Structural Insights into the Inactive Subunit of the Apicoplast-localized Caseinolytic Protease Complex of Plasmodium falciparum*

Background: In several organisms, caseinolytic proteases have active and inactive subunits termed ClpP and ClpR, respectively. Results: The x-ray structure of ClpR from Plasmodium falciparum (PfClpR) was solved. Conclusion: PfClpR monomer has a similar fold as ClpP but the PfClpR heptamer exhibits a more open ring than a ClpP heptamer. Significance: This is the first structure of a ClpR subunit. The ATP-dependent caseinolytic protease, ClpP, is highly conserved in bacteria and in the organelles of different organisms. In cyanobacteria, plant plastids, and the apicoplast of the genus Plasmodium, a noncatalytic paralog of ClpP, termed ClpR, has been identified. ClpRs are found to form heterocomplexes with ClpP resulting in a ClpRP tetradecameric cylinder having less than 14 catalytic triads. The exact role of ClpR in such a complex remains enigmatic. Here we describe the x-ray crystal structure of ClpR protein heptamer from Plasmodium falciparum (PfClpR). This is the first structure of a ClpR protein. The structure shows that the PfClpR monomer adopts a fold similar to that of ClpP, but has a unique motif, which we named the R-motif, forming a β turn located near the inactive catalytic triad in a three-dimensional space. The PfClpR heptamer exhibits a more open and flat ring than a ClpP heptamer. PfClpR was localized in the P. falciparum apicoplast as is the case of PfClpP. However, biochemical and structural data suggest that, contrary to what has been observed in other organisms, PfClpP and PfClpR do not form a stable heterocomplex in the apicoplast of P. falciparum.

The ATP-dependent caseinolytic proteases (Clp) 6 have essential roles in protein quality control by removing misfolded, damaged, and regulatory proteins. These proteases are highly conserved in prokaryotes, in the mitochondria of many eukaryotes, and in the chloroplast of plants (1)(2)(3)(4). The proteases consist of a proteolytic component for protein degradation and a chaperone of the AAA ϩ superfamily (5, 6) that uses ATP for substrate recognition, unfolding, and then transfer to the proteolytic component. Clp ATPase chaperones typically form hexameric complexes and contain, in addition to other domains, one (class II) or two (class I) nucleotide-binding AAA ϩ domains that have characteristic Walker A and Walker B nucleotide binding and recognition motifs (5,(7)(8)(9)(10).
The proteolytic component, termed ClpP, forms two heptameric rings resulting in a cylindrical structure (3,11). ClpP is a serine protease containing the canonical Ser-His-Asp catalytic triad. Substrates enter the proteolytic chamber through the two axial pores located on opposite sides of the cylinder; the generated peptides of about 7 to 8 residues (12,13) then exit the chamber through equatorial side pores that we proposed to form due to the dynamics in the structure of the ClpP complex (14 -20). The AAA ϩ chaperones catalyze the translocation of substrates through the ClpP axial pores and might also modulate the dynamics of the ClpP cylinder for substrate degradation and the exit of the generated peptides.
In organisms such as cyanobacteria and many plants, multiple isoforms of ClpP were found and a noncatalytic paralog of the protease, termed ClpR, was also identified (1,3,21,22). ClpR has high sequence similarity to ClpP but is noncatalytic because it lacks the catalytic Ser residue, however, it was shown to be a core component of the Clp protease complex in cyanobacteria (23,24) and plants (25). In these organisms, the Clp chaperone-protease complex typically consists of several paralogs of ClpP/ClpR and multiple ATPase chaperones. For exam-ple, the unicellular cyanobacterium Synechococcus elongatus has three ClpP paralogs and one ClpR. Three proteolytic core complexes were identified in S. elongatus: SeClpP3R, SeClpP1P2, and SeClpP1R that interact with different AAA ϩ chaperones (23). Additional studies showed that the SeClpP3R cylinder consists of two identical heptameric rings of three ClpP3 and four ClpR arranged in a specific layout (24).
