Biophysical, Biochemical, and Physiological Characterization ofChlamydomonas reinhardtii Mutants with Amino Acid Substitutions at the Ala251 Residue in the D1 Protein That Result in Varying Levels of Photosynthetic Competence*

The QBbinding site of the D1 reaction center protein, located within a stromal loop between transmembrane helices IV and V formed by residues Ile219 to Leu272, is essential for photosynthetic electron transport through photosystem II (PSII). We have examined the function of the highly conserved Ala251D1 residue in this domain in chloroplast transformants ofChlamydomonas reinhardtii and found that Arg, Asp, Gln, Glu, and His substitutions are nonphotosynthetic, whereas Cys, Ser, Pro, Gly, Ile, Val, and Leu substitutions show various alterations in D1 turnover, photosynthesis, and photoautotrophic growth. The latter mutations reduce the rate of QA to QB electron transfer, but this is not necessarily rate-limiting for photoautotrophic growth. The Cys mutant divides and evolves O2 at wild type rates, although it has slightly higher rates of D1 synthesis and turnover and reduced electron transfer between QA and QB. O2 evolution, D1 synthesis, and accumulation in the Ser, Pro, and Gly mutants in high light is reduced, but photoautotrophic growth rate is not affected. In contrast, the Ile, Val, and Leu mutants are impaired in photoautotrophic growth and photosynthesis in both low and high light and have elevated rates of D1 synthesis and degradation, but D1 accumulation is normal. While rates of synthesis/degradation of the D1 protein are not necessarily correlated with alterations in specific parameters of PSII function in these mutants, bulkiness of the substituted amino acids is highly correlated with the dissociation constant for QB in the seven mutants examined. These observations imply that the Ala251 residue plays a key role in D1 protein.

Photosystem II (PSII) 1 complexes of all oxygen-evolving organisms convert light energy into chemical free energy by transferring electrons from water to plastoquinone, associated with the release of molecular oxygen (1,2). The D1 protein of the PSII reaction center turns over in a light-dependent manner more rapidly than any other chloroplast protein and is a primary target for photoinhibitory damage (3,4). Based on homology with the crystal structure of the L protein from Rhodopseudomonas viridis (5), D1 is predicted to have five hydrophobic membrane-spanning helices (6). One of the stromally exposed regions of D1, extending from the C terminus of transmembrane helix IV through the N terminus of transmembrane helix V (IV-V loop), includes a stromal helix thought to lie parallel to the membrane surface (6). This loop, whose amino acid sequence is highly conserved among cyanobacteria, algae, and higher plants (7), is involved in binding both Q B , the second stable quinone acceptor in PSII, and several classes of herbicides that inhibit photosynthetic electron transport at the Q B docking site (8). Only a few of the many amino acid substitutions made in this region result in the loss of D1 function and photosynthetic capability (9,10), suggesting that most positions can tolerate considerable variation in R group conformation or charge and still permit D1 function. However, particular residues may be necessary for optimal PSII activity or may provide functional advantages under certain environmental conditions (11,12). Analysis of site-directed mutations in the region between Ile 248 and Ala 251 at one end of the parallel helix has led to the hypothesis that these residues may be buried in the thylakoid membrane (11), thus dividing the IV-V loop of D1 into two segments. One segment is postulated to be involved in Q B binding and herbicide resistance (from the parallel helix to the helix V) and the other possibly in D1 degradation (from helix IV to the parallel helix).
Substitution of Val for Ala 251 in the D1 protein in cyanobacteria and Chlamydomonas results in herbicide resistance and increased sensitivity to photoinhibition (13,14), reduced photosynthetic yield associated with slower photoautotrophic growth (11,15), and perturbation of the pattern of oxygen evolution in single turnover flashes (16). To examine further the role of the D1 Ala 251 residue in PSII function and light sensitivity, we generated 12 of the 19 possible amino acid substitutions at this position in Chlamydomonas reinhardtii.
Five of these substitutions (Arg, Asp, Gln, Glu, and His) were shown to lead to a nonphotosynthetic phenotype and to impaired D1 synthesis and accumulation (10). Seven (Cys, Gly, Ile, Leu, Pro, Ser, and Val), which retain photosynthetic function to various degrees, affect herbicide resistance and photoautotrophic growth rates to various degrees (17). In this paper, we examine D1 synthesis and turnover, photosynthetic rate, quantum efficiency, and electron transfer between Q A and Q B in these seven mutants. We also show that the bulkiness of the R group of the amino acid substituted in each of 12 Ala 251 mutants is correlated with the dissociation constant (K O ) for plastoquinone in the Q B pocket, consistent with the hypothesis that the Ala 251 residue of D1 is structurally involved in Q B binding and is likely to be buried in the thylakoid membrane (11).

MATERIALS AND METHODS
Strains and Culture Conditions-Wild type and mutant strains of C. reinhardtii designated "CC-" are available from the Chlamydomonas Genetics Center at Duke University (Dr. E. H. Harris, Box 91000, Duke University, Durham, NC 27708). Stocks were maintained on plates of solid TAP or HSHAYA media (18) under dim light (10 mol m Ϫ2 s Ϫ1 ). For physiological, biochemical, and biophysical analysis, cells were grown photoautotrophically in liquid minimal HS medium (18) bubbled with 5% CO 2 in air under low light (LL; 70 mol m Ϫ2 s Ϫ1 ) or high light (HL; 600 mol m Ϫ2 s Ϫ1 ) at 25°C. Cells were harvested in early exponential to midexponential growth (up to 2 ϫ 10 6 cells ml Ϫ1 ), but sufficient doublings were allowed for full adaptation to each environmental condition. In HL, these conditions are saturating for maximal rates of photoautotrophic growth of the wild type strain CC-125 (19).
Recovery of Same Site Suppressors-Photosynthetically competent same site suppressors arise spontaneously when cells of the nonphotosynthetic A251R* transformant (10) are kept under conditions of acetate starvation on plates in medium light (300 mol m Ϫ2 s Ϫ1 ) or HL following 2-3 weeks of growth on TAP plates (18). Under these conditions, the lawn of A251R* cells bleaches, and the suppressors appear as isolated green colonies. DNA from each of the photoautotrophically competent suppressors was extracted, and the region of the chloroplast psbA gene encoding the D1 loop between helices IV and V was amplified by polymerase chain reaction and sequenced as described previously (10). Four new amino acid substitutions at Ala 251 were recovered: A251C* (CC-3390, TGT), A251S* (CC-3392, AGT), A251P* (CC-3391, CCT), and A251L* (CC-3389, CTT) in addition to A251G*, which was already obtained by site-directed mutagenesis.
