Plasmodium falciparum Dynein Light Chain 1 Interacts with Actin/Myosin during Blood Stage Development*

Dynein light chain 1 (LC1), a member of the leucine-rich repeat protein family, has been shown to be engaged in controlling flagellar motility in Chlamydomonas reinhardtii and Trypanosoma brucei via its interaction with the dynein γ heavy chain. In Plasmodium falciparum, we have identified the LC1 ortholog, designated Pfdlc1. Negative attempts to disrupt the dlc1 gene by reverse genetic approaches in both P. falciparum and P. berghei suggest either its essentiality for parasite survival or the inaccessibility of its locus. Expression studies revealed high levels of DLC1 protein in late trophozoites and schizonts, pointing to an unexpected role of this protein in blood-stage parasites as they do not have flagella. Interactions studies and co-immunoprecipitation experiments revealed that PfDLC1 was able to bind to P. falciparum myosin A and actin 1. The PfDLC1 interacting domains present in P. falciparum myosin A and actin 1 were mapped to sequences containing SDIE and/or EEMKT motifs present in the upper 50-kDa segment of the myosin A head domain and in the subdomain IV of actin 1, respectively. Detection of PfDLC1 by fluorescence tagging and immunofluorescence staining using specific antibodies showed a cytoplasmic location similar to actin and immunofluorescence studies showed a co-localization of PfDLC1 and myosin A. Taken together, these findings suggest that PfDLC1 might play an important role in P. falciparum erythrocytic stages by its interaction with myosin A and actin 1, known to be essential for parasite development.

Dynein light chain 1 (LC1), a member of the leucine-rich repeat protein family, has been shown to be engaged in controlling flagellar motility in Chlamydomonas reinhardtii and Trypanosoma brucei via its interaction with the dynein ␥ heavy chain. In Plasmodium falciparum, we have identified the LC1 ortholog, designated Pfdlc1. Negative attempts to disrupt the dlc1 gene by reverse genetic approaches in both P. falciparum and P. berghei suggest either its essentiality for parasite survival or the inaccessibility of its locus. Expression studies revealed high levels of DLC1 protein in late trophozoites and schizonts, pointing to an unexpected role of this protein in blood-stage parasites as they do not have flagella. Interactions studies and co-immunoprecipitation experiments revealed that PfDLC1 was able to bind to P. falciparum myosin A and actin 1. The PfDLC1 interacting domains present in P. falciparum myosin A and actin 1 were mapped to sequences containing SDIE and/or EEMKT motifs present in the upper 50-kDa segment of the myosin A head domain and in the subdomain IV of actin 1, respectively. Detection of PfDLC1 by fluorescence tagging and immunofluorescence staining using specific antibodies showed a cytoplasmic location similar to actin and immunofluorescence studies showed a co-localization of PfDLC1 and myosin A. Taken together, these findings suggest that PfDLC1 might play an important role in P. falciparum erythrocytic stages by its interaction with myosin A and actin 1, known to be essential for parasite development.
Molecular motors that involve myosin, kinesin, and dynein complexes are considered to be the most important driving force responsible for the motility/mobility and the active transport of proteins and vesicles in eukaryotic cells. These motor proteins are powered by the hydrolysis of ATP and convert chemical energy into mechanical work. In Plasmodium falciparum, which is an obligate intracellular parasite, the active penetration across non-permissive biological barriers and host cells is powered by an actin/myosin-based motor complex located at the pellicle of parasites (1,2). Drugs that destabilize the actin filaments or inhibit actomyosin ATPase have been shown to be able to decrease dramatically the invasion of red blood cells by Plasmodium merozoites (3,4). Although the function of actomyosin motor in P. falciparum in mediating motility/invasion appears reasonably clear, the roles of dyneins are still poorly characterized, although they have been shown to play diverse and important roles in many physiological functions (5). Early work by Fowler et al. (6), using monoclonal antibodies directed against dynein heavy and intermediate chains purified from chicken brain, demonstrated that P. falciparum expressed the cross-reactive dyneins only in the late stages of schizogony and in purified merozoites.
More recently the use of a bioinformatics screen of the full P. falciparum genome, together with an in silico study on dynein light chains (7), revealed a total of 17 genes that are expected to encode 7 dynein heavy chains, 2 dynein intermediate chains, 1 dynein intermediate light chain, and 7 dynein light chains. Mass spectrometric proteome analysis of separated female and male of Plasmodium berghei gametocytes revealed the expression of 11 dynein chains predicted to be associated with the axoneme/ flagella of the male gametes (8,9). This includes the ␥ dynein heavy chain orthologs as well as several dynein light chains. In the case of P. falciparum sexual and asexual stages, proteome analysis indicates that at least three dynein heavy chains (MAL7P1.162, PF11_0240, and PF10_0224) and three light chains (PFL0660w, PF11_0148, and MAL8P1.46) are expressed exclusively in gametocytes (10). Taken together, these data are consistent with the notion that these dynein proteins could play their expected role in the flagellar motility of the male gametes, which is in agreement with previous observations made in other organisms (11,12).
In general, dyneins are known to be microtubules-based motors, ATP-driven, and belong to two main classes: the axonemal dyneins responsible for sliding microtubules against one another to generate flagellar and ciliar movements and the cytoplasmic dyneins involved in cargo transport along micro-tubules (13)(14)(15). The axonemal dynein complex is usually composed of ␣, ␤, and ␥ heavy chains, and the cytoplasmic dynein contains two identical heavy chains. To be efficient, the function of the dynein complex must be adequately controlled both in the cytoplasm to direct correctly cargo transport and in the flagella/cilia to obtain the right beat cycle. In this context, it has been shown that the ␥ heavy chain of the axonemal dynein binds with two distinct proteins, a Ca 2ϩ -binding EF-hand protein (16) and a dynein light chain 1 (LC1), 3 which is a member of the leucine-rich repeat (LRR) protein family (17), well known to play their functional roles through protein-protein interactions. The NMR structure of Chlamydomonas reinhardtii LC1 (CrLC1) suggested that its binding is near the ATP binding site of the ␥ heavy chain (XP_001702026), raising a potential role of LC1 in the regulation of the dynein motor activity (12). In addition cross-linking experiments revealed that LC1 also binds directly to an axonemal component of ϳ45 kDa, believed to be a putative microtubule-binding protein (12,17). More recently, Baron et al. (11) demonstrated that the loss of the flagellum dynein LC1 in Trypanosoma brucei by creating an LC1 knockdown mutant by RNA interference results in a defect of flagellar motility leading to an incomplete cell division. So far, the presence and/or implication of LC1 in cytoplasmic dynein complexes have not been reported.