The Clp core complex in the plastid of Arabidopsis thaliana consists of five different proteolytic ClpP proteins (ClpP1, ClpP3-ClpP6) and four different ClpR proteins (ClpR1-ClpR4) (25). Recently, the subunit stoichiometry of the chloroplast AtClpRP tetradecameric complex was characterized by mass spectrometry and was found to have one heptameric ring consisting of ClpP3, -4, -5, -6 in a 1:2:3:1 ratio and the other heptameric ring consisting of three ClpP1s and one of each of ClpR1, -2, -3, -4 (26). Analyses of various ClpP and ClpR mutants in plants showed that the proteolytic and nonproteolytic subunits have differential contributions to the function of the complex but all of them, except for ClpR1, are essential for embryo or seedling development (21,22).
Recent studies from our group (27,28) demonstrated that the Plasmodium falciparum protozoan parasite that causes malaria in humans contains four Clp ATPases, which we termed PfClpB1, PfClpB2, PfClpC, and PfClpM, one PfClpP (gene ID PF3D7_0307400; old gene ID was PFC0310c), and one PfClpR (gene ID PF3D7_1436800; old gene ID was PF14_0348). Expression of all Clp chaperones and proteases was confirmed in blood-stage parasites. PfClpB1, PfClpC, PfClpM, and PfClpP were experimentally found to be localized in the apicoplast. PfClpB2 (also known as Hsp101) was found in the parasitophorous vacuole, whereas PfClpR was only predicted to be targeted to the apicoplast but not experimentally confirmed. Unlike the plant plastids, the apicoplast is a nonphotosynthetic organelle that does, however, accommodate several important metabolic pathways essential for parasite survival. Hence, proteins in the apicoplast are potential targets for the development of novel antimalarial drugs. Indeed, inhibiting PfClpP using ␤-lactones showed that the protease plays an important role in development of functional apicoplasts (28). We earlier showed that both recombinant PfClpP and PfClpR can form separate homoheptameric rings as observed by size exclusion chromatography, analytical ultracentrifugation, and electron microscopy (27,28). However, the x-ray structure of PfClpP solved in our study showed that the protein oligomerizes as a compacted tetradecamer at high concentrations required for crystallography. No stable complexes were observed between PfClpP and PfClpR (27).
In this work, we describe the x-ray crystal structure of the PfClpR protein. This is the first structure of a ClpR protein. The structure shows that the PfClpR monomer adopts a fold similar to that of other ClpPs. However, arrangement of the PfClpR heptamer has some unique features including a more open and flat ring, as well as, the presence of a unique motif surrounding the inactive catalytic site. The biochemical and structural data indicate that both PfClpP and PfClpR are localized in the parasite apicoplast, but that they form either an unstable complex or separate complexes.

EXPERIMENTAL PROCEDURES
Parasite Culture, Plasmid Construct, and Parasite Transfection-Plasmodium falciparum strain 3D7 was cultured with human erythrocytes (4% hematocrit) in RPMI media (Invitrogen) supplemented with 10% Oϩ human serum using a protocol described previously (29). Parasite cultures were synchronized by repeated sorbitol treatment following the protocol of Lambros and Vanderberg (30). To generate a transfection vector construct, the full-length PfclpR gene (732 bp) was amplified from P. falciparum 3D7 genomic DNA. The amplified PCR product was then cloned in-frame to the N terminus of GFP into the vector pHH2 (31,32). The PfClpR-GFP fragment was subcloned into the XhoI site of transfection vector pARL1a ϩ (52) to yield construct pARL-ClpR-GFP. Synchronized P. falciparum 3D7 ring-stage parasites were transfected with 100 g of purified plasmid DNA (Plasmid Maxi Kit, Qiagen) by electroporation (33), and the transfected parasites were selected over 2.5 nM WR99210 drug.
