Molecular genetic analysis of plastocyanin biosynthesis in Chlamydomonas reinhardtii.

Five plastocyanin-deficient mutants were identified from a population of UV-mutagenized Chlamydomonas reinhardtii cells. Genetic complementation experiments indicated that four mutants represented alleles at the PCY1 locus (pcy1-2, pcy1-3, pcy1-4, and pcy1-5). Sequence analysis confirmed that two strains, pcy1-2 and pcy1-3, carry a frameshift (−1) and a nonsense mutation, respectively, while strains pcy1-4 and pcy1-5 synthesize an extended protein as a result of read-through mutations at the stop codon. The C-terminal extension does not affect synthesis or processing of the pre-proteins, but the polypeptides are rapidly degraded after the second (lumenal) processing event. The frameshift mutation in pcy1-2 results in loss of Pcy1 mRNA, as noted previously for strain ac208 (pcy1-1), but the abundance of Pcy1 mRNA in strain pcy1-3, which carries a nonsense mutation at codon 26, is unaffected relative to wild-type cells. The decreased abundance of frameshifted Pcy1 mRNA is attributed to increased degradation rather than decreased synthesis, since the mRNAs can be stabilized by treatment of cells with cycloheximide or anisomycin. The fifth strain has a wild-type plastocyanin-encoding gene, but the strain accumulates apoplastocyanin at the expense of holoplastocyanin. We suggest that the mutation identifies a new locus (PCY2) whose function is required for normal holoplastocyanin accumulation. Like ac208 (pcy1-1), several of the new mutants were suppressed spontaneously owing to accumulation of cytochrome c6 (a functional substitute for plastocyanin). The suppressor mutation(s) displayed Mendelian inheritance and segregated independently from the PCY1 locus, which confirms that regulation of Cyc6 expression is not tightly linked to plastocyanin function.


Five plastocyanin-deficient mutants were identified from a population of UV-mutagenized Chlamydomonas reinhardtii cells. Genetic complementation experiments indicated that four mutants represented alleles at the
. Sequence analysis confirmed that two strains, pcy1-2 and pcy1-3, carry a frameshift (؊1) and a nonsense mutation, respectively, while strains pcy1-4 and pcy1-5 synthesize an extended protein as a result of read-through mutations at the stop codon. The C-terminal extension does not affect synthesis or processing of the pre-proteins, but the polypeptides are rapidly degraded after the second (lumenal) processing event. The frameshift mutation in pcy1-2 results in loss of Pcy1 mRNA, as noted previously for strain ac208 (pcy1-1), but the abundance of Pcy1 mRNA in strain pcy1-3, which carries a nonsense mutation at codon 26, is unaffected relative to wild-type cells. The decreased abundance of frameshifted Pcy1 mRNA is attributed to increased degradation rather than decreased synthesis, since the mRNAs can be stabilized by treatment of cells with cycloheximide or anisomycin. The fifth strain has a wild-type plastocyanin-encoding gene, but the strain accumulates apoplastocyanin at the expense of holoplastocyanin. We suggest that the mutation identifies a new locus (PCY2) whose function is required for normal holoplastocyanin accumulation. Like ac208 (pcy1-1), several of the new mutants were suppressed spontaneously owing to accumulation of cytochrome c 6 (a functional substitute for plastocyanin). The suppressor mutation(s) displayed Mendelian inheritance and segregated independently from the PCY1 locus, which confirms that regulation of Cyc6 expression is not tightly linked to plastocyanin function.
Plastocyanin is a copper-binding protein whose function in the Z-scheme of photosynthesis is the catalysis of electron transfer from cytochrome f in the b 6 f complex to P700 ϩ in Photosystem I. The protein is an eight-stranded, anti-parallel ␤-barrel (ϳ100 residues) and is abundant in the lumenal compartment of the chloroplast (several million molecules per cell). Its single copper atom, liganded by a cysteinyl thiolate, a methionyl sulfur, and two imidazole nitrogens of histidines, is at the active site of the protein and is responsible for its redox function (reviewed in Ref. 2). The metal also serves to stabilize the folded conformation of the protein (3). Other structural features of the protein (whose functional importance has been confirmed by site-directed mutagenesis) include a pronounced negatively charged patch and a flat hydrophobic surface.
The plastocyanin biosynthetic pathway has served as a model for studies of nuclear gene expression in plants (reviewed in Ref. 4; see also Refs. 5-7), import and sorting of chloroplast proteins (reviewed in Refs. 8 -10), and metal-responsive gene expression (e.g. Refs. 11 and 12; reviewed in Ref. 13). In eukaryotic cells, plastocyanin is encoded by a single nuclear gene (PetE/Pcy1). 1 Expression of PetE in plants is restricted to photosynthetic tissues and is responsive both to light and to (as yet) unidentified signals derived from plastids (e.g. Refs. 14 and 15). These expression characteristics are common to most genes encoding proteins that function in photosynthesis. In some green algae and cyanobacteria, the expression of the Pcy1/petE gene can be regulated by availability of copper (a cofactor that is essential for function of the gene product). In these organisms, if sufficient amounts of copper are available in the micronutrient source, plastocyanin is synthesized and accumulates to the stoichiometry required for photosynthesis. On the other hand, if the organism faces conditions of nutritional deficiency (with respect to copper), plastocyanin accumulation is prevented and an alternate, hemecontaining, cytochrome is induced as a functional substitute. The regulatory components involved in copper-responsive plastocyanin accumulation are not known, but unlike the tissuespecific and light-responsive regulators, these components are expected to be specific for plastocyanin expression.
