INTRODUCTION

In all oxygen-evolving organisms, the reaction center of the photosystem II (PSII)¹ complex consists of the D1 protein, the structurally related D2 protein, cytochrome b₅₅₉, and at least two small proteins of unknown function (1). The rapidly turning over D1 protein continuously undergoes a cycle of damage, degradation, and replacement in response to photodamage from normal PSII photochemistry (2). The D1 protein, encoded by the chloroplast psbA gene, is synthesized as a membrane associated 33.5-kDa precursor with a C-terminal extension. After processing to the 32-kDa mature form, D1 undergoes several posttranslational modifications postulated to facilitate translocation and proper assembly into PSII centers in the lipid bilayer (3), where it binds chlorophyll, pheophytin, quinone, carotenoid, iron, and manganese (4). D1 is thought to have five hydrophobic membrane-spanning helices, with its N terminus facing the chloroplast stroma and its C terminus projecting into the thylakoid lumen (5, 6). One stromally exposed region of D1, extending from the C terminus of helix IV through the N terminus of helix V (IV-V loop), participates in binding both Q_B, the second stable quinone acceptor in PSII, and several classes of herbicides that inhibit photosynthetic electron transport (6). The IV-V loop region includes a stromal helix thought to lie parallel to the membrane surface (6) that divides this loop into two parts, one thought to be involved in D1 degradation in vivo and the other functioning in herbicide and quinone binding (7, 8).

Phylogenetic conservation of the IV-V loop among Cyanobacteria, algae, and higher plants (9) suggests that most of these amino acid residues should be essential. However, numerous deletions and amino acid substitutions in this region of D1 in Cyanobacteria do not abolish photosynthetic competence (10, 11, 12, 13, 14, 15). Substitution of Val for Ala²⁵¹, in the putative membrane parallel helix between transmembrane helices IV and V, is reported to confer resistance to certain PSII herbicides and to increase sensitivity to photoinhibition in Cyanobacteria and in Chlamydomonas (8, 16). This modification also affects the electron transfer reactions on both the acceptor and donor sides of PSII, resulting in a reduced electron transfer rate between Q_A and Q_B (17), a decrease of the affinity binding constant of Q_B for its site (7), and modification of the normal oscillatory pattern of oxygen evolution (8).

Since these results pointed to a relationship between photosensitivity and an alanine-to-valine substitution at position 251 of the D1 protein, we have isolated and characterized homoplasmic chloroplast transformants of Chlamydomonas reinhardtii with 12 of the 19 possible alterations at Ala²⁵¹ of D1. Our results show that the Cys, Gly, Ile, Leu, Pro, Ser, and Val mutants retain photosynthetic function to different levels and will be described in a forthcoming article,² whereas the Arg, Asp, Gln, Glu, and His mutants are nonphotosynthetic. In this article, we describe the capacity of these nonphotosynthetic mutants with respect to D1 synthesis, accumulation, and degradation. The Arg mutant makes a mature sized 32-kDa D1 protein with greatly reduced stability, whereas the other four mutants synthesize and accumulate a 33-34-kDa form of D1 of varying stability as well as a very unstable 24-25-kDa peptide that appears to be an aborted D1 translation product. Our results suggest that Ala²⁵¹ plays a key role in the overall tertiary structure of D1 and consequently in the function of PSII.