Isolation of Total DNA and RNA, cDNA Synthesis, and Quantitative Real-time PCR-The genomic DNA was isolated from an in vitro culture of P. falciparum following a standard protocol (34). Total RNAs were isolated from synchronized P. falciparum 3D7 parasite cultures using a mini RNA isolation kit (Qiagen). An aliquot of 50 ng of total RNA was used to synthesize cDNA using cDNA synthesis kit (Invitrogen) following the manufacturer's protocol. Gene-specific primers were designed using Beacon Designer 4.0 software, for PfclpR, PfclpP, eba175, and falcipain-2 (35), and 18 S rRNA control primers were used following Blair et al. (36). Quantitative real-time PCR was carried out in triplicate using the iCycler version 3.0 (Bio-Rad). Each reaction contained equal amounts of cDNA, 100 ng, of both the gene-specific primers, and 1ϫ SYBR Green PCR mix (Bio-Rad). Threshold cycle (C t ) values were calculated using iCycler software. Standard curves for each gene were obtained by using different dilutions of wild-type genomic DNA (100 to 1 ng) as template, and these standard curves were used to determine genome equivalents of C t values for the respective gene and 18 S rRNA in each RNA sample (36). Genome equivalents of each gene were normalized using that of 18 S rRNA for all the RNA samples.
Fractionation of Parasite Lysate-Parasites were isolated from tightly synchronized P. falciparum 3D7 cultures at different developmental stages by lyses of infected erythrocyte with 0.15% saponin. Cell-free parasite lysate was prepared from the parasite pellet and fractionated as described earlier (28). Briefly, the parasite pellet was suspended in 1ϫ PBS, lysed by three freeze-thaw cycles, and the lysate was clarified by centrifugation at 20,000 ϫ g for 30 min at 4°C. The cell-free extract (ϳ5 mg protein) was fractionated on a Superose 6 HR 10/30 column (GE Healthcare) equilibrated with the lysis buffer. Eluted fractions were analyzed by Western blot using the respective antibodies. Antibodies used include: rabbit anti-PfClpP (1:1,000), rabbit anti-PfClpR (1:1,000), mice anti-GFP (1:1,000); rabbit anti-histidine-rich protein II (HRPII) (1:2,000), secondary antibody (anti-rabbit or anti-mouse, 1:2,000) conjugated to horseradish peroxidase. Bands were visualized using the ECL detection kit (GE Healthcare).
Fluorescence Microscopy and Indirect Immunofluorescence Assay-P. falciparum culture transfected with pARL-ClpR plasmid was synchronized by two sorbitol treatments. Parasites at different developmental stages were collected from the culture for fluorescence microscopy and stained with DAPI at a final concentration of 2 g ml Ϫ1 and MitoTracker Red CMXRos (Invitrogen) at a final concentration of 20 nM in 1ϫ PBS for 15 min at 37°C. The parasites were viewed using a Nikon TE 2000-U fluorescence microscope.
Cryo-immunoelectron Microscopy-Immunoelectron microscopy was carried out on transgenic P. falciparum parasites expressing PfClpR-GFP at trophozoite stages as described earlier (28,37). Parasites were fixed in 4% paraformaldehyde, 0.04% glutaraldehyde in 1ϫ PBS at 4°C for 1 h and, subsequently, embedded in gelatin and infiltrated with a cryo-preservative and plasticizer (2.3 M sucrose, 20% polyvinyl pyrrolidone). After freezing in liquid nitrogen, samples were sectioned with a Leica Ultracut UCT cryo-ultramicrotome (Leica Microsystems, Bannockburn, IL) at Ϫ260°C. Ultra-thin sections were blocked with 5% fetal bovine serum and 5% normal goat serum in 1ϫ PBS for 30 min and, subsequently, stained with rabbit anti-GFP antibody (Abcam, 1:500 dilution in blocking buffer), washed thoroughly, and then incubated with 18-nm colloidal gold-conjugated anti-rabbit IgG for 1 h. Sections were stained with 0.3% uranyl acetate, 1.7% methyl cellulose and visualized under a JEOL 1200EX transmission electron microscope (JEOL USA, Peabody, MA). All labeling experiments were conducted in parallel with controls omitting the primary antibody or using preimmune sera as primary antibodies.
