Calcium-activated (p)ppGpp Synthetase in Chloroplasts of Land Plants*

The genetic system of chloroplasts, including the machinery for transcription, translation, and DNA replication, exhibits substantial similarity to that of eubacteria. Chloroplasts are also thought to possess a system for generating guanosine 5′-triphosphate ((p)ppGpp), which triggers the stringent response in eubacteria, with genes encoding chloroplastic (p)ppGpp synthetase having been identified. We now describe the identification and characterization of genes (OsCRSH1, OsCRSH2, and OsCRSH3) for a novel type of (p)ppGpp synthetase in rice. The proteins encoded by these genes contain a putative chloroplast transit peptide at the NH2 terminus, a central RelA-SpoT-like domain, and two EF-hand motifs at the COOH terminus. The recombinant OsCRSH1 protein was imported into chloroplasts in vitro, and genetic complementation analysis revealed that expression of OsCRSH1 suppressed the phenotype of an Escherichia coli mutant deficient in the RelA and SpoT enzymes. Biochemical analysis showed that the OsCRSH proteins possess (p)ppGpp synthetase activity that is dependent both on Ca2+ and on the EF-hand motifs. A data base search identified a CRSH homolog in the dicotyledon Arabidopsis thaliana, indicating that such genes are conserved among both monocotyledonous and dicotyledonous land plants. CRSH proteins thus likely function as Ca2+-activated (p)ppGpp synthetases in plant chloroplasts, implicating both Ca2+ and (p)ppGpp signaling in regulation of the genetic system of these organelles.

The hyperphosphorylated guanosine nucleotides ppGpp and pppGpp were initially identified as "magic spots" that accumulate during amino acid deprivation in Escherichia coli and induce rapid down-regulation of stable RNA synthesis (1). Subsequently, carbon, fatty acid, or iron limitation was also found to trigger the accumulation of (p)ppGpp in E. coli (2)(3)(4)(5). Specific targets of the global regulator (p)ppGpp have been identified among enzymes of nucleotide metabolism (6) and protein components of the apparatuses for transcription (7), translation (8), and DNA replication (9).
Two enzymes, RelA and SpoT, are responsible for fluctuations in (p)ppGpp concentration in E. coli. RelA is a ribosomeassociated protein and synthesizes (p)ppGpp in response to the increase in the proportion of uncharged tRNAs that results from amino acid starvation (10). In contrast, SpoT is a bifunctional (p)ppGpp synthetase and hydrolase and regulates (p)ppGpp levels in response to carbon source or fatty acid limitation (3,4,11). Many other bacteria possess only a single bifunctional (p)ppGpp synthetase and hydrolase, designated Rel (12). Rel proteins have been shown to be essential for the long term survival or virulence of pathogenic bacteria (13).
Stringent factor and RelA-SpoT-like (p)ppGpp synthetases have also been identified in chloroplasts (14 -17), and we previously demonstrated the presence of intrinsic (p)ppGpp synthetase activity in pea chloroplasts (18). Three RelA-SpoT homolog genes (RSH1, RSH2, and RSH3) have been identified in plants (14,17), and one such gene (RSH) has been identified in the alga Chlamydomonas reinhardtii (16). Enzymatic activity and specific localization to chloroplasts have been demonstrated for plant RSH2 and RSH3 and for algal RSH (16,17). Chloroplasts of land plants and algae are thought to originate from ancient cyanobacteria-like prokaryotes. Indeed, the genetic system of chloroplasts, including the apparatuses for transcription, translation, and DNA replication, is highly similar to that of bacteria. A stringent control system similar to that of bacteria has therefore been suggested to operate in chloroplasts of photosynthetic eukaryotes (19).
In a search for plant genes that encode RelA-SpoT-like proteins, we have now identified a previously unknown type of RSH gene for putative chloroplast-localizing proteins in rice (Oryza sativa). The deduced amino acid sequences of the novel RSH genes contain motifs similar to the EF-hand, a Ca 2ϩ -binding domain, at their COOH termini. We show that the (p)ppGpp synthetase activity of the gene products is regulated by Ca 2ϩ via the EF-hand motifs, and we therefore named them CRSH (Ca 2ϩ -activated RelA-SpoT homolog) to distinguish their mechanism of regulation from that of other RSH proteins. Our findings thus have implications for Ca 2ϩ signaling and Ca 2ϩdependent (p)ppGpp accumulation in plant chloroplasts.