In Vivo Pulse-chase Labeling of Chloroplast Proteins with [ 35 S]Sulfate-Pulse labeling of HL-grown cells was carried out as described in Ref. 19 with the following modifications. After a 15-min incubation with anisomycin (Pfizer) to inhibit cytoplasmic protein synthesis (final concentration, 140 g/ml), 30 l of carrier-free H 2 35 SO 4 (NEN Life Science Products) was added to a final concentration of 125 Ci/ml. Aliquots (0.4 ml) of algal suspension were removed at 5,8,11,14, and 17 min and injected into 6 ml of ice-cold acetone. The chase was initiated 5 min after the last time point in the pulse experiment by adding 0.995 l of HS medium containing 200 mM Na 2 SO 4 , 50 mM NaHCO 3 , and lincomycin (500 g/ml final concentration). After 5 min, the first aliquot (0.4 ml) was removed as the 0-min point of the chase. Subsequent samples were taken at 30, 60, 90, and 120 min. Protein solubilization, gel electrophoresis, and quantification of the labeled bands of interest, D1 and large subunit of ribulose-bisphosphate carboxylase oxygenase (LSU), were done as described in Ref. 10.
Photosynthetic O 2 Evolution, Photochemical Quenching, and Determination of Whole Chain Electron Transport-Photosynthetic O 2 evolution and chlorophyll a fluorescence were measured at 25°C (19,21) using a Clark-type O 2 electrode system (Hansatech Instruments, Sheffield, UK) and a pulse-amplitude modulation fluorometer (Heinz-Waltz PAM 101). Cells in early exponential growth (Ͻ0.1 A 750 ) were resuspended at 0.165 A 750 for HL-grown cells or 0.175 A 750 for LL-grown cells and dark-adapted for 15 min. Initial Fv/Fm values were determined immediately following dark adaptation (21), and quenching of Fm during the 1-s 100-kHz flash did not appear to be a problem as previously reported for the PAM method (22). Light response curves of photosynthetic O 2 evolution and chlorophyll fluorescence were generated by increasing incident irradiance stepwise from 10 to 900 mol m Ϫ2 s Ϫ1 . Chlorophyll content was determined according to Ref. 23. Photosynthetic data, normalized for chlorophyll or biomass and plotted against incident irradiance, were fitted by Marquardt nonlinear regression to the equation of Smith/Talling (24) to determine net light-saturated O 2 evolution (P max ). Relative apparent quantum yields (Ø rel ) were estimated by linear regression from the initial slopes of the O 2 evolution curves plotted against absorbed irradiance (21). Maximum quantum yields for O 2 evolution (Ø max ) were also calculated from the initial slopes of photosynthesis-irradiance curves, determined by nonlinear regression analysis, and the chlorophyll-specific optical absorption cross-sections according to Ref. 25. Photochemical quenching was calculated as described by Schreiber et al. (26). Values reported for P max , Ø rel , PAM Fv/Fm, and chlorophyll are means Ϯ S.E. of at least three determinations done on independently grown cultures. Values derived from fast repetition rate fluorometry and thermoluminescence measurements (see below) represent the means of two independent determinations.
Times for whole chain electron transport ( WC ) were deduced from the photosynthetic irradiance curves and knowledge of the functional absorption cross-section of PSII ( PSII(i) ) estimated by fast repetition rate fluorometry (see below). The intercept of the initial slope of the photosynthesis-irradiance curve (␣) and the maximum photosynthetic rate (P max ) can be defined as I K ϭ P max /␣ ϭ 1/( PSII(i) WC ). We calculated I K values for each mutant, and we derived WC based on measured values of PSII(i) . WC corresponds to the maximal time for transfer of an electron from water to a terminal electron acceptor (e.g. CO 2 ) at light saturation in the steady state.
Fast Repetition Rate Fluorometry (FRRF) Measurements-Chlorophyll a variable fluorescence was measured using a FRR fluorometer (22,27) that provides complex information about the properties of PSII available from chlorophyll fluorescence, i.e. photochemical yield, apparent PSII cross-sections, connectivity between PSII units, and kinetics of electron transfer from Q A Ϫ to Q B . Aliquots of cells from algal cultures grown under LL and HL conditions, at 1 g of total chlorophyll/ml, were dark-adapted for 2 min in a cylindrical cuvette of 5-mm inner diameter and illuminated by blue light-emitting diodes (450 nm), which were pulsed in flashlets of 1.5-s duration. Flashlets can be repeated at a variable rate, thus effectively providing either single-turnover excitation (100 flashes given at 0.8 MHz) or the multiple turnovers of PSII that are necessary to reduce fully the pool of secondary plastoquinone acceptors (e.g. 3000 flashlets given at 10 kHz) (22).
Thermoluminescence Measurements-Thermoluminescence (TL) and delayed luminescence (DL) were measured using a computerized laboratory-built apparatus (28). Five ml of algal suspension containing 5 g of chlorophyll were gently filtered on Millipore HA 25 filters (25-mm diameter, 0.8-m pore size) and placed on a temperature-regulated sample holder. Care was taken to keep the cells always covered by a layer of medium. Samples were placed in the dark for 180 s at 25°C. For TL measurements, the sample holder was cooled to 3°C, and, after temperature equilibration (60 s), the sample was excited by xenon flashes given at 1 Hz. The emitted luminescence signal was collected as a function of temperature during linear heating from 3 to 70°C (at a rate of 0.5°C/s), and the data were stored for further analysis. DL was measured at 25°C, from 0.2 to 100 s after the end of excitation by xenon flashes. During the measurement, the temperature of the sample was kept constant within 0.4°C.