Protein Expression and Purification-PfClpR constructs with N-terminal His 6 tag followed by a tobacco etch virus cut site were expressed in Escherichia coli SG1146, which lacks the gene coding for the endogenous E. coli ClpP. Cells were grown in Terrific Broth media containing 100 g/ml of ampicillin and 34 g/ml of chloramphenicol at 37°C. Once the A 600 reached 0.6, protein expression was induced by the addition of 1 mM isopropyl 1-thio-␤-D-galactopyranoside. Cells were incubated overnight at 18°C and then harvested by centrifugation and re-suspended in buffer A (50 mM Tris-HCl, pH 7.5, 300 mM NaCl, 10 mM imidazole, 10% (v/v) glycerol, and 1 mM ␤-mercaptoethanol). Cells were sonicated or subjected to French press and the protein was purified using a column of Ni 2ϩnitrilotriacetic acid-agarose beads (Qiagen). After loading the sample, the column was washed in three steps using buffer A with 20, 50, and 100 mM imidazole consecutively. The protein was eluted with 500 mM imidazole. The purest fractions were pooled and dialyzed in buffer B (50 mM Tris-HCl, pH 7.5, 25 mM NaCl, 10% (v/v) glycerol, and 1 mM DTT), and subjected to ion exchange chromatography using a MonoQ column. The purest fractions were pooled once more and subjected to size exclusion chromatography using a Superdex 200 HR 10/300 GL (Amersham Biosciences) column in buffer C (10 mM HEPES, pH 7.5, and 300 mM NaCl). Fractions were pooled, concentrated, and quantified using absorbance with the following extinction coefficients: 16,515 M Ϫ1 cm Ϫ1 for PfClpR(P61-E244) and 19,495 M Ϫ1 cm Ϫ1 for PfClpR(S49-E244) obtained from the ProtParam Tools on the ExPASy Server (38).
Crystallization, Data Collection, and Processing-The hanging-drop vapor-diffusion method of crystallization was used to obtain PfClpR crystals. Initially, commercial screening kits, such as Hampton Index screen and Nextal JCSGϩ Suite, were used to find a starting condition for crystallization. Single crystals were obtained for two different constructs of PfClpR: PfClpR(P61-E244) and PfClpR(S49-E244). Crystals of the first construct were obtained in crystallization conditions containing 0.1 M Tris-HCl, pH 8.5, and 20% (v/v) ethanol. Crystals of the second construct were obtained in a crystallization condition containing 0.25 M calcium acetate, pH 8.6, 7.5% PEG 20K, and 7.5% PEG MME500. 1 l of crystallization solution was added to 1 l of ϳ14 mg/ml of PfClpR protein. The reservoir contained 700 l of crystallization solution. Crystals typically grew in 2 days at room temperature. We were not able to grow crystals in cryoprotectant condition, hence, single crystals were transferred gradually to new drops with higher concentrations of glycerol or ethylene glycol. Several such crystals were tested for x-ray diffraction. The best crystals were flash-frozen in liquid nitrogen in the presence of 25% ethylene glycol for PfClpR(P61-E244) and 10% glycerol for PfClpR(S49-E244). Diffraction data at 2.8-Å resolution were collected for crystals of PfClpR(P61-E244) at the beamline 19-ID at the Advanced Photon Source, Argonne National Laboratory. Diffraction data at 2.9-Å resolution were collected for crystals of PfClpR(S49-E244) at the CHESS A1 beamline. Data were processed and scaled with HKL2000 (39). Phaser program (53) from the CCP4 crystallographic program suite was used to search for 7 copies within the asymmetric unit. The program found 3 copies and the symmetry between two subunits was used to generate the heptamer. PfClpR structure was visualized and built using COOT (40) and refinement was done with REFMAC (41) and Phenix (42). Crystallographic details and refinement statistics are summarized in Table 1. The coordinates have been deposited in the protein structure database under the accession numbers 4GM2 for PfClpR(P61-E244) and 4HNK for PfClpR(S49-E244).