The protein product of the PetE/Pcy1 gene is a precursor molecule, which is targeted after translation to the chloroplast. Upon import into the chloroplast, the pre-protein is processed to an intermediate form, which serves as a substrate for the ATP/secA-dependent thylakoid import apparatus (16,17). A lumen-facing thylakoid peptidase (18) cleaves the intermediate form to yield the mature polypeptide, which can associate with copper to form the functional holoprotein (19,20). Thus, the post-translational component of plastocyanin biosynthesis includes steps that are common in the biosynthesis of most, if not all, nucleus-encoded lumenal proteins, such as transport of the pre-protein across the envelope membranes into the chloro-* This work was supported in part by Grant 94-37306-0471 from the United States Department of Agriculture National Research Initiative. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
¶ plast (8,21). Transport across the thylakoid membrane, on the other hand, is a step that is shared with only a few other nucleus-encoded lumenal proteins (e.g. OEE1 and polypeptide F of PSI), 2 while copper ligation is expected to be a step that is unique to plastocyanin biosynthesis.
In Chlamydomonas reinhardtii as in other green algae, plastocyanin accounts for the bulk of the copper content of the cell. The green algae appear to lack other abundant copper enzymes such as Cu/Zn-superoxide dismutase (normally found in both the stromal and cytosolic compartments) and polyphenol oxidase (normally found in the thylakoid lumen). The organism thus lends itself to the study of copper-dependent biosynthetic processes in the chloroplast. We were interested in the identification of genes or genetic loci whose products were required in trans for holoplastocyanin formation and its copper-responsive accumulation in the thylakoid lumen. In bacterial experimental systems, a genetic approach proved very fruitful for the identification of components required for the assembly of metalloenzymes. For instance, studies of urease synthesis in Klebsiella aerogenes led to the identification of genes required for nickel transport and metabolism (22), and analysis of PSII mutants of cyanobacteria led to the identification of an operon encoding a putative manganese transporter (23). However, such an approach has not yet been fully exploited in a eukaryotic experimental system with its added level of biosynthetic complexity.
With a view to further dissecting the plastocyanin biosynthetic pathway, we sought to identify non-photosynthetic mutants of C. reinhardtii whose phenotype could be attributed to a deficiency in plastocyanin function. In this work, we describe the identification of five plastocyanin mutants from a collection of UV-induced non-photosynthetic C. reinhardtii strains. Four of the five mutants carry mutations in the Pcy1 gene encoding pre-apoplastocyanin, while the fifth may define a new locus (PCY2) required for plastocyanin function. 1

MATERIALS AND METHODS
Strains and Cell Culture-C. reinhardtii wild-type strain CC125 was obtained from the Chlamydomonas Genetics Center, Duke University, Durham, NC. Strain ac208 (pcy1-1) (see Footnote 1), isolated by Gorman and Levine in 1965 (49), has been characterized previously (24). Five plastocyanin-deficient mutants were generated by UV mutagenesis of strain CC125.
Cultures of wild-type strains were grown at 22°C in TAP medium (1) under fluorescent illumination (15-125 mol⅐m Ϫ2 ⅐s Ϫ1 ) with agitation (225 rpm). Mutant strains were grown under the same conditions except that the illumination was always reduced (15-25 mol⅐m Ϫ2 ⅐s Ϫ1 ). For some experiments, the mutant strains were cultured on the laboratory bench at room temperature under fluorescent house lights with occasional swirling, and transferred to an incubator (22°C, 15-25 mol⅐m Ϫ2 ⅐s Ϫ1 , 225 rpm) only 1-2 days before they were collected for biochemical analysis.
UV Mutagenesis and Enrichment for Mutants with Defects in Photosynthetic Electron Transfer-The wild-type strain was grown in TAP medium to a density of 2.8 -22 ϫ 10 6 cells/ml at a light intensity of 125 mol⅐m Ϫ2 ⅐s Ϫ1 , transferred to 15 ϫ 100-mm plastic Petri dishes (20 ml total volume) and exposed to UV irradiation (254 nm, ϳ 2 ϫ 10 2 microwatts⅐m Ϫ2 ) for 1-5 min. The cells were kept in suspension during irradiation by stirring or agitation on a vortex mixer. Samples were removed after the desired amount of irradiation and plated on TAP medium to test for survival (28 -58% after 2 min of irradiation) or subjected to metronidazole treatment to enrich for non-photosynthetic survivors (25).
One ml of irradiated cells (diluted to 1.4 -22 ϫ 10 5 total cells/ml) was plated on thin agar slabs over solid medium containing acetate (1.5% agar in TAP) and allowed to recover in the dark for 24 -48 h. The agar slabs were prepared by pouring 2% agar in TAP medium over sterile 85-mm diameter circles cut out of Miracloth (Calbiochem, San Diego, CA) and placed in sterile 100-mm Petri dishes. The solidified slabs (2 mm depth) were transferred to traditional agar plates. (In some experiments, the cells were illuminated at ϳ 50 mol⅐m Ϫ2 ⅐s Ϫ1 for 24 h prior to metronidazole treatment.) The agar slabs were transferred to fresh plates (1.5% agar in TAP) containing metronidazole (20 mM), incubated for an additional 24 -48 h under illumination (50 mol⅐m Ϫ2 ⅐s Ϫ1 ), transferred back to fresh plates lacking metronidazole, and incubated in either dim light (Ͻ5 mol⅐m Ϫ2 ⅐s Ϫ1 ) or medium light (50 mol⅐m Ϫ2 ⅐s Ϫ1 ) to allow survivors to grow. Colonies (between 15 and 272/plate) were apparent after 2-3 weeks. Since the enrichment procedure was conducted on solid medium with little opportunity for cells to divide after mutagenesis, each colony was expected to represent an independent mutation.