Thermodynamics and Kinetics of the PSII Acceptor Side-The kinetic and thermodynamic characteristics of the acceptor side of PSII were determined using the two-electron gate model (15). The dissociation constant for plastoquinone in the Q B pocket, K O , was determined ex-perimentally as the ratio of amplitudes of the first two monoexponential components (one ϳ300 s and two ϳ2 ms in wild type) of Q A Ϫ reoxidation in dark-adapted samples, following a single turnover (ST) flash. The semiquinone equilibrium constant K E was determined from fluorescence decay kinetics by calculating the fraction of reduced Q A Ϫ 25 ms following the ST flash (29). Thus, Ϫ ] 25 ms), and K app was calculated as K app ϭ K E /(1 ϩ K O ). Alternatively, K app was determined experimentally from analysis of the decay of DL in the 0.1-100-s time interval as the ratio of the number of centers recombining from Q B Ϫ (slow component, 1 ϭ ϳ6 s) to the number of centers recombining from Q A Ϫ (fast component, 1 ϭ ϳ0.7 s) or as a ratio of the half-times of these two decay components (15). Although each of these methods provided slightly different absolute values of K app (K app was highest when determined from [Q A Ϫ ] 25 ms), the relative changes of K app between wild type and each of the mutant strains were independent of the method used. The remaining kinetic parameters of the two-electron gate model (k on , k off , k BA , k AB ) were calculated by solving the set of equations in Ref. 15.
Estimation of the Number of PSII Centers in the Thylakoid-The numbers of PSII reaction centers were determined from O 2 flash yields (30,31). In wild type (CC-125), 2 chlorophyll a/O 2 averages 2014 mol/mol for cells grown at 70 mol quanta m Ϫ2 s Ϫ1 . Since each O 2 is generated from four sequential electron transfers through PSII, the numbers of PSII reaction centers can be calculated as 1 per 503 chlorophyll a (25). We extended this approach to all mutants by calculating the functional absorption cross-section of PSII (from both the FRR fluorescence and O 2 evolution saturation profiles).
Estimation of Carbon Assimilation-Cell carbon and nitrogen were determined by filtering replicate aliquots of known cell densities on precombusted Gelman A/E glass fiber filters. The filters were rinsed with distilled water, dried, and combusted in a Perkin-Elmer 241 Elemental Analyzer. Carbon-specific growth rates were calculated as described by Falkowski et al. (32).
Estimation of Amino Acid Bulkiness-In the absence of a robust, high resolution structural model for D1, molecular dynamic calculations of the predicted effects of amino acid substitutions at position 251 on Q B binding are highly speculative. For this reason, we calculated the bulkiness of the substituted amino acids at position 251 as the volume of space enclosed by their van der Waals surface (33). These calculations do not account for changes in binding properties associated with side chain charge distributions or tertiary structural interactions between amino acid 251 and other residues in the binding domain.

Isolation of Site-directed D1 Mutants and Recovery of Same
Site Suppressors with Varying Levels of Photosynthetic Competence-Three site-directed D1 mutants resulting from amino acid substitutions at residue Ala 251 (A251G*, A251I*, and A251V*) retain some capacity for photoautotrophic growth. Four additional mutants at this position (A251C*, A251S*, A251P*, and A251L*) plus A251G*, were recovered from the nonphotosynthetic A251R* transformant (10) as spontaneous same site suppressors capable of photoautotrophic growth. Each of the codon changes in these suppressors required only a single base pair alteration, while a true Ala 251 revertant of the A251R* strain, which was not obtained, would require two simultaneous base pair changes, a highly improbable event ( Fig. 1). All 40 of the photoautotrophically competent suppressors isolated from A251R* showed single base pair substitutions in the Arg 251 D1 codon (8 Cys,9 Ser,4 Pro,15 Gly, 4 Leu) when the DNA fragment encoding the IV-V loop was sequenced. Hence, restored photoautotrophic capacity of the A251R* isolates probably cannot arise from intragenic suppressor mutations affecting residues other than amino acid 251 in the D1 protein or from extragenic suppressors.
We also recovered spontaneous same site suppressors from two of the four other nonphotosynthetic mutants (A251Q* (CAA) and A251H* (CAC)). The other two nonphotosynthetic mutants (A251D* (GAC) and A251E* (GAA)) would have yielded wild type Ala revertants (GCC or GCA). Photoautotrophically competent strains with Pro (CCA, CCC) and Leu (CTA, CTC) substitutions were obtained in a 12:1 ratio from the A251Q* and A251H* mutants. The fact that we did not recover a Lys (AAA) revertant from A251Q* or Asn (AAC) and Tyr (TAC) from A251H*, suggests that the Asn, Lys, and Tyr alterations at D1 residue 251 may lead to a nonphotosynthetic phenotype. Although we did not verify this prediction by introducing the Asn, Lys, and Tyr mutations in vitro, the His substitution, shown by site-directed mutagenesis to result in a nonphotosynthetic phenotype (10), was not recovered from the A251R* transformant among the isolates capable of photoautotrophic growth (Fig. 1).
Levels of D1 and LHCP Proteins-Levels of the D1 and LHCP proteins of PSII have been examined in LL-and HLgrown cells by immunoblotting (Fig. 2, A and B), with the tubulin ␤-subunit used as a control for equal loading. Variations seen in the proportions of the 32-and 29-kDa D1 conformers might result from the specific amino acid substitutions introduced. Therefore, both bands were summed to give the total amount of D1 present, and these values were normalized to the amount of ␤-tubulin protein. When grown under LL, all mutants, with the exception of A251C*, showed a slight reduction (10 -20%) in D1 level, expressed as the sum of the 32-and 29-kDa conformers, compared with wild type (Fig. 2B). In contrast, the D1 level under HL growth varied considerably between mutants. Thus, although A251C* and A251I* did not differ markedly from wild type, A251G*, A251S*, and A251P* showed 16 -28% reduction from the wild type levels, and A251L* and A251V* appeared to contain elevated amounts of D1 (Fig. 2B). The 24 -25-kDa D1 polypeptides, thought to represent N-terminal aborted translation products in the A251Q*, A251E*, A251H*, and A251D* D1 mutants with nonphotosynthetic phenotypes (10), were absent in all seven D1 mutants capable of photoautotrophic growth. The level of D1 in wild type, A251C*, A251G*, A251S*, and A251P* under HL was 57-59% of that in the same genotypes under LL. In contrast, under HL A251I*, A251V*, and A251L* contained 68 -80% of the LL amount of D1.