In a screen for the genes related to LRR proteins interactome in P. falciparum blood stages, we cloned a candidate dynein LC1 ortholog, designated PfLRR7 (18). This gene was assigned in PlasmoDB as outer arm dynein light chain 2 (gene identifier MAL8P1.46). Using real-time PCR on cDNA of total RNAs from highly different blood parasite stages, we observed that the highest levels of transcripts were detected in both late trophozoites and schizonts (18). In the light of these studies, together with the fact that PfDLC1 protein is expressed in male gametes, it appears vital to unravel the function(s) of PfDLC1 in P. falciparum.
In this work, we characterize in detail the Pfdlc1 gene and show that the PfDLC1 protein is expressed in almost all stages throughout the blood parasite life cycle with a maximal level in late trophozoites and schizonts and localizes in the cytoplasm. Reverse genetic approaches revealed that the gene encoding LC1 could neither be tagged nor disrupted in both P. berghei and P. falciparum, suggesting that it has an indispensable role in the parasite blood stages. Further biochemical studies demonstrated that PfDLC1 gene product has the capacity to bind P. falciparum actin 1 and myosin A. These data suggest an unexpected role of PfDLC1 in the motor complex involving actin/ myosin A.

EXPERIMENTAL PROCEDURES
Preparation of Parasites-P. falciparum 3D7 clone was grown according to Trager and Jensen (19), in RPMI 1640 medium with 10% human AB ϩ serum, in the presence of O ϩ erythrocytes. Cultures were maintained at 37°C in a humidified atmosphere (5% CO 2 , 5% O 2 , and 90% N 2 ). Parasites were synchronized by a double sorbitol treatment as previously described (20). To isolate total RNA or proteins, parasitized erythrocytes were saponin-lysed (21) and either resuspended in TRIzol (Invitrogen) or in phosphate-buffered saline containing EDTA-free protease inhibitor mixture (Roche Applied Science). For some experiments, infected red blood cells were purified using Percoll-sorbitol density gradients with slight modifications (22). Protein extracts were prepared by five consecutive freeze/thawing cycles with intermediate homogenizing steps using a micro-pestle and 0.7-mm glass beads (Sigma) and subsequent centrifugation at 15,000 ϫ g for 30 min at 4°C. Protein concentrations in the supernatants were determined using a BCA kit (Thermo Scientific).
Complete Molecular Cloning and Analysis of PfDLC1-All primers used throughout this study are listed in supplemental Table 1. The encoding region of Pfdlc1 was initially obtained from first-strand cDNA derived from mRNA prepared from P. falciparum asynchronous culture. The PCR was performed with the F1 and R1 primers. To confirm the stop codon, 3Ј-rapid amplification of cDNA ends was carried out using the SMART kit (Biosciences Clontech). The 3Ј-end was obtained using the forward primer F2 and the adapter primer according to the manufacturer's instructions. To determine the start codon, two forward primers (F3 and F4) derived from the 5Ј-upstream genomic region were tested in PCR on cDNA with the reverse R2 primer derived from the coding region. Only F4 and R2 amplified one PCR product. In all experiments, The PCR products were cloned in TA cloning vector (Invitrogen) and sequenced. Comparative analysis of PfDLC1 protein was performed by DNA Star and ClustalW, Pfam database (www. sanger.ac.uk) and Conserved Domain Database (www.ncbi. nlm.nih.gov/Structure/cdd/wrpsb.cgi).
Generation of P. falciparum Transgenic Parasites-For an episomal expression of PfDLC1-GFP, the full-length coding region of Pfdlc1 was amplified by PCR using the 3D7 cDNA as a template and the primers number 7 and 8 containing BglII and AvrII restriction sites, respectively. The PCR fragment was cloned into TOPO-TA cloning vector (Invitrogen), and its nucleotide sequence was determined. The PCR product was then subcloned in-frame with GFP into pHH2 previously digested with BglII and AvrII. The plasmid carries the human dhfr gene for selection with WR99210.
The Pfdlc1 disruption plasmid (pCAM-Pfdlc1) was generated by inserting a PCR product corresponding to a 5Ј-portion from the Pfdlc1 open reading frame (540 bp) into the pCAM-BSD vector (a gift from Dr. Christian Doerig, INSERM U609, Glasgow, UK), which contains a cassette conferring resistance to blasticidin. The insert was obtained using 3D7 genomic DNA as template and the oligonucleotides F6 and R6, which contain PstI and BamHI sites, respectively. Attempts to check the accessibility of Pfdlc1 locus were performed by transfecting wild 3D7 parasites with 3Ј-tagging constructs. To this end, the 3Ј-end of the Pfdlc1 sequence (700 bp, omitting the stop codon) was amplified by PCR using 3D7 genomic DNA and the primers F7 and R7 containing PstI and BamHI restriction sites, respectively. The 3Ј-tagging plasmids were generated by inserting the PCR product into PstI and BamHI sites of the pCAM-BSDhemagglutinin (HA) or GFP plasmids. Transfections were carried out by electroporation of ring stage 3D7 parasites with 75-100 g of plasmid DNA, according to Sidhu et al. (23). WR99210 or Blasticidin (Invivogen) were added to a final concentration of 5 nM or 2.5 g/ml, respectively, 48 h after transfection to select for transformed parasites. Resistant parasites appeared after 3-4 weeks and were maintained under drug selection. Populations of stably transfected parasites were obtained after 6 weeks. To enrich the populations for integrants, three to four cycles of on/off drug were applied.
The expression of PfDLC1-HA or -GFP fusion protein was checked by Western blotting. Live parasites potentially expressing PfDLC1-GFP were analyzed by fluorescence microscopy as described above. Genotypes of Pfdlc1 knock-out parasites were analyzed by PCR using standard procedures with the primers number 13 specific for Pfdlc1 (derived from the 5Ј-non-translated region and not present in the construct) and 14 specific for the pCAM-BSD vector. Genotypes of Pfdlc1 knock-in were analyzed using the primer F6 and the primer number 15 derived from GFP sequence (P635) or the reverse primer number 16 corresponding to HA tag (P639).
