REP27, a Tetratricopeptide Repeat Nuclear-Encoded and Chloroplast-Localized Protein, Functions in D1/32-kD Reaction Center Protein Turnover and Photosystem II Repair from Photodamage 1[OA]

The goal of this research is elucidation of the molecular mechanism for the unique photosystem II (PSII) damage and repair cycle in chloroplasts. A frequently occurring, irreversible photooxidative damage inhibits the PSII charge separation reaction and stops photosynthesis. The chloroplast PSII repair process rectiﬁes this adverse effect by selectively removing and replacing the photoinactivated D1/32-kD reaction center protein (the chloroplast-encoded psbA gene product) from the massive ( . 1,000 kD) water-oxidizing and O 2 -evolving PSII holocomplex. DNA insertional mutagenesis in the model organism Chlamydomonas reinhardtii was applied for the isolation and characterization of rep27 , a repair-aberrant mutant. Gene cloning and biochemical analyses in this mutant resulted in the identiﬁcation of REP27 , a nuclear gene encoding a putative chloroplast-targeted protein, which is speciﬁcally required for the completion of the D1 turnover process but is not essential for the de novo biogenesis and assembly of the PSII holocomplex in this model green alga. The REP27 protein contains two highly conserved tetratricopeptide repeats, postulated to facilitate the psbA mRNA cotranslational insertion of the nascent D1 protein in the existing PSII core template. Elucidation of the PSII repair mechanism may reveal the occurrence of hitherto unknown regulatory and catalytic reactions for the selective in situ replacement of speciﬁc proteins from within multiprotein complexes. of the ECL (Pierce) detection kit following the manufacturer’s speciﬁcations. and western-blot cation

The goal of this research is elucidation of the molecular mechanism for the unique photosystem II (PSII) damage and repair cycle in chloroplasts. A frequently occurring, irreversible photooxidative damage inhibits the PSII charge separation reaction and stops photosynthesis. The chloroplast PSII repair process rectifies this adverse effect by selectively removing and replacing the photoinactivated D1/32-kD reaction center protein (the chloroplast-encoded psbA gene product) from the massive (.1,000 kD) water-oxidizing and O 2 -evolving PSII holocomplex. DNA insertional mutagenesis in the model organism Chlamydomonas reinhardtii was applied for the isolation and characterization of rep27, a repair-aberrant mutant. Gene cloning and biochemical analyses in this mutant resulted in the identification of REP27, a nuclear gene encoding a putative chloroplast-targeted protein, which is specifically required for the completion of the D1 turnover process but is not essential for the de novo biogenesis and assembly of the PSII holocomplex in this model green alga. The REP27 protein contains two highly conserved tetratricopeptide repeats, postulated to facilitate the psbA mRNA cotranslational insertion of the nascent D1 protein in the existing PSII core template. Elucidation of the PSII repair mechanism may reveal the occurrence of hitherto unknown regulatory and catalytic reactions for the selective in situ replacement of specific proteins from within multiprotein complexes.
Chloroplast development and differentiation in the unicellular green alga Chlamydomonas reinhardtii can take place either under autotrophic or (photo) heterotrophic conditions. Biogenesis and assembly of functional PSII, and of the other thylakoid membrane complexes, including PSI, the cytochrome b6-f complex, and the ATP synthase (Wollman et al., 1999;Minai et al., 2006), can take place in the dark in C. reinhardtii, provided that organic carbon, e.g. acetate, is supplied to the cells . Whether in the light or in the dark, assembled thylakoid membrane complexes are fairly stable and do not turnover with physiologically relevant rates. An exception to this rule is the D1/ 32-kD PSII reaction center protein (the psbA gene product), which is known for its frequent turnover in the light (Mattoo and Edelman, 1987;Vasilikiotis and Melis, 1994).
The PSII holocomplex performs the functions of light absorption and excitation energy transfer to P680, leading to water oxidation, the release of O 2 and protons, and the transport of electrons from water to reduce plastoquinone molecules in the thylakoid membrane. The transient formation of strong oxidants, the abundance of O 2 , and the presence of excitation energy are conditions that may lead to photooxidative damage (Ohad et al., 1984;Barber, 1994;Melis, 1999), causing irreversible inactivation in the electron-transport function of the D1 protein and inhibiting the function of PSII (Kok, 1956;Powles, 1984;Melis, 1991;Aro et al., 1993). An elaborate repair mechanism operates in organisms of oxygenic photosynthesis and restores the functional status of PSII. The PSII damage and repair cycle, as the phenomenon has come to be known , operates in tandem with photosynthesis and appears to be conserved in cyanobacteria, algae, and crop plants.
The rate constant of photodamage is proportional to the incident light intensity (Baroli and Melis, 1996;Tyystjarvi and Aro, 1996;Hakala et al., 2005;Ohnishi et al., 2005). Thus, photodamage can occur with a half time as slow as 24 h under very low light intensity conditions, or as fast as 30 min under bright sunlight Yokthongwattana and Melis, 2006). The rate-limiting step in the enzymatic repair of PSII occurs with a half time of about 2 h (Sundby et al., 1993;Vasilikiotis and Melis, 1994;Neidhardt et al., 1998).
This defines the overall rate constant for the repair process and is largely independent of the light intensity to which the photosynthetic organism is exposed.
In PSII repair-aberrant mutants of C. reinhardtii, water oxidation and electron transport capacity depend on the balance between de novo biogenesis/ assembly of PSII complexes (presence of acetate) and the rate of photodamage. Therefore, in repair-aberrant mutants, the degree of photoinhibition of photosynthesis can be manipulated by the relative dominance of these two players, i.e. de novo biogenesis/assembly and photodamage of PSII. In the presence of acetate and under weak light intensities, when the rate of de novo biogenesis/assembly of PSII is comparable to that of photodamage, intermediate levels of photoinhibition will manifest in relation to a wild-type control. Stronger light intensities, however, would make the rate of photodamage far exceed the rate of de novo biogenesis/assembly, causing a quantitative accumulation of photodamaged PSII and the end of detectable photosynthetic water oxidation activity.