MATERIALS AND METHODS

The 900-bp KpnI-EcoRI chloroplast DNA fragment from wild-type psbA containing the fourth exon (18) with adjacent intron sequences was subcloned into Bluescript KS vector (Stratagene) from P-528 (19). Single-stranded DNA prepared according to the method of Sambrook et al. (20) was mutagenized with the 32-nucleotide-long oligonucleotides psbAArg (5-CTTACGTAGCTCGTCATAAGTAAAAC-3), psbAGl₂Pr (5-CTTACGTAGCT(C/G)(A/C)ACATAAGTAAAAC-3) or psbAHisAsp (5-CTTACGTAGCT(C/G)ACCATAAGTAAAAC-3) using an oligonucleotide-directed in vitro mutagenesis system (version 2.1; Amersham Corp.). In addition to the mutations at position 251 (bold), a new SspI restriction site (underlined) was created by substituting a T for a C (bold) at position 247, which does not lead to an amino acid substitution (Fig. 1A). After transfection of XL1-Blue cells (Stratagene), bacterial colonies were screened by hybridization with the aforementioned radiolabeled oligonucleotides (20). Alterations at D1 positions 247 and 251 were verified by sequencing (Life Technologies, Inc., double-stranded DNA cycle sequencing system; and [³²P]ATP, DuPont NEN). Mutagenized 250-bp BstEII-EcoRI fragments covering the end of exon 4 and the beginning of intron 4 (Fig. 1A) were then inserted into P-528 to replace the corresponding wild-type sequence. The resulting plasmids, P-529 (Ala  Arg, A251R*), P-598 (Ala  Gln, A251Q*), P-599 (Ala  Glu, A251E*), P-601 (Ala  His, A251H*), and P-602 (Ala  Asp, A251D*), have amino acid alterations at position 251 and the SspI restriction fragment marker at position 247 (designated by the asterisk). The codons for the introduced amino acids at position 251 were chosen according to the codon usage of the D1 protein in Chlamydomonas and are therefore expected to be properly translated.

The NP strain (CC-2835) of Lers et al. (21) was used to create the mutant A251D*:NP by exchanging the last 180 bp of exon 5 from psbA of P-602 carrying the exon 4 Ala²⁵¹ to Asp change with the same fragment generated from P-388 (21), thus introducing the stop codon at Ser³⁴⁵ and the new AflII restriction fragment marker into the mutated psbA gene. The BsmI-BamHI fragment of P-602 was replaced by the BsmI-KpnI fragment of P-388, as explained below (see Fig. 1A). P-388 (21) was digested with KpnI, and the overhanging ends were filled in using the Klenow reaction (20). P-602 (see above) was digested with BamHI and filled in using Klenow as well. Both linearized plasmids P-388 and P-602 were then digested at the unique BsmI restriction site located in the middle of the fifth exon of psbA (Fig. 1A). The 11-kilobase fragment obtained from P-602 was then ligated to the 500-bp fragment derived from P-388. The chimeric plasmid was transformed into XL1-Blue cells to generate P-653. The Ala²⁵¹ to Asp and the Ser³⁴⁵ to stop codon changes (21) were verified by sequencing.

Each of the above plasmids carrying psbA mutations was introduced into wild-type strain CC-125 by cotransformation (22) together with P-228 (21), which possesses a point mutation in the 16 S rRNA gene conferring spectinomycin resistance (Fig. 1A). A wild-type control strain carrying both the spectinomycin resistance marker and the restriction fragment marker at position 247 of psbA was also created by transforming CC-125 (19) with P-528 and P-228. After transformation, cells were respread on Tris acetate-phosphate (TAP) medium (19) containing 50 µg/ml ampicillin and 100 µg/ml spectinomycin and placed in very dim light (5 µmol/m²/s photosynthetically active radiation). Transformants resistant to spectinomycin were screened for the presence of the SspI restriction site in PCR DNA products obtained using primers pBr1 (5-GCTGGTGTATTCGGTGG-3) and pBr3 (5-CTTGCGGGAAACTAACG-3) (Fig. 1A). Transformants carrying the SspI site were single cell cloned, and homoplasmicity for this marker (and very likely for the closely linked A²⁵¹ mutation) was verified by both restriction analysis of PCR products and Southern blots (Fig. 1B). The alteration at position 251 was established by DNA sequencing of PCR products using the primer pBr1. Isolation of total cell DNA from Chlamydomonas, Southern hybridizations, and PCR DNA amplifications were done as described by Lers et al. (21). The mutant A251D*:NP in which psbA contains the Ala²⁵¹ to Asp change and a stop codon at Ser³⁴⁵ was also created by cotransformation. Transformants resistant to spectinomycin were analyzed for the SspI and AflII restriction fragment markers, single cell cloned, and sequenced to verify that all mutations were present.

