Proteomic Characterization of the Chlamydomonas reinhardtii Chloroplast Ribosome

We have conducted a proteomic analysis of the 70 S ribosome from the Chlamydomonas reinhardtii chloroplast. Twenty-seven orthologs of Escherichia coli large subunit proteins were identified in the 50 S subunit, as well as an ortholog of the spinach plastid-specific ribosomal protein-6. Several of the large subunit proteins of C. reinhardtii have short extension or insertion sequences, but overall the large subunit proteins are very similar to those of spinach chloroplast and E. coli. Two proteins of 38 and 41 kDa, designated RAP38 and RAP41, were identified from the 70 S ribosome that were not found in either of the ribosomal subunits. Phylogenetic analysis identified RAP38 and RAP41 as paralogs of spinach CSP41, a chloroplast RNA-binding protein with endoribonuclease activity. Overall, the chloroplast ribosome of C. reinhardtii is similar to those of spinach chloroplast and E. coli, but the C. reinhardtii ribosome has proteins associated with the 70 S complex that are related to non-ribosomal proteins in other species. In addition, the 30 S subunit contains unusually large orthologs of E. coli S2, S3, and S5 and a novel S1-type protein (Yamaguchi, K. et al., (2002) Plant Cell 14, 2957–2974). These additional proteins and domains likely confer functions used to regulate chloroplast translation in C. reinhardtii.

In the chloroplast, where proteins of the photosynthetic apparatus and the carbon-fixing enzymes are synthesized, gene expression is primarily regulated during translation (1). Chloroplast translation has been thought to be similar to transla-tion in bacterial systems, mainly because of similarities in ribosomal RNA and the sensitivity of chloroplast ribosomes to bacterial antibiotics. These similarities support the endosymbiotic theory that chloroplasts originated from a photosynthetic prokaryote, cyanobacteria (2,3). It is now recognized that chloroplast gene expression and chloroplast translation are unique and quite different from bacterial systems (1, 4 -7). The chloroplast ribosome contains plastid-specific ribosomal proteins (PSRPs) 1 in addition to bacterial orthologs (5)(6)(7). It has been proposed that PSRPs may take part in the unique light-dependent aspects of chloroplast translation (8). Bacterial gene expression is strongly influenced by the rate of transcription, and translation and transcription are often closely coupled. In the chloroplast, transcription is often globally regulated, and mRNA accumulation can be unrelated to the rate of translation of a protein (reviewed in Refs. 1 and 9 -13). Translation of many chloroplast mRNAs is activated in response to light illumination, with little change in the corresponding mRNA levels (14 -19). A majority of the work on chloroplast translation has been carried out in the unicellular green alga, Chlamydomonas reinhardtii (reviewed in Refs. 13, 20, and 21), because it is amenable to both genetic and biochemical analysis (22,23). Identification of all of the proteins required for chloroplast translation in C. reinhardtii would facilitate our understanding of the mechanisms of chloroplast translation, photosystem biogenesis, plastid differentiation, and ultimately plant development and function.
Ribosomal proteins (RPs) of the small and large subunits from the C. reinhardtii chloroplast ribosome have been characterized by two-dimensional PAGE (24 -26). Schmidt et al. (26) designated these proteins S-1 to S-31 and L-1 to L-33, according to their estimated size on two-dimensional PAGE, and identified their sites of synthesis by in vivo pulse-labeling in the presence of inhibitors of cytoplasm (anisomycin) and chloroplast (lincomycin) translation. Immunological characterization of chloroplast RPs from C. reinhardtii was carried out with antisera available for 15 bacterial RPs (27,28) and suggested that chloroplasts contained orthologs for each of these bacterial proteins. However, to date the majority of pro-* This work was supported by Grant GM54659 from the National Institutes of Health, Contract DE-FG03-93ER70116 from the Department of Energy, and a grant from Syngenta Corp. (all to S. P. M.). 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.
teins of the C. reinhardtii chloroplast ribosome have not been identified.