In this study, DNA insertional mutagenesis of C. reinhardtii was employed to generate and isolate PSII repair mutants. The appropriate screening steps (Zhang et al., 1997) resulted in the isolation of rep27, a C. reinhardtii strain that is defective in photoautotrophic growth, which, however, greens normally in the presence of acetate and displays a lowered photosynthetic water oxidation and CO 2 conversion activity when grown under low irradiance conditions. Under moderate and high light intensity conditions, PSII activity is selectively abolished in this mutant, suggesting a defect in the repair from photodamage. Gene cloning and biochemical analyses in this PSII repair-aberrant mutant resulted in the identification of REP27, a nuclear gene encoding a putative chloroplast-targeted protein, which is specifically required for the completion of the D1 turnover process but is not essential for the de novo biogenesis/assembly of the PSII holocomplex in this model green alga. A working hypothesis model is presented, postulating a function of the REP27 protein in the psbA mRNA translation and/or cotranslational insertion of nascent D1 in existing PSII core templates.

Isolation and Characterization of rep27, a Putative PSII Repair Mutant
A putative PSII repair mutant, termed rep27, was isolated following DNA insertional mutagenesis and on the basis of a stringent two-step screening protocol (Zhang et al., 1997). Acetate requirement was the first screening step in this protocol. Whereas both wild-type and rep27 mutant strains grew well on acetate-containing Tris-acetate phosphate (TAP) media, only the wild type grew on minimal Tris-bicarbonate-phosphate (TBP) media (Table I; Fig. 1A). Cellular chlorophyll (Chl) content was similar in wild type and rep27 mu-tant, about 3.4 3 10 215 mol/cell when grown under 10 mmol photons m 22 s 21 (10 mE) and about 1.7 3 10 215 mol/cell when grown under 150 mmol photons m 22 s 21 (150 mE). Moreover, wild type and rep27 displayed similar Chl a/b ratios (data not shown). These results suggested that in the presence of acetate, growth, Chl accumulation, as well as the assembly and acclimation properties of the photosynthetic apparatus were not affected by the rep27 mutation. When cultivated in TAP media under either 10-or 50-mE conditions, the wild type displayed a slightly faster rate of growth than the rep27 mutant (Fig. 1B).
Pivotal in the successful screening of repair mutants in C. reinhardtii is the differential biogenesis/assembly of functional PSII units in the presence of acetate and the subsequent inability of the chloroplast to repair them once photodamage has occurred (Zhang et al., 1997). It is thus expected that, under low-light (10 mE) growth in the presence of acetate, the rate of de novo biogenesis/assembly of the PSII holocomplex would be faster than the rate of PSII photodamage. Under these conditions, repair mutants will possess functional PSII reaction centers, the abundance of which will decline with growth irradiance faster than that in the wild type. This differential irradiance response can be monitored from the in vivo variable to maximal (F v /F m ) fluorescence yield ratio and by the light-saturated rate of oxygen evolution in the two strains. The F v /F m ratio provides a convenient and direct method for the measurement of the photochemical charge separation efficiency of PSII reaction centers (Kitajima and Butler, 1975), while the light-saturated rate of oxygen evolution provides a measure of PSII electron transport capacity and, therefore, of the number of fully assembled/functional PSII units.
The F v /F m fluorescence yield ratio was measured in C. reinhardtii cells grown under different irradiance regimes (10-150 mE; Fig. 1C). The ratio was lower in the rep27 mutant than in the wild type under all growth irradiance conditions, suggesting fewer functional PSII reaction centers in the former than in the latter. This was confirmed by sensitive absorbance difference spectrophotometry (DA 320 ; Melis, 1989), which revealed a substantially lower concentration of the functional primary quinone acceptor (Q A ) in the mutant than in the wild type. For cells grown at 150 mE, analysis showed Q A /Chl (wild type) 5 1:560 mol:mol; Q A /Chl (rep27) 5 1:1,800 mol:mol, whereas P700/Chl ratios were similar at about 1:650 mol:mol in each of these the two strains. Accordingly, the operational PSII/PSI ratio was lowered in the rep27 mutant in an irradiance-dependent fashion (Table I). These results are consistent with the interpretation that the rep27 strain is a putative PSII repair mutant.
To independently assess the functional capacity of PSII, the light-saturated rate of photosynthesis (Pmax) was measured (Fig. 1D). Wild type and rep27 mutant were grown at three different irradiances (10, 50, and 150 mE) and Pmax was measured from the rate of oxygen evolution in a Clark-type oxygen electrode under saturating illumination conditions. Figure 1D shows that Pmax for the wild type increased with growth irradiance, as is expected upon acclimation of the photosynthetic apparatus to the level of irradiance (Anderson, 1986;Baroli and Melis, 1996). On the contrary, Pmax for the rep27 declined with growth irradiance and dropped below the isosbestic point when grown at 150 mE, consistent with the repair-aberrant phenotype of this mutant. Under higher than 150 mE growth irradiance conditions, photosynthetic O 2 evolution in the rep27 mutant became negligible, and the only gas exchange reaction recorded was that of cellular respiration that consumed oxygen from the medium. These results are consistent with the notion that, at irradiance levels greater than 150 mE, the rate of photodamage was substantially greater than that of de novo PSII biogenesis/assembly, resulting in a quantitative PSII photoinhibition. Survival of the organism under these conditions is ensured by the presence of acetate, which supports oxidative phosphorylation by the cells.
The F v /F m ratio of the rep27 mutant did not exceed 0.4, irrespective of whether cells were grown under low light or in the dark. At present, it is not clear why the rep27 cells have a lower than wild type F v /F m ratio. The possibility cannot be excluded that, although PSII biogenesis/de novo assembly and the PSII repair processes occur in different regions of the chloroplast thylakoid membrane and likely involve substantially different mechanisms, they may nevertheless utilize at least some common enzymatic reactions, possibly including the one catalyzed by the REP27 protein. Further work is needed to delineate these points.