Generation of Pbdlc1 Transgenic P. berghei Strain-The knock-in construct for P. berghei pSD141/CtPfdlc1 was generated by cloning 1009 bp from Pfdlc1 DNA fragment corresponding to the C-terminal part of dlc1 without the stop codon. This target fragment was amplified by PCR using primers 17 and 18 and P. berghei genomic DNA (gene identifier: PB000535.01.0) and cloned into KpnI and ApaI sites of pSD141 vector, respectively (24). To achieve knock-in in P. berghei the vector was linearized with HindIII to allow single homologous recombination event in Pbdlc1 locus (25,26). Transfection of parasites with pSD141/C-terminalPfdlc1 (construction PL1473) in line 507cl1 (mutant RMgm-7; www.pberghei.eu; PubMed ID number 16242190) and selection of pyrimethamine-resistant parasites was performed as described previously (27). Genotypes of selected parasites were analyzed by Southern analysis of Field inversion gel electrophoresis separated chromosomes as described (27).
Recombinant Protein Expression and Antiserum Production-The entire Pfdlc1 cDNA obtained by PCR with the primers F1 and R1 mentioned above was inserted into pQE30 expression vector (Qiagen) using BamHI and NotI sites. For the expression of the GFP protein, the plasmid was generated by inserting a PCR product corresponding to the full-length sequence of GFP into PQE30 vector. The insert was obtained using the pHH2 vector (a gift from Dr. S. Müller, Glasgow, UK) containing the GFP as template and the primers GFP1 and GFP2, which contain BamHI and HindIII restriction sites for subcloning in PQE30. All inserts were checked by DNA sequencing prior to expression of the recombinant proteins in Escherichia coli (strain M15). The expression was carried out according to manufacturer's instructions (Qiagen). The (His) 6-tagged proteins were purified through Ni 2ϩ chelation chromatography. The imidazole-eluted proteins were dialyzed against PBS containing 10% glycerol. The purity checked on SDS-PAGE gel was Ͼ95%. The PfDLC1 recombinant protein was further subjected to peptide mass fingerprint by matrix-assisted laser desorption ionization time-of-flight mass spectrometry to confirm its identity.
For antisera production, the purified recombinant PfDLC1 (rPfDLC1) or recombinant GFP was mixed with complete Freund's adjuvant (100 g per injection) and subcutaneously injected into rats. Animals were boosted twice at intervals of 4 weeks with the same quantity in Freund's incomplete adjuvant. The sera were obtained 2 weeks after the last boost and tested for their titers and specificity by ELISA and Western blotting against recombinant proteins. Preimmune sera were used as negative control. Purification of IgG from sera was performed using caprylic acid as described previously (28).
Detection of PfDLC1 in P. falciparum Erythrocytic Stages-For Western blots, 60 g/lane of P. falciparum-soluble proteins from synchronous and asynchronous cultures were separated on a 4 -12% SDS-PAGE and subsequently blotted onto nitrocellulose. For the detection of native PfDLC1, the blots were probed with primary rat anti-PfDLC1 at 1:50. A mouse monoclonal antibody specific to actin was used at 1:1000 (kindly given by Prof. Dominique Soldati, University of Geneva). The antiserum against PfMyoA described previously (4) was used at 1:1000 in this study. All secondary antibodies were purchased from Jackson ImmunoResearch Laboratories. Horseradish peroxidase-labeled anti-mouse IgG (1:1000 dilution), anti-rat (1:500), and anti-rabbit (1:500) were used as secondary antibodies followed by chemiluminescence detection (Santa Cruz Biotechnology, Santa Cruz, CA).
To further assess the localization pattern, cytoplasmic, nuclear, and membranes fractions of parasites from asynchronous culture were prepared as previously described (29) and analyzed by immunoblotting. The detergent-resistant pellet (cytoskeletal fraction) was then dissolved in 5% (w/v) SDS in 10 mM sodium phosphate, pH 7.4, and subjected to Western blot analysis.
Measurement of Binding of PfDLC1-Binding of PfDLC1 to P. falciparum extracts, actin (Sigma), myosin, or tubulin was assessed by ELISA. Plates were coated with 10 g/ml of each protein in PBS overnight at 4°C. Following washing with PBS-Tween 0.1%, the plates were blocked with PBS containing 1% gelatin for 1 h at room temperature. Coated plates were then incubated with different concentrations of biotinylated PfDLC1, which has been labeled with biotin-N-hydroxysuccinimide according to the manufacturer's instructions (Calbiochem). Incubation of biotin-PfDLC1 with the different proteins was performed in PBS-0.1% Tween at 37°C for 2 h. After 5 washes with PBS-Tween 0.1%, binding was detected using streptavidin-horseradish peroxidase. After an incubation period of 30 min and 5 washes, tetramethylbenzidine substrate (Uptima) was added, and the reaction was stopped by using 2 N HCl. The optical density was measured on an ELISA plate reader at 450 nm. In these experiments, bovine serum albumin, and red blood cells extracts were used as controls.
Inhibition of Binding of PfDLC1 to Rabbit Myosin by Rabbit Actin-Rabbit myosin was coated to ELISA plates at 10 g/ml (100 l/well) for 3 h at 37°C. After saturation with gelatin 1%, increased quantities of bovine serum albumin or rabbit actin were added, and the mixture was incubated overnight at 4°C. After 5 washing steps, biotinylated-PfDLC1 (400 ng/ml) was incubated for 3 h at 37°C. The reaction was revealed as described above.
Far Western Analysis-For far Western blots, P. falciparum proteins (20 g/lane) were run on 4 -12% SDS-PAGE, transferred to nitrocellulose, and blocked with PBS-5% milk for 3 h at room temperature. Blots were incubated with PfDLC1 (10 g/ml) in PBS-5% milk overnight at 4°C, washed with PBS-0.4% Tween, and incubated for 1 h with anti-His antibodies (Qiagen) diluted 1000-fold in PBS-5% milk containing 0.4% Tween and developed with the enhanced chemiluminescence system (Santa Cruz Biotechnology). In these experiments, a recombinant LRR protein of Schistosoma mansoni (30) was used as a control under the same conditions.
