Expression of a Mutant Form of Leishmania donovani Centrin Reduces the Growth of the Parasite*

Leishmania donovani, a protozoan parasite, causes visceral disease in humans. To identify genes that control growth, we have isolated for the first time in the order Kinetoplastida a gene encoding for centrin from L. donovani. Centrin is a calcium-binding cytoskeletal protein essential for centrosome duplication or segregation. Protein sequence similarity and immunoreactivity confirmed that Leishmaniacentrin is a homolog of human centrin 2. Immunofluorescence analysis localized the protein in the basal body. Calcium binding analysis revealed that its C-terminal Ca2+ binding domain binds 16-fold more calcium than the N-terminal domain. Electrophoretic mobility shift of centrin treated with EGTA and abrogation of the shift in its mutants lacking a Ca2+ binding site suggest that Ca2+ binding to these regions may have a role in the protein conformation. The levels of centrin mRNA and protein were high during the exponential growth of the parasite in culture and declined to a low level in the stationary phase. Expression of N-terminal-deleted centrin in the parasite significantly reduces its growth rate, and it was found that significantly more cells are arrested in the G2/M stage than in control cells. These studies indicate that centrin may have a functional role inLeishmania growth.

Leishmania donovani, a protozoan parasite and a member of the order Kinetoplastida, is the causative agent of visceral leishmaniasis in humans worldwide. The disease is also known as "kala-azar" in India and Nepal. Treatment for this disease involves chemotherapy using antimony-based drugs, which is less effective in immunocompromised individuals (1). To date no vaccine is available for this disease. L. donovani has a digenic life cycle. The flagellated form, called promastigote, resides extracellularly in the gut of a dipteran sand fly insect. The second form, amastigote, is found in the macrophages of the infected human host. The growth and differentiation of Leishmania involves both qualitative and quantitative changes in various biochemical parameters (2). In our previous studies, we have identified genes that may have a role in growth and differentiation of the parasite (3)(4)(5). Here we describe one such gene, centrin, and analyze its function with respect to growth of the parasite.
Centrins are cytoskeletal, calcium-binding (EF-hand) proteins that are localized in the microtubule-organizing center of eukaryotes (6). Centrins are one of the several regulatory proteins essential for duplication or segregation of the centrosome in higher eukaryotes and basal bodies in lower eukaryotes (7). In many organisms, more than one centrin isotype has been described e.g. three centrins in humans and mice. Though three centrin forms have been recognized in the unicellular algae Chlamydomonas, only one has so far been characterized (8). One subfamily of centrins, which includes human centrin 1 (HsCEN1), human centrin 2 (HsCEN2), and Chlamydomonas reinhardtii centrin (CrCEN1), is involved in centrosome segregation (9). The other subfamily, which includes human centrin 3 (HsCEN3) and yeast centrin (CDC31), is involved in centrosome duplication (10,11). Results from diverse experimental systems, mostly in yeast, suggest that different types of proteins like PKic1p, protein kinase (11,12), and Kar1p, a component of the half-bridge of the spindle pole body, (13,14) bind to centrins. Though centrin has been characterized in a variety of eukaryotes, it has not been reported in the order Kinetoplastida.
The parasites of the Kinetoplastida are responsible for a wide variety of diseases affecting humans, animals, and plants (15). Members of this group have been considered to be one of the earliest eukaryotes, developing conventional organelles, but sometimes with extreme features rarely seen in other organisms (15). The role of such unique structural and functional features of these organelles, like the cytoskeleton and the flagellar apparatus, in infectivity is still obscure. Hence, the identification of genes that enable Leishmania to grow and differentiate within the harsh and diverse environments (sand fly gut and human macrophages) continues to be our objective (3, 16 -18).
The nature and the role of either the microtubule-organizing center or the basal body apparatus in the primitive eukaryote Leishmania are still not known. None of the genes (e.g. centrin, calmodulin, and ␥-tubulin) that are associated with these organelles in higher eukaryotes has been characterized so far in this important human parasite. As a first step toward that, here we report our findings leading to cloning, sequencing, and characterization of a centrin gene from L. donovani (LdCEN). We have also undertaken a functional analysis of Leishmania centrin protein by performing deletion analysis of the centrin gene and expressing such truncated proteins in the parasites to test their potential role in the growth of the parasite.