Biochemical Analysis of Wild-Type and rep27 Strains Figure 2A (D1) shows that the rep27 mutant has lower than wild-type steady-state levels of the D1/32-kD reaction center protein in its thylakoid membranes, consistent with earlier finding on this mutant (Zhang et al., 1997). Figure 2A also shows that subunits of the PSII holocomplex, other than the D1 reaction center Table I. Differential growth, irradiance screening, and functional characterization of the C. reinhardtii wild-type and rep27 mutant strains Cell growth was tested in TBP or TAP as the sole carbon source. Chl and functional photosystem content was measured spectrophotometrically, as described (Melis, 1989 (e.g. D2, CP47, and PsbO), occur in equivalent amounts in the wild type and rep27 mutant. This observation offers additional evidence that de novo biogenesis/assembly of the PSII holocomplex is not impaired in the rep27 mutant. Rather, the effect of the mutation appears to be specific to the replacement (turnover) of the photodamaged D1. Figure 2B shows that subunits of non-PSII complexes in thylakoids, e.g.
Cyt b6-f, PSI, and ATP synthase, also occur in equivalent amounts in wild type and rep27 mutant. This observation strengthens the interpretation that biogenesis/ assembly of the thylakoid membrane and the electron transport chain are not impaired in the rep27 mutant. Somewhat enhanced HSP70B protein accumulation was observed in the rep27 mutant relative to that in the wild type (Fig. 2B). This could be attributed to the apparently inhibited and, therefore, unfinished repair in the rep27 mutant (accumulation of the PSII repair intermediate; Yokthongwattana et al., 2001).
Uncoupled psbA Transcription and D1 Translation in the rep27 Mutant One possible scenario for the presence of lower steady-state levels of the D1 protein in the rep27 mutant might be a defect in the incorporation of nascent D1 in the photodamaged PSII core complex. This hypothesis requires that psbA gene transcription and mRNA levels are about the same in wild type and rep27 mutant. Steady-state levels of psbA and psbD mRNA were measured by northern-blot analysis (Fig. 3A), showing that psbA mRNA levels in mutant and wild type were indistinguishable from each other. Interestingly, rep27 showed somewhat enhanced steady-state levels of psbD mRNA relative to the wild type, underlining the greater relative amounts of D2 protein in the mutant ( Fig. 3A) relative to that in the wild type.
The preceding suggested that a lower level of the D1 protein in the rep27 mutant, relative to that in the wild type, may be a consequence of the unimpeded degradation of photodamaged D1, coupled with the chloroplast inability to replace the lost D1 in the PSII-core complex. It should thus be possible to delineate between the de novo biogenesis/assembly of PSII from the selective D1 turnover in the chloroplast of the rep27 mutant. This was tested by comparative [ 35 S]-sulfate pulse-labeling experiments with the cw15 wild type and rep27 mutant (Fig. 3B). In such experiments, cycloheximide was added just prior to the application of the radioactivity to inhibit synthesis of cytosolic proteins. Figure 3B shows about equivalent amounts of large subunit of Rubisco, ATP synthase a-subunit (CF1-a), and D2 protein that were labeled in wild type and mutant. These provide a measure of the de novo biogenesis/assembly and accumulation of chloroplast components as a function of cell growth in the presence of acetate. Interestingly, in the rep27, but not in the wild type, radiolabeled D1 and D2 accumulated in about equal amounts, evidence that de novo biogenesis/ assembly of D1 is not affected in the mutant. However, the wild type showed a much greater accumulation of radiolabeled D1 than the mutant, consistent with the active turnover of this PSII reaction center protein.
These results are evidence of selective inhibition of the D1 turnover/PSII repair and lack of impairment in Figure 2. A, Western-blot analysis of 50-mE grown C. reinhardtii wild type and rep27 total protein cell extract. The steady-state level of PSII subunits was determined from the intensity of the antibody crossreaction in the western blots with specific polyclonal antibodies generated against D1, D2, CP47, and PsbO, respectively. B, Western-blot analysis of 50-mE grown wild type and rep27 total protein cell extract. The steady-state level of these thylakoid membrane proteins was determined from the intensity of the antibody cross-reaction in the western blots with specific polyclonal antibodies generated against PsaK, Cyt f, PetC, AtpA, and Hsp70B. Figure 3. A, Northern-blot analysis of total RNA from 50-mE grown wild-type and rep27 mutant strains separated onto 1.5% agarose gel. Following transfer to a nylon membrane, samples were treated with radiolabeled psbA and psbD cDNA. The C. reinhardtii rRNA loading control is shown at the bottom. B, Autoradiogram of [ 35 S]-radiolabeled proteins in C. reinhardtii. Chloroplast protein biosynthesis was conducted in vivo under 50 mE illumination for a period of 10 min (pulse) in the presence of [ 35 S]-Na 2 SO 4 . [ 35 S]-radiolabeled proteins were visualized upon exposure of the SDS-PAGE onto a radioactivity detector. The protein positions corresponding to D1, D2, Rubisco large subunit (LSU), and ATP synthase a-subunit (CF1) are denoted. Minor rep27-specific protein bands are denoted by asterisk, while black circles denote wildtype-specific bands.
the de novo biogenesis/assembly of the PSII complex during cell growth in the mutant. In sum, except for the remarkable disparity of D1 radiolabeling in wild type and rep27, the labeling of other PSII and thylakoid membrane proteins was found to occur in equivalent amounts in the two strains.
Deletion of Genomic DNA upon pJD67 Plasmid DNA Insertion in rep27 Backcrosses of rep27 (mt1) and strain CC1068 (mt2) revealed that the repair-aberrant rep27 phenotype cosegregated with the ARG7 tag that was employed to generate the insertional mutagenesis library (data not shown). To determine the copy number of pJD67 plasmid insertions in rep27, Southern-blot analysis with the wild type and rep27 genomic DNA was carried out, as follows. ARG7-NdeI primer set (Table II) was used to PCR amplify a 0.75-kb NdeI/NdeI DNA fragment, derived from the 3# end of the ARG7 plasmid, as indicated in Figure 4. This DNA fragment of the Arg-7 gene was employed as a probe in Southern-blot analysis of wildtype and rep27 genomic DNA, digested with NcoI or HpaI restriction enzymes. Figure 4A shows a single hybridization band between the probe and the wildtype genomic DNA corresponding to the endogenous inactive ARG7 gene. The rep27 showed polymorphism with two distinct hybridization bands between probe and mutant genomic DNA. In addition to the endogenous ARG7 gene, a single extra DNA band was observed, confirming a single pJD67 plasmid insertion in the rep27 mutant genome ( Fig. 4A; see also Yokthongwattana and Melis, 2006).