Localization of PfClpR in Transgenic
Parasites-Initially, we carried out experiments to confirm the apicoplast localization of PfClpR. According to PlasmoDB (43), the first 49 residues of PfClpR are predicted to harbor the hydrophobic N-terminal signal sequence and the apicoplast-targeting transit peptide sequence (Fig. 1A). A GFP-targeting approach was employed to study the localization of PfClpR in the parasite by expressing plasmid-borne full-length PfClpR protein with a C terminally fused GFP in transgenic parasites (Fig. 1A). Immunoblot analysis using GFP-specific antibodies confirmed the presence of the fusion protein in trophozoite-stage transgenic parasites and not in the wild type parasite lines (Fig. 1B). Western blot against HRPII was used as loading control (Fig. 1B). Fluorescent microscopic images of the transgenic parasites at early and late trophozoite stages showed that the PfClpR-GFP fusion protein was localized in an apicoplast-like cellular organelle (Fig. 1C,  green). To ascertain that the GFP fusion protein is not in the parasite mitochondria, the transgenic parasites were co-stained with MitoTracker dye, the mitochondria-specific live stain. The GFP fluorescence pattern (Fig. 1C, green) was observed in close association but distinct from mitochondrial staining (Fig. 1C,  red). The parasite nuclei were stained with DAPI (Fig. 1C, blue). To further ascertain the localization of PfClpR-GFP fusion protein in the apicoplast, immunoelectron microscopic studies were carried out. Labeling with anti-GFP antibody followed by gold-labeled secondary antibody demonstrated the localization of the fusion protein in the lumen of the apicoplast, which was identified as a four-membrane structure in the parasite (Fig.  1D). From these experiments, we concluded that PfClpR is indeed localized to the apicoplast.
Analysis of the Transcription and Translation of PfClpR in the Asexual Blood-stage Parasites-The transcriptional pattern of the PfclpR gene in different developmental stages of the asexual blood-stage parasites was assessed by quantitative real-time PCR using total RNA samples prepared from tightly synchronized P. falciparum 3D7 parasite cultures at 8, 16, 30, 40, and 48 h after invasion. These results showed maximum transcription of PfclpR in late-trophozoite and early-and late-schizont-stage parasites ( Fig. 2A). Maximum transcription of PfclpP was also observed in the same parasite stages. As control, the same set of samples were also analyzed by quantitative PCR for two other P. falciparum genes, erythrocyte binding antigen-175 (eba-175) and the cysteine protease falcipain-2. As expected, eba-175 transcript levels were found to be maximum in samples from late schizont-stage parasites, whereas falcipain-2 showed maximum transcript levels in trophozoite-stage parasites ( Fig. 2A).
Western blot analysis of total parasite lysates detected a band of ϳ23 kDa in the trophozoite and schizont-stage parasites using anti-ClpR antibodies (Fig. 2B). This band corresponds to the calculated molecular mass of the processed PfClpR corresponding to residues ϳ49 -244 (27) (Fig. 1A). No band was , pro-domain, which gets cleaved off during processing, and the putative protease domain. B, immunoblot analysis using GFP-specific antibodies and trophozoite-stage wild type (WT) and transgenic parasites expressing PfClpR-GFP. A band of ϳ48 kDa, representing the GFP fusion protein, is recognized by GFP-specific antibodies in the transgenics, but not in the wild type parasite lines. A parallel blot was probed with anti-HRPII antibodies to show equal loading. C, fluorescent microscopic images of transgenic parasites at early and late trophozoite stages expressing PfClpR-GFP fusion protein (green) co-stained for live mitochondria using MitoTracker Red (red). The parasite nuclei were stained with DAPI (blue) and slides were visualized by fluorescence microscope. Scale bar represents 2 m. D, localization of PfClpR by immunoelectron microscopy. Ultra-thin sections of transgenic P. falciparum parasites expressing PfClpR-GFP were labeled with anti-GFP antibody and gold-labeled secondary antibody. Labeling was observed in the apicoplast, which has the characteristic four membranes. Scale bar represents 250 nm. The Structure of ClpR from P. falciparum JANUARY 11, 2013 • VOLUME 288 • NUMBER 2 JOURNAL OF BIOLOGICAL CHEMISTRY 1025 detected using preimmune sera; in addition, the anti-PfClpR antibodies did not react with the lysate of uninfected RBCs (data not shown). Western blot against HRPII was used as loading control (Fig. 2B). As expected anti-PfclpP antibodies detected a band in the trophozoite-and schizont-stage parasites. Both the transcription and translation analyses results suggest that PfClpR is expressed in blood-stage parasites at trophozoite and schizont stages, which is similar to that of PfClpP (28).
