individual redox changes within the low potential chain. As a consequence, there has been much controversy about the role of the Q i site for cytochrome bf turnover and whether the Q cycle mechanism of electron transfer (described above) operates requisitely or facultatively (see Refs. 5, 11, 43, and 44 for reviews). Several interesting alternative mechanisms for electron and proton transfer have been proposed that might operate in bf complexes under particular environmental conditions. These include "semiquinone" (36), "proton pump" (45)(46)(47), and "bypass" (43) models that involve transfer of semiquinones or protons across the membrane or the transfer of two electrons from plastoquinol to two Rieske ISP clusters in a cytochrome bf dimer on the lumenal side of the membrane. Most of these mechanisms still require the operation of a quinone-reductase (Q i ) site, in some form, on the electronegative side of the membrane.
To help elucidate Q i site function in the cytochrome bf complex, we attempted to make the cytochrome bf complex of the cyanobacterium Synechococcus 7002 more like the bc complex. Because this project was initiated before the recent cyanobacterial (9) and algal (10) cytochrome bf structures, the rationale employed to design the mutation was based on homology to the cytochrome b protein of R. capsulatus (33). Cytochrome b 6 Arg 214 of Synechococcus corresponds to the conserved cytochrome b His 217 of R. capsulatus (and to cytochrome b His 202 of yeast shown in Fig. 8, right panel). The H217R mutation of R. capsulatus greatly increased (by ϳ10-fold) the binding affinity of semiquinone for the Q i site and thereby slowed Q i site turnover, increased the reduction level of heme b H , and slowed overall turnover of the bc complex.
Our corresponding R214H mutation in Synechococcus did not generate an antimycin-sensitive cytochrome bf complex, nor did it alter the spectra of the b 6 hemes, but it did markedly alter Q i site catalysis and the overall turnover of the cytochrome bf complex. In the cytochrome bf structures (9, 10), Arg 214 (equivalent to Arg 207 in both Mastigocladus and Chlamydomonas) lies within or close to the Q i pocket and the newly discovered c i (x) heme (Fig. 8, left and middle panels). Therefore, the removal of this positively charged residue might not only alter the affinity of the Q i site for plastoquinone but also the midpoint potentials of the redox cofactors within its vicinity. Our redox titration of the b hemes from the R214H mutant revealed midpoint potentials similar, within experimental error, to those from wild type Synechococcus. This seemed surprising in light of the relatively close proximity of Arg 207 (Synechococcus Arg 214 ) to the b H heme and the likely electronic coupling of the b H and c i (or x) hemes via a heme b H propionate and a water molecule in both the Mastigocladus and Chlamydomonas structures (9, 10). If the Q i site structure is more like that in Fig. 8 (left panel), then the interaction of the c i (x) heme propionates with Arg 207 (Synechococcus Arg 214 ) should modulate the positive charge of the arginine, and there might be less direct influence on heme b H . Similarly, if Arg 207 (Synechococcus Arg 214 ) has a direct role in binding plastoquinone, again as suggested in Fig. 8 (left panel), the effect of the R214H mutation on heme b H might be minimized. For these reasons, we believe that our data favor a direct role for Arg 207 (Synechococcus Arg 214 ) in plastoquinone binding at the Q i site, although we cannot exclude other possibilities. We were unable to directly assess the midpoint potential of the c i (x) heme in the mutant because of the absence of either a distinguishing optical signal, as discussed above, or a well defined EPR signal.
Thanks to its unique position in the structure, Arg 207 (Synechococcus Arg 214 ) may have additional roles in Q i site catalysis. Besides plastoquinone binding and stabilization of a negative charge on heme c i (x) (see below), it may participate in proton transfer to the Q i quinone, either by providing a direct path for protons or indirectly by participating in a network of residues involved in protonation. Such protonation networks are well established in the Q B pocket of the bacterial reaction center (48,49). In the cytochrome bf structures, Arg 207 (Synechococcus Arg 214 ) may be hydrogen-bonded to plastoquinone via a water molecule. As detailed by Kurisu et al. (9) and Stroebel et al. (10), heme c i (x), and the Q i pocket are relatively accessible to the stromal, aqueous phase being bounded mostly by the N-and C-terminal loops of cytochrome b 6 and the Nterminal extension of subunit IV. Thus, Arg 214 could be a link in a short water chain such as that in cytochrome bc Q i (50 -52) or the reaction center Q B sites (49,53).