Southern hybridization and PCR analyses were conducted to test for the presence of the pBluescript origin of replication and ampicillin resistance, to be used in a plasmid rescue effort for the cloning of the genomic DNA flanking the pJD67 insertion site (Gumpel and Purton, 1994). Such analysis revealed that a significant portion of the 5# end of the plasmid was deleted, including the important ori and amp sites (Fig. 4B). Deletion of the 5# end of the transforming linearized plasmid is a common occurrence in DNA insertional mutagenesis (Tetali et al., 2007). As an alternative approach by which to clone the genomic DNA flanking the pJD67 insertion site, a thermal asymmetric interlaced (TAIL)-PCR was employed with rep27 genomic DNA (Liu et al., 1995;Dent et al., 2005). TAIL-amplified PCR products were blasted against the Ver. 3.0 annotated C. reinhardtii genomic database (http://genome. jgi-psf.org/Chlre3/Chlre3.home.html). We found that the 3# end of the pJD67 plasmid interrupted the coding sequence of an exostosin-like gene (open reading frame [ORF]0 in Fig. 4B) at contig 25 on scaffold 8.  To define the opposite (5# end) site of the pJD67 insertion locus, we designed six marker primer sets for PCR-based diagnostic analysis (RT0-RT5; Fig. 4B; Table  II). Reverse transcription (RT)-PCR results with the RT0 to RT5 set of primers revealed that a 25-kb fragment of the C. reinhardtii genomic DNA, encompassing the RT1-RT4 loci, was missing and apparently deleted during the pJD67 insertional event (Fig. 4B). Four ORFs were annotated in the missing fragment of the Chlamydomonas genomic DNA. Of those, ORF1 (Fig.  4B, shaded) turned out to be the gene necessary and sufficient to alleviate the acetate-requiring phenotype of the rep27 strain (please see below).
The TAIL-PCR-amplified genomic DNA flanking the insertion site was subsequently used as a probe for the screening of a C. reinhardtii bacterial artificial chromosome (BAC) genomic library, obtained from the Clemson University Genomics Institute (https://www.genome. clemson.edu/). Four BAC clones (27N13, 10C5, 10A5, and 10K12) hybridized to the TAIL-PCR-amplified genomic DNA fragment (data not shown). Restriction patterns of these BAC clones showed that they were identical to each other and contained the missing 25-kb genomic DNA fragment of rep27 (data not shown).
Complementation of rep27 with BAC Clone and Subcloned Genomic DNA BAC clone 10A5 was used to complement the rep27 acetate-requiring phenotype. Cotransformation of the rep27 mutant with the BAC clone 10A5 DNA was implemented with the pSL18 plasmid, conferring paramomycin resistance (Depege et al., 2003). Transformants were selected on TAP-agar plates containing 5 mg/mL paramomycin. To test whether BAC clone 10A5 com-plements the acetate-requiring phenotype of rep27, each BAC transformant (approximately 200 colonies in total) was replica plated onto paramomycin containing TAPagar and TBP-agar media. One such transformant, which showed robust and unequivocally photoautotropic growth in TBP media (Fig. 5A, BAC-T), was recovered and further analyzed by PCR (Fig. 5B). A randomly selected BAC clone 10A5 transformant (BAC-C), which did not grow on TBP media, was chosen as a negative control for the molecular and PCR analysis. Genomic DNA from BAC-T and BAC-C were used as templates in these PCR amplifications. Figure 5B shows that RT0 and RT5 primer sets, probing a DNA region outside the deleted 25 kb in the rep27 strain, generated PCR products, consistent with the presence of this genomic DNA in the mutant. The RT1 primer set failed to generate PCR products in either BAC-T or BAC-C transformants, suggesting that ORF0 may not be responsible for the autotrophic growth properties of the BAC-T strain. RT2 primer set successfully generated PCR products from the BAC-T genomic DNA only, but not from the BAC-C DNA, suggesting that ORF1 is present in the BAC-T and might encode a protein that confers autotrophic growth properties of the BAC-T strain (Fig. 5B). RT3 primer set successfully generated PCR products in both BAC-Tand BAC-C transformants. Conversely, RT4 primer set generated PCR products in the BAC-C but not in the BAC-T transformant, suggesting that ORF3 may not be responsible for the autotrophic growth properties of the BAC-T strain. These results point to ORF1 as the likely gene candidate that confers autotrophic growth properties to the rep27 mutant. The results also provide testimony to the complex nature and unpredictable outcome of C. reinhardtii complementation by BAC clone DNA. Because the mutant phenotype of rep27 was successfully complemented by BAC clone 10A5, PstI partially digested 10A5 DNA fragments were ligated onto the pBR322/PstI vector to isolate subclones, which contained the various deleted ORFs (ORF0-ORF3). Four different subclones, each containing ORFs from the deleted 25-kb region, were identified by DNA hybridization and direct sequencing analysis (Fig. 4B, subclones 1A, 8, 4A, and 2A). Transformants of the rep27 mutant with each of these subclones were tested for autotrophic growth on TBP minimal media. Only subclone 4A, containing a 6.2-kb PstI partial-genomic DNA fragment that includes ORF1, successfully rescued the acetate-requiring phenotype of the rep27 mutant. Transformation of the rep27 mutant with any other subclone (1A, 8, and 2A) failed to rescue the acetate-requiring phenotype of rep27 (data not shown). Accordingly, ORF1 was designated as the putative REP27 gene.

The REP27 Gene Product Is Required for the PSII Repair
For further molecular and biochemical analysis, three independent transformant lines with subclone 4A were selected (rep27-complemented [rep27-comp] strains). In the subsequent physiological and biochemical analyses, and to alleviate crowding and enhance clarity of the presentation, results from only one line are reported. However, all three lines displayed similar properties to those of the rep27-comp strain shown.