The colonies were suspended in 50 l of minimal medium (lacking acetate) (1) in sterile 96-well microtiter dishes, and tested for their acetate-requiring phenotype by replica-stamping cells on solid (1.5% agar, 150-mm plates) minimal medium or TAP medium (two plates). One TAP plate (used as a master plate) was maintained in dim light, while the other two were incubated in brighter light. Growth was assessed after 2 weeks. Colonies that did not grow on minimal medium or grew much slower than wild-type cells on minimal medium were picked from the master plate, restreaked to isolate single colonies, and rescreened for acetate-requirement.
To test for plastocyanin accumulation, each acetate-requiring strain was grown as a lawn on a Petri dish (100 mm). The cells were collected by scraping them off with a razor blade, and resuspended in 50 l of 10 mM sodium phosphate (pH 7.0). The soluble cell extract was prepared and analyzed as described below.
Genetic Analysis of Plastocyanin-deficient Mutants-Each mutant strain was back-crossed to the wild-type strain (CC124). The procedures for generating gametes, mating, germination, and isolation of spores, have been described (1). Colonies resulting from a single spore were tested for their acetate-requiring phenotype, and transferred to liquid medium to grow cultures which were analyzed for plastocyanin and cytochrome c 6 accumulation by immunoblot analysis of soluble protein extracts (see below). For complementation tests, progeny isolates, derived from the back-cross of each mutant strain, were crossed with ac208 (pcy1-1) strains to obtain sheets of young zygotes as described previously (26); the resulting diploid zygotes were analyzed for their fluorescence induction kinetics (27) in comparison to zygotes that were homozygous for the pcy1-1 (ac208) allele.
Sequence Analysis of the Plastocyanin Gene-Total DNA was isolated from 100 ml of late log phase cultures of each mutant strain as described in Ref. 28. Fragments of interest were amplified as described previously (24), except that 100 ng of total DNA was used for each reaction instead of 500 ng. After amplification, the products were separated by electrophoresis, the fragment of interest was purified (29) and the entire amount was used as a template for sequencing reactions. The sequence of the genomic fragment from position Ϫ57 to ϩ907 (24), corresponding to the entire plastocyanin-encoding sequence, was determined on both strands for the gene from each mutant strain. In the case of pcy2-1, the amplified fragments were cloned and the cloned DNA was sequenced. The equivalent DNA fragments from wild-type strain CC125 were sequenced and analyzed in parallel lanes for comparison with the sequence determined from the mutant strains. Any differences in sequence were confirmed by repetition of the analysis. Except for mutations as discussed under "Results," no other differences in sequence were noted.
Preparation and Analysis of Proteins-Cells from liquid cultures in late log phase were collected by centrifugation (3,000 ϫ g, 5 min), washed once in 10 mM sodium phosphate (pH 7.0), and resuspended in a minimal volume (5 ϫ 10 8 /ml) of the same solution. Cells were lysed by two cycles of slow freezing to Ϫ80°C followed by thawing to room temperature. The soluble cell extract was separated from the insoluble fraction by centrifugation (15,850 ϫ g) in a microcentrifuge at 4°C, and stored at Ϫ80°C until they were analyzed. Protein concentration was estimated by the Coomassie dye binding method as described by the manufacturer (Pierce). The proteins were separated by non-denaturing or denaturing gel electrophoresis as described previously (30) and were visualized by staining with Coomassie Blue R-250. Alternatively, specific proteins were detected by immunoblot analysis. The primary antisera were diluted as follows: anti-plastocyanin (1:500 -1:1000), anticyt c6 (1:500), anti-OEE1 (1:500 -1:1000), anti-CF 1 (1:10,000), and anti-PsaF (1:1000). Bound primary antibody was detected with an alkaline phosphatase-conjugated secondary antibody and a chromogenic substrate (Bio-Rad).
RNA Isolation and Analysis-For the initial analysis of plastocyanin-deficient mutants, RNA was isolated from 50 -100 ml of C. reinhardtii cultures according to previously described procedures, and analyzed by Northern blot hybridization or by translation in vitro (30 -32). To assess the effect of cycloheximide on transcript stability (Fig. 3), 25 ml of cells were sampled from 100-ml cultures of the various strains (grown in TAP medium, ϳ 27 mol m Ϫ2 s Ϫ1 , 22°C, 180 rpm). Plastocyaninencoding mRNAs were identified by hybridization to a 577-base pair EcoRI fragment (cDNA insert from pTZ18RCrPC6-2), and mRNAs encoding the small subunit of ribulose-bisphosphate carboxylase/oxygenase were identified by hybridization to a radiolabeled fragment from plasmid pM1 corresponding to the cDNA sequence (33,34).