Isolation and Characterization of Rice CRSH Genes-Standard recombinant DNA techniques were performed basically as described (21). The PCR was performed with a rice cDNA library (20) and oligonucleotide primers (OsREL1F, 5Ј-CAGG-AAGGACACAAAAAATTATGTCAC-3Ј; OsREL1R, 5Ј-CGTC-GAGAAGATGCATGAG-3Ј) based on a rice expressed sequence tag that shows sequence similarity to bacterial relA and spoT (GenBank TM accession number D48993). The PCR products were cloned into the pCRII vector (Invitrogen) and sequenced. Amplified DNA fragments were also purified and used to prepare fluorescein isothiocyanate-labeled probes with an ECL random prime labeling system (GE Healthcare). Screening of a Nipponbare cDNA library in phage gt11 for OsCRSH1 cDNA clones and isolation of such clones were performed as previously described (20).
Complementary DNA for OsCRSH3 was constructed as follows. The cDNA clone AK110850, which encodes an NH 2 -terminal portion of OsCRSH3, was obtained from the National Institute of Agricultural Resources (NIAR, Tsukuba, Japan). The missing portion of the open reading frame (ORF) was predicted from the corresponding genomic sequence of rice chromosome 5 (accession number AC137614). The 1598-bp portion of the ORF contained in clone AK110850 was amplified by PCR with the forward primer CRSH3F (5Ј-ATAACTAGTAT-GGCGAACGCCGGTGTCAA-3Ј) and the reverse primer CR3MR (5Ј-ATGATGGCCTTGAGCCTCTT-3Ј). The portion of the ORF for the COOH-terminal region of OsCRSH3 was amplified by PCR with a template cDNA mixture (20) prepared from 20-day-old Nipponbare seedlings, the forward primer CR3MF (5Ј-GGTGATCAAGGCCATGTGGA-3Ј), and the reverse primer CR3RR (5Ј-TCTTGCGGTAGACGTGG-ATG-3Ј). The resulting products were then subjected to a further PCR amplification with the primers CR3MF and CR3R (5Ј-ATAAGATCTTCAAACGAGGACCAACTTGT-CGCTGAGGTTCTTGCGGTAGACGT-3Ј). The new PCR products were then mixed with those obtained from clone AK110850 and subjected to PCR with the primers CRSH3F and CR3RR. The resulting product containing the full-length ORF was digested with SpeI and BglII and cloned into the corresponding sites of the vector pEU3b (22), generating pECRSH3.
Northern Hybridization-Isolation of total RNA from 20-day-old Nipponbare seedlings and Northern hybridization were performed as previously described (20). For preparation of an OsCRSH1 riboprobe, a 910-bp EcoRV-EcoRI fragment of OsCRSH1 cDNA was cloned into the HincII-EcoRI sites of pSPT19 (Roche Applied Science), and a digoxigenin-labeled antisense riboprobe was generated from the resulting plasmid by run-off in vitro transcription with SP6 RNA polymerase (Roche Applied Science). Riboprobes specific for rice SIG1 (OsSIG1) and actin2 (Rac2) mRNAs (internal standards) were also synthesized and tested as described (20).
Cell-free Protein Synthesis-A series of cDNAs encoding intact or NH 2 -terminally truncated forms of OsCRSH1 was constructed by PCR with a plasmid harboring the full-length cDNA as template, a gene-specific forward primer (FNF for the full-length protein, 5Ј-CCATGGCGACCGCCGCGA-3Ј; DN41F for a protein lacking the 41 NH 2 -terminal amino acids, 5Ј-CCATGGCCCAAGCAGCCGCCGCG-3Ј; DN65F for a protein lacking the 65 NH 2 -terminal residues, 5Ј-CCATGGC-GGAGGGAGGAGGGCGG-3Ј), and the reverse primer FCR (5Ј-AAAGTCGACTTAGCTTACAAGCAGTTTGTCGCCC-3Ј). The amplified fragments were digested with NcoI and SalI and then cloned into the corresponding sites of pEU3b. The resulting plasmids were designated pECR1FL, pECR1⌬N41, and pECR1⌬N65, respectively. A pEU3b-based vector for an OsCRSH1 mutant lacking both the 41 NH 2 -terminal and 128 COOH-terminal residues was similarly constructed by PCR with the plasmid encoding the full-length protein, the forward primer DN41F, and the reverse primer DC128R (5Ј-AAAGTCG-ACTTAGCTGTCTCCTTGATCACTAGCTG-3Ј). The resulting vector was designated pECR1⌬N41⌬C128. After verification of the inserted sequences, the various plasmids for the truncated proteins were used as templates for in vitro transcription with SP6 RNA polymerase as described (23).
For construction of cell-free expression vectors for OsCRSH2 and OsCRSH3, cDNA fragments corresponding to the putative mature form of each protein were amplified by PCR with a plasmid containing the full-length ORF (clone AK058438 for OsCRSH2; pECRSH3 for OsCRSH3), the forward primer CR2DN14F (5Ј-ATAACTAGTATGGCGGC-CGCGGTGGCGCCTGA-3Ј) for OsCRSH2 or CR3DN12F (5Ј-ATAACTAGTATGGCCGTCGCCATTGACGCACC-3Ј) for OsCRSH3, and the reverse primer CR2R (5Ј-ATAAGATCTT-CAGCCGGAGACGAGCTTGT-3Ј) for OsCRSH2 or CR3R for OsCRSH3. The PCR products were digested with SpeI and BglII and cloned into the corresponding sites of pEU3b, generating pECRSH2⌬N14 and pECRSH3⌬N12, respectively.