Recently, sequence databases for the nuclear and chloroplast genomes of Chlamydomonas have become available (29 -31). 2 Although similarity searches of open reading frames (ORFs) within these databases can help us speculate on the composition of chloroplast RPs, these databases will not allow us to identify actual RPs, especially proteins unique to this alga. Actually, ORFs can be precisely annotated only after protein sequencing and protein characterization. Precise identification of all the RPs in the C. reinhardtii chloroplast ribosome based on a precise proteomic characterization would give a much clearer picture of how algal RPs are similar, and different, from those of bacteria and higher plants.
In our recent study, we characterized the small subunit of the C. reinhardtii chloroplast ribosome via a proteomic approach. By using LC-MS/MS and LC/LC-MS/MS, we identified a novel S1 domain-containing protein (named PSRP-7) and unusually large bacterial orthologs of S2, S3, and S5 (7). Structural predictions, based on the crystal structure of the Thermus thermophilus 30 S subunit, suggest that the additional domains of S2, S3, and S5 are located adjacent to each other on the solvent side of the ribosome near the binding site for the S1 protein. We proposed that these additional domains interact with the S1 protein and PSRP-7 to function in mRNA recognition and translation initiation in C. reinhardtii chloroplast.
In this paper, we report the proteomic characterization of the chloroplast ribosomal large subunit and the complete 70 S ribosome. All of the proteins identified in the large subunit have bacterial orthologs and higher plant homologs. Overall, the large subunit proteins are very similar to those of Escherichia coli and spinach chloroplast in terms of size, isoelectric point, and amino acid sequence. We have classified all of the C. reinhardtii large subunit proteins in accordance with plastid ribosomal protein (PRP) nomenclature (5)(6)(7). In addition, proteomic and immunological analyses revealed that the 70 S ribosome of C. reinhardtii chloroplast contains two additional proteins of 38 and 41 kDa (RAP38 and RAP41), and these two proteins are not components of either the 30 S or 50 S subunits. Nucleotide sequencing and phylogenetic analysis revealed that RAP38 and RAP41 are paralogs of an ancestral nucleotide binding protein related to a bacterial epimerase/dehydratase and to a higher plant (spinach) chloroplast RNA-binding protein CSP41. This analysis has shown that the chloroplast translation machine of C. reinhardtii is compositionally and structurally similar to those of higher plants and eubacteria, but that C. reinhardtii ribosomes contain additional domains and proteins previously unidentified in other 70 S ribosomes.

EXPERIMENTAL PROCEDURES
Nomenclature-We describe C. reinhardtii chloroplast ribosomes, the large subunits, and the small subunits as 70 S, 50 S, and 30 S, respectively. This is to keep the nomenclature consistent with their higher plant and bacterial counterparts, although the sedimentation values of ribosomes and subunits from this algal chloroplast have been reported to be somewhat higher (32,33). Likewise, we use the PRP nomenclature system adapted for plant PRPs, based on sequence similarity to bacterial RPs (5,6). That is, chloroplast/plastid orthologs of E. coli L1-L36 are to be designated PRP L1-L36. In accordance with the approved Commission on Plant Gene Nomenclature designation for plant genes, gene names are written in italics, with nuclear genes having capital first letters, and organelle genes having lowercase first letters (34). For example, the gene for nuclear-encoded PRP L1 is PrpL1, whereas the gene for plastid-encoded PRP L2 is prpL2. We propose to use the cytoplasmic ribosomal protein (CRP) nomenclature for the plant and algal homologs of rat CRPs (35) to distinguish them from PRP and mitochondrial RP. Prefixes PRP or CRP may be omitted when it is obvious what specified RP is being discussed.
Preparation of C. reinhardtii Chloroplast Ribosomes and Subunits-C. reinhardtii strain cc3395 (cw15/Arg7) was used for this study. The chloroplast ribosomes and subunits were prepared as described in our previous report (7), which is a modified procedure of nitrogen bomb extraction (36) and successive sucrose gradient centrifugation (37). Details on centrifugal conditions are described in the figure legends. RPs were extracted from purified ribosomes or subunits by an acetic acid extraction method as described (6).