Transformed cells were grown in liquid TAP medium at 25 °C (19) bubbled with filtered air under low light (LL; 15 µmol/m²/s). For determination of the mutant strain autotrophic growth, cells were spotted at the same cell density on plates of TAP and minimal HS media (19) and incubated under medium light (300 µmol/m²/s). Resistance to metronidazole was assessed as described by Smith and Kohorn (23) under medium light.

Cells were harvested in log phase, and thylakoid membranes were purified (24). Thylakoid association of the 24-25-kDa fragment was assessed using dissociating buffers, such as 2 M NaCl, 20 mM dithiothreitol, Na₂CO₃ (pH 11), and 6.8 M urea (24). Proteins from whole cells or thylakoid membranes were precipitated in ice-cold acetone for 1-2 h, pelleted at 12,000 × g for 20 min, washed in 80% ice-cold acetone (v/v), dried, and resuspended in 1 volume H₂O and 1 volume lysis buffer (2% LDS (unless otherwise specified), 5 mM EDTA and 40 mM Tris-HCl, pH 7.4). Samples were boiled for 30 s to 3 min, and 1 volume sample buffer (Laemmli (25) containing LDS instead of SDS) was added, heated for an additional 5 min, and pelleted for 6 min in a microcentrifuge. The supernatants were loaded on an equal protein basis (unless otherwise specified) onto denaturing (8 M urea) or nondenaturing 10-17.5% LDS-PAGE together with the Life Technologies 10-kDa protein ladder and run at a constant 100 V for 6 h at room temperature.

For immunodetection, proteins were electrotransferred to nitrocellulose membranes (BA85, Schleicher & Schuell) and detected by an enhanced chemiluminescence Renaissance kit (Dupont NEN) according to the manufacturer's instructions, using antisera at the following dilutions: D1, 1:1000; tubulin, 1:2000 and OEE1, 1:500. Immunoprecipitation with D1 antiserum using protein A-Sepharose beads (Repligen) was carried out as described by Schmidt et al. (26).

Transformants were grown under LL at 25 °C in liquid TAP medium in which the MgSO₄ was replaced with an equimolar amount of MgCl₂, and the acetate concentration was reduced by 3-fold (5.8 mM). Cells in early log phase (A₇₅₀, 0.08-0.12) were harvested by centrifugation (3,000 × g for 6 min). Pelleted cells were resuspended at 0.2 A₇₅₀ in TAP medium lacking sulfate and having further reduced acetate concentration (4.35 mM), which was found to decrease the background of nonspecific labeling. Five-ml aliquots of the cell suspension were placed in individual wells of a six-well microtiter plate (Falcon 3046) and agitated on an orbital shaker under LL at 25 °C for 1 h to equilibrate. Anisomycin (Pfizer) was then added to the culture to a final concentration of 140 µg/ml to inhibit cytoplasmic protein synthesis. After a 15-min incubation, 25 µl of carrier-free H₂³⁵SO₄ (Dupont NEN) was added to a final concentration of 125 µCi/ml. At each time point (4, 8, 12, and 20 min), 0.5-ml aliquots were removed, injected into 6 ml of ice-cold acetone, and stored on ice for 1-2 h. Protein solubilization and gel electrophoresis were carried out as described above. Gels were stained with Coomassie Brilliant Blue, dried, and analyzed using a Molecular Dynamics PhosphorImager. Bands were quantified using ImageQuant software, version 3.2, with backgrounds for representative areas in each lane subtracted manually. Labeled polypeptides were also detected by autoradiography of the dried gels.

Five min after taking the last time point in the pulse experiment, 1.2 ml of TAP medium containing 200 mM Na₂SO₄ and 5.8 mM acetate was added to initiate the chase. After 15 min, the first aliquot (0.5 ml) was removed and considered as the 0 min time point of the chase. The other time points were taken at 20, 40, 60, 90, 120, and 180 min. In most chases, anisomycin (final concentration, 140 µg/ml) and lincomycin (final concentration, 550 µg/ml) were added to inhibit cytoplasmic and chloroplast protein synthesis.