HL-grown cells of wild type, A251C*, A251I*, A251L*, A251P*, and A251V* contain 29 -35% less LHCP compared with cells of the same genotypes grown under LL (Fig. 2B). This reduction in LHCP in the A251A* wild type and in A251C* 2 B. Förster, unpublished observations. The Ser, Gly, Cys, Pro, and Leu mutations predicted were all recovered as photoautotrophically competent suppressors. In contrast, the Arg and His mutants, which were shown previously to be nonphotoautotrophic when created by site-directed mutagenesis (10), were not recovered among the suppressors.
correlates with a reduction in total chlorophyll. In contrast, A251G* and A251S* appear to be strongly impaired in their ability to down-regulate levels of both LHCP and total chlorophyll when grown in HL. The relationship between downregulation of chlorophyll and LHCP in the A251I*, A251L*, A251P*, and A251V* mutants is less clear. Down-regulation of chlorophyll content occurs rapidly when LL-grown wild type cells of C. reinhardtii are shifted to HL (20), and down-regulation of LHCP in the related alga Dunaliella shifted from LL to HL depends on the redox state of plastoquinone (34).
Synthesis and Degradation of the D1 Protein-Wild type and the four mutants tested (A251C*, A251G*, A251L*, and A251P*) synthesized D1 at a substantially higher rate than LSU under HL during the 17-min labeling period and degraded D1 faster than LSU during the 2-h chase. Synthesis of D1 in A251C* and A251L* occurred at rates equal to or greater than wild type (Fig. 3), whereas D1 labeling in A251G* and A251P* was reduced by about one-third compared with wild type (Fig.  3 legend; 1.9 and 2.0 ϫ 10 5 versus 3.0 ϫ 10 5 ). Degradation of the newly synthesized D1 during the 2-h chase was elevated in A251G*, A251L*, and A251P* compared with wild type and A251C*. In the case of A251L*, the high D1 degradation rate was correlated with an increased D1 synthesis rate. A251P* had the lowest rate of D1 synthesis and the highest rate of D1 degradation of all of the mutants examined. Rates of LSU degradation were comparable with wild type in A251G* and A251P* (Fig. 3 legend) and lower than wild type in A251C* and A251L*. No labeling of 24 -25-kDa bands typical of the aborted D1 translation products found in the nonphotosynthetic D1 mutants (10) was observed.
Photosynthetic O 2 Evolution and Whole Chain Electron Transport-The relative apparent quantum yield (Ø rel ), maximum quantum yield (Ø max ), and maximum rates of light-saturated photosynthetic O 2 evolution (P max ) of the aforementioned strains are compared in Table I. Mutants A251C*, A251G*, A251P*, and A251S* grown under LL have P max values somewhat higher than wild type, whereas P max values for A251I*, A251L*, and A251V* are lower than wild type. All mutants grown in HL except A251C* have lower Ø rel and Ø max values than wild type, with A251I*, A251V*, and A251L* lowest of all. With the exception of the A251C* mutant, which has rates of O 2 evolution equal to or higher than A251A*, all of the other photoautotrophic mutants appear to be impaired when grown under HL (Table I) and failed to acclimate to HL on a chlorophyll basis. Whereas wild type and A251C* show a 50% reduction in chlorophyll per unit biomass, and higher chlorophyll a/b ratios in HL, other mutants retain chlorophyll per unit biomass ratios and chlorophyll a/b ratios similar to those in LL.
Photoautotrophic Growth and Carbon Accumulation-LL and HL photoautotrophic growth rates for the seven Ala 251 mutants with varying levels of photosynthetic competence have been described elsewhere (17). Substitution of Cys, Gly, Pro, or Ser for Ala does not affect the photoautotrophic growth rates under LL or HL, whereas Ile, Val, and Leu substitutions lead to reduced growth rates under both conditions (17). This slower growth is not correlated either with decreases in whole chain ( WC ) or PSII ( PSII ) electron transfer (Table II) or with the total carbon per cell and the quantum requirement per carbon fixed (Table I). Therefore, the rates of these processes per se in the mutants do not appear to be limiting growth.
Efficiency of Primary Photochemical Charge Separation-The photochemical competence of each mutant grown under LL and HL was assessed as variable fluorescence measured as the single turnover of PSII using FRRF and by saturating, multiple turnover flashes with the PAM system. In general, the two Fv/Fm values obtained for each genotype grown under either LL or HL were in reasonable agreement (Table II), suggesting that potential quenching problems inherent to the PAM method (22) did not occur during the initial 1-s flash used for determination of dark-adapted Fm. The PAM Fv/Fm values for the wild type control A251A* were comparable with previously published values for wild type Chlamydomonas (19, 21). Mu-FIG. 2. Accumulation of the D1 and LHCP proteins in the A251A* wild type and the seven D1 mutants at residue 251 capable of photoautotrophic growth. Cells were grown in liquid minimal medium in LL (70 mol m Ϫ2 s Ϫ1 ) or HL (600 mol m Ϫ2 s Ϫ1 ), total cell protein was isolated, cells were solubilized with LDS, and equal total protein amounts were resolved on 10 -17.5% LDS-polyacrylamide gel electrophoresis. A, total cell proteins electroblotted to nitrocellulose membranes, probed with antibodies raised against D1 or LHCP, the ␤-subunit of tubulin, and the OEE1 oxygen-evolving protein of PSII were detected by enhanced chemiluminescence. The ␤-tubulin antibodies were used to verify equal loading. Under these conditions, D1 migrated as a doublet consisting of the mature 32-kDa protein and a 29-kDa band, which is probably a conformer of the intact 32-kDa protein (45). As previously observed (10), migration of the D1 protein on LDS gels can vary according to the amino acid substituted for Ala at position 251. The Gly, Ser, and Pro substitutions appear to reduce the electrophoretic mobility of D1 slightly, whereas the other substitutions have no visible effects on D1 migration. Three LHCP isoforms were recognized by the LHCP antibody. B, relative amounts of D1 and LHCP proteins quantified on multiple exposures of x-ray film with Image Scion software. Protein accumulation in each of the mutants was measured as pixel density on x-ray film summed for the two D1 bands or the three LHCP bands, normalized to accumulation of ␤-tubulin in each genotype, and expressed relative to the respective D1 and LHCP bands for wild type. D1 or LHCP accumulation in HL as a percentage of the values in LL is also given for each genotype. tants A251G*, A251P*, and A251S* showed a lower quantum yield of photochemistry (Ø max ϭ Fv/Fm) than wild type and the other four mutants when grown under LL or HL and measured with either technique (Table II). Fv/Fm values obtained for the mutants A251G*, A251P*, and A251S* grown under HL approach those measured for the nonphotosynthetic Ala 251 mutants that range from 0.13 to 0.46 (10).
