The crk3 Gene of Leishmania mexicanaEncodes a Stage-regulated cdc2-related Histone H1 Kinase That Associates with p12 cks1 *

A cdc2-related protein kinase gene,crk3, has been isolated from the parasitic protozoanLeishmania mexicana. Data presented here suggests thatcrk3 is a good candidate to be the leishmanial cdc2 homologue but that the parasite protein has some characteristics which distinguish it from mammalian cdc2. crk3 is predicted to encode a 35.6-kDa protein with 54% sequence identity with the human cyclin-dependent kinase cdc2 and 78% identity with theTrypanosoma brucei CRK3. The trypanosomatid CRK3 proteins have an unusual, poorly conserved 19-amino acid N-terminal extension not present in human cdc2. crk3 is single copy, and there is 5-fold higher mRNA in the replicative promastigote life-cycle stage than in the non-dividing metacyclic form or mammalian amastigote form. A leishmanial suc-bindingcdc2-related kinase (SBCRK) histone H1 kinase, has previously been described which binds the yeast protein, p13 suc1 , and that has stage-regulated activity (Mottram J. C., Kinnaird, J., Shiels, B. R., Tait, A., and Barry, J. D. (1993) J. Biol. Chem. 268, 21044–21051). CRK3 from cell extracts of the three life-cycle stages was found to bind p13 suc1 and the leishmanial homologue p12 cks1 . CRK3 fused with six histidines at the C terminus was expressed in L. mexicana and shown to have SBCRK histone H1 kinase activity. Depletion of histidine-tagged CRK3 from L. mexicana cell extracts, by Ni-nitrilotriacetic acid agarose selection, reduced histone H1 kinase activity binding to p13 suc1 . These data imply that crk3 encodes the kinase subunit of SBCRK. SBCRK and histidine-tagged CRK3 activities were inhibited by the purine analogue olomoucine with an IC50 of 28 and 42 μm, respectively, 5–6-fold higher than human p34 cdc2 /cyclinB.

vator in yeast and animals and plays a crucial role in controlling the cell cycle in all eukaryotes (1,2). In mammals, a large number of CDKs have been described, some of which have been shown to be directly involved in controlling cell cycle transitions, e.g. CDKs 1-4 and CDK6 (1,3), while others have less defined roles, e.g. CDK5 (4,5). In contrast, yeast have only one CDK (Schizosaccharomyces pombe cdc2 or Saccharomyces cerevisiae CDC28), which is responsible for transition through both G 1 /S and G 2 /M boundaries (6,7). The activity of CDKs is controlled post-translationally by a number of different mechanisms: the association of the kinase subunit with its positive regulatory partner, one of the family of cyclins; phosphorylation of conserved sites which can stimulate or inhibit kinase activity; or binding of an inhibitory partner, a CDK inhibitor protein such as members of the p16 ink4 or p21 cip1 families (1). Another group of proteins known to interact with cdc2 are the cdc2 kinase subunit proteins, such as human p9 cks1 and p9 cks2 (8) and S. pombe p13 suc1 (9,10). The function of these proteins in the kinase complex is as yet unknown but the high specificity with which they interact with cdc2 has been exploited in the affinity purification of cdc2 from a variety of organisms on p13 suc1 -Sepharose (9,11,12).
Members of the order Trypanosomatidae, which includes Leishmania and the closely related parasite Trypanosoma brucei, possess a large family of cdc2-related kinase genes, crk1-4 (13)(14)(15)(16), and thus in this respect appear to be more similar to multicellular organisms than to yeast. This is not surprising since, although it is unicellular, Leishmania has a complex life cycle during which the cell cycle is switched on and off as the parasite differentiates between proliferative forms (the promastigote in the sandfly midgut and the amastigote in the mammalian host) and the cell cycle arrested metacyclic stage, which is the human infective form found in the fly proboscis. Thus, there is an integral link between the cell cycle and the developmental life cycle of this parasite. The large evolutionary divergence between protozoa and yeast or higher eukaryotes has made it difficult to determine which of the trypanosomatid crk genes might be the functional cdc2 homologue, as each of the genes have a similar level of homology to yeast or human cdc2, and they do not complement yeast cdc2 temperaturesensitive mutants (15,16).
In Leishmania, one cdc2-related kinase gene has been described, crk1 (15). The CRK1 protein was found in all leishmanial life cycle stages, but its histone H1 kinase activity was only detected in the promastigote and metacyclic forms, indicating post-translational control of kinase activity (15). Evidence from gene-targeted disruption experiments indicated that the Leishmania mexicana crk1 gene is essential to the promastigote form of the parasite (17). A distinct histone H1 kinase activity has been identified biochemically from L. mexicana (15,18). This activity, known as the suc-binding cdc2-related kinase (SBCRK), is thought to be a candidate for the functional cdc2 homologue from L. mexicana since it shares a number of attributes with cdc2 from other eukaryotic species: it binds the S. pombe protein p13 suc1 and its L. mexicana homologue p12 cks1 ; it phosphorylates the cdc2 substrate histone H1, and its activity correlates with the division status of the parasite, being active in the proliferative life cycle stages (promastigote and amastigote) and inactive in the cell cycle-arrested metacyclic stage (15,18). In the closely related protozoan T. brucei, three crk genes have been cloned and analyzed (16). tbcrk1-3 encode proteins of 34, 39, and 35 kDa, respectively, which share 49 -54% sequence identity with human cdc2.