Chlorophyll concentration was measured in 80% acetone extracts (21). Photosynthetic O₂ evolution and chlorophyll fluorescence parameters were determined as described by Lers et al. (21), except that the measurements were carried out in TAP medium, and no bicarbonate was added to the chamber. Dark adapted (F₀), maximum (F_m), and variable (F_v = F_m  F₀) chlorophyll a fluorescence values were determined after the cells were dark adapted for at least 15 min to allow relaxation of fast fluorescence-quenching components.

RESULTS

Mutated psbA genes encoding changes of D1 Ala²⁵¹ to Ala (A251A*, codon GCT = wild-type control), Arg (A251R*, CGT), Gln (A251Q*, CAA), Glu (A251E*, GAA), His (A251H*, CAC) and Asp (A251D*, GAC) were introduced into wild-type C. reinhardtii strain CC-125 by cotransformation (22) with a selectable spectinomycin resistance marker in the 16 S rRNA (21) (Fig. 1A). Southern analysis of total cell DNA from transformant subclones demonstrated that all chloroplast genomes (~70/plastid) carried the new SspI site closely linked to the mutation at position 251 (Fig. 1B). Fragments of about 300 and 500 bp were seen in SspI digests of total cell DNA probed with the pBr1-pBr3 PCR product when the SspI restriction fragment marker was present at position 247, whereas an 800-bp fragment occurred in wild-type DNA lacking the SspI restriction site (Fig. 1B).

Substitutions of Arg, Gln, Glu, His, or Asp for Ala²⁵¹ of the D1 protein result in transformants with a nonphotosynthetic phenotype that require an exogenous acetate carbon source for growth (Table I). They are light insensitive (up to 600 µmol/m²/s) and also grow in the presence of the suicide compound metronidazole (19). None of these mutants is capable of photosynthetic O₂ evolution, suggesting that their PSII centers are inactive even after dark adaptation and exhibit the high fluorescence phenotype typical of strains impaired in PSII function (27). All five mutants show an elevated level of dark adapted chlorophyll a fluorescence (F₀) as well as a reduction in F_v/F_m (Table I), an indicator of the photochemical efficiency of PSII. Damage to PSII centers is known to result in the rise of the initial F₀ level (27), whereas differences in F₀ and F_v/F_m between the mutants suggest that their PSII function is altered to various extents.

Table I. Analysis of growth, O2 evolution (Pmax) and chlorophyll a fluorescence parameters F0 and Fv/Fm of transformants homoplasmic for mutations at the Ala251 residue Genotype Growtha Photosynthesisb HS TAP MET Pmax (µmol/mg Chl/h) F0 (arbitrary units) Fv/Fm WT +c +   120 25.0 0.77 A251R*   + + 0 51.5 0.24 A251Q*   + + 0 53.0 0.23 A251E*   + + 0 96.0 0.13 A251H*   + + 0 61.0 0.46 A251D*   + + 0 68.0 0.41 a  Cells were either grown on minimal HS medium (autotrophic growth), on acetate-containing TAP medium (mixotrophic growth), or on TAP medium containing 20 mM metronidazole (MET) at different light intensities from 70 to 600 µmol/m2/s. b  Cells were pregrown in TAP medium at LL (15 µmol/m2/s). c  +, growth; , no growth.