EXPERIMENTAL PROCEDURES
In Vitro Culture of Parasites-L. donovani isolated from either the Indian kala-azar patient (K80) or the cloned line designated by the World Health Organization as MHOM/SD/62/1S-C1 2D (SD) (16) was used in all the experiments. Promastigotes and the axenic amastigotes were grown and harvested as described previously (16).
Cloning of Centrin and Sequence Analysis-A cloned arbitrarily primed polymerase chain reaction fragment (AP-PCR-8), amplified specifically from an Indian kala-azar parasite DNA using AP-PCR primer 9 (5), hybridized differentially to 2.1 kb 1 RNA from promastigotes compared with axenic amastigotes. This DNA was used as a probe to screen a genomic cosmid library of L. donovani (19). A positive cosmid clone was sequenced in the region of the AP-PCR8 fragment. Sequencing revealed that the 3Ј-end of AP-PCR8 fragment overlapped with the 5Ј-untranslated region of an ORF. The encoded protein of 149 amino acids was identified through BLAST search as a homolog of the centrin gene from other organisms. The authenticity of the start site of the centrin ORF was confirmed by performing a reverse transcription-PCR, using an internal reverse primer (5Ј-TTCGCAACCTCCTTCAAG) and a forward primer (5Ј-ACTAACGCTATATAAGTA) designed from the Leishmania-specific mRNA splice leader sequence (20). The reverse transcription-PCR-amplified products were cloned into pCRII-TOPO vector and sequenced with the M13 forward and reverse primers. Multiple sequence alignment of centrins from various organisms was conducted in MacVector 7.0 program and used to determine cluster relationships among the sequences and to construct dendrograms representing cluster relationships (21).
Isolation of Genomic DNA and Southern Blot Analysis-Total Genomic DNA was isolated from either promastigotes or axenic amastigotes according to the methods described in the manual for GENOME DNA isolation kit from BIO 101 Inc. The DNA was digested with restriction endonucleases EcoRI, NcoI, and SalI and separated on 1% agarose gels. Southern blot analysis of digested DNA was done as described previously (22).
Isolation of RNA and Northern Blot Analysis-Total RNA was isolated from promastigote and axenic amastigote cultures of L. donovani using RNA STAT-60 according to the manufacturer's instructions (Tel-Test, Inc. Friendswood, TX). Total RNA (10 g) was analyzed by Northern blot as described (23). Both Northern and Southern blots were hybridized with a 32 P-labeled centrin coding region (450 base pairs) probe. The membranes were exposed and scanned on the Phosphor-Imager system (Molecular Dynamics Amersham Pharmacia Biotech Piscataway, NJ). The intensity of the hybridized bands was quantitated using ImageQuant software version 1.1 (Molecular Dynamics).
Plasmid Constructs and Expression of Recombinant Wild Type and Mutant Centrin Forms-Plasmid constructs to express either the wild type or the mutant forms of LdCEN in either Escherichia coli or L. donovani cells are described in Fig. 1. The full-length open reading frame of wild type centrin and its various mutant forms were PCRamplified with primers as described in Fig. 1 and ligated individually in pCR-T7/CT TOPO vector (Invitrogen) and transformed into competent E. coli BL21 (DE3) PlysS (Invitrogen). The authenticity of each of the constructs and the PCR-amplified fragments was confirmed by DNA sequencing. Reverse primers that were used to amplify wild type or mutant centrin forms to be expressed in the Leishmania parasite contained a hemagglutinin (HA) tag sequence (24) added in frame with centrin at the 3Ј-end. All the amplified products were digested with SpeI and ligated at the same site of pKSNEO (25) and transfected into the parasite. To clone LdCEN in bacterial expression vector (pQE-70) the PCR-amplified fragment was digested with SphI and BglII and cloned into SphI and BamHI site of pQE-70 in frame with the histidinetag (His 6 ) at the 3Ј-end and transformed into E. coli M15 cells as described (26). The protein was expressed from E. coli, purified through Ni-agarose ion-exchange column chromatography according to the manufacturer protocol (Qiagen Inc., Valencia, CA), and the protein was used to generate polyclonal antibody in rabbits (26).