Cell-free protein synthesis was performed with wheat embryo extracts (Cell-Free Sciences, Yokohama, Japan) according to two different methodologies. For the synthesis of 35 S-labeled proteins, the "batch method" was performed as described (23) in the presence of [ 35 S]methionine and [ 35 S]cysteine (total of 14.8 mCi/ml) with the use of a Pro-mix L-[ 35 S] in vitro cell labeling kit (GE Healthcare). For preparation of protein for enzyme assays, the "bilayer method" (22) was performed. GST-OsCRSH1 fusion proteins synthesized by the bilayer method were isolated on a glutathione affinity column (GE Healthcare) and treated with Precision protease on the column to yield the corresponding OsCRSH1 protein. The solubility of proteins synthesized in the presence of [ 14 C]leucine (0.074 mCi/ml; GE Healthcare) was assessed by SDS-PAGE both of the total reaction mixture and of the supernatant obtained after centrifugation of the mixture at 22,000 ϫ g for 20 min at 4°C. The gel was dried and exposed to a BAS-III imaging plate (Fujifilm, Tokyo, Japan), and protein-associated radioactivity was detected with a BAS-2500 analyzer (Fujifilm) and quantitated with Image Gauge version 3.41 software (Fujifilm).
Site-directed Mutagenesis-Site-directed mutagenesis of OsCRSH1 was performed by in vitro overlap extension PCR as described (25). The gene was mutated so that glutamate in the ϪZ position of either Ca 2ϩ binding loop (see Fig. 1C) was replaced with glutamine. Two separate amplification reactions were performed to amplify the two halves of OsCRSH1 with the use of four primers. The outside-forward primer (DN41FB) was paired with a middle-reverse mutation primer (EF1R for the NH 2 -terminal Ca 2ϩ binding loop, 5Ј-CCTCTGTGAGCTGC-TCAATACTGAT-3Ј, with the substituted nucleotide under-lined; EF2R for the COOH-terminal Ca 2ϩ binding loop, 5Ј-GAATGCTTCGAATTGATCGGAGCTCAAGG-3Ј) to generate the first half of the gene; the outside-reverse primer (FCR) was paired with a middle-forward mutation primer (EF1F for the NH 2terminal Ca 2ϩ binding loop, 5Ј-ATCAGTATTGAGCAGCTCA-CAGAGG-3Ј; EF2F for the COOH-terminal Ca 2ϩ binding loop, 5Ј-CTTGAGCTCCGATCAATTCGAAGCATTCC-3Ј) to generate the second half. The reactions were performed with 5 ng of pECR1⌬N41 as the template in a final volume of 50 l containing 1ϫ Pyrobest Buffer II (TaKaRa), 0.2 mM of each deoxynucleoside triphosphate, 0.025 unit of Pyrobest DNA polymerase (TaKaRa), and 0.2 M of each primer. The amplification protocol comprised 20 cycles of denaturation at 98°C for 15 s, annealing at 60°C for 35 s, and elongation at 72°C for 3 min, followed by a final elongation step at 72°C for 10 min. The DNA fragments bearing overlapping sequences introduced by the two middle primers were then mixed together and subjected to three cycles of denaturation at 98°C for 15 s, annealing at 60°C for 30 s, and elongation at 72°C for 3 min. The resulting products were further amplified by PCR with the outside-forward primer (DN41FB) and the outside-reverse primer (FCR); the protocol comprised 20 cycles of denaturation at 98°C for 15 s, annealing at 60°C for 35 s, and elongation at 72°C for 3 min, followed by a final elongation step at 72°C for 10 min. The final products were digested with SalI and cloned into the SmaI-SalI sites of pGCR1⌬N41. The resulting plasmids for GST fusion proteins of OsCRSH1(⌬N41) with mutated NH 2 -and COOH-terminal Ca 2ϩ -binding loops were designated pGCR1⌬N41EF1 and pGCR1⌬N41EF2, respectively.