The reduction in photochemical efficiency in HL suggests a loss of functional PSII reaction centers. Indeed, calculations show that A251G*, A251P*, and A251S* have ϳ2.5-fold fewer functional PSII reaction centers than wild type under HL, while A251I* and A251L* have 1.3-1.7-fold fewer than wild type (Table II). The lower concentration of PSII centers in HL-grown cells corresponds with the reduced probability of FIG. 3. Synthesis and degradation of the D1 and LSU proteins in the A251A* wild type and representative D1 mutants at residue 251 capable of photoautotrophic growth. Log-phase cells of wild type (A251A*) and the A251C*, A251G*, A251P*, and A251L* mutants grown in reduced sulfate medium under 600 mol m Ϫ2 s Ϫ1 light (HL) were preincubated in 140 g/ml anisomycin to block cytoplasmic protein translation, and 5-ml aliquots were pulse-labeled over a 17-min time course with 125 Ci of H 2 35 SO 4 /ml under HL. The algal suspensions were then chased over a 120-min time course with cold sulfate and 500 g/ml lincomycin to inhibit further chloroplast protein synthesis. Labeled proteins were solubilized in LDS, separated by 10 -17.5% gradient LDS-polyacrylamide gel electrophoresis, and visualized by autoradiography. Rates of D1 and LSU synthesis and degradation shown as pixels per unit of time were measured directly from the gels using a Molecular Dynamics PhosphorImager and Image-Quant software. Synthesis was linear over the 17-min labeling period, whereas degradation fit a first order exponential decay function. Curves for D1 and LSU synthesis and degradation in the A251C* and A251L* mutants and the A251A* wild type are shown. Rates of synthesis/ degradation in A251G* were as follows: D1 ϭ 2.0 ϫ 10 5 /Ϫ17.2 ϫ 10 Ϫ3 ; LSU ϭ 0.8 ϫ 10 5 /Ϫ2.9 ϫ 10 Ϫ3 . In A251P* they were as follows: D1 ϭ 1.9 ϫ 10 5 /Ϫ23.4 ϫ 10 Ϫ3 ; LSU ϭ 0.9 ϫ 10 5 /Ϫ3.1 ϫ 10 Ϫ3 .  Table II). With the exception of the A251C* mutant, which behaves like wild type, the connectivity is reduced 1.6 -1.7-fold in A251S*, A251I*, and A251V* and is virtually absent in A251P*, A251G*, and A251L*. Furthermore, a decrease in the effective absorption cross-section of PSII (functional size of PSII antenna) of 13% can be seen in A251C*, A251I*, A251L*, and A251V* ( PSII(i) , Table II). In contrast, PSII(i) in A251G* and A251P* is slightly elevated and is unaffected in A251S*.
Photochemical Fluorescence Quenching-Photochemical fluorescence quenching can be related to the level of reduction of Q A , the first stable acceptor of PSII, at a given irradiance. The irradiance dependence of qP derived during the photosynthetic light response curves of selected mutants is shown in Fig. 4. As observed for photosynthetic O 2 evolution, A251C* is the only mutant that has qP values similar to wild type when grown under both LL and HL. In contrast, A251G* (and A251P*, A251S*) and especially A251L* (as well as A251I* and A251V*) exhibit a dramatic reduction in qP, particularly when measured at low incident irradiances. The low values of qP in darkadapted mutants other than A251C* and the A251A* wild type presumably reflect the higher proportion of nonfunctional (photochemically incompetent) PSII centers in these strains. The coefficient 1 Ϫ qP is directly proportional to the average reduction state of Q A . At steady-state irradiance, qP is related to the cross-section of PSII and the rate of oxidation of Q A Ϫ by the equation (1 Ϫ qP)(I) ϭ ( PSII ϫ I)/( PSII ϫ I ϩ 1/ PSII ) for any given light intensity (I). From Fig. 4 we can deduce that A251A* wild type and A251C* show much lower reduction states of Q A than any of the other mutants at both growth irradiances. The more reduced Q A pools in these genotypes might potentiate photoinhibitory damage and lead to loss of functional PSII reaction centers.

Ability of the Reaction Center to Transfer Electrons from Q A to Q B and Stability of Q B
Ϫ in the D1 Binding Pocket-The effect of each amino acid substitution on the rate of electron transfer from Q A to the plastoquinone pool and the stability of Q B Ϫ in the quinone binding pocket of the D1 protein can be assessed from a combination of variable fluorescence and luminescence measurements. The ratio of the maximum fluorescence yield induced by a multiple turnover flash (MTF) to that observed following a single turnover flash (STF) is indicative of the rate of reduction of secondary plastoquinone acceptors in PSII (i.e. Q B and the mobile plastoquinone pool) by Q A (Table III). MTF/ STF ratios ranged from 1.45 in the A251A* wild type control to 1.10 in A251L*, which appears to be most impaired in transferring electrons from Q A to Q B . Among the photosynthetically competent mutants, the correlations between MTF/STF ratios and the percentage of PSII centers containing reduced Q A Ϫ measured 25 ms following a saturating flash (Table III) 4. Light response of photochemical fluorescence quenching (qP) of the A251A* wild type and three representative D1 mutants at residue 251 capable of photoautotrophic growth. Cells were grown in liquid minimal medium in LL (70 mol m Ϫ2 s Ϫ1 ) or HL (600 mol m Ϫ2 s Ϫ1 ). Curves for A251V* and A251I* were very similar to that for A251L*, whereas the curves for A251S* and A251P* were similar to that for A251G*. impaired mutants (A251I*, A251V*, and A251L*) show values of 24 -42%. A251V*, which is strongly reduced in P max like A251I* and A251L*, shows a smaller reduction in MTF/STF and [Q A Ϫ ] 25 ms than the other two mutants. The proportion of the PSII reaction centers with Q A Ϫ in the nonphotosynthetic mutants (10) is especially high (89 -96%), which suggests that almost no electrons are transferred to Q B in these mutants at physiologically significant rates.