Radiolabeling of Cells and Immunoprecipitation-Two different procedures were used for analyzing plastocyanin synthesis. To examine the time course of precursor processing (Fig. 4), the cells were radiolabeled as described previously (35) with some modifications. Cells were cultured in copper-free, reduced sulfate (95 M sulfate) TAP medium to densities of 2-18 ϫ 10 6 /ml and collected by centrifugation (3,800 ϫ g, 5 min). The pelleted cells were washed once with 0.5 volume of copperfree, sulfate-free TAP medium, and resuspended in the same medium to a final concentration of 1 ϫ 10 8 /ml. The concentrated cells were allowed to recover for 1 h in an incubator (22°C, 250 rpm, 90 mol⅐m Ϫ2 ⅐s Ϫ1 ). Fifteen minutes before the addition of radioisotope, CuCl 2 was added to a final concentration of 6 M. Radiolabeling was initiated by the addition of Na 2 35 SO 4 (1,488 Ci/mmol, DuPont NEN) to a final concentration of 1 mCi/ml. The cells were maintained in a water bath (18°C) under illumination (90 mol⅐m Ϫ2 ⅐s Ϫ1 ) with periodic agitation by hand during the course of the labeling experiment. Cells were sampled by removing 0.2-ml aliquots into 1 ml of ice-cold acetone. The sample was immediately vortexed vigorously and left at 0 C for 30 min before the precipitate was collected by centrifugation (12,000 ϫ g, 5 min). The pellet was air-dried and resuspended in 100 l of solution A (2% SDS, 60 mM Tris-Cl (pH 8.6), 60 mM dithiothreitol, 5 mM ⑀-aminocaproic acid, 5 mM benzamidine, 1 mM EDTA, and 1 mM phenylmethylsulfonyl fluoride) by vortex agitation and heating (90°C, 5 min). Insoluble material was removed by centrifugation. Plastocyanin was immunoprecipitated from the supernatant (95 l) as described previously (35) except that cytochrome c 6 was preimmunoprecipitated from the samples. (The remaining 5 l was diluted 10-fold and used to assess radioisotope incorporation as trichloroacetic acid-precipitable counts.) The immunoprecipitate was solubilized in solution A and subjected to a second round of immunoprecipitation. The final IgGSORB pellet from the second immunoprecipitation was resuspended in 60 l of sample buffer for electrophoresis on denaturing gels containing SDS, and boiled for 5 min to release the bound proteins. The IgGSORB was removed by centrifugation, and the entire supernatant was loaded on the gel.
To examine the stability of plastocyanin in vivo (Fig. 5), the cells were grown in copper-deficient, reduced sulfate medium as described above, and labeled as described previously (36). Copper (6 M) was added as the chloride salt, where indicated. Labeled cells (pcy2-1 at 2.4 ϫ 10 6 /ml or CC125 at 5.9 ϫ 10 6 /ml) were sampled (150 ml) and collected by centrifugation (3,800 ϫ g, 5 min). The pelleted cells were washed with 15 ml of a cold solution containing 10 mM sodium phosphate (pH 7.0) and 10 mM Na 2 SO 4 , collected by centrifugation (3,400 ϫ g, 2 min), and resuspended in 150 l (pcy2-1) or 200 l (CC125) of 10 mM sodium phosphate (pH 7.0). The cells were lysed by freeze-thaw cycles (3 times), cell debris was removed by centrifugation, and plastocyanin was immunoprecipitated from a portion of the supernatant (ϳ14 -20 l, corresponding to an equivalent amount of trichloroacetic acid-precipitable radioactivity) as described previously, except that 50 mM dithiothreitol was used instead of 2-mercaptoethanol in the initial denaturing solution.
Other Materials-Agar was purchased from either JRH Biosciences (Lenexa, KS) or from Life Technologies, Inc., and metronidazole was from Sigma. All other materials have been specified or are described in the cited publications.

Identification of Plastocyanin-deficient Mutant Strains-A
collection of non-photosynthetic strains that survived UV mutagenesis was generated, and each strain was tested for plastocyanin accumulation. Soluble extracts were prepared from a lawn of cells scraped off a single agar plate. The plastocyanin content in the extracts was assessed by immunoblot analysis (Fig. 1, A and B). Of 1122 strains tested, four were found to be completely deficient in plastocyanin accumulation, pcy1-2, pcy1-3, pcy1-4, and pcy1-5; one additional strain appeared to accumulate apoplastocyanin at the expense of holoplastocyanin (pcy2-1). Each mutant was also checked for the accumulation of other proteins of the photosynthetic apparatus, including OEE1 (a lumenal polypeptide associated with PSII; Fig. 1C), cyt c 6 (a lumenal polypeptide that replaces plastocyanin; Fig.  1D), cyt f (not shown), and the ATP synthase (not shown) in order to identify mutants that might be affected in biosynthetic processes that were required for other thylakoid membrane proteins. The abundance of these proteins was unaffected, which suggested that the defect in the plastocyanin-deficient strains was likely to be restricted to plastocyanin accumulation. The record of mutagenesis experiments indicated that each strain arose from an independent experiment.