RT-PCR with the RT2 primer set confirmed the presence of the REP27 transcript in the rep27-comp strains (data not shown). Because the rep27-comp strains were selected on the basis of growth on TBPagar minimal media, which lack acetate, it was initially assumed that transcription of REP27 is necessary and sufficient to rescue the acetate-requiring phenotype of the rep27 mutant. To further test this hypothesis, the F v /F m fluorescence yield ratio and photosynthetic activity in wild-type, rep27 mutant, and rep27-comp strains were measured under a variety of growth conditions. A phenotype of the rep27 mutation is the lower-thanwild-type F v /F m ratio under all growth irradiance conditions (Fig. 1C), consistent with the interpretation that the rep27 strain is a PSII repair-aberrant mutant (Zhang et al., 1997). The effect of light-shift experiments on the F v /F m value of the rep27-comp strains was compared to that of the wild type and rep27 mutant (Fig. 6A). Cells grown under 10 mE were transferred to approximately 1,000 mE. Chl fluorescence yield parameters were measured as a function of exposure time to 1,000 mE (Fig. 6A). Prior to the light shift, F v /F m for the wild type and rep27-comp strain was high (about 0.78), whereas it was considerably lower for the rep27 mutant (0.37). F v /F m for rep27 decreased to nearly zero within 1 h under a 1,000-mE irradiance, whereas the F v /F m for the wild type and rep27-comp strains declined to about 0.45 in the same period of time. These results are consistent with the notion that rep27 is a PSII repair-aberrant mutant and that complementation of this strain with a wild-type copy of the REP27 gene rescues the mutation.
The activity of photosynthesis and the response of wild-type, rep27, and rep27-comp strains to irradiance were further evaluated from the measurement of the light-saturation curve of photosynthesis (Fig. 6B). In the wild-type and rep27-comp strains, photosynthetic activity increased as a function of light intensity with identical slopes (quantum yield of photosynthesis) and reached a Pmax of approximately 40 mmol O 2 (mol Chl) 21 s 21 at about 1,000-mE light intensity. This pattern was nearly indistinguishable among wild-type and rep27-comp strains and independent of growth irradiance in the 10 to 50 mE region. Contrary to the wild-type and rep27-comp strains, the rep27 mutant showed a quantum yield of photosynthesis (initial slope in the photosynthesis versus irradiance curve) that was only about 50% of that in the wild-type and Figure 5. A, Growth of wild type, rep27 mutant, and BAC clone 10A5 complemented rep27 strains on TBP minimal media under 50-mE irradiance. BAC clone 10A5, which contains the 25-kb deleted genomic DNA fragment, successfully rescued the acetate-requiring mutant phenotype in rep27 (BAC-T). BAC-C was randomly selected as a negative control for BAC-T. B, PCR characterization of BAC-Tand BAC-C genomic DNA. PCR amplification of BAC-T and BAC-C was performed with a variety of different marker primer sets (RT0-RT5), as indicated. the rep27-comp strains (Fig. 6B), consistent with the approximately 50% lower F v /F m and with the notion of a state of chronic photoinhibition in the mutant. It is also of interest to observe that a lower Pmax of approximately 15 mmol O 2 (mol Chl) 21 s 21 in the rep27 mutant was reached at about 1,000-mE light intensity. Upon exposure to progressively higher light intensities, Pmax declined in the rep27, a consequence of the progressively faster rates of PSII photodamage coupled with the mutant's inability to complete the repair reactions. It is also of interest to note that this photoinhibitory effect was more pronounced with the mutant strain grown at 50 rather than 10 mE (Fig. 6B, white versus black triangles).
The preceding analysis suggested that rep27 undergoes a normal biogenesis/assembly of functional PSII units when grown in the presence of acetate, but cannot repair them when PSII photodamage occurs. There is thus a competition between the de novo biogenesis/assembly of functional PSII units in the presence of acetate and the irreversible inhibition of PSII due to photodamage in the rep27 mutant. This interpretation requires that either photodamaged and inactive D1 accumulates in the thylakoid membrane or, if the lesion is in a step subsequent to inactive D1 degradation, a depletion of the D1 protein in the mutant relative to that in the wild type. To delineate between these two different mechanistic alternatives, western-blot analysis was applied to assay steadystate levels of D1 protein in wild-type, rep27, and rep27-comp strains (Fig. 6C). It is seen that rep27 has substantially lower steady-state amounts of the D1 protein relative to the wild type, whereas the rep27-comp strain displayed levels of D1 that were comparable to those of the wild type. These results strengthened the notion that the rep27 lesion does not prevent degradation of the photodamaged and inactive D1, but it does interfere with the next step, i.e. the biosynthesis and/or cotranslational insertion of the nascent D1, thereby impeding the completion of the D1 protein turnover and the PSII repair process.

REP27 Gene Identification and Structure
To identify the transcription start site of the REP27 gene, 5# RNA ligase mediated-RACE was employed and resulted in the identification of 455 coding region nucleotides, in addition to those reported by the U.S. Department of Energy Joint Genome Institute (JGI) for ORF1 (Fig. 4). These occurred in the upstream region of the JGI-reported ORF1 translation initiation site. A fulllength REP27 cDNA was subsequently isolated by RT-PCR amplification using 27P-F3 and 27E11-R1X primer set (Table II), which confirmed the extended REP27 coding sequence. The REP27 cDNA nucleotide sequence was deposited in the GenBank with accession number EF127650, predicted to encode a 449-amino acid putative precursor protein (Fig. 7A), with a molecular mass of 49.5 kD for the precursor polypeptide and 45.6 kD for the mature protein. The REP27 genomic Figure 6. Physiological and biochemical analyses of C. reinhardtii wildtype, rep27, and rep27-comp strains. A, Strains were grown on TAP under 10 mE and transferred to 1,000 mE for different periods of time. F v /F m was plotted as a function of incubation time in 1,000 mE. Values shown are the average of two independent experiments. B, The light-saturation curves of photosynthesis in wild-type, rep27, and rep27-comp strains grown in TAP under 10-or 50-mE conditions. Registration of the dark respiration in the cell suspension was followed by 5-min measurements of the rate of oxygen evolution at 500, 1,000, 1,500, 2,000, and 2,500 mE. The values shown are the average of two independent experiments. C, Western-blot analysis of total cell protein extracts from wild-type, rep27, and rep27-comp strains loaded on an equal Chl basis. Proteins were probed with specific anti-D1 polyclonal antibodies.
DNA was also predicted to contain a minimum of 11 exons and 10 introns.