Immunoblot Analysis of Leishmania Centrin Protein (LdCenp)-Antisera raised against two recombinant human centrins (HsCen2p and HsCen3p, both rabbit polyclonal; gifts from Dr. Michel Bornens, Institute Curie, Paris, France), C. reinhardtii centrin (CrCenp1 (20H5) mouse monoclonal; gift from Dr. J. L. Salisbury, Mayo Clinic Foundation, Rochester, MN), and LdCenp (rabbit polyclonal; prepared by Spring Valley Labs, Woodbine, MD) were used to determine immunogenic cross-reactivity of the recombinant LdCenp. 100 -200 ng of recombinant centrin protein was separated in a 12 or 15% SDS-PAGE gel, transferred to nitrocellulose membranes, and analyzed by Western blot (27) using the centrin antibodies (dilutions: anti-LdCenp Ab 1/1000, anti-HsCen2p Ab 1/2000, anti-HsCen3p Ab 1/250, and anti-CrCen1p Ab 1/1000). The proteins were visualized using the SuperSignal Chemiluminescent substrate system (Pierce). To analyze the endogenous centrin and determine its size, mid-log culture of either promastigotes or axenic amastigotes were resuspended in 20 mM HEPES buffer, protein concentration was determined by BCA (Pierce), and 20 g of total cell protein was analyzed on SDS-PAGE by Western blot analysis using various centrin antibodies.
Calcium Binding of LdCenp-The binding assay for 45 Ca was performed as described (17,28). Briefly, the various recombinant proteins were subjected to SDS-PAGE and transferred onto a nitrocellulose membrane. The membrane was incubated with 1 Ci/ml 45 Ca in buffer containing 60 mM KCl, 5 mM MgCl 2 , and 10 mM imidazole-HCl, pH 7.0, for 10 min at the room temperature, washed with distilled water, dried, and exposed to x-ray film. To confirm the calcium binding and conformational changes of LdCenp in vivo, metal chelator EGTA (100 mM) was added to the parasite's cell lysate in 10 l of reaction volume, incubated for 5 min at room temperature, and processed for Western blot analysis on a non-denaturing gel according to Novex TM , San Diego, CA using anti-LdCenp antibody.
Immunofluorescence Analysis-L. donovani promastigotes were fixed in suspension in 4% (w/v) paraformaldehyde in PBS (50 mM Na 2 HPO 4 , 150 mM NaCl, pH 7.4) for 20 min at room temperature, washed three times in PBS, and allowed to attach to glass slides. After air drying, the slides were first immersed in ice-cold methanol (Ϫ20°C) Culture and Transfection of the Parasites-Mid-log phase promastigotes (2-4 ϫ 10 7 cells/ml) were harvested by centrifugation at 3000 g for 10 min at 4°C. Cell pellets were washed in ice-cold PBS and electroporated with the DNA using conditions as described previously (29). Transfected promastigotes were selected with minimal doses of G418 (20 g/ml). The drug-resistant cells were used in all subsequent experiments.
Flow Cytometry-Promastigotes from early exponential cultures were collected, fixed in 70% ethanol, and washed with PBS. The fixed cells were treated with 100 g/ml ribonuclease in PBS for 5 min at room temperature and stained with 50 g/ml propidium iodide (Sigma) in PBS for 15 min on ice and analyzed on a FACScan flow cytometer (Becton Dickinson Immunocytometry Systems, San Jose, CA) and CELLQuest software. For each sample 20,000 fluorescent events were measured, and the data was analyzed using the Modfit Lt. Software was Verity Software House, Inc. (Topshan, ME).