Assay of (p)ppGpp Synthetase Activity-ATP:GTP 5Ј-pyrophosphotransferase assays were performed in a final volume of 25 l containing 2 mM ATP, [␥-32 P]ATP at 10 Ci/ml (3000 Ci/mmol; GE Healthcare), 1.3 mM GTP, and synthesized protein in Buffer C (50 mM Tris acetate, pH 7.8, 0.5 mM magnesium acetate, 60 mM potassium acetate, 30 mM ammonium acetate, 1 mM dithiothreitol, 0.5 mM phenylmethylsulfonyl fluoride) and in the absence or presence of CaCl 2 . The reaction was performed for 1 h at 30°C and was stopped by the addition of 1 l of 88% formic acid. After the further addition of 12.5 l of phenol:chloroform:isoamyl alcohol (25:24:1, v/v/v), the mix-ture was agitated and then centrifuged at 10,000 ϫ g for 5 min at 4°C. The aqueous phase was transferred to another tube, and a 3-l portion was spotted onto a polyethyleneimine-cellulose thin layer sheet (Merck). For one-dimensional analysis, 1.5 M KH 2 PO 4 was used as the chromatographic solvent, as described (18). Two-dimensional analysis was performed essentially as described previously (18) with solvent Tb (0.75 M Tris, 0.45 M HCl, 0.5 M LiCl) for the first dimension and solvent Sb (prepared by dissolving 74 g of (NH 4 ) 2 SO 4 , 0.4 g of (NH 4 )HSO 4 , and 4 g of EDTA (disodium salt) in 100 ml of H 2 O) for the second dimension. Each chromatogram was exposed to a BAS-III imaging plate, and the associated radioactivity was detected with a BAS-2500 analyzer and quantitated with Image Gauge version 3.41 software.
Assay of (p)ppGpp Hydrolase Activity-The 32 P-labeled pppGpp was synthesized in vitro by using purified OsCRSH1(⌬N41) in the presence of 500 M CaCl 2 as described in the method for (p)ppGpp synthetase assay. Purification of the 32 P-labeled pppGpp was carried out as described (18). Hydrolysis mixtures (25 l) contained 50 mM Tris acetate, pH 7.8, 0.5 mM magnesium acetate, 60 mM potassium acetate, 30 mM ammonium acetate, 1 mM dithiothreitol, 0.5 g of enzyme, and 3 l of the purified pppGpp with or without the addition of 1 mM MnCl 2 . The mixtures were incubated at 30°C for 2 h, and the reaction was stopped by the addition of 1 l of 88% formic acid. After the further addition of 12.5 l of phenol:chloroform: isoamylalcohol (25:24:1, v/v/v), the mixture was agitated and then centrifuged at 10,000 ϫ g for 5 min at 4°C. The aqueous phase was transferred to another tube, and a 5-l portion was subjected to the one-dimensional TLC analysis.
Expression and Purification of SpoT Protein-A DNA segment of E. coli spoT gene coding for the NH 2 -terminal fragment (1-380 amino acids) of the SpoT protein was amplified by PCR using primers SPQEF (5Ј-GGATCCTTGTATCTGTTTGAA-AGCCTGAA-3Ј) and SPQER (5Ј-AAGCTTTTAAAACGAA-CTACCGGCGCTCT-3Ј). The obtained DNA fragment was digested with BamHI and HindIII and cloned into the BamHI-HindIII sites of pQE30 (Qiagen), yielding pQESP⌬C. E. coli strain JM109 transformed with pQESP⌬C was incubated in 100 ml of LB at 37°C with vigorous shaking. When the optical density (600 nm) was reached at 0.5, IPTG was added to the culture at a final concentration of 1 mM and further incubated at 37°C for 3 h. The cells were collected by centrifugation at 6,000 ϫ g for 15 min at 4°C and then resuspend in TALON buffer A (50 mM sodium phosphate, pH 7.0, 0.5 M NaCl, 0.1% Triton X-100, 20% glycerol). The cells were lysed by sonication, and the cell debris was removed by centrifugation at 12,000 ϫ g for 15 min at 4°C. The lysate was added to TALON Metal affinity resins (Clontech, Mountain View, CA) for 1 h at room temperature, and the resin was then transferred to a column and washed with TALON wash buffer (50 mM sodium phosphate, pH 7.0, 0.5 M NaCl, 17 mM imidazole, 0.1% Triton X-100, 20% glycerol). The SpoT⌬C protein tagged with 6ϫHis at its NH 2 terminus was eluted from the column with TALON elution buffer (50 mM sodium phosphate, pH 7.0, 0.5 M NaCl, 180 mM imidazole, 0.1% Triton X-100, 20% glycerol). The purity was checked by SDS-PAGE analysis, and the fraction containing SpoT⌬C was stored until use at Ϫ30°C after the addition of equal volume of 100% glycerol.