REP27 appears in the JGI Chlamydomonas Genome Project database as an unknown gene whose product contains a single tetratricopeptide repeat (TPR) motif with no assigned function. However, our complete cDNA sequence showed that this gene also contains a plastid-targeted transit peptide and includes an additional TPR motif, both of which are missing from the JGI annotation. In agreement, database searches revealed the occurrence of REP27 orthologs in the genomic DNA of other plant species, including the transit peptide and the two pairs of the TPR domains, which appear to be highly conserved (Fig. 7A).
Amino acid sequence analysis by ChloroP1.1 software (Emanuelsson et al., 1999) predicted that REP27 contains a chloroplast transit peptide. Similarly, application of the HMMTOP v2.0 software (Tusnady and Simon, 2001) pointed to the existence of two putative transmembrane helices in REP27 (amino acid regions Figure 7. Deduced amino acid sequence alignments (ClustalW analysis; Higgins et al., 1994) of REP27 and its homologs. A, Database searches revealed that the deduced amino acid sequence of rice (Os01g0358300), Arabidopsis (At1g02910), and Ostreococcus (CAL55849) gene products are orthologs to that of REP27. Asterisks mark absolutely conserved residues among the four protein sequences. High and low similarity residues are marked by double or single dots, respectively. Vertical arrow indicates the predicted cleavage site of the chloroplast transit peptide targeting sequence. Solid lines and dotted lines over the amino acid sequences indicate the TPR motifs and predicted transmembrane helices, respectively. B, Phylogenetic tree showing the relative evolutionary proximity of the REP27 ortholog (Os01g0358300, At1g02910, and CAL55849) and paralog proteins ( TMH1 and TMH2 including amino acids 180-202 and 217-239, respectively; Fig. 7A), which may suggest that REP27 is a thylakoid membrane integral protein.
Application of Inter-ProScan 13.1 software analysis (Quevillon et al., 2005) identified two distinct TRP motifs in REP27 (TPR1A/TPR1B and TPR2A/TPR2B; Fig. 7A). The helix-rich TPR motifs have been described in the literature to form a compact unit of two helices interacting with each other in the antiparallel direction. It is thought that such conserved TPR motifs are involved in a variety of critical protein-protein interactions in the living cell (for review, see Lamb et al., 1995) and are consistent with the putative role of the REP27 in the D1 turnover and PSII repair.
A search of the National Center for Biotechnology Information database revealed the occurrence of six REP27 homologs, including two copies each in rice (Oryza sativa), Arabidopsis (Arabidopsis thaliana), and Ostreococcus, a small unicellular green alga whose genomic DNA sequence and annotation recently became available (Derelle et al., 2006). Moreover, a search of the C. reinhardtii genomic DNA database revealed the occurrence of a REP27 homolog in this green alga. Accordingly, it appears that two copies of REP27 homologs are present in each of these photosynthetic organisms. These can be divided into two distinct groups: REP27 orthologs (Os01g0358300, rice; At1g02910, Arabidopsis; and CAL55849, Ostreococcus) contain the chloroplast transit peptide and two TPR motifs (Fig.  7A), while REP27 paralogs (Os04g0507100 for rice; At2g28740 for Arabidopsis; CAL58275 for Ostreococcus; and C_142189 for Chlamydomonas) are devoid of both a transit peptide and the TPR domains (data not shown). REP27 had about a 35% amino acid sequence identity with its orthologs (CAL55849, Os01g0358300, and At1g02910). However, it had limited similarity and that was only in the C-terminal half of the protein with the paralogs, which were devoid of transit peptide and TPR motifs.
The REP27 protein appears to exist only in granacontaining organisms of oxygenic photosynthesis. The phylogenetic comparison diagram in Figure 7B shows that REP27, CAL55849, Os01g0358300, and At1g02910 are grouped together and, therefore, closely related. Likewise, REP27 paralog proteins Os04g0507100, At2g28740, CAL58275, and C_142189 are closely related among them and distantly related to the former. This may suggest a distant phylogenetic relationship between the two groups of genes and proteins and points to the possibility that REP27-like genes in photosynthetic eukaryotes exist in pairs, one copy encoding a putative chloroplast-targeted protein, the other encoding a nonplastid counterpart.
Induction of REP27 Gene Expression by Irradiance and Requirement for D1 Protein Turnover But Not for de Novo PSII Biogenesis/Assembly To test whether expression of the REP27 gene is constitutive or induced by irradiance, steady-state REP27 mRNA levels were measured before and after a shift in light intensity. Northern-blot analysis with a probe specific for the REP27 gene showed that 1,000 mEtreated wild-type cells had about 5 times more transcript accumulation than that of 50 mE-acclimated control cells (Fig. 8A). This observation is consistent with the well-documented, irradiance-induced increase in repair gene expression, including ELIP/Cbr (Fig. 8B) and HSP70B . Such increase in gene expression is though to be a meaningful response of the cells to the increased rates of photodamage and the need for enhanced rates of repair with increasing irradiance (Melis, 1999;Yokthongwattana and Melis, 2006).

DISCUSSION
It is known that specific D1-less (DpsbA) or D2-less (DpsbD) mutants fail to assemble the entire PSII core complex (Bennoun et al., 1986;Erickson, 1986;Cohen et al., 2001;Baena-Gonzalez et al., 2003). The rep27 is distinct from such mutants, as it undergoes biogenesis/ assembly of a functional PSII holocomplex but fails to perform the D1 reaction center protein turnover. Therefore, under weak growth irradiance, rep27 displays a limited water oxidation and O 2 evolution capacity. Under moderate to high rates of photodamage, PSII activity ceases to exist.
Due to incomplete annotation of the C. reinhardtii genomic DNA sequencing in the REP27 locus, the true transcription initiation site for this gene was not available. 5# RACE PCR was performed in this work, resulting in the proper identification of the transcription initiation site and translation start locus of REP27, which is 261 bp upstream from the tentative translation start site reported by the JGI. A search of the GenBank database identified a pair of REP27 homologs in photoautotrophic eukaryotes, including Chlamydomonas, Arabidopsis, rice, and Ostreococcus (Fig. 7). REP27 orthologs contained both the transit peptide for chloroplast targeting of the protein, the two apparently conserved TPR motifs, and the two transmembrane domains. REP27 paralogs lacked both the transit peptide for chloroplast targeting of the protein and the two apparently conserved TPR motifs and are apparently cytosolic proteins.