Fluorescence decay measurements by FRRF estimate the rate of electron transfer to Q B . All mutants have slower electron transfer at PSII ( PSII ) than A251A*, and the most strongly impaired mutant is A251L* (Table II). Times for whole chain electron transport ( WC ) can be deduced for wild type and each of the photoautotrophic mutants from their O 2 evolution curves combined with fluorescence properties measured by FRRF and their functional absorption cross-sections of PSII (25). With the exception of the A251C*, A251I*, A251L*, and A251V* mutants grown under HL, the WC values for all mutants grown under LL or HL are lower than wild type (Table  II). In all genotypes except A251L* grown under LL and HL, WC is much greater than PSII , indicating that electron transfer between Q A and Q B does not limit whole chain electron transport.
TL analysis reveals that mutants with slower electron transfer from Q A to Q B show a marked decrease in the temperature of the TL B band (Table III)

, indicative of increased instability of Q B
Ϫ in the Q B niche. The data fall into two groups with the shift in the TL B band temperature from 30°C in the A251A* wild type control to ϳ13°C in A251G* and other mutants that are relatively unimpaired photosynthetically without much change in PSII . In contrast, the TL B band remains at 9 -10°C while PSII ranges from 0.65 to 3.5 ms in A251I*, A251V*, and A251L*. This change of the TL B band corresponds to a decreased change in activation energy for charge recombination between electron equivalents on the acceptor side and the S 2 /S 3 states on the donor side PSII reaction center. The increased propensity for a back-reaction in the mutants is also revealed by the marked reduction in the half-time for luminescence (DL t1 ⁄2 ) as shown in Table III.

Apparent Equilibrium Constant Kinetics and Thermodynamics of Electron Transfer between Q A
Ϫ and Q B -With the exception of A251P* and A251S*, all other mutants show significantly reduced apparent redox equilibrium constants for the electron resonance between Q A and Q B (K app ) as calculated according to the "two-electron gate" model (15). Lower values of K app are indicative of lower overall rates of reduction of Q B by Q A Ϫ (Tables II and III). Moreover, the relationship between qP at LL (Fig. 4) for each of the mutants reflects the overall equilibrium rate constant K app (Table III). In the nonphotosynthetic mutants (10), K app approaches 0, confirming that electron transfer to secondary quinones in these strains is blocked at the primary quinone acceptor Q A .
The accessibility of the Q B pocket and/or the energy of bonds formed to ligate Q B were estimated by the calculation of the kinetic parameters k on and k off , which reflect the rates of association and dissociation of Q B with its binding site on the D1 protein, respectively (Table III). The wild type A251A* has k on and k off values of 0.48 and 0.23, respectively, indicating stable and efficient binding of Q B (15). Although A251S* and A251G* have k off reduced relative to k on as in wild type, suggesting that Q B binding in these strains is stable, their lower k on values indicate a reduced efficiency of Q B binding to its D1 site compared with wild type. Values of k off for the other five mutants capable of photoautotrophic growth are at least as high as their k on values, again suggesting a high instability of Q B in its binding pocket. Furthermore, the k on values indicate that the accessibility of Q B to its site is significantly reduced in all mutants except A251C*. No binding of Q B to its site on the D1 subunit could be detected in the five nonphotosynthetic mutants (10).
Retardation of Q A Ϫ reoxidation is correlated with the calculated bulkiness of the substituted amino acids, based on their increased side chain volume length. For example, when the plastoquinone dissociation constant (K O ) was plotted against residue 251 bulkiness, a linear regression coefficient of 0.91 was obtained (Fig. 5), consistent with this residue being directly involved in plastoquinone binding. Indeed, residue 251 is predicted to be a member of the minimum set of amino acids forming the Q B pocket of D1 (35). Data in Table III show that the K O dissociation constants for wild type and the seven mutants with varying photosynthetic capacity do not differ by more than a factor of 5. The accuracy of current modeling procedures does not allow a more quantitative interpretation of the effect of amino acid bulkiness without a detailed molecular structure of the protein complex.

TABLE III
Biophysical characterization of the A251A* wild type and the Ala 251 mutants using FRRF, DL, and TL MTF/STF represents the reduction level of the plastoquinone pool of LL-grown cells. HL-grown cells were not analyzed in detail because preliminary biophysical characterization showed no significant differences from LL-grown cultures (data not shown). Q A Ϫ is the proportion of RCII with Q A Ϫ in the dark in which Q A Ϫ becomes completely oxidized under ideal conditions. The TL B band peaks at the temperature (activation energy) required for recombination of S 2/3 states (donor side of PSII) and Q B (acceptor side of PSII) after trapping the sample in the state of charge separation at PSII. K app is the apparent equilibrium constant, indicative of the distribution of an electron between Q A and Q B at equilibrium. k off is the rate of Q B dissociation from its binding site on the D1 subunit. K O , Q B dissociation constant, k on is the rate of Q B binding to its site. k on and k off indicate accessibility of the pocket and/or energy of bonds formed to ligate Q B .

DISCUSSION
We have used a combination of biophysical, biochemical, and physiological assays to characterize the effects of seven amino acid substitutions at the Ala 251 residue of the D1 protein that result in varying levels of photoautotrophic competence. These seven mutants and the five nonphotosynthetic Ala 251 D1 mutants we characterized previously (10) can be grouped into four phenotypic classes with respect to D1 synthesis and steady state level, photosynthetic capacity, photoautotrophic growth rate, and herbicide resistance (Fig. 6): 1) similar to wild type (A251C*), 2) impaired in photosynthesis but not in photoautotrophic growth (A251G*, A251P*, and A251S*), 3) markedly impaired in photoautotrophic growth and photosynthesis and herbicide-resistant (A251I*, A251L*, and A251V*), and 4) not photosynthetically competent (A251R*, A251D*, A251Q*, A251E* and A251H*). The effects of the individual amino acid substitutions on D1 synthesis/accumulation, growth, and carbon accumulation, O 2 evolution, composition of the PSII reaction centers, and Q A to Q B electron transfer will be considered below.