Immunoblot analysis of total cell extracts reveals that D1 accumulation in the mutants is perturbed to varying degrees with respect to wild type (Fig. 2, A and B). Antisera against the tubulin  subunit and OEE1, one of the three extrinsic proteins of the oxygen-evolving complex, were used as controls for equal loading. A251R* is the only mutant that accumulates nearly normal amounts (75%) of the 32-kDa mature form of the D1 protein (Fig. 2C). In contrast, the A251Q*, A251E*, A251H*, and A251D* mutants accumulate a slower migrating 33-34-kDa form of D1 (P33-34) (Fig. 2B) that is distinct from the D1 precursor. Indeed, a A251D* transformant that carries the NP preprocessed mutation (A251D*:NP mutant) lacking the C-terminal D1 extension (21) still synthesizes P33-34 (Fig. 2D). Variations in D1 size were confirmed in a comparative pulse-labeling experiment (see below and Fig. 4B). Our inability to detect the P33-34 peptides by immunoblot analysis of deletion mutants in the psbA gene encoding D1 (CC-1078 and FUD 7; Ref. 19) proved that these peptides are not contaminant proteins recognized by the D1 antibody (data not shown) but, rather, variant forms of the D1 protein.

The A251Q*, A251E*, A251H*, and A251D* mutants also accumulate varying amounts of a 24-25-kDa peptide (P24-25), which is recognized (Figs. 2 and 3) and immunoprecipitated (data not shown) by the D1 antibody. When 1% LDS (final concentration) is used for protein solubilization, P24-25 is barely detectable, and a band migrating at the same position as the mature D1 protein can be seen (Fig. 2A). This latter band disappears, and P24-25 is intensified when the solubilization buffer contains 2% LDS, suggesting that accumulated P24-25 might be tightly associated with another small peptide of about 8-9 kDa (Fig. 2, compare A and B). Since we do not observe this association during pulse-chase labeling experiments, it must occur at a later time (see "Discussion"). The similar upshift of both P33-34 and P24-25 in the Glu and Asp mutants compared with the respective peptides in the Gln and His mutants is likely due to the negative charge of the amino acid introduced in the former mutants.

P33-34 accumulates up to ~45-50% of the wild-type D1 level in the A251H* and A251D* mutants but only reaches 9 and 15% of the wild-type D1 level in the A251Q* and A251E* mutants, respectively (Fig. 2C). P24-25 as well as P33-34 is found in purified thylakoids of these mutants (shown for the A251D* mutant in Fig. 3), demonstrating that it can assemble into the photosynthetic membrane. Furthermore, extraction of polypeptides from thylakoids by various treatments (see "Materials and Methods") showed that P24-25 is intrinsic and as resistant to extraction as mature wild-type D1 and P33-34 from the other mutants (data not shown). Since our D1 antibody was raised against a synthetic peptide containing amino acids in the luminal loop between helices I and II of the D1 protein, we know that the 24-25-kDa peptides represent N-terminal fragments.

To determine whether the reduced accumulation of the D1 protein in the Arg, Gln, Glu, His, and Asp mutants was due to a defect in synthesis or to a high turnover rate, Chlamydomonas transformants were pulse labeled with [³⁵S]H₂SO₄ and subjected to a cold sulfate chase. Synthesis of D1 is impaired to varying degrees depending on the amino acid present at position 251 of the D1 protein (Fig. 4A). Comparison of the migration of labeled D1 protein from the mutants and wild type demonstrated that A251R* synthesizes mature sized D1, whereas A251Q*, A251H*, A251E*, and A251D* make the slower migrating P33-34 (Fig. 4A), as was observed on the immunoblot (see Fig. 2A). When migration of newly synthesized D1 proteins in the Glu, Gln, His, and Asp mutants was compared with those of wild-type and the preprocessed NP mutant (21) in which D1 is directly synthesized in its mature form, only the slower migrating form of D1 was detected (Fig. 4B).

These experiments also provided information on the origin of P24-25. The A251H* and especially the A251E* and A251D* strains synthesize predominantly P24-25, whereas the A251Q* strain synthesizes a lesser amount of this peptide. A251Q* and A251H* synthesize a peptide migrating at 24 kDa, whereas the A251E* and A251D* mutants make a 25-kDa peptide. P24-25 is detected very early during the pulse-labeling experiment, before any P33-34 is seen, suggesting that it is probably not an N-terminal D1 degradation product. Furthermore, we never observed labeling of the smaller C-terminal peptide of 7-8 kDa, which would be expected if P33-34 was cleaved to yield P24-25.