Identification of Binding Partners of PfDLC1 in P. falciparum Extracts-For co-immunoprecipitation assays, 50 l of anti-PfDLC1, monoclonal antibody anti-actin, anti-PfMyoA antibodies, anti-GFP, or control serum were incubated for 4 h at 4°C with 100 l of protein G-Sepharose in 0.5 ml of TNE (50 mM Tris, 0.1% Nonidet P-40, 2 mM EDTA). 1 mg of total protein extracts from wild-type or GFP transgenic parasites sorted by fluorescence-activated cell sorting was then added to the mixture. After an overnight incubation at 4°C, Sepharose beads were extensively washed in TNTE (TNE plus 0.1% Triton X-100). In all experiments, bound proteins were eluted with loading buffer (200 mM Tris, 2% SDS, 5% glycerol, 1% ␤-mercaptoethanol), separated by SDS-PAGE, transferred to nitrocellulose filters, and subjected to immunoblot analysis with PfDLC1, GFP, actin, and PfMyoA antibodies.
The peptides used were listed in the supplemental Table 2. Binding of PfDLC1 to synthetic peptides was assessed by ELISA as described above with the exception that the biotin-conjugated PfDLC1 was incubated for 3 h at 37°C.

Molecular Characterization of Pfdlc1
Gene-We initially reported the cDNA corresponding to the predicted encoding region of dlc1 of P. falciparum (gene identifier MAL8P1.46, GenBank TM accession number AY898272). In this work, the 3Ј-rapid amplification of cDNA ends confirmed the predicted stop codon and revealed the presence of 133 nucleotides in the 3Ј-non-translated region before the poly(A) tail. To determine the correct start codon, we used different genomic primers derived from the upstream region of the predicted start codon and performed PCR on total cDNA in the presence of primers derived from the coding Pfdlc1 sequence. By this approach, we showed an open reading frame of 597 bp with a 5Ј-non-translated region of 345 bp. The comparison of cDNA-encoding sequence with the sequenced PCR product obtained from the genomic DNA indicated that the gene is composed of 6 exons interrupted by 5 introns, which are in agreement with the gene organization reported by PlasmoDB (Fig. 1).
The deduced amino acid sequence of the open reading frame corresponds to 199 aa with a 23-kDa predicted molecular mass. In reciprocal BLASTP searches of the GenBank TM database, the highest score was obtained with the flagellar outer arm dynein light chain of Chlamydomonas reinhardtii (40% identity and 66% similarity, accession number Q9XHH2) ( Fig. 2A). The aligned region of PfDLC1 spanned almost the complete sequence, covering residues 5-192 ( Fig. 2A). Proteins family bank analysis confirmed that PfDLC1 belongs to the LRR protein family and displayed 6 LRR motifs that fit with the consensus sequence LXLXXNXL (where L can be replaced by I or V and X any aa) (Fig. 2B). Using the Conserved Domain program, we observed that the modeled three-dimensional structure of PfDLC1 protein was similar to the structure of CrLC1 (PDB, 1DS9LC1) where the central LRR regions form a cylindrical spiral composed of six ␤-␤␣ motifs (Fig. 2C). It is interesting to note that, although the cylinder of CrLC1 is flanked by two helical subdomains, PfDCL1 structure exhibits only one helical subdomain at the N-terminal side of the protein. This analysis indicated that the major difference between the two proteins is the absence of a helical structure at the C-terminal side of the PfDLC1 (Fig. 2C).
Genetic Manipulations of dlc1 in P. falciparum and in P. berghei-To investigate the role of DLC1 in the Plasmodium life cycle, we attempted to disrupt the Pfdlc1 gene using the pCAM vector system (23). We transfected wild-type 3D7 parasites with a plasmid containing a 5Ј-fragment derived from the Pfdlc1 and the blasticidin S deaminase gene, conferring the resistance to blasticidin (supplemental Fig. S1A). In two independent experiments, genotype analysis by specific PCR of stable transfectant parasites did not reveal the interruption of the gene (data not shown) however, the wild endogenous gene was still detectable in genomic DNA. The plasmid remained episomal even after prolonged drug cycling (Ͼ8 months).
To check the accessibility for recombination of the genomic Pfdlc1 locus, we tried to modify the locus without causing lossof-function of the gene product (supplemental Fig. S1B). We transfected wild-type 3D7 parasites with a plasmid containing the 3Ј-end of the Pfdlc1 coding region fused to HA epitope or GFP (supplemental Fig. S1B). Unexpectedly, we never observed in the blasticidin-resistant population an integration of the Pfdlc1-HA gene (diagnosis by specific PCR, immunofluorescence assays, and Western blotting) even after drug cycling and maintenance of the cultures (Ͼ8 months) or Pfdlc1-GFP gene (diagnosis by specific PCR and the presence of fluorescent parasites). Only PCR products corresponding to the non-integrated episome and to the unmodified genomic locus were detected (data not shown). As positive control of the transfection experiments, we used another P. falciparum gene (gene identifier PFL1360c, previously described (18)) and showed its integration in P. falciparum by Western blot using anti-HA antibodies (data not shown). These results excluded any technical problem related to electroporation or prolonged parasite culture. The absence of transgenic parasites for tagged Pfdlc1 may indicate that the plasmid was unable to recombine with the endogenous locus, or that the addition of any tag to the protein alters its function, which affects bloodstage parasites.
Based on the fact that PfDLC1 is conserved in Plasmodium species, including P. berghei (supplemental Fig. S2), we attempted to modify the Pbdlc1 gene (PB000535.01.0) in this rodent parasite, which is more amenable to reverse genetic approaches with a higher transfection efficiency than P. falciparum (33). We used the "knock-in" approach by transfecting P. berghei parasites with the construct PL1473 containing the 3Ј-end of the Pbdlc1 coding region fused to a Myc epitope (supplemental Fig. S3A). Integration of this construct would result in replacement of the endogenous copy with a c-myc-tagged version of Pbdlc1 via a single crossover integration event. In two independent transfection experiments (exp. 1373 and 1374), we were not able to select parasites with the expected integration of the construct into the Pbdlc1 locus on chromosome 7 (supplemental Fig. S3B). Failure to detect integration indicates that, as in P. falciparum, the dlc1 gene is refractory to genetic manipulations that interfere with the production of a functional gene product.