The Arg, Asp, Gln, Glu, and His Substitutions-Previously, we showed that these five amino acid substitutions at the Ala 251 D1 residue led to a nonphotosynthetic phenotype, markedly impaired D1 synthesis and accumulation, and severely affected primary photochemical charge separation efficiency (10). Here, we report that electron transfer between Q A and Q B is almost completely blocked in these five mutants (Table III). No binding of Q B in its niche on the D1 protein could be detected in these strains. Electrons that can be extracted from water oxidation are blocked at the primary quinone acceptor Q A , which has been suggested to lead to a doubly reduced Q A 2Ϫ (36). This could increase the possibility for recombination of the radical pair P 680 ϩ /Phe Ϫ and generate highly toxic 1 O 2 via the 3 P 680 state (37). The conformation of the IV-V loop of the D1 protein in these mutants is presumably structurally altered compared with wild type (10).
Importance of the Ala 251 Residue Deduced from Suppressor Analysis of the Nonphotosynthetic A251R* Mutant-Among about 40 isolates recovered after acetate starvation of the nonphotosynthetic A251R* mutant and sequenced, only same site suppressors were obtained. This strongly suggests that other mutations in the D1 protein or in other chloroplast proteins are unable to compensate for the structural modification caused by insertion of a bulky, charged amino acid at position 251. These results reinforce the hypothesis that the amino acid residue present at position 251 plays a critical role in conformation and function of the IV-V loop (10,16,17,35).
The Cys Substitution-For most of the parameters analyzed, the A251C* mutant appears indistinguishable from the wild type A251A* control (this paper and Ref. 17). However, detailed analysis of the electron flow between water and the plastoquinone pool in the thylakoid membranes indicated a noticeable reduction in the rate of Q A Ϫ reoxidation and a destabilized S 2 Q B Ϫ (S 3 Q B Ϫ ) state, which can be attributed to a shift of the apparent equilibrium constant (K app ) between Q A Ϫ Q B and Q A Q B Ϫ (Table  III). Although the Cys substitution has some impact on the electron transfer between Q A and Q B , this mutant is able to grow and evolve O 2 as well as wild type under LL and HL (Table I), photoacclimates properly during growth in HL by down-regulating its LHCP and chlorophyll levels ( Fig. 2B and Table I), and does not show resistance to PSII herbicides (17). Based on the MTF/STF ratio in A251C* (Table III), the ability of this mutant to reduce fully its plastoquinone pool, as well as its overall rate of electron transport from water to the terminal acceptors, is similar to wild type. Cysteine possesses a reversibly reducible SH group, which might help mediate electron flow between Q A and Q B . This could explain why the overall photosynthetic output and biomass production in this mutant are indistinguishable from those of wild type. D1 turnover in A251C* appears to be increased slightly (Fig.  3), although D1 accumulation is comparable with that of the A251A* control strain (Fig. 2). Although no obvious signs of photoinduced PSII damage have been detected under our experimental conditions (Tables I-III), the slight increase in the rates of D1 synthesis (3.5 ϫ 10 5 versus 3.0 ϫ 10 5 ) and turnover (Ϫ13.9 ϫ 10 3 versus Ϫ13.5 ϫ 10 3 ) observed in this mutant compared with wild type (Fig. 3) might nevertheless indicate an elevated repair of photodamaged PSII centers. Alternatively, the Cys substitution might affect the susceptibility of D1 to normal degradative processes. In either case, if the rate of D1 repair is sufficient to match the rate of D1 degradation caused by the Cys substitution, the number of functional PSII centers would be expected to remain unaffected.
The Gly, Pro, and Ser Substitutions-The A251G*, A251S*, and to a lesser extent A251P* mutants are not only markedly impaired in the electron transfer between Q A and Q B but also in efficiency of primary photochemical charge separation (Tables II and III). These defects on the acceptor and donor sides of D1 lead to an overall reduction of the photosynthetic function (Table I) and to an increased PSII excitation pressure (Table II and Fig. 4). The mutants also fail to down-regulate their chlorophyll content normally when grown in HL (Table I) and have been shown to contain a larger functional antenna per PSII reaction center and lower probability of excitation energy transfer between the reaction centers than wild type (Table II). This observation is corroborated by immunoblot analysis showing that LHCP accumulation was also not down-regulated to the level of the wild type control under HL (Fig. 2).
The redox state of the plastoquinone pool in the thylakoid membrane has been shown to regulate expression of the nuclear cab gene encoding LHCP by molecular signaling (34,38). Reduced plastoquinone was postulated to repress cab gene transcription via a signaling pathway involving a phosphorylated factor coupled to the redox state of plastoquinone through a chloroplast kinase. Substitutions of Gly, Pro, and Ser for Ala 251 slow the electron transfer rate between Q A and Q B (Table III)  vesting light ameliorates the potentially destructive redox chemistry occurring in chloroplasts exposed to bright light, the question arises as to why the A251G*, A251P*, and A251S* mutants with their large antennae grow as well as wild type under HL (17). Decreases in PSII centers per cell seen in these three mutants (Table II) might be a consequence of photodamage or might reflect another adaptation of the cell to the HL environment. Photoinhibition has been defined as a light-dependent decrease in maximum photosynthetic efficiency (39), which results in damage and degradation of the D1 protein. For example, in A251G* and A251P*, D1 synthesis in short term pulse-chase experiments is reduced by about one-third, whereas D1 turnover is increased by 27 and 73%, respectively, compared with wild type (Fig. 3). These mutants also tend to accumulate less D1 than wild type (Fig. 2). Alternatively, photoinhibition may be considered a mechanism that matches overall photosynthetic electron flow to the photon flux rate through the down-regulation of PSII photochemistry (40 -42). In fact, the decrease in the number of PSII reaction centers per cell observed in A251G*, A251P*, and A251S* (Table II) might reflect an adjustment of the photosynthetic electron transport system to high PSII excitation pressure. The 3.3-h photoautotrophic doubling times obtained with wild type C. reinhardtii, A251C* and the A251G*, A251P*, and A251S* mutants in microtiter plates under optimum conditions (saturating light and CO 2 ) (17) are the shortest known for this alga. Hence, our observation that the growth rate of the latter three mutants is not photosensitive under these conditions suggests that photosynthetic electron transfer from Q A to the plastoquinone pool is not rate-limiting for overall photosynthetic electron flow at light saturation. The calculated time for whole chain electron transport ( WC , Table II) shows no correlation with K app (Table III) in these mutants. Sukenik et al. (43) have shown that 1/ WC is correlated with carboxylation in the Calvin cycle rather than electron transport reactions within the thylakoid membranes. A similar discrepancy between photoautotrophic growth and photosynthetic efficiency has been observed in certain cyanobacterial mutants altering the CP47 polypeptide of PSII (44).