Proteomic Analyses-SDS-PAGE was performed as described by Laemmli (38), using 1.5-mm-thick, 12% acrylamide gel. The molecular weight markers used were the BENCHMARK™ prestained protein ladder (Life Technologies, Inc.). For LC-MS/MS analysis, 10 pmol of total protein from the 50 S subunit (TP50) and 70 S ribosome (TP70) were separated by SDS-PAGE and then stained with Coomassie brilliant blue R-250 as described (7). The gels were sectioned into pieces as shown in Fig. 1E, and each gel piece was further fragmented into 1-mm 2 pieces and transferred into one well of a 96-well plate. The plates were transferred to a Massprep digestion robot (Micromass, Beverley, MA) for destaining (39), reduction/alkylation (iodoacetamide), and in-gel digestion with trypsin or endoproteinase Lys-C (40). After digestion, tryptic peptides were extracted from the gel pieces on the Massprep robot. The extracted peptides were then subjected to LC-MS/MS equipped with a microbore HPLC system (Surveyor, ThermoFinnigan, San Jose, CA), a Surveyor autosampler, and a ThermoFinnigan LCQ-Deca ion trap mass spectrometer (ThermoFinnigan) as described previously (7). Spectra were scanned over the range 400 -1400 mass units. Automated peak recognition, dynamic exclusion, and product ion scanning of the top two most intense ions were performed using the Xcalibur software as described previously (41).
Direct analysis of total large subunit proteins was performed using ϳ10 g of TP50. The TP50 was reduced and alkylated by iodoacetamide and then digested with Endoproteinase Lys-C (Boehringer Mannheim) and Porozyme Trypsin Beads (Perseptive Biosystems) as described in our previous report (7). The tryptic peptide mixture was analyzed by multidimensional protein identification technology (MudPIT) as described previously (42)(43)(44).
MS/MS data obtained were analyzed using SEQUEST, a computer program that allows the correlation of experimental data with theoretical spectra generated from known protein sequences (45,46). In this work, the general criteria for a preliminary positive peptide identification for a doubly-charged peptide were a correlation factor Ͼ2.5, a ␦ cross-correlation factor Ͼ0.1 (indicating a significant difference between the best match reported and the next best match), a minimum of one tryptic peptide terminus, and a high preliminary scoring. For triplycharged peptides, the correlation factor threshold was set at 3.5. All matched peptides were confirmed by visual examination of the spectra. All spectra were searched against a FASTA-format database generated from Chlamydomonas ESTs and ORFs in the C. reinhardtii plastid genome (SWISS-PROT) (29).
Western Blot Analysis-Samples were run in SDS-PAGE under reducing conditions and transferred to nitrocellulose. Membranes were blocked in TBST (10 mM Tris-HCl, pH 8.0, 0.9% NaCl, and 0.1% Tween 20) containing 5% skim milk and 0.03% NaN 3 and then incubated overnight in the same solution containing primary antibodies. Anti-Chlamydomonas reinhardtii RAP38 polyclonal antibodies were raised in rabbits (The Scripps Research Institute, Protein and Nucleic Acid Core Facility) and used at 1:2000 dilutions. Anti-spinach chloroplast RP S1 and L2 polyclonal antisera were provided by Dr. Subramanian and used at 1:2000 dilutions. Western blots were developed using nitroblue tetrazolium and 5-bromo-4-chloro-3-indolyl phosphate (Sigma).
DNA Sequencing-Nucleotide sequencing was carried out with capillary electrophoresis technology using an ABI 3700 DNA analyzer (Nucleic Acid Core Facility, The Scripps Research Institute). The reactions were performed using thermal cycle sequencing conditions with fluorescently labeled terminators.
Computational Analyses-EST contig (assembled EST) was obtained from ChlamyDatabase using a WU-BLAST search. 3 BLAST (National Center for Biotechnology Information) was used for general sequence searches. Homology comparison was done using BLAST 2 SEQUENCES (National Center for Biotechnology Information). Multiple sequence alignments were performed using CLUSTAL X (47) and refined manually for representation. Isoelectric points and sequence masses were calculated by the ProtParam in the ExPasy proteomic tools. 4 Prediction of cleavage sites for chloroplast transit peptides were obtained using the ChloroP program (48). Phylogenetic analysis was performed by a neighbor-joining method using PAUP* version 4.0 Beta, and the reliability of the created tree was estimated by bootstrapping (49).