Expression of PfDLC1 and Its Subcellular Localization-To examine if P. falciparum expresses PfDLC1 protein and to determine its subcellular localization, an antiserum was raised against the recombinant PfDLC1 expressed in E. coli as His-tagged protein. When analyzed by immunoblot, the PfDLC1 antiserum reacted with a single protein in a whole P. falciparum extracts of mixed cultures of ϳ23 kDa (Fig. 3A), the expected molecular mass of PfDLC1. Next, the presence of PfDLC1 was searched in different parasite compartments (nuclear, cytoplasm, and pellet extracts (membrane)) using subcellular fractionation methods. Fig. 3 (B-D) clearly revealed that PfDLC1 were present in both cytoplasmic and membranes fractions (Fig. 3, B and C, lane 4) and not in the nuclear fraction (Fig. 3D). The quality control of the subcellular fractionation was checked for the presence of nuclear marker (histone 4), cytoplasmic marker (SOD1) and membrane marker (MSP1). The immunoblots performed with different antibodies on the three fractions revealed the lack of detectable cross-contaminations between extracts. These results indicated that PfDLC1 is likely localized in both parasite cytoplasm and membranes (pellet). Binding of PfDLC1 to P. falciparum Extracts-Because PfDLC1 belongs to the LRR protein family known to be involved in protein-protein interactions, we explored the capacity of recombinant PfDLC1 (rPfDLC1) to bind to the whole P. falciparum-soluble extract. The recombinant protein was expressed at 16°C, allowing the isolation of soluble product in a state of high purity (Ͼ95%, data not shown). The rPfDLC1 labeled with biotin was added in different concentrations to wells coated with P. falciparum extracts prepared as described under "Experimental Procedures." The results presented in Fig.  4A revealed that rPfDLC1 specifically binds to P. falciparum extracts when compared with controls (bovine serum albumin or uninfected RBCs). This clearly indicates the presence of PfDLC1-binding proteins in intra-erythrocytic blood parasites.
Despite the clear evidence of PfDLC1 binding to P. falciparum extracts, we were unable to detect specific PfDLC1-binding protein(s) from streptavidin columns using biotin-PfDLC1 as a bait. To bypass this constraint, to confirm the specificity of binding and to identify the number and molecular mass of PfDLC1-binding protein(s), we performed far Western blot experiments. Data presented in Fig. 4B showed that two pro-teins of ϳ43 kDa and Ͼ62 kDa reacted with biotin-PfDLC1. In control incubations in which PfDLC1 was omitted (lane 1) or replaced by a recombinant biotin-labeled LRR protein of S. mansoni (30) (lane 2), no binding was observed.
Next, we checked whether PfDLC1 may interact with proteins of high molecular mass given the fact that DLC1 has been described to interact with dynein heavy chains (Ͼ500 kDa) and that the genes of these chains are present in the genome of P. falciparum. To this end, total extracts were separated on a gradient of 3-8% SDS-PAGE. The far Western analysis did not reveal any interaction with proteins of high molecular masses, but the interaction with the two proteins mentioned above was still observable (Fig. 4C).
Identification of PfDLC1-binding Proteins-The results presented above indicated that PfDLC1-binding proteins are present in P. falciparum extracts. Based on the molecular masses of one of the two proteins (ϳ43 kDa) interacting with PfDLC1 with the fact that direct and indirect studies indicated the co-presence of CrLC1 with actin in the inner arm dynein complex (34,35), we surmised that PfDLC1 could interact with P. falciparum actin. To address this point, we explored the capacity of PfDLC1 to bind to highly purified rabbit actin (shares 83% identity with P. falciparum actin, PFL2215w). In parallel, we used two other proteins involved in the protein motor complex and cytoskeleton, the rabbit myosin (31% identity, with P. falciparum myosin A PF13_0233) and the tubulin purified from bovine brain. As shown in Fig. 5A, recombinant PfDLC1 bound significantly to actin in a dose-dependent manner, suggesting that actin could be one of the partners of PfDLC1. Unexpectedly binding experiments showed that rPfDLC1 was also able to interact with rabbit myosin. Under the same conditions, we did not observe any significant binding with RBC extracts or purified tubulin-containing ␣ chain (GenBank TM number 3E22_A) and ␤ chain (GenBank TM number AAT84374), which exhibit 84% identity with P. falciparum ␤-tubulin (PF10_0084) and 82% identity with P. falciparum ␣-tubulin (PFD1050w), respectively. To assess a possible interaction of PfDLC1 with both rabbit myosin and actin proteins, we performed binding assays of PfDLC1 to myosin in the presence of actin. Data presented in supplemental Fig. S4 indicated that the preincubation of myosin with actin dramatically affected the binding of PfDLC1 to myosin. The capacity of FIGURE 2. Analysis of amino acid sequence and three-dimensional structure model of PfDLC1. A, the PfDLC1 has been aligned with the LC1 from C. reinhardtii (CrLC1) using Clustal multiple alignments. The conserved residues are shown in blue, similar residues are shown in green, and identical residues are in red. B, alignment of the six LRR motifs from CrLC1 and PfDLC1. Individual residues forming ␤ sheets and ␣ helices are color-coded to correspond with the structure of CrLC1 (12) (␤ sheets, magenta and blue; helices, red; other residues, green). Conserved hydrophobic positions and the almost invariant Asn residue are blocked in yellow and pink, respectively. C, structural model of the PfDLC1 based on the crystal structure of CrLC1 (PDB: 1DS9). The site used was swissmodel.expasy.org.
PfDLC1 binding decreased with the increase of actin concentration indicating that the interaction of PfDLC1 to myosin is dependent upon the level of actin present. These data suggest that PfDLC1 could not bind to the complex actin-myosin in vitro.