Cloning and Sequence Analysis of the L. donovani Centrin
Gene-We used an arbitrarily primed polymerase chain reaction (AP-PCR) approach, as previously described (5), to isolate genes from L. donovani that are differentially expressed during the growth and differentiation of this parasite. A 1-kb AP-PCR fragment was generated specifically using DNA isolated from L. donovani parasites collected from an Indian Kala-azar patient. The fragment was cloned and used as a probe to analyze total RNA from the promastigotes and axenic amastigotes of L. donovani. A 2.1-kb RNA hybridizing to the AP-PCR fragment was found to be expressed significantly more in promastigotes than in axenic amastigotes (data not shown). Based on such differential expression, the AP-PCR fragment was used to screen a L. donovani cosmid DNA library. Sequence analysis of a positive clone showed an ORF that overlapped with the 3Ј-end of the 1-kb AP-PCR sequence. Homology search of the open reading frame revealed that it has significant similarity with centrin proteins from many organisms. The complete nucleotide and deduced amino acid sequence of the L. donovani centrin (LdCEN) gene is shown in Fig. 2. The authenticity of the start site of the centrin's ORF was confirmed by performing a reverse transcription PCR, using an internal reverse primer and the splice leader forward primer. The nucleotide sequence of the reverse transcription PCR clones confirmed that the identified ORF occurs in a mature transcript of the LdCEN gene. The sequences of the amplified fragments also indicated the presence of at least two splice sites in the 5Ј-untranslated region of the gene (Fig. 2).
Centrins and calmodulins, another closely related Ca 2ϩbinding protein, have in general four EF-hand (Ca 2ϩ binding) domains; however, the number of functional EF-hand domains vary among the centrins (30,31). Sequence motif analysis of L. donovani centrin protein (LdCenp) predicted only two Ca 2ϩ binding sites (EF-hand 1 and 4) (Fig. 2). In addition the Ld-Cenp was also found to possess hydrophobic amino acids in their ␣-helices around both the EF-hands 1 and 4 ( Fig. 2) as have been observed with other centrins. The amino acid sequence of LdCenp was analyzed by ClustalW alignment with centrins from different organisms, as shown in Fig. 3A. The calculated percent similarity shows that it is closer to HsCen2p (61%), HsCen1p (60%), or Trichomonas centrin (60%) than other centrins. The NЈ-terminal non-conserved domain of centrins, which is variable in length, is considered to be responsible for the functional diversity of centrins (32,33). Interestingly, L. donovani centrin has a significantly small N-terminal region compared with centrins from other species (Fig. 3A).
A neighbor-joining systematic tree based on centrins of various eukaryotes was constructed, to study the phylogenetic relationship of LdCenp with centrins of other organisms. Two distinct clusters were seen in the tree (Fig. 3B). One cluster had CrCen1p, mouse centrin, HsCen2p, and HsCen1p, and the second cluster had Giardia centrin, CDC31p, and HsCen3p. However, centrins of Paramecium and Leishmania branched off independently from the common ancestor of all the centrins (Fig. 3B).
Northern and Southern Blot Analyses of LdCEN-To analyze LdCEN mRNA levels in both promastigotes and axenic amastigotes, a Northern blot analysis was performed using total RNA obtained from the mid-log cultures of these two parasite stages. The 32 P-labeled LdCEN probe recognized two bands (Fig. 4A): a major ϳ1.7-kb band that expressed equally in both promastigotes and axenic amastigotes and a 2.1-kb band expressed significantly more in promastigotes than in axenic amastigotes.
To determine the copy number of LdCEN in the genome, Southern blot analysis of the total genomic DNA from L. donovani promastigotes was done using the centrin coding region as a probe (Fig. 4B). DNA digested individually with either SalI or EcoRI, whose sites are not present in the coding region of the centrin gene, gave a single band of ϳ4.1 or 4.3 kb, respectively. On the other hand, digestion of DNA with NcoI, which has two  sites at positions 1 and 203 in the LdCEN coding region, resulted in three bands of ϳ0.2, 3.8, and 6.5 kb (Fig. 4B, lane 2). These results are consistent with LdCEN being a single copy gene.