Since plastocyanin is not essential for photosynthesis in copper-deficient cultures of C. reinhardtii where cyt c 6 functions in place of plastocyanin, the mutants were expected to exhibit a conditional (copper-dependent) acetate-requiring phenotype if the defect were confined to plastocyanin function. Accordingly, each strain was checked for its acetate-requiring phenotype in copper-deficient medium. Indeed, each strain grew well on copper-deficient minimal medium but grew poorly or not at all on copper-supplemented minimal medium. On copper-supplemented acetate-containing medium, each strain exhibited re-  and D) conditions. The separated proteins were transferred to polyvinylidene difluoride membranes as described previously (30). Proteins of interest were detected by immunodecoration with antisera raised against plastocyanin (A and B), OEE1 (C), and cyt c 6 (D). For panels A, B, and C, cell extracts were prepared from coppersupplemented cultures of the relevant strains (lanes 1-7) whereas for panel D, cell extracts were prepared from copper-deficient cultures (lanes 1-7). Lane 8 in each panel represents a suitable negative control. For panels A and B, lane 8 contains extracts from copper-deficient cultures of wild-type cells, which accumulate little or no plastocyanin, while for panel D, lane 8 contains extracts from copper-supplemented wild-type cells, which do not accumulate cyt c 6 (36). For panel C, lane 8 contains extracts from strain FUD44, which does not accumulate OEE1 (59). The arrows in panel A mark the position of migration of apo-and holo-plastocyanin. For pcy2-1, thylakoid membranes were also tested by immunoblot analysis to confirm that plastocyanin was quantitatively released into the soluble phase during fractionation. sistance to metronidazole which confirmed that their non-photosynthetic phenotypes resulted from a defect in the electron transfer apparatus (37). Fluorescence induction kinetics revealed normal PSII function (not shown). These results confirmed that for each mutant strain, the acetate-requiring phenotype could be attributed to a loss of plastocyanin function.
Each mutant strain (derived from a mtϩ wild-type strain) was subsequently back-crossed to the wild-type strain, and resulting spores were analyzed for plastocyanin accumulation. Since the strains had a tendency to accumulate suppressor mutations, the spores were analyzed for plastocyanin accumulation rather than for their acetate-requiring phenotype (data not shown). The results showed that the plastocyanin-deficient phenotype could be ascribed to a single locus, which displayed Mendelian inheritance. The suppressor phenotype, resulting generally from copper-independent accumulation of cyt c 6 (38), segregates independently from the original mutation, which supports the contention that Cyc6 expression is not tightly linked to plastocyanin function.
In previous work (24), we had analyzed a plastocyanin-deficient mutant, strain ac208 (pcy1-1), and determined that its phenotype resulted from a frameshift mutation in the Pcy1 gene. To test whether the new mutants represented alleles at the PCY1 locus or a new locus, each strain was tested for its ability to complement pcy1-1. Four strains, pcy1-2, pcy1-3, pcy1-4, and pcy1-5, clearly failed to complement pcy1-1. Accordingly, we concluded that those strains were defective in the expression of the Pcy1 gene. The results of the complementation tests between pcy2-1 and pcy1-1 were difficult to interpret unambiguously, owing to the very weak acetate-requiring phenotype of pcy2-1. The restoration of photosynthetic activity in pcy2-1/pcy1-1 zygotes was therefore only marginally greater than the leaky photosynthetic activity observed for pcy2-1. Thus, the complementation assay was at the limit of detection, but it suggests that the two mutations are not in the same gene.
Analysis of the Pcy1 Gene in Plastocyanin-deficient Mutants-Northern analysis of total RNA isolated from each strain indicated that strain pcy1-2 lacked Pcy1 mRNA, while each of the other strains accumulated normal amounts of the mRNA, and of a size comparable to that of the wild-type mRNA ( Fig. 2A). The loss of plastocyanin-encoding mRNA in strain pcy1-2 was proposed to result from a cis-mutation affecting either transcription or RNA stability (see below). In the case of strain pcy1-3, the Pcy1 mRNA could not be translated to yield a product containing plastocyanin sequences, which suggested that the gene might carry a nonsense (or frameshift) mutation (Fig. 2B). For pcy1-4 and pcy1-5, the translation product appeared to be significantly longer (ϳ3 kDa) than the translation product from wild-type Pcy1 mRNA (Fig. 2B, compare lanes 4  and 5 to lane 6). Since the mRNAs were comparable in size, the increased size of the polypeptide was attributed to a readthrough mutation. As expected from its phenotype and the genetic analysis (described above), Pcy1 mRNA from strain pcy2-1 was indistinguishable from wild-type with respect to its size and coding information.
The pcy1-2, pcy1-3, pcy1-4, and pcy1-5 alleles were sequenced to localize the mutations. As suspected, pcy1-3 had a nonsense mutation at the 26th codon, while pcy1-4 and pcy1-5 were confirmed to carry read-through mutations. The TAA stop codon was altered to a Tyr-encoding TAT codon in pcy1-4 and a Leu-encoding TTA codon in pcy1-5. This adds 32 codons to the reading frame, corresponding to an increase in molecular mass of about 3 kDa, which is compatible with the size of the in vitro translation product from the RNA (Fig. 2B). In the case of pcy1-2, sequence analysis revealed a frameshift mutation at the 59th codon. This may account for the low abundance of Pcy1 RNA; less than 1% of wild-type levels (estimated by comparing the hybridization signal to that observed for a 1:128 dilution of wild-type RNA). The pcy1-1 allele, which also carries a frameshift mutation, has similarly reduced levels of mRNA (24).