Because PfDLC1 has been shown to interact with actin and myosin, we then investigated whether this interaction could be detectable in parasite extracts. To test this, immunoprecipitation experiments with anti-actin, anti-myosin A, or anti-PfDLC1 antibodies were performed. The antibodies used for co-immunoprecipitation were detected in all lanes, which provide a useful control for similar total antibodies loading (data not shown). Immunoblot analysis of PfDLC1 immunoprecipitates showed that P. falciparum actin (PfACT1) and P. falciparum myosin A (PfMyoA) (Fig. 5B, PfDLC1 antibodies) co-immunoprecipitated with the PfDLC1 (lane 6) from the parasite extracts. Similar results were observed when we co-immunoprecipitated with anti-actin or anti-myosin A antibodies (lanes 4 and 5). Taken together, our data demonstrate that PfDLC1 physically interacts with P. falciparum actin and myosin A.
Identification of PfDLC1 Interacting Motifs in P. falciparum Actin and Myosin A-The different approaches used in the binding experiments likely suggest the presence of similar motifs in P. falciparum myosin A and P. falciparum actin as well as in those corresponding to the rabbit that may explain their binding to PfDLC1. The screening by BLAST searches combined with visual inspection revealed that myosin A and actin of both species display three identical/similar motifs GILTL, ESDIE, and EEMKT (Fig. 6). To investigate the role of these motifs in the binding to PfDLC1, 8 peptides derived from PfMyoA or PfACT1 comprising the motifs and their respective flanking residues (15 aa in length) and 3 control peptides were synthesized and used in ELISA binding experiments. As shown A, parasite protein extracts were separated on 4 -12% SDS-PAGE and subsequently blotted to nitrocellulose. The blots were probed with preimmune serum (lane 1) or with anti-PfDLC1 rat antibodies (lane 2) followed by an incubation with an anti-rat IgG horseradish peroxidase-conjugated. The blots were revealed by ECL under autoradiography conditions. B-D, cytoplasm, nuclear, and membranes fractions were prepared as previously described (29). Lanes 1 and 2 correspond to blots incubated with rat and rabbit preimmune sera, respectively. Lane 3 represents a blot using rat anti-SOD1 antibodies. Lane 4 corresponds to the blot probed with rat anti-PfDLC1 antibodies. Lane 5 corresponds to a blot probed with rabbit anti-histone 4 antibodies (49). Lane 6 corresponds to a blot probed with rabbit anti-MSP1 antibodies. in Fig. 6, a clear interaction was observed with the peptides P2, P5, P6, P7, and P8. The intensity of the signal observed was dose-dependent of PfDLC1 added. With respect to peptides P1, P4, and P9, no significant binding was observed when compared with control peptides (P10, P11, and P12). From these data, it appears that the PfDLC1 interacting domains present in myosin and actin fall into SDIE and/or EEMKT motifs. Under identical experimental conditions, no significant interaction of PfDLC1 was observed with GILTL motif positioned in either PfMyoA or PfACT1 environments. The SDIE residues on PfMyoA protein are present in the upper 50-kDa domain between the switch I and II loops (supplemental Fig. S5). To map these residues, the chicken skeletal myosin was used as reference, because the head domain of this myosin has been crystallized (36). In addition, we aligned the PfACT1 protein sequence with the rabbit skeletal muscle actin for which the atomic structure is available (37). The alignment indicates that the SDIE and EEMKT residues present in the PfACT1 protein sequence are located in the subdomain IV of actin (supplemental Fig. S6).
Expression and Localization of PfDLC1-In previous studies, we have demonstrated that Pfdlc1 transcripts increased during intra-erythocytic parasite growth (18). This observation led us to follow up the expression of PfDLC1 protein by immunoblot assays. To accomplish this, Western blot experiments were performed on 60 g per lane of total proteins prepared from different stages of synchronized cultures of P. falciparum. Results presented in Fig. 7A did reveal that the gene product levels of PfDLC1 varied according to the stage tested with maximal levels being detected in late trophozoites/schizonts. With respect to the expression of P. falciparum myosin A (Fig. 7A), we started to detect the protein in late trophozoites with the highest quantity in schizonts. On the contrary, immunoblot analysis showed that actin gene product was equally expressed during the asexual cycle of P. falciparum (Fig. 7A).
Next we analyzed the localization of PfDLC1 protein in live 3D7 parasites transfected with pHH2 construct mediating the expression of full-length GFP-fused PfDLC1. GFP-tagged protein was successfully expressed in the presence of the endogenous protein, and the integrity of the fused protein was maintained as observed by Western blot analysis (Fig. 7B, lane 2). A single band with molecular mass of 48 kDa was observed, which is the expected molecular mass of the GFP-tagged PfDLC1. To compare the level of expression of endogenous PfDLC1 with that of the GFP-tagged PfDLC1 protein, we performed a Western blot analysis on protein extracts obtained from highly fluorescence-activated cell-sorted GFP fluorescent parasites. Western blot analysis using the anti-PfDLC1 antibodies indicated that the GFP-tagged PfDLC1 was highly expressed when compared with the level of expression of endogenous PfDLC1 (supplemental Fig. S7A). The main location of the transgenic fusion protein matched that of the endogenous protein detected with the anti-PfDLC1 antibodies (Fig. 7C). Nevertheless, the co-localization of the endogenous and the GFP-tagged DLC1 does not imply that the interaction of the latter with its partners is not affected. To address this point we performed co-immunoprecipitation experiments using the same extracts derived from the GFP-positive parasites sorted by fluorescence-activated cell sorting. We used either the anti-GFP antibodies or the anti-PfDLC1 antibodies to assess the interaction of the tagged and the untagged PfDLC1 with PfACT1. Following co-immunoprecipitation, we performed a Western blot analysis with the anti-actin antibodies. Results obtained have revealed that the quantity of PfACT1 protein pulled down with the GFP-PfDLC1 protein was much lower when compared with the amount of the PfACT1 protein precipitated with the anti-PfDLC1 antibodies (supplemental Fig. S7B). Despite the overexpression of the GFP-PfDLC1 protein, a tiny amount of PfACT1 was interacting with the GFP-tagged PfDLC1 protein.