Our data showed a considerable kinetic isotope effect on both the reduction and oxidation kinetics of the b 6 hemes in wild type Synechococcus cells (Fig. 6). This implies that steps in both plastoquinol deprotonation at the Q o site and plastoquinone protonation at the Q i site affect the overall reactions at these sites as shown previously for chloroplasts (37,54). The R214H mutant showed a small isotope effect on b 6 heme reduction but virtually none on b 6 heme oxidation, indicating that an electron transfer step has been slowed at least as much as any of the protonation steps. These data do not preclude a role for Arg 214 FIG. 7. Redox titration of membranes isolated from the R214H mutant. Titrations were performed in 50 mM MOPS, pH 7.0, 50 mM KCl buffer in the Ϫ200 mV to ϩ35 mV range as described under "Materials and Methods." A shows spectra obtained in the cytochrome b 6 spectral region from membranes poised at the potentials listed below. B shows the peak absorbance values for cytochrome b 6 (at 564 nm) as a function of the applied potential. The midpoint potentials of the two b 6 hemes were calculated either by fitting the data in B to the sum of two n ϭ 1 Nernst equations or by globally fitting the spectra in A to Nernst equations as described under "Materials and Methods." The two procedures gave similar results. The absorbance changes are expressed as ⌬I/I as in Fig. 5. ƒ, Ϫ200 mV; ‚, Ϫ175 mV; E, Ϫ145 mV; Ⅺ, Ϫ115 mV; ࡗ, Ϫ85 mV; , Ϫ55 mV; OE, Ϫ25 mV; •, 5 mV; f, 35 mV.
in Q i plastoquinone protonation in the native cytochrome bf complex.
Based on our findings, the main role of cytochrome b 6 Arg 214 (Arg 207 in the cytochrome bf structures) seems to be in modulating the binding affinity of the Q i site for its plastoquinone substrate. In this sense, the R214H mutation of Synechococcus, which appears to weaken quinone binding, is indeed complementary to H217R of R. capsulatus where arginine at position 217 binds quinone more strongly than the native histidine.
Structural Features of Cytochrome b 6 Arg 214 -Based on their characterization of the mutant phenotype, Gray et al. (33) proposed that His 217 binds quinone at the cytochrome bc Q i site in R. capsulatus via a hydrogen bond to one of the quinone carbonyls analogous to the interaction of a histidine with Q B in the bacterial reaction center (55). Indeed, this predicted role of His 217 has been largely confirmed by the structures of mitochondrial cytochrome bc complexes (17)(18)(19)(20)56). However, it remains controversial whether His 217 (or the equivalent histidine in other bc complexes) directly participates in binding forms of the Q i -quinone or does so via a linking water molecule. Some cytochrome bc structures (18,57) and EPR investigations (52, 57) support the former, whereas others (51,56), where the distance between this histidine and quinone is too great for hydrogen bonding, favor the latter. These structural differences may reflect dynamic, functional aspects of cytochrome bc Q i domains as proposed in recent models of ubiquinone reduction and protonation (51,52,57). Fig. 8 (right panel) illustrates the juxtaposition of His 202 (equivalent to R capsulatus His 217 ) and ubiquinone via hydrogen-bonds to a linking water molecule in the Q i pocket of the yeast cytochrome bc structure (56).
There are clear differences among the three Q i site structures shown in Fig. 8. In the Chlamydomonas structure, Arg 207 (Synechococcus Arg 214 ) lies close to the c i heme (Fig. 8, left  panel). One of the arginine amino nitrogens is within easy hydrogen bonding distance of a c i heme propionate oxygen. The Q i site plastoquinone has not been unambiguously identified in the Chlamydomonas bf structure, but the quinone ring likely resides near the face of heme c i (10). 2 This is represented by a red ring (viewed largely edge-on) in Fig. 8 (left panel). Because the quinone oxygens have not been assigned, likely hydrogen bonding interactions cannot be precisely deduced from the structure. However, depending on the orientation of these oxygens, the Arg 207 (Synechococcus Arg 214 ) amino nitrogens could either be within direct, 3 Å, hydrogen bonding distance of the Q i plastoquinone (as illustrated in Fig. 8, left panel) or be linked via an intervening water molecule as in the His 202ubiquinone linkage in the yeast Q i site (Fig. 8, right panel, and Ref. 50).