Isolation of the C. reinhardtii Chloroplast Ribosomes and
Subunits-Chloroplast ribosomes were purified from the postmitochondrial S-40 fraction of total cell extracts by successive sucrose gradient centrifugation, according to the method of Chua et al. (37), to obtain biologically active ribosomes. The chloroplast 70 S ribosomes were first separated from the cytoplasmic 80 S ribosomes in a 10 -40% sucrose gradient (Fig. 1A) and then purified on a second sucrose gradient (Fig. 1B). About half of the 70 S ribosomes loaded were dissociated into subunits, because of the fragile nature of the 70 S ribosome of the C. reinhardtii chloroplast (37,50,51). Purified 70 S ribosomes were collected and dissociated into 30 S and 50 S subunits in a 10 -30% sucrose gradient containing dissociation buffer (Fig.  1C). The purity of the 70 S ribosomes, and 30 S and 50 S subunits, was assessed by the RP pattern observed on SDS-PAGE ( Fig. 1, D and E) and by rRNA analysis on agarose gels (data not shown).
In-gel Digestion and Liquid Chromatography-Tandem Mass Spectrometry-Total proteins (TP30, TP50, and TP70) extracted from the 30 S subunits, 50 S subunits, and 70 S ribosomes, respectively, were resolved by SDS-PAGE and sectioned into gel pieces (Fig. 1E). Proteins in each section were digested with trypsin or Lys-C, and the generated peptide fragments were subjected to LC-MS/MS analyses (see "Materials and Methods" for details), as done previously for the 30 S subunit proteins (7).
An example of the LC-MS/MS analysis of the peptides derived from the chloroplast 70 S ribosome is shown in Fig. 2, demonstrating identification of the L1 protein. In the first step of LC-MS, the trypsin fragments from section 13 of the TP70 gel yielded the mass chromatogram shown in Fig. 2A. A peptide eluting at 26.21 min yielded the MS spectrum shown in Fig. 2B. In the following step of MS/MS, the abundant precursor ion of m/z 737.64 generated collision-induced dissociation spectra shown in Fig. 2C. Subsequent SEQUEST analysis of the collison-induced dissociation spectrum identified the peptide sequence to be DAGADVVGGDDLIEK, the precursor ion mass (M ϩ H) ϩ to be 1474.56 (i.e. m/z 737.64 is doubly charged ion), and the sequence belonged to an EST (AV620102), as listed in Table I. A BLAST search indicated AV620102 to be a homolog of chloroplast and bacterial RP L1. An EST contig (ACE number 20011023.1212.1) in the ChlamyEST database was identified by WU-BLAST search probing with AV620102 and confirmed as encoding a full-length chloroplast precursor protein homologous to the bacterial L1 protein. In the same way, other proteins from the TP70 and TP50 fractions were identified by LC-MS/MS analyses. As a complementary analysis, we also applied MudPIT for 50 S protein identification by the same procedure that we used for 30 S protein identification (7). MudPIT identifies protein components directly from an enzymatic digest of the entire large protein complex (here, it is TP50), using LC/LC-MS/MS without resolving proteins by poly- 4 Internet address: www.expasy.ch/tools/protparam.html. acrylamide gel electrophoresis (42)(43)(44). Table I Table I also shows their precursor ion masses, accession numbers of the identified ESTs or chloroplast ORFs, the protein identified and peptide origin, the SDS-PAGE gel section numbers, and the enzyme used or MudPIT.