These results suggest that the fusion of the GFP to the PfDLC1 protein affects its binding to partners. Further examination of PfDLC1 location showed a homogenous distribution in the cytoplasm of the parasite (Fig. 7, C and  D). It seems that the protein is not exported to the nucleus of P. falciparum or to the RBC cytoplasm or to the RBC surface (Fig.  7, C and D). The expression of PfDLC1 in younger stage transfectant as well as in the wild-type stage (young ring) was undetectable (Fig. 7D). Interestingly, and in support of the interaction of PfDLC1 with actin, we found that the P. falciparum actin is distributed in the parasite cytoplasm (Fig. 7E). This is in good agreement with previous published results about the location of P. falciparum actin (38,39). With respect to the distribution of myosin A along with PfDLC1, we performed co-localization experiments as the available antibodies were raised in different species. Results presented in Fig. 7F did show a partial overlap of myosin A staining with that of PfDLC1 in schizonts.

DISCUSSION
Our study indicates that the PfDLC1 gene product belongs to the proteins of the dynein motor complex, previously described to play a role in the regulation of flagellar motility in algae and trypanosomes. In T. brucei, the loss of the PfDLC1 ortholog, TbLC1, leads to reverse flagellar beat and backward movement of the parasite. Here we report the unexpected finding that the PfDLC1 is expressed in P. falciparum blood parasite stages that lack flagella, indicating that the function of this protein in not restricted to a regulation of the motility related to dynein motor complex.
The Pfdlc1 locus spans 1.3 kb of genomic DNA on chromosome 8 and consists of 6 exons. The predicted full length of Pfdlc1 encodes a protein of ϳ23 kDa, which is consistent with the size of Pfdlc1 gene product identified in this study. This size corresponds well with those of already identified dynein light chain 1 in C. reinhardtii and T. brucei. To evaluate the function of PfDLC1 in P. falciparum, particularly in flagellar motility, disruption of the PfDLC1 gene was investigated. Although stable transfectants were obtained, no mutants could be selected with a disrupted dlc1 locus. The refractory of the dlc1 locus to gene knock-out attempts suggests either its essentiality for parasite blood stage growth and survival or the inaccessibility of its locus for genetic modifications. Therefore, we have tried to generate mutants by introducing a GFP or HA tag at the C terminus of the dlc1 gene. Despite obtaining stable transfectants after prolonged periods of drug selection, we were unable to detect any GFP-or HA-tagged protein in live parasites or in protein extracts. Because of the failure of the generation and selection of P. falciparum mutants expressing tagged versions of DLC1, we decided to try to tag this protein in P. berghei (PfDLC1 is conserved in Plasmodium species) (supplemental Fig. S2). Also in this parasite we have been unable to select mutants expressing a (c-myc)-tagged version of DLC1. The inability to select for mutant parasites with a tagged DLC1 may indicate that the locus is refractory to the genetic modifications. However, the fact that we were unable to select mutants with a disrupted dlc1 gene and that we have been unable to generate mutants expressing the tagged version of DLC1 is suggestive for an essential role of this protein for blood stages growth and survival. Another possible explanation could be a defect in binding of tagged versions to its partners. We demonstrated that the fusion of a GFP tag to the PfDLC1 protein impairs its interaction at least with PfACT1 (supplemental Fig. S7B). This observation may explain why the knock-in attempts were unsuccessful, and may suggest that an intact PfDLC1 protein is needed for an efficient binding to its partners and for the asexual red blood cell cycle. The failure to generate mutants with a disrupted or tagged dlc1 locus is not an absolute proof for the essential nature of the DLC1 protein for the asexual blood-stage parasites. Ideally, both the essential function and the precise timing of when this protein is critical would have been investigated using methods that permit the manipulation of Plasmodium protein expression. Although a few conditional protein expression systems have been reported in Plasmodium research, none of these have been proven to be, as yet, robust systems for regulating Plasmodium protein expression to study gene function in blood stages (40). Therefore, analysis of the precise function and deployment of DLC1, using reverse genetic methodologies, awaits the development of optimized conditional mutagenesis systems.
PfDLC1 emerges as a potential gene product involved in protein-protein interaction, because it comprised six LRR motifs covering almost the whole protein. Interaction studies revealed that these motifs are crucial for the capacity of binding of the LRR protein family. Based on the structure of CrLC1, the model of PfDLC1 showed that its LRR motifs adopt curved parallel ␤ strand lining on the concave side while the helices flank its convex side. In the case of PfDLC1, it is important to point out the absence of ␣8 and ␣9 helices present in the C-terminal region of CrLC1. These helices are thought to play a role in the regulation of dynein motor activity through contact with the ATP-hydrolyzing site at 2 basic residues situated at the carboxyl end of the protein (12). A more recent work devoted to study the structure-activity relationships of CrLC1 showed that ␣8 and ␣9 helix structures are essential for its transport to the flagella and for its binding to the dynein ␥ heavy chain (41). The absence of these two helices as well as one of the two basic residues in PfDLC1 strongly suggests a distinct role in intraerythrocytic stages of P. falciparum. The following two arguments support the idea for a divergent or expanded role of some P. falciparum dynein proteins. First, the P. falciparum ortholog to ␥ dynein heavy chain (gene identifier PFI0260c; aa sequence 368 -2711: 40% identity and 60% similarity with aa sequence 2813-5732 of Chlamydomonas ␥ dynein heavy chain, XP_001702026) expected to bind to PfDLC1 was not detectable either by transcriptomic or proteomic analysis of blood parasite stages, whereas the latter approach allowed the detection of three other dynein heavy chains exclusively in the sexual stages (8). Second, using GFP transgenic P. berghei parasites Khan et al. observed that the promoter of the P. berghei dynein ␥ heavy D, expression and localization of PfDLC1-GFP throughout the erythrocytic cell cycle of P. falciparum. Parasites were transfected as described under "Experimental Procedures," and live transfectants were analyzed by fluorescence microscopy. E, immunofluorescence localization of P. falciparum actin in fixed transgenic parasites (late trophozoites). F, co-localization of PfDLC1 and PfMyoA. Smears of erythrocytes infected with 3D7 strain were fixed and probed with rat anti-PfDLC1 followed by anti-rat IgG-fluorescein isothiocyanate (45) and rabbit anti-PfMyoA, followed by anti-rabbit IgG-Alexa Fluor 568 (red). chain (gene identifier PB000791.03.0, present also in P. falciparum PFI0260c) was uniquely activated in male gametes, likely suggesting the specific expression of the corresponding gene at this stage (8). In this study, proteome analysis revealed that a PfDLC1 ortholog (PB000535.01.0) is expressed in male gametes. With respect to P. falciparum, and according to the available transcriptome analysis provided by PlasmoDB, the Pfdlc1 mRNA seems to be expressed in male gametes. It will be important to study the expression and the localization of PfDLC1 protein in male gametes and to investigate whether this protein is indeed expressed in flagella and interacts with one of the Plasmodium dynein motors.