Degradation of Frame-shifted Pcy1 RNA-In previous work, we had suggested that the reduced abundance of the frameshifted RNA in pcy1-1 might result from increased degradation of that message. In this work, we identify another mutation (pcy1-2) with essentially the same phenotype. Messenger RNAs carrying early termination codons resulting from nonsense or frameshift mutations have been shown in many systems to be subject to nucleolytic degradation in vivo (e.g. Refs. 39 -43). Although the mechanism by which the non-functional mRNAs are recognized and degraded is not known, cycloheximide treatment is known to stabilize these decay-prone messages and some involvement of the translation apparatus has been suggested (Refs. 44 and 45; reviewed in Ref. 46). To test whether the same mechanism(s) might operate in C. reinhardtii, Pcy1 mRNA levels were measured after treatment of pcy1-1 and pcy1-2 with cycloheximide (Fig. 3). Indeed, cycloheximide treatment, which did not affect the abundance of other messages (e.g. RbcS mRNA), increased the abundance of Pcy1 transcripts. Cycloheximide treatment cannot restore the level of frameshifted Pcy1 RNA to wild-type levels because the drug eventually inhibits transcription and hence RNA accumulation (see wild-type samples at 180 min). Anisomycin treatment likewise had a similar effect. (The change in size of the messenger RNA is attributed to decreased polyadenylylation in cycloheximide-treated cells (47).) This result confirms that the Pcy1 gene is transcribed in strains carrying the pcy1-1 and pcy1-2 alleles, since the increased amount in cycloheximide-treated cells must result from de novo synthesis.
Degradation of Extended Plastocyanins in pcy1-4 and pcy1-5-Immunoblot analysis of cell extracts from strains carrying FIG. 2. Accumulation of plastocyanin-encoding mRNA in the mutant strains. A, Northern hybridization analysis. Total RNA was isolated from copper-supplemented cells and separated by electrophoresis in formaldehyde-containing agarose gels (5 g/lane). The separated RNA was transferred to nylon membranes (GeneScreen, DuPont NEN) by capillary action and probed with radiolabeled DNA fragments encoding C. reinhardtii plastocyanin. The blot was exposed to x-ray film for 69 h. B, in vitro translation and immunoprecipitation of plastocyanin-precursors. Total RNA (5 g) was translated in vitro in the presence of radiolabeled methionine in a system derived from wheat germ extract (Promega Corp., Madison, WI). Polypeptides that were immunoreactive with anti-plastocyanin were precipitated and analyzed by electrophoresis on SDS-containing polyacrylamide gels followed by fluorography. Radiolabeled polypeptides (Life Technologies, Inc.) were used as molecular weight markers in order to estimate the apparent molecular weights of the translation products. the pcy1-4 and pcy1-5 alleles did not identify plastocyaninrelated products (Fig. 1, A and B), which suggested that the extended polypeptides (Fig. 2B) must be degraded. To confirm that the mutated mRNAs were translated in vivo, plastocyanin synthesis was assessed in pcy1-4 and pcy1-5 cells during a 5-min labeling period, and to test whether the extended polypeptides enter the post-translational pathway for pre-apoplastocyanin, the fate of the newly synthesized products was followed during a "chase." Three plastocyanin-related species are routinely detected after a brief period of labeling (Fig. 4,  lane 0). In wild-type cells, these correspond to the precursor form ("p"), and intermediate form ("i") generated by the first processing protease, and a mature form ("m") generated by the lumen-facing thylakoid processing peptidase. The p and i forms are rapidly converted to the mature form (lanes 1, 3, and 5). The same pattern of three species (p, i, m) is evident in the immunoprecipitate from pcy1-4 and pcy1-5 extracts. The size of each species in pcy1-4 and pcy1-5 is larger, and corresponds to the predicted increase in size as a consequence of the readthrough mutation. The long exposure time required to visualize mature plastocyanin synthesis in pcy1-5 permits also the visualization of small amounts of contaminating material with mobility similar to i in the immunoprecipitate (open arrowhead). The processing of the precursor forms in pcy1-4 and pcy1-5 occurs in approximately the same time frame as in wild-type cells. Thus, the C-terminal extensions do not appear to affect the translocation and processing pathway in vivo. The mature extended products do not accumulate owing to their rapid degradation. Interestingly, the estimated half-lives of the extended proteins are significantly different (t 1/2 ϳ 20 min for the pcy1-4 product, t 1/2 ϳ 3 min for the pcy1-5 product) although the two proteins differ in sequence at only one position. The extension was not expected a priori to have a drastic effect on the structure of the ␤-barrel, but the folding pathway may well be affected.
Biochemical Characterization of pcy2-1-The strain carrying the pcy2-1 allele accumulates apoplastocyanin at the expense of holoplastocyanin. This could result either from inhibition of holoplastocyanin formation or from post-synthesis loss of copper in vivo owing to destabilization of holoplastocyanin. However, the reason why the holoform of plastocyanin might be thermodynamically destabilized is not obvious. Analysis of thylakoid membranes does not reveal any major differences in the abundance of thylakoid membrane proteins, e.g. core polypeptides of the cyt b 6 f complex or PSI (data not shown). 3 The PsaF polypeptide, which cross-links to plastocyanin (48), is present at the same abundance in wild-type versus pcy2-1 cells as is cyt f, the electron donor to plastocyanin (not shown). Inhibition of holoplastocyanin formation could be attributed to a defect in copper transport into the lumen or a defect in a trans-acting factor required for copper-plastocyanin assembly, e.g. a copper "chaperone." To assess whether copper metabolism in strain pcy2-1 was normal, we examined two previously characterized copper-responsive processes: transcription of the Cyc6 gene (32) and degradation of apoplastocyanin in the thylakoid lumen (20,36). The Cyc6 gene is activated if the cells perceive internal copper deficiency, and apoplastocyanin is degraded if copper is not available in the lumen for holoprotein formation. The expression of the Cyc6 gene was de-activated at roughly the same level of medium copper in pcy2-1 versus wild-type cells, which indicates that copper was available to the regulatory molecules within the cell (data not shown). Likewise, plastocyanin degradation in the lumen is prevented by copper addition, and to the same extent as in wild-type cells (Fig. 5, compare lanes 3  and 4 of wild-type versus pcy2-1 cells). This result suggests that externally added copper is available in the lumen in pcy2-1 as in wild-type cells. Thus, copper transport either into the cell or into the lumen is not affected in pcy2-1, and the defect in holo-plastocyanin accumulation in pcy2-1 cannot be attributed to depletion of copper from the thylakoid lumen. However, pcy2-1 may well be deficient in an assembly factor required for folding of plastocyanin or for delivery of lumen copper to the active site of apoplastocyanin. 3 Andrew Webber, personal communication.