An Arabidopsis T-DNA mutant was isolated in which the REP27 ortholog gene was interrupted by the T-DNA insertion. This mutant, termed lpa1 for low PSII accumulation (Peng et al., 2006), had many characteristics similar to those reported in this work for the rep27 mutant. However, contrary to the conclusions reached in this work, Peng et al. (2006) assigned a PSII biogenesis function to the LPA1 protein in Arabidopsis, as opposed to the specific D1 turnover function of the REP27 gene in this work.
TPR motif-containing proteins are widely encountered among a variety of organisms, including nonphotosynthetic bacteria, cyanobacteria, yeast (Saccharomyces cerevisiae), fungi, plants, animals, and humans (Blatch and Lassle, 1999). The number of TPR motifs is also variable, and no position preference exists with respect to their location along the primary polypeptide sequence. TPR domains are highly degenerate 34-amino acid repeats that are often arranged in tandem arrays, known to mediate protein-protein interactions (Lamb et al., 1995). Since the discovery of a TPR motif in soybean (Glycine max), several cases of TPR-containing proteins in eukaryotic chloroplasts have been reported and characterized (Hernandez Torres et al., 1995). For example, the subunits of the translocation machinery in the chloroplast envelope (Tic40 and Toc64) contain a single and three TPR motifs, respectively (Sohrt and Soll, 2000;Chou et al., 2003). Recent studies revealed that TPR domains in Toc64 are responsible for the recognition of precursor proteins targeted to the chloroplast (Qbadou et al., 2006). Moreover, a TPR motif in Tic40 is believed to form a cochaperone with HSP70B and/or ClpC (Chou et al., 2003), possibly helping in the translocation of proteins into the chloroplast and/or cleavage of the transit peptide sequence from the imported precursor protein.
The work presented in this article suggests that the REP27 protein is essential for the D1 reaction center protein turnover, probably facilitating translation and/ or insertion of the nascent D1 in the vacated PSII reaction center template. It is of interest to speculate about the mechanistic role of TPR motifs in this process. TPR motifs are thought to be important factors in both transcription and translation processes. For example, Chlamydomonas Mbb1 and its Arabidopsis ortholog, HCF107, contains 10 and 11 TPR motifs, respectively, thought to be responsible for the mRNA stability and translation initiation of the plastidic psbH (Vaistij et al., 2000;Sane et al., 2005). NAC2, which has nine TPR motifs, is required for stable accumulation of the psbD mRNA in Chlamydomonas (Boudreau et al., 2000). A regulator of Chl biosynthesis, FLU, harbors two TPR motifs in its C terminus, which are required for interaction with glutamyl-tRNA reductase (Meskauskiene et al., 2001;Meskauskiene and Apel, 2002). A novel chloroplast protein, TCP34, with three TPR motifs, was localized in the fraction of transcriptionally active chloroplast chromosome, indicating that TPR motifs possessed the ability to bind plastid DNA (Weber et al., 2006). The Ycf3 mutant analysis also showed that TPR motifs are involved in the PSI stability via a direct interaction with PsaA and PsaD (Naver et al., 2001).
It is well known that molecular chaperones, such as the cytoplasmic HSP70 and HSP90, interact with a number of cochaperones that contain TPR motifs (for review, see Blatch and Lassle, 1999). In the chloroplast, Tic40 may bind to chloroplast ClpC and HSP70B (Chou et al., 2003). In the green algae Dunaliella salina, the presence of chloroplast-localized HSP70B was detected as a component of the transiently forming 320-kD PSII repair intermediate Melis and Nemson, 1995;Yokthongwattana et al., 2001). It is possible (currently under investigation) that REP27 is also a component of this PSII repair intermediate, in a structure that may serve to preserve the conformation of the disassembled PSII core complex, from which the D1 protein has been temporarily vacated. Therefore, a working hypothesis is that the TPR motifs contained by the REP27 protein act as a receptor of the nascent D1 and are required for the proper guidance and insertion of the de novo-synthesized D1 protein in the existing PSII template. The REP27 and HSP70 proteins may act coordinately in the scaffold of the PSII repair intermediate in Figure 9. Schematic of the putative function of the REP27 protein in the D1 protein turnover and PSII repair process. REP27 is shown to play a role in the cotranslational insertion of a nascent D1 (pD1) in the D1-less PSII template during the PSII repair cycle. This step follows the PSII disassembly and specific degradation of photodamaged D1 in the thylakoid membrane. Figure 8. A, Northern-blot analysis probed with C. reinhardtii REP27specific DNA. Wild-type cells, grown under 50-mE conditions (L), were incubated under high light (1,000 mE, H) for 1 h. Agarose gel lanes were loaded with 4 or 8 mg of total RNA from the 50 mE-grown (L) and 1,000 mE-exposed (H) cells, transferred to nylon membrane, and probed with a radiolabeled full-size REP27 cDNA probe. A (bottom), rRNA loading control. B, Northern-blot analysis probed with C. reinhardtii ELIPspecific DNA. Wild-type and rep27 mutant grown under 50-mE conditions were incubated under high light (1,000 mE) for 1 h. Agarose gel lanes were loaded with 4 mg of total RNA from the 50 mE-grown (L) and 1,000 mE-exposed (H) cells, transferred to nylon membrane, and probed with a radiolabeled full-size ELIP cDNA probe. Ribosomal RNA loading controls are shown at the bottom. a process that facilitates the stability of the D1-less PSII complex and guides psbA cotranslational insertion of the de novo-synthesized D1 protein in the PSII template. Figure 9 shows a preliminary schematic, aiming to illustrate the step(s) in the PSII repair process and the possible mechanism of action for the REP27 protein.