Antigenic Similarity of LdCenp with Centrins from Other Organisms-To test whether LdCenp has any antigenic similarity with centrins from other organisms, a Western blot containing the recombinant LdCenp expressed in E. coli was performed with antisera raised against human centrin 2, human centrin 3, C. reinhardtii centrin 1, and LdCenp (Fig. 5A). The recombinant LdCenp cross-reacted with anti-LdCenp Ab and anti-HsCen2p Ab, but did not react with either anti-HsCen3p Ab or anti-CrCen1p Ab (Fig. 5A). To further confirm that LdCenp is expressed in the parasite and to analyze whether there are more than one form of LdCenp that may cross-react with the various anti-centrin antibodies, we tested these antibodies in a separate Western blot using cell lysates from midlog promastigotes and axenic amastigotes (Fig. 5B). Anti-Ld-Cenp Ab and anti-HsCen2p Ab reacted with a 17-kDa protein in both the lysates (Fig. 5B, lanes 1-4). However, anti-Hs-Cen2p antibody reacted with an additional protein of ϳ25 kDa (Fig. 5B, lanes 3 and 4). On the other hand, anti-HsCen3p and anti-CrCen1p antibodies reacted with different size proteins (ϳ19 kDa and ϳ18 kDa, respectively) in Leishmania cell lysates. (Fig. 5B, lanes 5-8). The intensity of cross-reacting centrin bands was similar in promastigotes and axenic amastigotes. These results suggest that additional centrin-like proteins may exist in the parasite.
EF-Hand 4 Is the High Affinity Ca 2ϩ Binding Site in LdCenp-Amino acid sequence analysis of LdCenp and its comparison with known centrins revealed that it has two putative Ca 2ϩ binding EF-hand domains (1 and 4). We tested whether LdCenp binds to Ca 2ϩ in vitro and defined which of the predicted sites interact with Ca 2ϩ . To do so we constructed three LdCenp mutants that had either the first, the fourth, or both putative Ca 2ϩ binding domains deleted (Fig. 6A), and expressed the full-length and mutant centrins in E. coli (Fig. 6B). The proteins were tested for their binding to 45 Ca in vitro. Full-length and N-terminal-deleted (LdCenp ⌬N) centrins bound a similar level of 45 Ca (Fig. 6C, lanes 1 and 3). However, the C-terminal-deleted centrin (LdCenp ⌬C) bound significantly less Ca 2ϩ (Fig. 6C, lane 2), and the LdCenp that lacked both the C-and the N-terminal regions (LdCenp ⌬NC) did not bind calcium at all (Fig. 6C, lane 4). The PhosphorImager quantitation of 45 Ca-bound bands revealed that the LdCenp ⌬N bound 16-fold more 45 Ca than the LdCenp ⌬C. The Ca 2ϩ binding to centrin was specific, since bovine serum albumin, a non-Ca 2ϩ -binding protein, did not bind to 45 Ca in this assay (data not shown). These results demonstrate that the LdCenp has two Ca 2ϩ binding sites and that Ca 2ϩ binds preferentially to the C-terminal site.
Both N-and C-terminal Regions Are Necessary for the Ca 2ϩdependent Protein Conformation-Recent reports using either NMR spectrum for C. reinhardtii centrin (34) or CD spectroscopy and size exclusion chromatography for Scherffelia dubia and human centrins (7) showed Ca 2ϩ induced conformational changes in the proteins. Having shown that the LdCenp binds to Ca 2ϩ , we analyzed the Ca 2ϩ -induced folding conformation of LdCenp. Plasmid constructs containing either full-length centrin (pKSNEO LdCEN) or N-terminal-deleted centrin (pKSNEO Ld-CEN ⌬N) and C-terminal-deleted centrin (pKSNEO LdCEN ⌬C) along with vector control (pKSNEO) were transfected into wild type L. donovani promastigote cells. Proteins were extracted from all the transfected parasites at mid-log stage and analyzed by Western blot using anti-LdCenp Ab (Fig. 7A). Though the level of expression of each form of recombinant centrin differed, each corresponded to the predicted molecular weight and was sufficiently abundant for further analysis. These protein extracts were incubated in the presence or absence of the metal chelator EGTA, separated in a non-denaturing gel, and analyzed by Western blot using antibody against either Ld-Cenp, which recognizes the endogenous centrin of the control parasite, or antibody against HA tag sequence, which recognizes the episomally expressed centrins in the transfected parasites (Fig. 7B). Endogenous centrin after treatment with EGTA migrated faster in the gel than untreated (Fig. 7B, lanes  1 and 2). The same EGTA-induced shift occurred with the transfected over-expressed full-length centrin (Fig. 7B, lanes 5  and 6). However, treatment with EGTA did not affect the mobility of either LdCenp ⌬N or LdCenp ⌬C (Fig. 7B, lanes  7-10). These results suggested that binding of Ca 2ϩ to both binding sites induced a conformational change of this protein.