PfClpR Forms a Multisubunit Complex-To determine the size of the native protein complex of PfClpR, the parasite lysate was subjected to size exclusion chromatography using a Sepharose-6 column. The eluted samples were analyzed by Western blot using anti-PfClpR, as well as anti-PfClpP antibodies. The native PfClpR was detected in fractions of molecular mass of ϳ440 kDa (Fig. 2C). On the other hand, PfClpP eluted as a much larger complex of about 660 kDa. We had found earlier that the mature, processed PfClpP and PfClpR have similar masses of about 25 kDa (27,28). Hence, PfClpP seems to be part of a larger complex that, unexpectedly, does not include PfClpR. This suggests that PfClpP and PfClpR do not form a stable protein complex in the parasite or that they form separate complexes.
Crystallization and Structure Determination of PfClpR-Because no structure has been reported for ClpR from any organism and to better understand the function of this inactive protease, we carried out experiments to solve the x-ray structure of PfClpR. We had shown earlier that the mature PfClpR consists approximately of residues 49 -244 (Fig. 3A). Based on the crystal structure of ClpP from different species, residues from Ser 49 to Ile 58 of PfClpR might form flexible axial loops, which would present a challenge to crystallization. Therefore, attempts were made to crystallize both PfClpR(S49-E244) as well as PfClpR (P61-E244). The constructs contained a His 6 tag and a cut site for the tobacco etch virus protease (MGSSH 6 SS-GRENLYFQG) prior to Ser 49 or Pro 61 ; these residues were not removed in the crystallization trials. Single crystals were obtained and diffracted at 2.8-Å resolution for PfClpR(P61-E244) and 2.9 Å for PfClpR(S49-E244) ( Table 1). The structure was solved by molecular replacement using PfClpP protomer as template. When PfClpP heptamer was used as a template, the molecular replacement program failed to find a solution because PfClpP and PfClpR heptamers have different conformations (see below).
Overall Structure of PfClpR-Using gel filtration, analytical ultracentrifugation, and electron microscopy, we previously confirmed that PfClpR exists mainly as a heptamer in solution (27). For both constructs, PfClpR was also crystallized as a 7-fold symmetric single ring of ϳ106 Å in diameter and a central pore of ϳ30 Å (Fig. 3, B-D). The PfClpR(S49-E244) asym-metric unit included two 7-fold heptamers (chains A to N) related by a translational symmetry (not a double ring structure). The asymmetric unit of PfClpR(P61-E244) includes a 7-fold heptamer (chains A to G) with the same crystal packing as for PfClpR(S49-E244).
Overall, PfClpR(P61-E244) and PfClpR(S49-E244) structures are similar (Fig. 4) showing that the presence of ethanol in the crystallization condition of PfClpR(P61-E244) does not drastically affect the PfClpR structure. The root mean square deviation for the protomers is 0.66 Å and for the heptamers is 1.4 Å. Below, we will mainly refer to PfClpR (P61-E244).
This first atomic structure of a ClpR is similar to other ClpP structures with a root mean square deviation for PfClpR(P61-E244) of 1.5 Å to PfClpP and 2 Å to Escherichia coli ClpP (EcClpP). Each PfClpR subunit has an ␣/␤-fold made up of six

TABLE 1 Crystallographic data collection and refinement statistics
Values in parentheses refer to the highest resolution shell. The various crystallographic parameters are defined as follows: R merge , ⌺͉I i Ϫ ϽIϾ͉/⌺I i , where I i is the intensity of the i-th observation, ϽIϾ is the mean intensity of the reflection, and the summation extends over all data. R cryst , ⌺͉F obs Ϫ F calc ͉/⌺F obs , where F obs and F calc represent, respectively, the observed and calculated structure factors. R free was calculated using 5% of the observed reflections excluded from refinement. Excluded data were randomly selected.