The Large Subunit Proteins-From the chloroplast 50 S subunit, 28 proteins were identified including orthologs of bacterial L1, L2, L3, L4, L5, L6, L9, L10, L11, L12, L13, L14, L15, L16, L17, L18, L19, L20, L21, L22, L23, L24, L27, L28, L31, L32, and L35 and a homolog of spinach PSRP-6. An overview of the characteristics of the C. reinhardtii 50 S subunit proteins is shown in Table II comparing the C. reinhardtii 50 S PRPs with 50 S RPs from Arabidopsis chloroplast, Synechocystis sp. PCC 6803, and E. coli. The accession numbers and gene allocations for C. reinhardtii 50 S PRPs and percent similarity of 50 S RPs found in C. reinhardtii chloroplast, Arabidopsis chloroplast, Synechocystis, and E. coli are shown. The mature protein sizes and isoelectric points (pI) were predicted after removal of the predicted transit peptide or the N-formyl methionine and are summarized in Table III. Apart from proteomic analyses, we independently searched for homologs of the spinach 50 S PRPs and E. coli 50 S RPs from the Chlamydomonas databases and identified 31 potential 50 S PRP genes, suggesting the potential occurrence of L33, L34, and L36 proteins that were not identified by our proteomic analysis (see Tables II and III). These small and highly basic proteins likely escaped mass spectrometric detection (m/z 400 -1400) because of digestion by trypsin or Lys-C generating peptides too small to be unequivocally sequenced using this method. No homologs of bacterial L29, L25, L30, and spinach PSRP-5 were identified in our proteomic analysis or from computation analysis of the complete chloroplast genome sequence and the currently available EST databases. This may be attributable to either incomplete EST databases or to the absence of these genes and proteins in this alga. L25 and L30 are likely absent from the chloroplast ribosome as discussed previously (6).
Proteins Unique to the 70 S Ribosome: Identification of 38and 41-kDa Ribosome-associating Proteins-In spinach, a plastid ribosome recycling factor (pRRF) was identified in the chloroplast 70 S ribosome in stoichiometric amounts with other proteins of the 30 S and 50 S subunits. The pRRF was not present in either of the isolated 30 S or 50 S subunits (6). and y-ion species (y2, y3, y6 -11, and y13), for which the sequence is indicated. Therefore, the protein composition of the 70 S ribosome is not necessarily equal to the sum of the 30 S proteins and the 50 S proteins. For this reason, we also analyzed proteins in the C. reinhardtii 70 S subunit. We had previously identified two proteins in the TP70 pattern resolved by SDS-PAGE that were not seen in the TP30 or TP50 patterns (7). Here we confirmed that the 38-and 41-kDa proteins cosedimented with the plastid 70 S ribosomes but not with the plastid 30 S or 50 S subunits nor with cytoplasmic 80 S ribosomes (Fig. 1D). Because the 38and 41-kDa proteins were associated with the 70 S ribosome, we have designated these proteins RAP38 and RAP41 (Ribosome Associated Protein of 38 kDa and 41 kDa). As shown in Fig. 3, Western blot analysis of proteins from whole cells and from isolated 30 S, 50 S, and 70 S ribosomes revealed that RAP38 is associated with purified 70 S subunits but not with either of the purified dissociated subunits (30 S or 50 S). Anti-S1 and anti-L2 antibodies were used as controls to show that the ribosomal preparations contained the RP expected for each fraction. The SDS-PAGE profile of TP70 (Fig. 1, D and E) also shows that the RAP38 and RAP41 are present in similar amounts as other RPs (e.g. PRP S1 or PSRP-3), judged by the intensity of the Coomassie staining of these proteins. Thus, RAP38 and RAP41 appear to be present in stoichiometric amounts on the ribosome.