D1 synthesis in the A251R* mutant occurs at ~50% of the wild-type rate (Table II). In the A251Q*, A251H*, A251E*, and A251D* mutants, synthesis of P33-34 is greatly reduced (6-15% of the wild-type D1 rate), and synthesis of P24-25 ranges from 7 to 107% of the wild-type rate of D1 synthesis. The A251H*, A251E*, and A251D* mutants synthesize P24-25 at a much higher rate than P33-34 (Table II and Fig. 4A). If the rates of synthesis of P24-25 and P33-34 are summed for each mutant, they amount to 18-122% of the wild-type rate of D1 synthesis.

Table II. Effect of the D1 Ala251 mutations on the synthesis of the D1 and D1-related proteins Cells were pulse-labeled as described in Fig. 4. Rates of synthesis were calculated from regressions (four points) of the D1 and D1-related volume-integrated pixel densities obtained after quantification with the Phosphorlmager system. Rates of synthesis were normalized to synthesis of WT D1. The ratio of rates of synthesis P24-25 to P33-34 is also given for each mutant. Normalized rates of synthesis WT A251R* A251Q* A251H* A251E* A251D* D1 (% of WT) 100 51.1 NA NA NA NA P33-34 (% of WT D1) NAa NA 10.8  6.0  15.0  7.1 P24-25 (% of WT D1) NA  NA  7.0 34.6 107.2 53.2 P33-34 + P24-25 (% of WT D1) NA  NA 17.8 40.6 122.2 60.4 Ratio of rates of synthesis WT A251R* A251Q* A251H* A251E* A251D* P24-25/P33-34 NA NA 0.7 5.8 7.2 7.5 a  NA, not applicable.

Stability of D1 and D1-related peptides (P33-34 and P24-25) was determined by pulse-chase experiments (Fig. 5). In the A251Q* and A251H* mutants P33-34 is as stable as wild-type D1 (a greater than 6-h half-life), whereas its stability in the A251D* and A251E* mutants is only one-third to one-half of that of wild-type D1. In the A251R* mutant, D1 has only a 45-min half-life. Where present, most of the P24-25 is very unstable (half-life, 10-20 min) and disappeared extremely rapidly without any concomitant increase of label in P33-34. A small fraction of the P24-25 synthesized seems to be stable over the 3-h chase and might explain the accumulation seen on immunoblots (Fig. 2B). Since P24-25 appears before any detectable P33-34 during the pulse (Fig. 4A) and does not chase into P33-34 or vice versa (Fig. 5), this polypeptide is neither a D1 breakdown product nor a translation intermediate that will eventually give rise to full-length D1. Instead it appears to be an aborted D1 translation product, which terminates at a particular domain (see "Discussion").

DISCUSSION

Although the loop between transmembrane helices IV and V of the D1 protein is highly conserved among Cyanobacteria, algae, and plants (9), most amino acid substitutions and several deletions created in this region, alone or in combination, still retain a certain level of photosynthetic capacity (Fig. 6). With the exception of a few deletions, only three single amino acid substitutions (Y254S, F255W, and L271S; Ref. 12) in the quinone binding region of the cyanobacterial D1 protein (residues 248-271) have been reported to lead to a nonphotosynthetic phenotype. In contrast, we find that replacement of Ala²⁵¹ by a charged or very polar and bulky amino acid disrupts D1 function completely (Table I), whereas substitution of seven other uncharged and smaller amino acids at this position still permits photoautotrophic growth.² These results suggest that residue 251 in the parallel helix between transmembrane helices IV and V is critical for D1 function. Similar major changes in amino acid shape and charge elsewhere in this loop do not result in a nonphotosynthetic phenotype (e.g. mutants S232D, V249D, and G256K in Fig. 6).

Chloroplast transformants of Chlamydomonas having Arg, Gln, Glu, His, or Asp substituted for Ala at residue 251 in the D1 protein are incapable of photosynthetic O₂ evolution and defective in photosynthetic electron transport, as indicated by their resistance to metronidazole. Furthermore, their high F₀ values are consistent with the other photosynthetic parameters in suggesting that the mutant PSII centers are much less efficient as energy traps than in the wild type (27).