FIG. 3. Stabilization of frameshifted mRNA by cycloheximide.
RNA was isolated from cycloheximide-treated (10 g/ml for the indicated times) or untreated cultures of strains CC125 (WT), pcy1-1 and pcy1-2 (3-4.4 ϫ 10 6 cells/ml), and analyzed after separation on formaldehyde-containing agarose gels (10 g/lane) for accumulation of Pcy1 mRNA by Northern hybridization. A radiolabeled DNA fragment (6 ϫ 10 8 cpm/g) corresponding to the cDNA sequence (34) was used as the probe. The film was developed after 18 h of exposure.  (60). For panels C and D, the arrows are labeled p, i, and m to indicate the position of migration of the equivalent forms of the longer translation product expected for pcy1-4 and pcy1-5. The gel was exposed to Kodax XAR5 film for 7 days (for panels A-C) or 56 days (for panel D). For the strain carrying the pcy1-5 allele, the polypeptide that migrates slightly more slowly than the i form of plastocyanin is a contaminant (open arrowhead). In another experiment, that band was not seen. The long exposure time, necessitated by the apparently very short half-life of plastocyanin in this mutant, reveals contaminants that are not visible in panels A-C.

DISCUSSION
Mutations at the PCY1 Locus-To dissect the plastocyanin biosynthetic pathway, we screened mutagenized cells for their ability to accumulate plastocyanin. The population of approximately 1.1 ϫ 10 3 candidate strains was enriched for mutants with defects in photosynthetic electron transfer by metronidazole treatment. Each candidate was analyzed by immunochemical methods to identify plastocyanin-deficiencies, and five such mutants were identified. Four of the five represent alleles at the previously identified PCY1 locus encoding preapoplastocyanin (24,49). The mutations include the repertoire expected from UV mutagenesis, consisting of nonsense (pcy1-3), frameshift (pcy1-2), and missense (pcy1-4, pcy1-5) mutations. The two read-through mutations (pcy1-4 and pcy1-5) result from independent events at the same codon and occur in a region of T-rich sequence, which is likely to be a target for UV-induced mutations.
Although we had expected to identify mutations that affected the trans-thylakoid transport of plastocyanin, none such were noted. Plastocyanin, OEE1, and possibly cyt f enter the lumen via a common Sec-dependent translocation pathway (16,50). Mutants in this pathway should exhibit a pleiotropic deficiency in these components of the thylakoid membrane. Thus, every candidate acetate-requiring colony was screened for plastocyanin as well as cyt f accumulation, and each plastocyanin-minus or cyt f-minus strain was screened for OEE1 abundance. Despite the large number of colonies screened, we did not identify any strains with pleiotropic deficiencies in lumenal proteins. One possibility is that such strains might be so severely compromised that they do not survive the mutagenesis and enrichment procedure. Another, perhaps more likely, possibility is that the collection analyzed in this work may not have included translocation pathway mutants. Such mutants of maize are pleiotropically deficient in thylakoid membrane complexes and exhibit a pale green phenotype (51)(52). This collection of acetate-requiring strains did not include non-green or pale mutants and may have therefore excluded translocation pathway mutants. The frequency with which we recovered PCY1 alleles suggests that we have saturated this system for the identification of plastocyanin-deficient mutants.
Degradation of Non-functional RNAs-Of the various alleles at the PCY1 locus, pcy1-1 and pcy1-2 result in greatly diminished amounts of Pcy1 mRNA. Both mRNAs contain frameshift mutations. The degradation of frameshifted or nonsense-containing mRNAs is a well documented phenomenon in various systems, referred to as mRNA surveillance (referenced above and reviewed in Ref. 53). It has been suggested that eukaryotic mRNAs are subjected to proofreading, perhaps even in the nucleus or during transit from the nucleus, and are degraded if they carry mutations that affect translation of the reading frame (reviewed in Refs. 46 and 53). The mechanism of recognition is not understood, but some relationship with the translation apparatus is evident. In general, the more 5Ј mutations appear to have a greater destabilizing effect. The phenotypes of pcy1-1 and pcy1-2 can be accommodated within this model. A similar phenomenon was noted in strain pc-1 for the Lpcr1 transcript, which has a frameshift at the fourth codon resulting from a 2-nucleotide deletion (54). On the other hand, the dissimilar phenotype of pcy1-3 is surprising. Even though the pcy1-3 mRNA carries an early nonsense codon, which occurs at a more 5Ј position than does the frameshift in pcy1-2, the nonsense-containing RNA of strain pcy1-3 is not degraded and accumulates to wild-type levels ( Fig. 2A). Thus, mRNA degradation in this surveillance pathway must involve features in addition to the position of the nonsense codon.