RESULTS
Isolation and Characterization of OsCRSH cDNAs-In a search for (p)ppGpp synthetase genes in land plants, we found a rice expressed sequence tag (GenBank TM accession number D48993) that encodes part of a putative polypeptide with sequence similarity to RelA and SpoT of E. coli, and we subsequently isolated a cDNA containing the corresponding fulllength ORF from a rice cDNA library in gt11 (20). After we deposited the sequence of this cDNA in DDBJ (GenBank TM accession number AB042936), the same cDNA sequence appeared (AK121808) in the rice full-length cDNA data base (cdna01.dna.affrc.go.jp/cDNA). A data base search identified a similar cDNA sequence in Arabidopsis thaliana (At3g17470). The predicted proteins encoded by the rice and Arabidopsis cDNAs each comprise 583 amino acids and show sequence similarity to RelA-SpoT family proteins (Fig. 1A). To compare the deduced amino acid sequences of the plant proteins with those of other RelA-SpoT family members, we aligned the regions implicated in (p)ppGpp hydrolase and (p)ppGpp synthetase activities (Fig. 1B). The plant sequences also include an NH 2 -terminal extension predicted to contain a chloroplast-targeting signal as well as two EF-hand motifs in the COOH-terminal region (Fig. 1, A and C). On the basis of these structural features and subsequent functional characterization of the encoded protein, we designated the isolated rice gene as OsCRSH1 (O. sativa Ca 2ϩ -activated RelA-SpoT homolog 1). A further search resulted in the identification of two additional homologous genes (GenBank TM accession numbers AK058438 and AK110850) in the rice expressed sequence tag data base. Whereas AK058438 contains a full-length ORF, the sequence of AK110850 was annotated as a cDNA in the reverse direction, with the DNA sequence corresponding to the COOH-terminal region of the encoded protein being missing. Given that the genomic sequence of rice chromosome 5 contains the genes corresponding to AK058438 and AK110850, we isolated the full-length cDNAs for both genes by PCR from a cDNA mixture prepared from polyadenylated RNA of rice leaves. The deduced amino acid sequences of both genes include a putative chloroplast-targeting signal in the NH 2 -terminal region, a central RelA-SpoT-like domain, and EF-hand motifs in the COOHterminal region (Fig. 1). These two genes were thus designated OsCRSH2 (GenBank TM accession number AK058438) and OsCRSH3 (GenBank TM accession number AB298325). The extent of nucleotide sequence identity between OsCRSH1 and either OsCRSH2 or OsCRSH3 cDNAs is 66.3 and 65.3%, respectively, whereas the corresponding values for amino acid sequence identity are 62.4 and 64.8%. All three genes are present on rice chromosome 5. An unrelated gene (expressed sequence tag clone AK105485) is positioned between OsCRSH1 and OsCRSH2 (data not shown). The genome sequence of A. thaliana appears to contain only a single gene, located on chromosome 2, that is homologous to the OsCRSH genes. It is therefore likely that the three OsCRSH genes arose as a result of triplication of the ancestral gene during rice evolution. Localization and Processing of OsCRSH1-To examine the function of the NH 2 -terminal sequence of the putative OsCRSH1 protein, we performed an in vitro chloroplast import assay as described previously (26). The full-length protein was synthesized with a wheat embryo cell-free protein synthesis system in the presence of [ 35 S]methionine and [ 35 S]cysteine. Incubation of the 35 S-labeled protein (ϳ64 kDa) with intact pea chloroplasts resulted in its processing to a smaller form (ϳ60 kDa) that was resistant to incubation with the protease thermolysin ( Fig. 2A). The resistance of processed protein to thermolycin indicates that mature OsCRSH1 protein was taken up into chloroplasts. These results thus indicated that OsCRSH1 was imported into chloroplasts, likely as a result of its putative NH 2terminal signal sequence (Fig. 1A).
We next compared the molecular size of the processed form of OsCRSH1 with that of NH 2 -terminal truncation mutants also synthesized with the cell-free translation system. The size (ϳ60 kDa) of the processed form of OsCRSH1 was similar to that of an OsCRSH1 mutant lacking the 41 NH 2 -terminal amino acids (Fig. 2B), suggesting that the cleavage site is located near the residue, alanine 42, in the polypeptide sequence (AB042936). The NH 2 -terminal sequence of the OsCRSH1(⌬N41) mutant is also similar to that of the predicted cleaved form of the Arabidopsis AtCRSH protein (cleavage site between arginine 15 and serine 16, At3g17470). On the basis of these results, we used the ⌬N41 form of OsCRSH1 as the mature form of the protein in subsequent experiments.
Functional Complementation of an E. coli Mutant by OsCRSH1-To determine whether OsCRSH1 functions as a RelA-SpoT homolog, we cloned the cDNA encoding the putative mature form of the protein into the expression vector pET11d to yield pETCR1. Transformation of E. coli CF1652(DE3) (relA Ϫ ), CF1678(DE3) (relA Ϫ , spoT Ϫ ), and W3110(DE3) (wild type) with pETCR1 or pET11d as described (16) revealed that the rice protein restored the growth of the relA, spoT double mutant on MOPS minimal agar medium containing 0.4% glucose and 50 M IPTG (Fig. 3). In contrast, transformation with pETCR1 did not suppress the phenotype of the relA single mutant (data not shown). These results therefore suggested that OsCRSH1 possesses only a low level of (p)ppGpp synthetase activity similar to that of E. coli SpoT.