In the Mastigocladus structure, the Arg 207 amines face away from heme x (c i ) and the assigned location of the Q i plastoquinone (Fig. 8, middle panel). In this position, direct hydrogen bonding to plastoquinone or heme x is not possible. The shortest distance from the arginine nitrogens to heme x propionates is ϳ8 Å, and the distance to the heme b H propionates is slightly greater. In this conformation, an interaction of Arg 207 via an intervening water molecule or that of a different rotamer of the arginine with heme x might still be possible.
The observed structural differences of the cytochrome bf Q i domains may reflect: (i) evolutionary and functional differences between chloroplast and cyanobacterial cytochrome complexes; (ii) interesting conformational changes that occur during Q i site catalysis, as in the Q B sites of reaction centers (58), the quinol-oxidase (Q o ) domains of bc complexes (12), and those postulated for bc Q i domains (51,52,57); or (iii) artifacts resulting from local differences in resolution and interpretation 2 D. Picot, personal communication. of electron densities. The phenotype observed here is more consistent with the description of the Q i site obtained from Chlamydomonas. Nevertheless, our data cannot rule out other possible interpretations such as a very tight interaction of Arg 214 with the c i (x) heme. Further analysis will be needed to resolve this question.
Electron Flow through Inhibited Q i Sites in Cytochrome bf Complexes-We have shown that the Synechococcus R214H mutation had an effect like that of NQNO, which slows b 6 heme oxidation and partially slows the turnover of the cytochrome bf complex (29). Surprisingly, the addition of NQNO to the R214H mutant did not further slow the b heme oxidation kinetics (Fig.  6B). This could imply that Arg 214 provides a crucial binding site for NQNO and that the inhibitor binds poorly to the mutant complexes. Alternatively, electron transfer from heme b H to the Q i plastoquinone might already be largely (or perhaps completely) blocked either in the R214H mutant or by NQNO at 4 M. The addition of NQNO to the mutant would then have no further impact. Turnover of the cytochrome complex under these conditions (ϳ30% that of the wild type) could only occur if there were a bypass or "electron leak" pathway that allows electrons to be diverted from Q i plastoquinone reduction.
The cytochrome bf structures reveal accessibility of the c i (x) heme to the stromal compartment, raising the possibility of Q i site access to ferredoxin or ferredoxin NAD(P)H oxidoreductase in a cyclic electron pathway around photosystem I (9,10,59). This being the case, the slow electron leak observed here might not reflect inefficient plastoquinone reduction in poorly inhibited Q i sites but perhaps a type of "reverse cyclic" pathway where electrons from heme c i (x) (or a modified lower potential heme c i or x) could reduce ferredoxin NAD(P)H oxidoreductase or ferredoxin. In recent work, Osyczka et al. (15) have discussed how electron transfer in the bc complex might operate in reverse depending on equilibrium midpoint potentials and sizes of electron pools as documented, for example, in the "reverse," uphill electron transfer through the bc complex of Thiobacillus ferrooxidans (60).
Conclusion-Accumulating evidence indicates that under typical conditions in wild type organisms, electron transfer to the cytochrome bf Q i site should proceed from heme b L Ͼ heme b H Ͼ heme c i (x) Ͼ Q i plastoquinone. As mentioned, the cytochrome bf structures present compelling evidence for electronic coupling of hemes b L and c i (x). It seems unlikely that the c i (x) heme placement evolved for uniquely structural rather than functional reasons. We suggest that Synechococcus cytochrome b 6 Arg 214 (Mastigocladus or Chlamydomonas Arg 207 ) plays a central role in Q i plastoquinone binding. Moreover, data presented here demonstrate for the first time, that modification of the Q i site of the cytochrome bf complex dramatically alters overall photosynthetic efficiency. This finding supports previous indications that the quinone-reductase reaction is continuously engaged during steady state photosynthesis (39), thus modulating H ϩ /e Ϫ stoichiometries and efficient photosynthetic CO 2 assimilation.