Cycloheximide, which stabilizes decay-prone mRNAs in other systems, has a similar effect in C. reinhardtii. The mechanism by which cycloheximide functions is not completely understood. One possibility is that cycloheximide inhibits the decapping reaction (a prerequisite for endonucleolytic cleavage in this pathway), and another possibility is that cycloheximide affects the translation-associated step in the pathway (45). The stabilizing effect of cycloheximide (Fig. 3) confirms that the reduced abundance is attributed to mRNA decay rather than an effect on transcription and provides perhaps an example of the existence of an mRNA surveillance mechanism in Chlamydomonas.
Degradation of Plastocyanin-The outcome of the readthrough mutations in pcy1-4 and pcy1-5 is polypeptides that are extended by 32 amino acids. These polypeptides are synthesized in vivo, and the intensity of label incorporation into the primary translation product (p) suggests that the rate of synthesis is not significantly affected in the mutant strains. The precursor species appear to enter the post-translational import pathway, and the time course of processing suggests that the C-terminal extension does not hinder the progress of the precursor through the pathway. However, the product of the maturation pathway (m) is rapidly degraded. Since the second processing event occurs on the lumenal face of the thylakoid membrane (18), the m species is expected to be localized to the lumen. We therefore expect that the degradation machinery lies in the lumen. We cannot rule out the possibility that the i species is also a substrate for degradation because the proportion of the i species is somewhat reduced in the mutant strains relative to the p species. If i were a substrate for degradation (owing perhaps to improper folding) as well as for processing to m, then a reduced abundance of i is to be expected.
The proteolytic system responsible for clearing the extended proteins from the lumen must be extraordinarily active, since the half-life of the protein in pcy1-5 is on the order of a few minutes (Fig. 4). In previous work, we have shown that unassembled lumen proteins, e.g. apoplastocyanin, apocyts c 6 and f, are also degraded very rapidly with half-lives comparable to that of the extended pcy1-4 plastocyanin (ϳ10 -20 min) (35,36,55). Since lumen proteases have not yet been identified, the mechanism of substrate recognition and the pathway of proteolysis is completely unknown.
PCY2 Locus-Of the various plastocyanin-deficient strains, only one has a phenotype that can be attributed to a mutation outside the PCY1 locus. Sequence analysis confirms that pcy2-1 has a wild-type Pcy1 gene. We therefore propose a new locus, PCY2, which we suggest is required for the stable accumulation of holoplastocyanin. One possibility is that this locus defines a FIG. 5. Plastocyanin synthesis and accumulation in strain pcy2-1 is indistinguishable from that in wild-type cells. Mature plastocyanin was immunoprecipitated from soluble extracts (corresponding to 5 ϫ 10 5 cpm for wild-type cells and 1 ϫ 10 6 cpm for pcy2-1 cells) of radiolabeled cells. Samples were labeled (10 Ci/ml) for 15 min either in the presence (lanes 1 and 2) or absence of copper (lanes 3-6) as indicated at the top of the figure, and sampled immediately after dilution of the isotope (lanes 1, 3, and 5 marked 0 min) or after 90 min (lanes 2, 4, and 6). The second row of labels indicates whether copper was added to the copper-deficient culture during the 90-min chase. metal transporter required for the uptake of copper into the cell or into the lumen (where plastocyanin assembly can occur). This possibility is precedented by the phenotypes of putative manganese transporter mutants in cyanobacteria (23) and copper transporter mutants in yeast (56,57). However, analysis of internal copper levels by examination of well characterized copper-responsive processes (i.e. regulation of Cyc6 expression or stabilization of apoplastocyanin) suggests that the phenotype cannot be ascribed to a defect in copper transport. Another possibility is that the PCY2 locus encodes a trans-acting factor required for assembly of holo-plastocyanin. This factor may serve to catalyze prolyl trans to cis isomerization, which is required for the folding of plastocyanin (3), or delivery of copper to the active site of plastocyanin. The fact that metal insertion into apo-plastocyanin is selective in vivo (31) argues in favor of catalysis of assembly. The involvement of a copper chaperone in copper protein assembly is precedented by the case of tyrosinase assembly in Streptomyces antibioticus where the MelC1 protein functions to provide copper to apotyrosinase (58). A third explanation for the phenotype might be that the coppercontaining form of plastocyanin is (thermodynamically) destabilized in pcy2-1 owing to a deficiency in one of its electron transfer reaction partners. Although the levels of cyt f and the PSI reaction center polypeptides appear to be normal in pcy2-1, this remains a formal possibility because the structure of the thylakoid membrane and the protein-protein interactions between complexes are not yet well understood.
Plastocyanin-independent Photosynthesis-The acetate-requiring phenotypes of some of the plastocyanin-deficient mutant strains appeared to be easily suppressed. Analysis of the suppressed strain revealed that suppression was generally accompanied by copper-insensitive expression of the Cyc6 gene (e.g. Ref. 38). The issue of whether the Cyc6 gene was merely responsive to medium copper concentrations or whether the Cyc6 responded more directly to plastocyanin function or abundance is of continuing interest. If Cyc6 expression resulted as a consequence of plastocyanin deficiency, the copper-independent expression of Cyc6 might be expected to segregate with the plastocyanin-deficiency. However, genetic analysis of several suppressed strains (data not shown) reveals that Cyc6 expression segregates independently of the PCY1 locus. This supports the model that Cyc6 activation occurs in direct response to copper deficiency rather than in response to plastocyanin deficiency.