Analysis of OsCRSH Gene Expression-We next investigated the pattern of OsCRSH1 expression in rice plants by Northern hybridization analysis. The expressions of OsSIG1, which encodes plastid RNA polymerase sigma-factor and shows specific expression in green tissues (20), and Rac2, which is constitutively expressed in whole tissues (20), were analyzed as well. A major transcript of 2.1 kb was detected in green tissues and roots, with the amount of this mRNA being greater in the former than in the latter (Fig. 4A). Given that the similarity in the nucleotide sequences of OsCRSH1, OsCRSH2, and OsCRSH3 might result in cross-hybridization of the OsCRSH1 probe in Northern analysis, even under the high stringency conditions   DECEMBER 7, 2007 • VOLUME 282 • NUMBER 49 used, we examined the expression of all three genes by RT-PCR analysis. The abundance of OsCRSH2 and OsCRSH3 mRNAs in both shoots and roots was much lower than that of OsCRSH1 mRNA (Fig. 4B), suggesting that OsCRSH1 is the most highly expressed gene among the three homologs in the rice tissues tested.

Calcium-activated (p)ppGpp Synthetase
Assay of (p)ppGpp Synthetase Activities of OsCRSH1 in Vitro-To examine further the enzymatic function of OsCRSH1 as a (p)ppGpp synthetase, we cloned the cDNA for the mature form of the protein into the vector pEGST to yield pGCR1⌬N41. The GST-OsCRSH1(⌬N41) protein encoded by this plasmid was synthesized by in vitro transcription and translation and isolated with a glutathione column. The GST portion of the fusion protein was removed on the column by Precision protease, and the released OsCRSH1(⌬N41) portion was isolated at Ͼ95% purity and assayed for (p)ppGpp synthetase activity in the absence or presence of Ca 2ϩ . The protein exhibited pppGpp synthetase activity only in the presence of Ca 2ϩ at concentrations of Ն100 M (Fig. 5A). The predominant 32 P-labeled reaction product was further characterized by two-dimensional TLC (Fig.  5B) and confirmed as pppGpp (18).
The low levels of pppGpp accumulation that were observed in the synthetase reactions under low Ca 2ϩ concentration (Fig.  5A) might be due to the pppGpp hydrolase activity of the OsCRSH1. To assess this possibility, we next examined the hydrolase activity of OsCRSH1. As shown in Fig. 5C, the Mn 2ϩ -dependent pppGpp hydrolase activity was exhibited by E. coli SpoT, which was tested as a positive control enzyme for pppGpp hydrolase, but the hydrolase activity was not observed in the reactions of OsCRSH1. These results thus indicated that OsCRSH1 possesses Ca 2ϩ -activated pppGpp synthetase activity.
To determine whether the Ca 2ϩ dependence of this enzymatic activity is attributable to the COOH-terminal EF-hand motifs of OsCRSH1, we next constructed single-site mutants in which Ca 2ϩ binding to the individual EF-hands would be expected to be disrupted (27). The mutants OsCRSH1(⌬N41)-EF1 and OsCRSH1(⌬N41)-EF2 thus contain glutamine instead of glutamate at the ϪZ position of the corresponding Ca 2ϩbinding loop (Fig. 1C). Neither mutant exhibited Ca 2ϩ -dependent (p)ppGpp synthetase activity (Fig. 5D), indicating that both EF-hand motifs are necessary for this activity.
To examine whether the COOH-terminal region containing the two EF-hand motifs is inhibitory for the pppGpp synthetase activity of OsCRSH1, we prepared a mutant form of the mature protein (OsCRSH1(⌬N41⌬C128)) that lacks the 128 COOH-terminal amino acids (including both EFhands). The mutant protein did not exhibit (p)ppGpp synthetase activity in the absence or presence of Ca 2ϩ (Fig. 5D), showing that the COOH-terminal region is necessary for enzymatic function.