Further, deletion of the N or C terminus resulted in a loss of the Ca 2ϩ dependent conformational change.
Parasite Growth Regulated Expression of LdCenp-Centrins have been implicated to have an essential function during the cell division cycle (35). As a first step toward understanding the role of LdCenp in Leishmania growth, we explored whether there is a correlation between the expression of LdCenp and the parasite growth. The level of expression of both centrin mRNA (the 1.7-kb) and protein was measured at different stages of promastigote and axenic amastigote growth. Quantitation of Northern (Fig. 8A) and Western blots (Fig. 8B) indicated that the level of LdCEN mRNA and protein, in both the promastigotes and axenic amastigotes, was maximal in the exponentially growing culture. The levels of both mRNA and protein steadily declined as the parasites progressed from latelog to stationary phase, thereby suggesting that the expression of centrin correlates with the growth rate of L. donovani.
LdCenp Localizes at the Basal Body Region of the Promastigotes and Axenic Amastigotes-To ascertain the localization of centrin in L. donovani, we carried out immunofluorescence analysis. Paraformaldehyde-fixed mid-log phase promastigotes and axenic amastigotes were stained with anti-LdCenp Ab and DAPI and examined by confocal microscopy. The intensity of fluorescence by anti-LdCenp Ab was mostly concentrated in the anterior part of both the parasitic forms. In addition, a dense fluorescent spot was seen in the area close to the DAPI-stained kinetoplast, (Fig. 9). However, the stationary phase cells of either promastigotes or axenic amastigotes showed no significant staining by anti-LdCenp Ab (data not shown). The basal body has been shown to be localized at the flagellar root that remains tightly associated to the kinetoplast in Leishmania (15). The results showed that LdCenp is predominantly localized close to the kinetoplast of both growing promastigotes and axenic amastigotes and may be associated with the basal body of Leishmania.
Differential Localization of the Mutant Centrins-To identify the domain of LdCenp, which is responsible for its targeting to the basal body, we analyzed log-phase Leishmania parasites that over-expressed either complete, N-deleted or C-deleted forms of centrin. Transfected parasites were stained with anti-HA tag Ab to reveal the over-expressed centrins. Parasites expressing full-length and carboxy-end-deleted centrins showed immunostaining throughout the parasite (Fig. 10, panels 4 and 8) possibly due to the high expression of both these centrin forms compared with control cells (Fig. 7A). On the contrary, in a significant number (ϳ40%) of cells expressing LdCenp ⌬N, the over-expressed protein was predominantly localized toward the posterior region (Fig. 10, panel 6). The redistribution of LdCenp ⌬N toward the posterior region in these cells suggests that the N-terminal region may have the localization signal that could target the protein toward the basal body region.
LdCenp Lacking the N-terminal Region Has a Dominant Negative Effect on the Growth of the Parasite-To determine the role of centrin in the growth of L. donovani, growth of the transfected parasites was analyzed in culture. The promastigotes that were transfected with either vector alone or the LdCenp ⌬C construct grew almost at the same rate (Fig. 11A). The full-length centrin over-expressing line showed slightly slower growth rate compared with the control cells; however, reached the same stationary plateau. On the other hand, parasites expressing LdCEN ⌬N displayed a 2-fold reduction in the growth rate compared with the other transfectants based on the slope of the curve during log phase (Fig. 11A). The maximum density of the LdCenp ⌬N-expressing cells at the stationary stage was also low compared with the control cells. Similar suppression of growth due to LdCenp ⌬N was also seen in the axenic amastigotes (data not shown). To determine the cause of slow growth of LdCenp ⌬N-expressing cells, the midlog stage promastigote cultures were subjected to cell cycle analysis using flow cytometry. Results showed a decrease of 23% in S-phase cells and an increase of 26% in G 2 -phase cells in the LdCenp ⌬N-transfected line compared with the control line (Fig. 11B). There was no significant difference in the number of S-phase cells expressing either full-length centrin or LdCenp ⌬C. However, ϳ10% more full-length centrin-expressing cells remained in G 2 -phase than in the control (Fig. 11B). These results suggest that the slower growth observed in LdCenp ⌬N expressing cells could be due to a significant number of cells remaining for a longer period of time in the G 2 /M phase of the cell cycle compared with control cells.