No homolog of spinach pRRF was identified from the C. rein-hardtii 70 S ribosome. However, pRRF is likely present as a non-RP in C. reinhardtii, because the corresponding EST can be identified by computational search of the C. reinhardtii EST database. Homologs of 19 CRPs S6, S8, S17, S25, L3, L7, L11, L13a, L17, L1a, L23a, L24, L26, L27, L27a, L32, L36, P0, and P1 were also identified from our TP70 fractions by the sensitive LC-MS/MS analysis. Some of these proteins were also identified in the TP30 (7) and TP50 fractions (this study). These proteins are contaminations of cytoplasmic 80 S ribosomes and their subunits because all 19 proteins were identified in purified 80 S ribosomes. 5 None of these 19 CRPs show putative chloroplast targeting sequences. Table I, five tryptic peptides of RAP38 and a Lys-C peptide of RAP41 were identified from section 8 and section 7 of the TP70 gel. The peptides belonged to ESTs AV620219 and AV621571 encoding RAP38 and to EST AV635725 encoding RAP41. We obtained the longest EST clone for each protein (AV620219 for RAP38; AV635725 for RAP41) from the Kazusa DNA Research Institute and sequenced the entire EST. The sequence data of RAP38 and RAP41 were submitted to the 5 K. Yamaguchi, P. A. Haynes, and S. P. Mayfield, unpublished data.  (52), and hypothetical proteins of cyanobacteria (see Table II for percent sequence similarities). RAP38 and RAP41 also showed a distant but still significant sequence similarity with several sugar-nucleotide epimerases/dehydratases of archaea, eubacteria, and eukarya. To examine phylogenetic and evolutional relationships between RAP proteins and their homologs, we performed an analysis using a neighbor-joining method, and the reliability of the created tree was estimated by bootstrapping (Fig. 4). The distance tree was constructed by rooting protein members of epimerase/dehydratase family as the outgroup, because proteins belonging to this enzyme family are well conserved between eubacteria and eukaryotes, including one of the earliest lineage in Methanosarcia (an archaebacterium). The tree shows three distinct clades: 1) RAP41 homologs, 2) RAP38 homologs, and 3) the epimerase/dehydratase family. No RAP homologs are identifiable in the complete genome sequences of E. coli or Methanosarcia. In contrast, a hypothetical protein encoded by a single gene (NP_440784) can be identified as a RAP38 homolog in the complete genome sequence of the cyanobacterium, Synechocystis sp. PCC6803. Similarly, another cyanobacterium, Nostoc sp. PCC 7120, has a RAP38-related gene (NP_488871). The Arabidopsis genome contains both a RAP38 homolog (At1g09340) and a RAP41 homolog (At3g63140). RAP41 and its homolog were identified only in algae and plants and were not found in photosynthetic bacteria. Thus, phylogenetic analysis suggests that RAP38 and RAP41 are paralogs of a cyanobacterial ancestor diverged from bacterial epimerase/dehydratase, and that RAP41 is closely related to spinach CSP41. This phylogenetic analysis is consistent with a previously reported analysis that was predicted by a a Cytosolic precursor and plastid pro-protein sequences deduced from nucleic acid sequences. b Predicted by ChloroP program (48). c Predicted by penultimate amino acid residues (66). d From predicted mature protein sequence. e Estimated from major peptides found in the TP50/70-gel section (see Fig. 1E and Table I). f Unidentified proteins by LC-MS/MS analyses. g ND, not determined due to MudPIT identification. h NI, not identified on this proteomic analysis. i Predicted by mass spectrometric results (see supplemental figure). j N-terminal sequence reported (60).

RAP38 and RAP41 Are Homologs of a Cyanobacterial Ancestor Related to Bacterial Epimerase/Dehydratase-As summarized in
conserved motif analysis, ancient divergence from a common ancestor of epimerases/dehydratase and an mRNA-binding protein with ribonuclease activity (53). Fig. 5 shows the amino acid sequence alignment of RAP38, RAP41, and their homologs, along with the motif structures as determined in the three-dimensional analysis of epimerases and dehydrogenases (53). The N-terminal ␤A-␣B-␤B region of the sugar-nucleotide epimerase/dehydratase involved in binding the adenine of NAD (54) is significantly conserved with RAPs and their homologs. Unexpected PRP Gene Allocation and Evolution in C. reinhardtii-In algae and plants, the PRP genes are partially distributed in the plastid genome and partially in the nuclear genome. In higher plants, the PRP genes are distributed between the two genomes, with approximately one-third of ϳ60 PRP genes found in the chloroplast and the remaining two-thirds in the nucleus (5, 6, 55). The transfer of PRP genes to the nuclear genome from the plastid genome occurred during plant/alga evolution. Unexpectedly, our proteomic analyses of the C. rein-hardtii chloroplast ribosome revealed that the number of transferred PRP genes in the nucleus of this alga is greater than that in higher plants. We reported previously that the C. reinhardtii PRP genes for S15 and S16 are nuclear genes, whereas these genes are plastid genes in higher plants (7). Likewise, C. reinhardtii PRP genes for L22, L32, and L33 are allocated in the nucleus, whereas these genes are maintained in plastids in Arabidopsis and other higher plants (Table II). In C. reinhardtii, the prpL21 gene is not present in plastid genomes, whereas our proteomic analysis identified a nuclear PrpL21 gene (Tables I  and II). In higher plants, the prpL21 gene has been lost from plastid genomes after the divergence from bryophytes, and a nuclear gene of mitochondrial origin has replaced the chloroplast gene (56). This is not the case for the PrpL21 nuclear gene of C. reinhardtii, in which phylogenetic analysis suggests that the gene originated from an ancestral plastid gene sharing significant homology with rpL21 of cyanobacteria and other algal plastid genomes (data not shown).