The Ile, Leu, and Val Substitutions-Replacement of Val for Ala 251 , the only substitution at this position previously characterized in cyanobacteria and in C. reinhardtii, has been reported to result in resistance to certain PSII herbicides and increased the sensitivity to photoinhibition (13,14). The cyanobacterial mutation was subsequently shown to affect the apparent equilibrium constant on the acceptor side (11) and the properties of the oxygen-evolving complex on the donor side (16). The phenotype of the new A251V* mutant of C. rein-hardtii characterized in this paper is similar to that of the two aforementioned Val mutants. In addition, we created and characterized Ile and Leu substitutions that fall into the same phenotypic category (this paper and Ref. 17). Based on herbicide resistance, we find the A251I* and A251V* mutants to be virtually identical, while the A251L* mutant results in a more extreme herbicide resistant and photosynthetic defective phenotype.
A251I* and A251L*, and to a lesser extent A251V*, are severely impaired in the Q A 3 Q B electron transfer as indicated by a very slow reoxidation rate of Q A Ϫ , a high instability of Q B Ϫ , and a large shift in the equilibrium constant K app (Table  III) but do not show any impairment in the primary charge separation (Table II). In contrast to the Gly, Pro, and Ser substitutions, the reduced photosynthetic yield in A251I*, A251L*, and A251V* is associated with slower photoautotrophic growth and increased herbicide resistance (17). Furthermore, these defects are observed at both LL and HL, and they differ from those in the herbicide-resistant Ser 264 to Ala D1 mutant, which shows reduced growth only in HL (17).
Electron transfer on the acceptor side of the D1 protein is greatly impaired in A251I*, A251V*, and A251L* (Table III), leading to high PSII steady state levels of reduced Q A Ϫ as indicated by reduced qP values (Fig. 4). Given our finding that A251L* accumulates at least as much D1 as wild type (Fig. 2), the 30% increase in the rate of D1 synthesis seen in the pulse experiments appears to offset the 47% increase in the rate of degradation measured in chase experiments compared with wild type (Fig. 3).
Properties of the Amino Acid at Position 251 in Relation to D1 Structure and Function-In cyanobacteria, amino acids 248 -251 of the IV-V loop of the D1 protein are thought to be buried in the thylakoid membrane, dividing the loop into two functional domains (11,16). Our grouping of the 12 Ala 251 mutants of C. reinhardtii into four phenotypic classes in terms of their effects on photosynthetic efficiency, photoautotrophic growth, and herbicide resistance (Fig. 6) relative to the nature of the substituted R group is consistent with the present structural models of the IV-V loop. Insertion of a charged or polar and relatively bulky amino acid at this position (e.g. Arg, Asp, Gln, Glu, or His) would be predicted to destroy the conformation of the D1 IV-V loop because of interactions of the R groups of these amino acids with polar heads of the lipid bilayer of the thylakoid (10). This is consistent with the total blockage of the electron flow between Q A and Q B and the nonphotosynthetic phenotype observed in these five mutants.
Substitution of Ala 251 with an uncharged amino acid having a bulky side chain (Fig. 5, e.g. Ile, Leu, or Val) does not disrupt D1 function completely but impairs photosynthetic electron transfer to various extents that appear to correlate with the position of the methyl group on the R group. Leu, whose methyl group is attached at C-3, has a more severe phenotype than Ile and Val, which have methyl groups attached at C-2. In contrast, substitution of small uncharged amino acids for Ala 251 (e.g. Cys, Gly, Pro, or Ser) has no effect on the mutants' photoautotrophic growth or herbicide sensitivity (17) and photosynthetic O 2 evolution, but all four mutations affect the function of the Q B binding niche to varying degrees. In the case of the Gly, Pro, and Ser substitutions, the acceptor side of PSII is affected. The effect of the Ser and Pro substitutions may be related to the polarizable hydroxyl group and five-ring structure, respectively, in the R groups of these two amino acids. That Gly, which has a hydrogen attached to the carbon backbone instead of the methyl group in the normal Ala residue, perturbs function suggests that steric interactions on the side chain are critical for plastoquinone binding. Interestingly, substitution of Cys with a sulfhydryl group, results in the fewest biophysical, biochemical, and physical consequences of the 12 mutations examined. The fact that Cys can exist in either reduced or oxidized form might help mediate proton and/or electron transfer to plastoquinone in the Q B pocket.
In the case of the seven Ala 251 mutants that retain the capacity for photoautotrophic growth, one has difficulty in correlating growth rate, carbon accumulation, and photosynthetic efficiency under LL and HL with the capacity for electron transfer between Q A and Q B . Furthermore, no correlation has been found between Q B Ϫ stability and the rate of acceleration in D1 turnover. Hence, Chlamydomonas may possess mechanisms for ameliorating the effects of reduced photosynthetic performance caused by certain of the Ala 251 mutations in the D1 protein and thus optimizing their fitness under particular environmental conditions. To identify other proteins and pathways involved in these processes, nuclear suppressors of the photosensitive A251I* and A251L* mutants have been isolated that have an elevated tolerance to very high light (2000 mol m Ϫ2 s Ϫ1 ), which is photoinhibitory both to these mutants and to wild type. Detailed characterization of these suppressors is now in progress.