DISCUSSION
Centrin is a calcium-binding protein involved in the contractile function of the cytoskeletal structures in eukaryotes (32,36). We report here the first cloning of such a gene among Kinetoplastids, whose members are considered to be primitive eukaryotes (15). The cloned gene, LdCEN, from the L. donovani parasite is more homologous to centrin than the other closely related Ca 2ϩ binding EF-hand proteins such as calmodulins. Typically, centrin is distinguished structurally from calmodulin by a longer N-terminal variable region. This longer N terminus is thought to confer functional diversity to centrins (7,32). However, this region is smaller in LdCEN than all other centrins and calmodulins. The short N terminus may reveal its ancestral position in the evolution of centrin. LdCenp crossreacts with the antibody raised against human centrin protein (HsCen2p). Multiple centrins are localized at the centrosome and expressed in all the dividing cell types (36). Specifically HsCen2p expression level increases during ciliogenesis in the tracheal epithelial cells, and its involvement in cell division as well as cellular motility has been demonstrated (37,38). Similarly, LdCenp was found to be localized at the basal body of the parasite, and its expression was high when cells were actively dividing and significantly reduced in the resting cells. This correlation of expression pattern suggests that Leishmania centrin may have a role in the growth of the parasite. However, it is also speculated that it may not have a role in the cellular movement, because this protein is equally expressed in the non-motile axenic amastigote stage and is present in lower concentration in the stationary stage promastigote cells, which are highly motile.
The actual number and position of the Ca 2ϩ binding sites varies from centrin to centrin. In HsCen2 only the fourth EFhand binds calcium (13), whereas, in LdCenp both EF-hand 1 and 4 bind to Ca 2ϩ . EF-hand 4 binds to calcium 16-fold more than EF-hand 1 as evidenced from our calcium binding study on the various recombinant mutant centrin proteins. At the molecular level, the calcium binding generates conformational changes in CrCen1p (7) and CDC31p (14). The Ca 2ϩ -dependent polymerization of the green algae S. dubia centrin protein results in a filamentous network (7). A similar Ca 2ϩ -induced protein conformational change of LdCenp was also demonstrated in the present study. The implications of this conformational change for the LdCenp remain to be studied.
At least three major types of centrin have been described by others. Type 1 (mouse centrin 1) is expressed during spermatogenesis suggesting a role in meiosis (39). The second type of centrin (HsCEN2) protein is up-regulated during ciliogenesis of mammalian cells (37) and has been shown to play a role in cellular motility and microtubule severing (38,40). The third type of centrin, similar to yeast protein CDC31 and HsCEN3 appears to play a role in centrosomal duplication (41). The antisera against HsCen2p reacted with a protein equivalent in size to the recombinant LdCenp and with a larger size protein in the L. donovani lysate. Anti-HsCen3p Ab and anti-CrCen1 Ab reacted with proteins in the Leishmania lysates yet they differed in size from LdCenp. These results suggest the existence of more than one centrin isotype in L. donovani. The reason for the molecular heterogeneity in LdCenp protein and the functional analysis of the different putative forms of Ld-Cenp remain to be analyzed, though each of them may serve a different function as described for human centrins (32,35).