During preparation of the various forms of OsCRSH1, we noticed that the yield of OsCRSH1(⌬N41) was lower than that of the two EF mutants (data not shown). We therefore monitored the cell-free synthesis of [ 14 C]leucine-labeled OsCRSH1(⌬N41) and OsCRSH1(⌬N41)-EF1 in the presence of various concentrations of the Ca 2ϩ chelator 1,2-bis(o-aminophenoxy)ethane-N,N,NЈ,NЈ-tetraacetic acid (BAPTA) to deplete endogenous Ca 2ϩ . The low yield of OsCRSH1(⌬N41) under the standard condition for cell-free synthesis was increased in the presence of BAPTA (Fig. 5E). In contrast, the high yield of OsCRSH1(⌬N41)-EF1 under the standard condition was not increased further by depletion of Ca 2ϩ . Given that stoichiometrically equal amounts of ATP and GTP are converted to pppGpp in the ATP:GTP 5Ј-pyrophosphotransferase reaction, these results suggested that OsCRSH1(⌬N41) is partially activated in the reaction mixture for cell-free protein synthesis, resulting in depletion of the energy source (ATP and GTP) for protein synthesis and a lower protein yield.
Translation-coupled (p)ppGpp Synthetase Activity of OsCRSH Proteins-To examine the enzymatic function of the other two rice CRSH homologs, OsCRSH2 and OsCRSH3, we performed the (p)ppGpp synthetase assay coupled with cell-free protein synthesis. The putative mature forms of the three rice CRSH proteins were thus synthesized and assayed in the presence of [␥-32 P]ATP and in the absence or presence of BAPTA. Protein synthesis was monitored under identical conditions with the exception that [␥-32 P]ATP was replaced with [ 14 C]leucine. All three rice proteins showed (p)ppGpp synthetase activity in the absence of BAPTA, and the activity of each protein was reduced by depletion of Ca 2ϩ (Fig. 6A). We also observed that three rice CRSH proteins had shown different sensitiv- The amplification efficiency for each target DNA fragment was also monitored by PCR for 25 or 30 cycles with the same primer sets and plasmids containing OsCRSH1, OsCRSH2, or OsCRSH3 cDNAs as templates.
ities to BAPTA, suggesting sequence-dependent diversities of their EF-hand functions. Conversely, the synthesis of each protein was increased by the presence of BAPTA (Fig. 6B).
These results thus indicated that the three rice CRSH proteins possess similar Ca 2ϩ -activated (p)ppGpp synthetase activities in vitro. FIGURE 5. Characterization of the (p)ppGpp synthetase activity of OsCRSH1 in vitro. A,OsCRSH1(⌬N41) was synthesized and purified as a GST fusion protein, and the GST moiety was removed before assay of (p)ppGpp synthetase activity in the presence of [␥-32 P]ATP and the indicated concentrations of CaCl 2 . The reaction mixture was analyzed by one-dimensional TLC and autoradiography as described under "Experimental Procedures. " The positions of the origin and signals corresponding to 32 P-labeled pppGpp and ATP are indicated. B, the reaction mixture for assay of the (p)ppGpp synthetase activity of OsCRSH1(⌬N41) in the presence of 500 M CaCl 2 was analyzed by two-dimensional TLC. C, purified OsCRSH1(⌬N41) or E. coli SpoT protein was incubated with the purified 32 P-labeled pppGpp in the absence or presence (1 mM) of Mn 2ϩ . The reaction mixture was analyzed by one-dimensional TLC and autoradiography. As a control, reaction mixture containing no enzyme was analyzed. The positions of the origin and signals corresponding to 32 P-labeled pppGpp and pyrophosphate are indicated. D, OsCRSH1(⌬N41), OsCRSH1(⌬N41)-EF1, OsCRSH1(⌬N41)-EF2, and OsCRSH1(⌬N41⌬C128) proteins were synthesized, purified, and assayed for (p)ppGpp synthetase activity as described in A. E, effect of Ca 2ϩ depletion on in vitro translation of OsCRSH1 proteins. OsCRSH1(⌬N41) and OsCRSH1(⌬N41)-EF1 were synthesized as GST fusion proteins with a cell-free system and by the bilayer method in the presence of [ 14 C]leucine and the indicated concentrations of BAPTA. The OsCRSH1 proteins were purified, released from their GST tags, and analyzed by SDS-PAGE and autoradiography. The position of the 14 C-labeled OsCRSH1 proteins is indicated by the arrowhead.

DISCUSSION
Recent studies have demonstrated the existence of ppGpp in plants (15), the presence of (p)ppGpp synthetase activity in chloroplasts (18), and the nuclear encoding of genes for chloroplastic (p)ppGpp synthetase (16,17). We have now identified plant genes for a novel type of RelA-SpoT homolog, designated CRSH, that localizes to chloroplasts and shows Ca 2ϩ -dependent (p)ppGpp synthetase activity. The COOH-terminal region of CRSH proteins contains two typical Ca 2ϩ -binding (EF-hand) motifs (28). Expression of the putative mature form of the rice CRSH protein OsCRSH1 in an E. coli mutant deficient in RelA and SpoT restored the ability to grow on minimal medium, indicative of functional complementation. Biochemical analysis showed that the (p)ppGpp synthetase activity of rice CRSH proteins is dependent on Ca 2ϩ and that such Ca 2ϩ -dependent activity of OsCRSH1 was abolished by mutation of the Ca 2ϩ binding loop of either EF-hand motif in the COOH-terminal region. Deletion of this COOH-terminal region also abolished the (p)ppGpp synthetase activity of OsCRSH1 in vitro, suggesting not only that the two EF-hand motifs are essential for activation by Ca 2ϩ but that this region of the protein also provides structural support for (p)ppGpp synthetase activity.