DISCUSSION
We previously characterized the C. reinhardtii 30 S ribosomal subunit via a proteomic analysis and identified 21 proteins: 19 E. coli orthologs, a spinach PSRP-3 homolog, and a novel protein, PSRP-7, that contains two S1 domains (7). We identified unusually large orthologs of bacterial S2 (57 kDa), S3 (76 kDa), and S5 (84 kDa) that are significantly larger than their counterparts in higher plants and bacteria. PSRP-7 is also an unusually large RP. The total mass of C. reinhardtii 30 S proteins was estimated to be ϳ600 kDa, which is 1.7-fold and 1.4-fold greater than those of E. coli 30 S (350 kDa) and spinach 30 S (430 kDa) proteins. To fully understand the chloroplast translation machinery, we have now characterized the components of the 50 S subunit and the 70 S ribosome of the C. reinhardtii chloroplast. From the 50 S subunit, 28 proteins were identified: 27 E. coli orthologs and a homolog of spinach PSRP-6. As shown in Fig. 6, the molecular masses of C. reinhardtii 50 S PRPs range from 4.3 kDa (L36) to 30.8 kDa (L2), and most of the 50 S PRPs are similar in size and sequence to their counterparts in spinach and E. coli (Table III and supplemental figure). The C. reinhardtii PRP L28 is twice as large as its spinach and E. coli counterparts because of the presence of a C-terminal extension. Other large subunit RPs show short insertion sequences or extension at their N-/C-termini (see supplemental figure), but these short extensions do not affect the net mass of the large subunit. The total mass of the C. reinhardtii 50 S PRPs is estimated to be ϳ490 kDa, which is similar to those of E. coli 50 S (440 kDa) and spinach 50 S (530 kDa) subunits. Previous studies have examined immunological similarity of the chloroplast RPs of C. reinhardtii with RPs of spinach chloroplasts, the cyanobacterium Anabaena, and E. coli (27,28).  (57). In general, the greatest immunological similarity is found between Anabaena and Chlamydomonas, and the similarity is greatest for chloroplast-synthesized proteins of the large ribosomal subunit. As shown in Table II, a majority of the C. reinhardtii chloroplast RPs have greater sequence similarity with cyanobacterium (Synechocystis) than with higher plants (Arabidopsis) or eubacterium (E. coli). Table II also shows that plastid-encoded proteins are more conserved between different organisms than nuclear-encoded proteins. Our results (this study and Ref. 7) have confirmed previous observations on the conservation of proteins within 70 S ribosomes and have expanded on these studies to include a complete molecular description of the entire set of RPs from C. reinhardtii chloroplasts.
Four of the C. reinhardtii PRPs are more than twice as large as the spinach and E. coli counterparts because of either Nterminal extensions (NTEs), insertion sequences, or C-terminal extensions (C-terminal extension): S2 (NTE), S3 (insertion sequence), PSRP-3 (NTE), and L28 (C-terminal extension). Recently, sensitive profile search techniques, such as PSI-BLAST, have identified a novel conserved domain, TRAM, predicted to be an RNA-binding domain common to tRNA uracil methylation and adenine thiolation enzymes (58). A TRAM domain is also identifiable from eukaryotic translation initiation factor-2␤ in Thermoplasma. Two TRAM domains are found at the N terminus of PRP S2 (ORF570) in C. reinhardtii. Accordingly, we speculate that the NTEs, C-terminal extensions, and insertion sequence domains in C. reinhardtii PRPs S2, S3, PSRP-3, and L28 may take part in regulation of translation, perhaps by directly binding chloroplast mRNAs with TRAM domains or other uncharacterized RNA-binding domains.