In our phylogenetic analysis, LdCenp branched off independently from a common centrin ancestor, while most other centrins can be grouped into two clusters. These clusters interestingly also correlate with their distinct biological functions as observed (9 -11). HsCen1p, HsCen2p, and CrCen1p of one cluster have been involved in the segregation of centrosome, whereas CDC31p and HsCen3p of the other cluster have been involved in centrosome duplication during cell division (9 -11). Despite Leishmania centrin's early divergence, the unique binding of the anti-HsCen2p Ab suggests structural relatedness of LdCenp and Hscen2p. This similarity to HsCen2p may not be reflected in the phylogenetic analysis due to amino acid differences in regions other than the common antigenic epitopes. Similar inconsistencies between phylogenetic trees and biochemical analysis have been observed by others (42,43). Parasites were processed for immunofluorescence using mouse anti-HA tag Ab to stain the episomally expressing centrins. Texas red antimouse IgG as a secondary antibody stains the cells red (2, 4, 6, and 8). Cells were viewed by confocal microscopy. C, F, ⌬N, and ⌬C are same as mentioned in Fig. 7A.   FIG. 11. A, effect of the expression of various mutant centrin forms on the growth of the parasite. The graph shows the growth of the parasite (promastigotes) transfected with either full-length (F) or mutant centrins (⌬N and ⌬C) or vector control constructs (C). The results were the mean of four independent growth experiments. B, flow cytometric analysis of the DNA content of L. donovani lines expressing various centrin mutants. Promastigotes transfected with either full-length (F) or mutant centrins (⌬N and ⌬C) or vector control constructs were grown to early-exponential phase, fixed, stained with propidium iodide, and subjected to fluorescence-activated cell sorter analysis. For each sample 20,000 fluorescent events were measured. The difference in the number of cells between each centrin over-expressing sample and the control was expressed as % change and shown for each phase of the cell cycle. The data were averaged from three independent experiments and plotted with standard deviation.
Whether the structural similarity reflects a common function remains to be analyzed for LdCenp.
Centrin's association with growth in several eukaryotes has been studied (8). Injecting recombinant Chlamydomonas centrin or human centrin 2 in two cell stage frog embryos delayed cleavage and promoted the formation of abnormal blastomeres (35). Expression of both antisense and sense transcripts of centrin arrested spermiogenesis in Marsilea vestita (44). In the present study, we analyzed the role of LdCenp in the parasite growth by over-expressing the full-length and the mutant centrins. Though full-length and C-terminal-deleted centrins did not alter the growth of the parasite, N-deleted centrin did reduce the growth by 2-fold over the control in a dominant negative fashion. The slow growth of this culture was correlated with the significant number of cells remaining for a longer period in G 2 /M phase of the cell cycle, suggesting centrin function is crucial for completion of mitosis. The importance of the N-terminal sub-domain for centrin function was emphasized similarly by observing slower growth rate in the yeast cells expressing centrin (CDC31), which had the N-terminal region replaced with that of S. dubia centrin (7). Secondly, yeast centrin interacts with the cellular protein Kar1p at its C-terminal region (7,14). In the present study in Leishmania, LdCenp could be similarly interacting with other cellular components after binding to Ca 2ϩ , and LdCenp ⌬N could compete with the endogenous centrin for this interaction. Such an interaction, having no functional N-terminal region, would probably affect the role of centrin responsible for the growth of the parasite. The second abnormality noticed with such a slow growing parasite was the localization of the LdCenp ⌬N in the cells. The protein was seen everywhere in the cell, though a significant number of cells (ϳ40%) expressing LdCenp ⌬N showed localization predominantly at the posterior region. This suggests that the basal body localization signal for LdCenp may be in its N-terminal region. Alternatively LdCenp may interact with another cellular protein through its N terminus for proper localization as observed for yeast centrin CDC31p. CDC31p interacts with Kar1p, a spindle pole body component that helps to localize CDC31p to the spindle pole body. No CDC31p was detected at this organelle in cells lacking Kar1p (45,46). However, the exact mechanism by which LdCenp ⌬N localizes to the posterior end of a significant number of cells remains to be investigated. Nonetheless, possible co-localization of LdCenp with the basal body, the correlation of centrin expression with Leishmania growth and alteration of growth rate, and alteration of the cell cycle upon expression of LdCenp ⌬N may suggest a role for centrin in Leishmania cell division. In conclusion, understanding the role of centrin in Leishmania growth could provide ways to alter the growth of the parasite and clues for the development of attenuated Leishmania vaccine candidates.