JOURNAL OF BIOLOGICAL CHEMISTRY 1027
repeats of the ␣/␤ unit (␣A/␤1, ␣B/␤2, ␣C/␤3, ␣D/␤4/␤5, ␣F/␤7, ␣G/␤8) forming the head domain and an additional protruding ␣/␤ unit (␣E/␤6) forming the handle region (Fig. 3, A  and B). Residues Pro 61 -Gln 174 and Lys 211 -E 244 comprise the head domain, whereas residues Ser 175 -Glu 210 form the handle region. As is typically found in other ClpPs, the head domain comprises 8 ␤ strands arranged into two layers of ␤ sheets. On one side, the ␤ sheets pack against a layer of ␣ helices from the same subunit whereas, on the other side, they pack against a layer of ␣ helices of the neighboring subunit within the heptamer (Fig. 3C).
We were not able to build the axial loop (Ser 49 -Ile 60 ) of PfClpR(S49-E244) because of weak electron density. Also, the handle region is either absent or is missing some residues in all the subunits of this construct. For PfClpR(P61-E244), the final models of subunits A and E include all residues, Pro 61 -Lys 244 . Residues Ser 175 -Lys 192 , Gln 174 -Ile 189 , Gln 174 -Glu 193 , Gln 174 -Asn 186 , and Tyr 176 -Ile 187 in the handle region are missing in subunits B, C, D, F, and G, respectively. As shown in Figs. 3D and 5, the two handle regions that we could trace in PfClpR(P61-E244) have different orientations due to crystal packing.
The Ser-His-Asp catalytic triad is missing in PfClpR and is replaced by residues Gly 148 -Asn 173 -Asp 223 , which we will call the inactive catalytic triad, which are located in a cleft between the head domain and handle region similar to the location of the catalytic triad in ClpPs (Figs. 3, B and C, and 6A). The head domain of PfClpR includes an insertion formed by the 10 residues Asp 111 -Gly 120 not present in ClpPs (Fig. 3A); we name Asp 109 -Gly 120 as the R-motif. The equivalent segments in A. thaliana and cyanobacterium Synechococcus have been modeled as an extension of helix C followed by a loop (24,25). In our PfClpR structure, the R-motif adopts a structured ␤ turn that is largely equivalent in all 7 subunits (Fig. 3D) and protrudes from the head domain without affecting the ␣/␤-fold (Fig. 3D). It also includes an extension of helix C (residues NG). The R-motif is adjacent to the presumed active site and faces the internal chamber of the heptamer (Fig. 6A).
In ClpP, each subunit has a hydrophobic groove forming a continuous substrate-binding surface leading to the catalytic triad (44,45). An examination of surface topology and surface potential of PfClpR reveals that this groove is present in the PfClpR structure and is formed by residues Ile 121 , Ile 149 , Ile 152 , Leu 172 , Asn 173 , Ser 175 , Val 201 , Ile 205 , and Leu 220 (Fig. 6, A and  B). The presence of the R-motif results in the formation of a deep cleft without significantly restricting accessibility to the groove from the internal chamber of the PfClpR heptamer (Fig.  6B). This is clearly observed if the PfClpR structure is overlaid with the structure of the Helicobacter pylori ClpP protomer   that is in complex with a heptapeptide NVLGFTQ (PDB code 2ZL2). The heptapeptide is found to nicely fit into that deep cleft of PfClpR (Fig. 6B). The amino acid sequence of the R-motif is highly conserved in the genus Plasmodium with a consensus sequence of: D(I/L)(D/E)NNKI(I/V)NLNG. This suggests that the motif may play an essential role related to the function of this protein.
The Oligomerization of PfClpR-The PfClpR(P61-E244) heptamer is stabilized by extensive interactions among the head domains with about 30% of the total surface area making contacts within the heptamer. Compared with EcClpP and PfClpP structures, the PfClpR(P61-E244) heptamer is compressed by about 10 Å along the z axis and is wider by about 14 Å along the x axis (Fig. 7A). This is the result of a twist of each subunit in the PfClpR(P61-E244) heptamer by about 15°outwards leading to a more open structure of the PfClpR(P61-E244) heptamer (Fig. 7, B and C). As a consequence of the open shape of PfClpR(P61-E244), the surface area of the heptameric complex is larger than that of PfClpP or EcClpP: ϳ56,630 Å 2 for PfClpR(P61-E244), ϳ46430 Å 2 for PfClpP, and ϳ50,970 Å 2 for EcClpP. This observation leads us to conclude that a flat ring of PfClpR is probably not favorable for an interaction with the PfClpP ring, unless a conformational change in PfClpR is induced. This might explain our inability to observe a stable PfClpRP complex using the purified proteins (27).