Two nuclear-encoded proteins have been identified from the 70 S ribosome that were not found in either of the subunits. These two ribosome-associated proteins were designated RAP38 and RAP41. Western blot analysis of total proteins and proteins from isolated ribosomal subunits using RAP38 antisera revealed that RAP38 is associated with 70 S ribosomes only and not with the individual subunits. These data give rise to at least two different hypothesizes: that RAP38 and RAP41 have a role in nontranslating 70 S ribosomes, or that RAP38 and RAP41 participate in translation but associate with the 70 S ribosome after the subunits have initiated translation. The role of these proteins is being studied using genetic knock-outs and in in vitro and in vivo experiments. In the proteomic characterization of the spinach 70 S chloroplast ribosome (6), RAP38-and RAP41-like proteins were not identified, although a pRRF was identified as a 70 S-specific protein. Identification of RAP38 and RAP41 from the 70 S subunit is entirely unexpected, because these two proteins share sequence similarity with spinach CSP41, a chloroplast RNA-binding protein with endoribonuclease activity. How does the C. reinhardtii chloroplast ribosome contain RAP38 and RAP41, whereas the spinach chloroplast ribosome does not? In C. reinhardtii, RAP38 and RAP41 have C-terminal extensions of 16 and 83 amino acids, respectively, compared with spinach CSP41. These additional domains may allow RAP38 and RAP41 to bind the C. reinhardtii 70 S ribosome with high affinity. Several additional domains in the 30 S and 50 S proteins unique to the C. reinhardtii (e.g. NTE of S2, C-terminal extension of L28) may also interact with the RAP proteins to allow binding to the ribosome.
The Arabidopsis genome contains both RAP38 and RAP41 homologs, whereas the cyanobacterial lineage has only a RAP38-like protein. Phylogenetic analysis suggests that RAP38 and RAP41 are paralogs of a cyanobacterial protein, but this protein has not yet been biochemically characterized. Sequence and motif homologies between CSP41, RAP38, and RAP41 suggest that similar RNA-binding and nuclease activities can be expected for these proteins as well. Histidine residues are known to be important for catalytic function in ribonucleases, and three histidine residues (His-34, His-130, and His-140) found in the catalytic domain (amino acid 1-191 re-  Table III. gion; Ref. 59) of CSP41 are conserved between CSP41 homologs and RAP proteins, whereas these histidines are not conserved in epimerases/dehydratases. In spinach, CSP41, CSP29, and CSP55 are thought to form a stem-loop RNA-protein complex that is required for processing of the petD mRNA at the 3Ј end (reviewed in Ref. 11). Similarly, 5Ј UTR processing of some chloroplast mRNAs is necessary for translation in algae and higher plants. We have shown previously that processing of the C. reinhardtii psbA mRNA 5Ј UTR depends upon factors mediating ribosome association (60), whereas failure to process the C. reinhardtii petD mRNA 5Ј UTR is also associated with the loss of translation (61,62). In barley chloroplast, rbcL mRNA with an unprocessed 5Ј UTR is not associated with polyribosomes (63). We propose that RAP38 and RAP41 are functional counterparts of CSP41 and may be involved in mRNA processing, directly or indirectly with other factors and/or RPs.
The proteomic, bioinformatics, and biochemical analyses presented here have shown that the chloroplast translation machine of C. reinhardtii is compositionally and structurally similar to those of higher plants and eubacteria, but that C. reinhardtii ribosomes contain additional domains and proteins previously unidentified in other 70 S ribosomes. Many of these additional domains reside on the 30 S subunit of the ribosome, the subunit responsible for mRNA discrimination during translation initiation. The part of the ribosome responsible for peptidyl bond formation between amino acids, the 50 S subunit, is more conserved. These results are not unexpected, because peptide bond formation would be expected to be a conserved enzymatic function, whereas mRNA discrimination would be anticipated to be organism dependent. Defining specific roles for individual proteins and domains in chloroplast translational regulation should now be greatly facilitated by having the complete set of RPs defined.