Immunoblot analysis (Fig. 2A) as well as pulse-labeling experiments (Fig. 4) clearly show that the Gln, His, Glu, and Asp mutants accumulate a modified form of the D1 protein (P33-34) that has a slightly reduced electrophoretic mobility but still localizes to the thylakoid membranes (Fig. 3). P33-34 was shown not to be the 33.5-kDa D1 precursor accumulated as a consequence of impaired C-terminal processing (Fig. 2C). A slower migrating form of D1 (D1*) resulting from N-terminal phosphorylation occurs in higher plants after photoinhibitory treatments (29, 30) but is never the exclusive form accumulated in vivo or in vitro, as is the case for P33-34 in these mutants under nonphotoinhibitory conditions. Phosphorylation of D1 has never been observed in C. reinhardtii (31) despite a conserved Thr residue present at its N terminus (18). For these reasons, P33-34 is likely to result from structural modifications of D1 other than phosphorylation that might protect it from proteolytic attack in the same manner as the N-terminal phosphorylation of D1 does in higher plants (32).

The discrepancy between P33-34 turnover and steady state levels in some of the mutants suggests that different forms of P33-34 with different half-lives might coexist in the thylakoid membrane. For example, in the Gln and His mutants, P33-34 is poorly synthesized but seems to be stable over 3 h in the pulse-chase labeling experiment. Although almost no P33-34 accumulates in the Gln mutant, a substantial amount is observed in the His mutant. Similar differences between the half-life of D1 determined by pulse-chase labeling and its accumulation analyzed by immunoblot have been reported previously for some cyanobacterial mutants (33). Stability of other photosynthetic membrane proteins such as Cytochrome b₅₅₉ has also been shown to be affected by conformational parameters as well as by their membrane association (34).

The Gln, His, Glu, and Asp mutants also synthesize and accumulate P24-25, an immunologically related D1 peptide, as well as P33-34. Although this N-terminal D1 peptide is intrinsic to thylakoid membranes (Fig. 3), we have no evidence that it is assembled into PSII complexes. Since the label in P24-25 does not chase into P33-34 (Fig. 5), we hypothesize that this polypeptide is an aborted D1 translation intermediate. The rapid disappearance of P24-25 during a 3-h chase in the presence of lincomycin indicates that this polypeptide is rather unstable, and continuous translation is required for its accumulation.

A small fraction of newly synthesized P24-25 seems to be stable (Fig. 5) and might correspond to the fraction detected on immunoblots (Fig. 2, A and B). Most of the P24-25 accumulated appears to be associated with another small protein of 8-9 kDa, giving rise to a dimer product running at about 32 kDa. This 8-9-kDa protein might be a PSII component that promotes assembly of P24-25 into the reaction center and protects it from degradation. Cross-linking of the  subunit of cytochrome b₅₅₉ and the D1 protein has been shown to occur in isolated PSII reaction centers after a light treatment (35). Integration of the chlorophyll a- and b-binding protein CAB-7p into photosystem I proceeds through a membrane intermediate that is originally cleavable and becomes protease-resistant during assembly (36). Thus stability of many photosynthetic membrane proteins seems to be promoted by their integration into photosynthetic complexes.

In barley, D1 is synthesized by thylakoid-bound polysomes, and translation intermediates of 15-25 kDa can be chased into full-length D1 (37). Six ribosome pausing sites, A-F, which correlate with D1 translation intermediates, are thought to be important for cotranslational binding of cofactors to D1 and for the insertion of the five transmembrane helices of this protein into the chloroplast thylakoids (38). Because the D1 proteins of C. reinhardtii and barley are highly similar in structure (>94% amino acid identity; Ref. 9), the P24-25 peptide we detect in the Gln, Glu, His, and Asp mutants might originate from an abortion of translation in the C-terminal portion of the D1 IV-V loop at the last (pause F) or next to last (pause E) ribosome pausing site. Since these mutants also synthesize varying amounts of P33-34, we suggest that the nascent D1 polypeptide either: (i) is cleaved proteolytically during ribosome pausing at the E or F sites, yielding the unstable P24-25; or (ii) undergoes a conformational change or posttranslational modification making it resistant to cleavage in the protease-sensitive loop connecting helices IV and V and yielding P33-34. Since D1 residue 251 influences Q_B binding and herbicide resistance (8), the inability of the mutants to bind quinone during D1 synthesis might render the protein sensitive to proteolytic cleavage downstream of Ala²⁵¹. Based on the molecular weight of P24-25, the predicted cleavage points would be between residues 258 and 269 near the C-terminal end of the IV-V loop.