DISCUSSION
In eukaryotes, the 26 S proteasome plays an essential role in controlling the levels of key regulatory proteins and in degrading abnormal polypeptides; it also plays a key role in cell cycle progression and the regulation of numerous transcription factors (46). In prokaryotes, mitochondria, and plant chloroplasts, these tasks are carried out by the Clp chaperone-protease system (47). We had earlier identified and characterized the Clp chaperones and proteases in P. falciparum (27,28,48). In this study, we have experimentally confirmed that, like PfClpP, PfClpR is localized in the apicoplast of the parasite (Fig. 1, C and  D). Furthermore, both proteins were found to be expressed in blood-stage parasites at trophozoite and schizont stages (Fig. 2,  A and B, and Ref. 27 and 28) suggesting that they might function together. However, it was surprising to find that the two proteins do not co-elute when a P. falciparum lysate is fractionated (Fig. 2C).
We also determined the crystal structure of PfClpR protein, which exists as a heptamer in solution (27). Although the structure of PfClpR monomer is similar to that of PfClpP (Figs. 3D and 7C), the dimensions and shapes of PfClpP and PfClpR heptameric rings are different (Figs. 4 and 7, A and B). The structure we obtained for PfClpR showed a flattened heptameric complex compared with that of ClpP (Fig. 7A) that is unlikely to be able to interact stably with an equivalent heptameric PfClpP by the intercalation of the handle regions as typically observed in ClpP tetradecamers. These results correlate with the presence of separate PfClpR and PfClpP complexes in the parasite as observed in the size exclusion fractionation experiment (Fig.  2C). However, at this stage, we cannot exclude the possibility that a AAA ϩ chaperone or some other cofactor might induce a conformational change in PfClpR or PfClpP to allow for the formation of a transient tetradecameric complex consisting of two homo-heptameric rings PfClpR 7 -PfClpP 7 . The presence of a substrate might also stabilize a PfClpR 7 -PfClpP 7 complex. Alternatively, a functional hetero-heptameric PfClpRP complex in the apicoplast may be present at very low levels to be detected by the fractionation experiment of Fig. 2C; however, we have not yet been able to form such a complex in vitro. It should be noted that PfClpP formed a tetradecamer in the crystal but a heptamer in solution (27), indicating a concentrationdependent oligomerization.
If indeed PfClpR and PfClpP do not form a complex in the apicoplast of Plasmodium, then this suggests the presence of a separate inactive PfClpR complex with distinct activities from PfClpP. PfClpR might indirectly regulate the activity of PfClpP by competing for interactions with cognate AAA ϩ proteins. Alternatively, the PfClpR ring might have a chaperone-like activity. This is suggested by the presence of the deep hydrophobic clefts formed by the R-motif and the substrate-binding surface leading to the inactive catalytic triad (Fig. 6). Exposed hydrophobic segments in unfolded or misfolded proteins can bind in these clefts. However, it should be pointed out that these possible alternative functions for PfClpR would be specific for apicomplexa and would not be consistent with the fact that hetero-oligomeric complexes have been observed in plant plastids and cyanobacteria.
It should be noted that, whereas the R-motif seems to be present in most of the cyanobacterial and plant ClpRs, the amino acid sequence of the motif is not conserved across species. Furthermore, some ClpRs have other insertions as well. For example, SeClpR has an insertion between ␤1 and ␣B, which is not present in plant or Plasmodium ClpRs (24,25,27). The role of such insertions is not known; these insertions might be required to further optimize or modify the function of the protein in a species-specific manner.
In conclusion, our study provides the first structure of a ClpR protein. The structure helps to further our understanding of the role that ClpR plays in different organisms. However, the function of PfClpR in Plasmodium and the possible presence of a PfClpRP complex remains enigmatic and requires additional studies.