The above observations suggest that PfDLC1 could act independently of dynein complex. An important finding made in this study is that the rPfDLC1 was found to be able to bind to purified rabbit actin and myosin, which exhibit 83 and 31% identity, respectively, with the orthologs in P. falciparum. More importantly, we were able to isolate PfDLC1-PfACT1 and PfDLC1-PfMyoA complexes from P. falciparum extracts of enriched trophozoites/schizonts parasites by co-immunoprecipitation and immunoblot experiments. These data clearly demonstrated the interactions of PfDLC1 with actin 1 and myosin A. The existence of a direct protein-protein interaction between PfDLC1 and PfMyoA and PfACT1 was further confirmed by identifying the PfDLC1-interacting domain. This was performed by the use of synthetic peptides sharing identical or similar motifs. Binding studies revealed that PfDLC1 was able to bind to ESDIE and EEMKT motifs derived from either myosin A or actin I, respectively. This fact provides structural basis for the proposed interaction with these two proteins and suggests that the LRR domain of PfDLC1 may participate in versatile protein-protein interactions. Interestingly, the peptide P2 (position 340 -354 in PfMyoA protein sequence) is located in the head domain of PfMyoA, in the upper 50-kDa segment between the switch I and II loops. This region is involved in actin binding, and the regulation of the contractile cycle of myosins (42,43). This finding may suggest that the PfDLC1, by binding to the upper 50-kDa segment present in the myosin head motor domain, may interfere with the open or closed conformation of the switch I loop for the regulation of the contractile myosin cycle, which is a nucleotide release-dependent mechanism mediating the interaction between actin and myosin (42). The SDIE and the EEMKT residues were mapped in the subdomain IV of PfACT1 (supplemental Fig. S6). The subdomains II and IV represent the pointed end of the actin molecule (37). Interestingly, the tropomyosin binds the two residues Ala-230 and Leu-236 located in the subdomain IV of Dictyostelium actin (44). In the presence of tropomyosin, the mutation of both residues (A230Y and L236A) of Dictyostelium actin enhances the myosin ATPase activity and increases the velocity of the in vitro motility assay at low head myosin motor concentrations (44). In the light of these findings, we can speculate that the PfDLC1 protein is involved in modulating the critical concentration of the pointed end of actin or may participate in the transition of filaments from a closed state to an open state. Inspection of the Plasmodium database revealed that there are 11 proteins that display an ESDIE motif and 10 proteins with SSDIE motifs. Although the far Western blot analysis revealed the binding with only two proteins, further studies will be required to establish the binding network of PfDLC1 in P. falciparum.
The observation, that P. falciparum expressed a PfDLC1 gene product together with the finding indicating that actin and myosin are PfDLC1 parasite targets, led us to further study the temporal and spatial expression of the three gene products during asexual growth. P. falciparum actin protein did not vary significantly along asexual development, in line with previously reported mRNA and protein expression data. In contrast, P. falciparum myosin A gene product appears to be differentially regulated during the progression of parasite cycle, with a clear accumulation of the protein in late schizonts while the protein was undetectable in rings and early trophozoites. With respect to PfDLC1, it was found that the gene product was detectable in the different developmental stages. Maximal accumulation of PfDLC1 protein occurred at the late trophozoites/schizonts when the transcript levels have been also shown to peak (18).
Interestingly, subcellular fractionation, PfDLC1-GFP overexpression, and localization analysis of wild-type PfDLC1 revealed its presence in the cytoplasmic compartment. PfDLC1 protein was also found to be present in the pellet of blood parasite-insoluble extracts (mainly trophozoites/schizonts membranes). Although we detected PfDLC1 in ring extracts by immunoblot, we were unable to detect it by immunofluorescence in ring stage parasites. This finding may suggest contamination of ring extracts with schizont extracts (purity of synchronization ϳ95%) or that the labeling signal in ring stage is too weak to be detected by IFA. It should be noticed that actin was mainly localized in the cytoplasm of all blood parasite stages, in agreement with the immunoblots and previously reported localization studies (38). Regarding the P. falciparum myosin A, we only observed its presence in the schizonts around the periphery of each merozoites (supplemental Fig. S8). This is in total agreement with data reported earlier (4). All together, these observations likely suggest that the interaction of PfDLC1 with actin could take place during almost the complete life cycle of blood parasites, whereas its interaction with myosin A may be restricted to late trophozoites and schizonts/merozoites where the protein started to be expressed. This work has provided evidence that PfDLC1 is expressed in P. falciparum intra-erythrocytic blood stages and is able to bind independently to P. falciparum myosin A and actin I. It is well known that apicomplexan motility is powered by the actin-myosin XIV motor complex composed of the myosin A heavy chain, the myosin light chain (MLC1 or MTIP), the glideosome-associated protein 45 (GAP45) and glideosome-associated protein 50 (GAP50) (45)(46)(47). Three of these proteins (Myo A, MTIP or MLC1, and GAP45) assemble into a soluble precomplex in the parasite cytoplasm. The pre-complex is exported, most likely by vesicular trafficking, from the endoplasmic reticulum to the inner membrane complex, where it is anchored to the membrane by GAP50 (46,48). The interactions between PfDLC1, PfMyoA, and PfACT1 in P. falciparum asexual stages may suggest that PfDLC1 is part of new and nonconserved (DLC1 is absent from other apicomplexa genera) complexes with an undefined function. From genetic studies, it is tempting to speculate that PfDLC1 plays an important role in asexual development of the malaria parasite by interacting with two proteins Myo A and actin 1 that are known to be essential for parasite survival. Further work is now required to determine the exact role of PfDLC1 during the intra-erythrocytic cell cycle of P. falciparum.