During acceptor side photoinhibition, a D1 N-terminal peptide of 23 kDa is observed both in vitro (39) and in vivo (28) under certain physiological conditions. This D1 breakdown product should not be equated with the P24-25 D1 translation product we observe in the Gln, Glu, His, and Asp mutants for several reasons. First, the mutant cells were not subjected to any photoinhibitory treatment. Second, the 24-25-kDa peptide appears before we can detect P33-34 during a pulse-labeling experiment (Fig. 4A). Third, the synthesis rate of this peptide is greater than the synthesis rate of P33-34, and the ratio of P24-25 to P33-34 synthesis rates remains constant throughout the time course of a pulse experiment (data not shown). In the case of a degradation product, this ratio would have been expected to increase as a function of time (28).

The nonphotosynthetic Arg mutant synthesizes a mature sized 32-kDa D1 protein but neither P24-25 nor P33-34 (Figs. 2A and 4). If the D1 protein in this mutant is properly incorporated into PSII complexes of the thylakoid membrane with normal cofactor interactions, its capacity for photosynthetic electron transfer must be blocked for other reasons. Perewoska et al. (8) showed that the Ala²⁵¹ to Val change in Cyanobacteria has a long range side effect on the donor side of the protein in destabilizing the oscillatory pattern of oxygen evolution (S states). The A251R* mutation could have severe detrimental effects on the photochemistry of PSII without destroying the structure of the IV-V loop. Surprisingly, no breakdown products resulting from either acceptor side or donor side photoinhibition (39) were detected in a pulse-chase experiment, although the D1 protein in the Arg mutant was found to be very unstable (half-life, ~45 min) under these LL conditions (Fig. 5). As discussed earlier, forms of D1 with different half-lives coexisting in the thylakoids of this mutant could explain the high level of D1 accumulation seen on immunoblots, since synthesis is only 50% of that of wild type, and newly synthesized D1 appears rather unstable. Although the predominant form has a short half-life, a small fraction could be stable and accumulate in the membranes.

In conclusion, we find that substitution of amino acids that are charged and/or have a very long side chain for D1 residue Ala²⁵¹ results in a nonphotosynthetic phenotype unlike most other amino acid substitutions and many deletions created in the IV-V loop of the D1 protein. The Gln, Glu, His, and Asp mutants represent the first instances in which a single amino acid substitution in D1 has been observed to result in premature termination of its translation. In contrast, the Arg mutant synthesizes nearly normal amounts of a mature sized D1, which is nonfunctional and appears rather unstable. Although all the mutant forms of the D1 protein were shown to be intrinsic in the thylakoid membrane, we do not know whether they assemble into PSII complexes. Collectively, these results suggest that the Ala²⁵¹ residue of D1, in the putative parallel helix between transmembrane helices IV and V where the quinone binds, plays a critical role in the structural conformation of the IV-V loop and therefore in the proper electron flow between Q_A and Q_B. Etienne and Kirilovsky (7) have suggested in Cyanobacteria that D1 amino acids 248-251 at one end of this parallel helix may be buried in the thylakoid membrane, thus dividing the IV-V loop into two parts, one involved in Q_B binding and the other in D1 degradation. Our findings support this hypothesis, since substitution of charged or very polar amino acids for Ala²⁵¹ is likely to modify this association severely due to interactions with polar head groups of the lipid bilayer.