With respect to the NH 2 -terminal hydrolase-like domain of the CRSH proteins, their sequences showed no conserved signature corresponding to the feature of the HD domain that is conserved in bacterial hydrolase-active RSH proteins (12). Our experimentalresultindicatedthatOsCRSH1doesnothaveMn 2ϩdependent pppGpp hydrolase activity, and it was consistent with the sequence feature of CRSH proteins.
Evidence indicates that Ca 2ϩ functions as a second messenger in many aspects of plant physiology (29,30). With regard to identification of the sensors of the cellular Ca 2ϩ signal, gene mining revealed ϳ250 genes for EF-hand-containing proteins in the Arabidopsis genome (31,32). These proteins include enzymes, contributors to transcription or translation, participants in protein-protein or protein-nucleic acid interaction, as well as a large number of molecules of unknown function, suggesting that Ca 2ϩ -mediated signaling regulates many processes in plants and that EF-hand-containing proteins are key mediators of Ca 2ϩ action.
In chloroplasts, Ca 2ϩ is required for several important processes. It is thus essential for the function of the oxygen-evolving complex of photosystem II in the thylakoid lumen (33). Stromal Ca 2ϩ also serves a regulatory role in photosynthesis via activation of calmodulin-dependent NAD kinase (34,35). In addition, trafficking of Ca 2ϩ across the thylakoid membrane regulates the light-dependent activity of enzymes of the Calvin-Benson cycle (36). A Ca 2ϩ -dependent ATPase-like protein and a Ca 2ϩ -H ϩ antiporter have been identified in the plastid envelope (37) and thylakoid membrane (38), respectively. Furthermore, both circadian oscillation of Ca 2ϩ concentration in chloroplasts (39) as well as a marked increase in stromal Ca 2ϩ concentration after the transition of plants from light to dark (40) have been described. The peak of this latter effect was observed 20 -25 min after the transition and represented a change in the free Ca 2ϩ concentration of the stroma from 150 nM to between 5 and 10 M (40). In the present study, we found that activation of the (p)ppGpp synthetase activity of OsCRSH1 required a Ca 2ϩ concentration of 100 M in vitro. We previously showed that the (p)ppGpp synthetase activity of intact pea chloroplasts requires the 70 S ribosome (18), suggesting that additional factors such as the 70 S ribosome might increase the activity of OsCRSH1 or render it sensitive to Ca 2ϩ at lower concentrations. An increase in Ca 2ϩ concentration in the chloroplast stroma may thus activate the (p)ppGpp synthetase activity of CRSH proteins and thereby regulate the genetic system of the chloroplast in a manner similar to that apparent in the bacterial stringent response.
The existence and conservation of multiple types of chloroplastic RelA-SpoT homolog, including RSH1 to Ϫ3 and CRSH1 to Ϫ3, suggest that (p)ppGpp signaling plays an important role in regulation of the genetic system of plastids in land plants. Whereas RSH genes are conserved among bacteria, land plants, and algae, CRSH genes appear to be present only in land plants. We suggest that Ca 2ϩ -activated (p)ppGpp signaling may have arisen together with the evolution of Ca 2ϩ signaling in the chloroplasts of land plants. Such coevolution FIGURE 6. Translation-coupled (p)ppGpp synthetase activity of OsCRSH proteins. A, the putative mature forms of OsCRSH1 (OsCRSH1(⌬N41)), OsCRSH2 (OsCRSH2(⌬N14)), or OsCRSH3 (OsCRSH3(⌬N12)) were synthesized with a cell-free system according to the bilayer method and in the presence of [␥-32 P]ATP and the indicated concentrations of BAPTA. As a negative control, the reaction was also performed with mRNA for green fluorescent protein (GFP) in the absence of BAPTA. After translation, the reaction mixtures were processed and analyzed for (p)ppGpp synthesis as in Fig. 5A. B, the same template mRNAs used for cell-free translation in A were subjected to cell-free protein synthesis by the bilayer method in the presence of [ 14 C]leucine and the indicated concentrations of BAPTA. The reaction mixtures (lanes T) and the soluble fractions thereof isolated by centrifugation (lanes S) were analyzed by SDS-PAGE and autoradiography. The arrowheads indicate the 14 Clabeled OsCRSH proteins. might have enabled land plants to better adapt to adverse circumstances, such as environmental stress, physical injury, and pathogen attack.