Consistent with the evidence presented in this work, de novo biogenesis/assembly of PSII in the presence of acetate is possible in C. reinhardtii, leading to the formation of functional PSII. Photodamage by excess irradiance leads to irreversible inhibition in the function of the D1 protein, occurring in the appressed thylakoid membranes of the chloroplast. PSII disassembly, followed by degradation of the inactive D1, takes place in stroma-exposed thylakoids (Melis, 1991(Melis, , 1999Ossenbuhl et al., 2002), leading to a D1 vacancy from the PSII reaction center complex. The latter serves as a template for the insertion of a de novo-synthesized nascent D1 protein, a process catalyzed by the REP27 protein. According to this working hypothesis, the REP27 protein is essential for the incorporation of the nascent D1 but not for the de novo biogenesis/ assembly of the PSII holocomplex. This model is consistent with the psbA mRNA association with thylakoid membranes, which is accentuated under conditions of enhanced photodamage and repair (Kettunen et al., 1997), underlining a cotranslational assembly of the D1 protein during the PSII repair (Zhang et al., 1999(Zhang et al., , 2000. In the rep27 mutant, incorporation of the nascent D1 in the PSII template during the D1 turnover process is inhibited, leading to the phenomenology described in this work.

Strains and Media
Chlamydomonas reinhardtii mutants were generated by transformation (Kindle, 1990;Tam and Lefebvre, 1993) of an Arg auxotroph, strain CC-425 (arg7-8 cw15 mt 1 sr-u-2-60; Chlamydomonas Center, Duke University), with pJD67 plasmid DNA (linearized with HindIII) containing the complementing argininosuccinate lyase gene (ARG7), as described by Davies and Grossman (1994) and Davies et al. (1996). Procedures for the screening, isolation, and maintenance of putative repair mutants, including the rep27 strain, have also been described (Zhang et al., 1997). Wild type and rep27 of the green alga C. reinhardtii were grown mixotrophically in a TAP medium (Gorman and Levine, 1965), either in liquid cultures or on 1% agar plates. To test photoautotrophic growth of wild-type and complemented strains, cells were grown on TBP minimal media, in which sodium bicarbonate (25 mM, pH 7.4) replaced the acetate as the growth carbon source (Polle et al., 2003).
Culture density was measured by cell counting using a Neubauer ultraplane hemacytometer and a BH-2 light microscope (Olympus). Cells were grown to the early exponential growth phase (about 1-2 3 10 6 cells/mL) prior to harvesting and measurements of photosynthesis.
Chlamydomonas BAC genomic DNA library filters and clones were acquired from the Clemson University Genomics Institute (https://www. genome.clemson.edu/).

Nucleic Acid Isolation and Analyses
To purify the genomic DNA of C. reinhardtii, cells were grown in liquid cultures, harvested by centrifugation, and resuspended in a cetyl-trimethylammonium bromide buffer containing 2% cetyl-trimethyl-ammonium bromide, 100 mM Tris-HCl, pH 8.0, 1.4 M NaCl, 20 mM EDTA, and 2% b-mercaptoethanol.
Total RNA was isolated from the isopropanol pellet of the cell extract using Invitrogen's Trizol Reagent and by following the manufacturer's recommended procedure. 5# RACE by PCR was performed using FirstChoice RNA ligase mediated-RACE kit (Ambion) by following the manufacturer's recommended procedure.
Southern-and northern-blot analyses were carried out according to standard protocol (Sambrook et al., 1989). The high light induction of REP27 gene expression was tested upon shifting cultures from 50 to 1,000 mmol photons m 22 s 21 for 1 h prior to RNA isolation and by using the full size of the cDNA fragment as a probe in the measurement of transcript abundance.
Cloning of the Flanking Regions of the pJD67 Insertion Site C. reinhardtii genomic DNA flanking the plasmid insertion site was amplified using a TAIL-PCR procedure, optimized for Chlamydomonas genomic DNA, by a modification of the method described (Dent et al., 2005). The primers used for the TAIL-PCR are listed in Table II. Briefly, flanking genomic DNA was amplified by PCR from the region adjacent to the inserted pJD67 plasmid that was used for DNA insertional mutagenesis (Zhang et al., 1997). For HindIIIdigested pJD67 transformants, specific primers for primary, secondary, and tertiary reactions were designed and named TPB1, TPB2, and TPB3, respectively (Table II). Two arbitrary degenerate primers were tested for amplification, TPdeg1 and TPdeg2, as previously described (Dent et al., 2005). The general TAIL-PCR protocol of Liu et al. (1995) was used with some minor modifications for the various PCR amplification reactions. Nucleotide sequences of the resulting PCR products were obtained via an ABI3100 sequencer. Chlamydomonas genomic sequence information was obtained from the Chlamydomonas Genome Project Web site (http://genome.jgi-psf.org/ Chlre3/Chlre3.home.html).

Insertional Mutagenesis and Chlamydomonas Transformation
Generation of a C. reinhardtii insertional mutagenesis library and screening for the identification of specific PSII repair mutants was implemented as previously described (Zhang et al., 1997). Complementation of the rep27 putative repair mutant was achieved by cotransformation of the mutant with BAC clone 10A5 and with subclones derived from this BAC with plasmid pSL18 (Depege et al., 2003), using the conventional glass bead transformation method (Kindle, 1990). pSL18 contains the paromomycin resistance gene (selectable marker) operated under the control of the C. reinhardtii Hsp70A and RbcS2 promoters (Sizova et al., 2001) and linked to the PsaD promoter and terminator that can be used to express ORFs in C. reinhardtii (Depege et al., 2003). C. reinhardtii transformants carrying the above-mentioned BAC clone(s) were selected on TAP (acetate-containing) or TBP (minimal media) plates containing 5 mg mL 21 paromomycin (Sigma Chemical).

Photosynthesis Measurements
Procedures for the measurement of the F v /F m variable-to-maximal fluorescence yield ratio and for the Pmax have been described (Zhang et al., 1997). To test for the effect of high light exposure on photosynthesis and PSII parameters, cw15, rep27, and rep27-comp strains were incubated under 1,000 mE for variable periods of time. The photosynthetic activity of strains cw15, rep27, and rep27comp lines was estimated from the light-saturation curve of photosynthesis, measured as the oxygen evolution activity of the cells at different actinic intensities. Measurements of the light-saturation curve of photosynthesis commenced with the registration of dark respiration in the cell suspension and were followed by measurements of the rate of oxygen evolution at 500, 1,000, 1,500, 2,000, and 2,500 mE. Registration of the rate (slope) of oxygen evolution at each light intensity step was recorded for about 5 min. Therefore, in the 30-min duration of these measurements, samples were exposed to progressively higher light intensities (0-2,500 mE).