Here, we describe a second crk gene from L. mexicana, crk3. This gene is single copy, is expressed in all leishmanial life cycle stages, and is closely related to the T. brucei crk3 gene. CRK3 has stage-regulated histone H1 kinase activity that binds p13 suc1 , features of the leishmanial SBCRK and mammalian cdc2.
Cloning and Sequencing of the L. mexicana crk3 Gene-A fragment of the crk3 gene was obtained by PCR using degenerate oligonucleotide primers designed to conserved regions of cdc2-related kinases and L. mexicana genomic DNA as described previously for the T. brucei crk3 gene (16). The full-length crk3 gene was isolated by screening an L. mexicana DashII genomic library as described previously (19). A 2-kb HindIII fragment containing the complete crk3 open reading frame (pGL89) was subcloned into Bluescript for sequencing.
Southern and Northern Blot Analyses-Southern and Northern blots were carried out as described previously (15). Wild-type L. mexicana DNA was isolated using a Nucleon kit (Scotlab, Coatbridge, Scotland). Poly(A) ϩ mRNA was isolated from the three life cycle stages of L. mexicana (mid-log promastigotes, stationary phase metacyclics, and axenic amastigotes) using the Oligotex Direct kit (QIAGEN Inc.). Two micrograms of poly(A) ϩ mRNA from each life cycle stage was separated by formaldehyde-agarose gel electrophoresis and blotted onto a nylon membrane (Hybond-N, Amersham Pharmacia Biotech) (15). Northern blots were hybridized at 42°C for 20 h in 5ϫ saline/sodium/phosphate/ EDTA, 50% deionized formamide, 5ϫ Denhardt's solution, 1% SDS and washed at 65°C in 0.2ϫ SSC, 0.1% SDS. 30 ng of a 1.2-kb EcoRI/ HindIII fragment from the plasmid clone pGL89 was labeled by random priming (Stratagene) and used to probe the Northern or Southern blots. A 600-base pair ␣-tubulin probe was used as a control (19).
Preparation of Antisera-A peptide corresponding to the C-terminal 12 amino acids of the predicted CRK3 sequence (CALQHPWFSDLRW) was synthesized (Genosys Biotechnologies Inc.). This peptide was conjugated to bovine serum albumin and used to immunize animals using standard immunization protocols (20).
Preparation of p12 csk1 and p13 suc1 Beads-p12 cks1 and control beads were prepared as described previously (18). Recombinant p13 suc1 , with a six-histidine tag at the C-terminal end, was produced in a similar manner. The S. pombe suc1 gene was inserted into the pQE60 (QIA-GEN Inc.) vector by constructing unique NcoI and BglII sites at the 5Ј and 3Ј ends of the coding region using PCR. The oligonucleotide primers used in the PCR reactions were as follows: 5Ј primer, 5Ј-GCCCATGG-CGAAAAGTGGTGTGC-3Ј; and 3Ј primer, 5Ј-GCAGATCTACCACCCC-GTTGTTGACT-3Ј (restriction sites are underlined). The PCR reactions were for 15 cycles of 94°C for 1 min, 42°C for 1 min, 50°C for 1 min, and 72°C for 2 min using 20 ng of pRK172 plasmid (21) as template and Pfu DNA polymerase (Stratagene). PCR products of approximately 350 base pair were digested with NcoI and BglII, gel purified, and cloned into pQE60. One of the resultant clones (pGL70) was sequenced to ensure that no PCR-induced mutations had been introduced and then transformed into M15[pREP4] cells. Recombinant p13 suc1 was then purified by nickel chelate (Ni-NTA) affinity chromatography and crosslinked to agarose beads (5 mg/ml), as described previously for p12 cks1 (18).
Binding of p12 cks1 to Leishmanial Proteins-50 g of promastigote S-100 lysate (approximately 2 mg/ml) and 1 g of 6-histidine-tagged CRK3 protein per lane were separated by SDS-PAGE and blotted onto PVDF membrane. The membranes were then incubated with either blocking solution or blocking solution plus 0.2 mg/ml recombinant p12 cks1 protein, for 90 min at room temperature. The blocking solution consisted of 10% horse serum, 5% low fat dried milk in TBST (10 mM Tris, pH 8.0, 150 mM NaCl, 0.05% Tween 20). The membranes were then washed with TBST to remove unbound p12 cks1 protein. The position of bound p12 cks1 was detected by incubation with an anti-p12 cks1 antibody at 1:50 dilution followed by an anti-rabbit secondary antibody and detection using the chemoluminescence system (18).
Immunoblotting-200 l of S-100 lysates of wild-type L. mexicana (1.3 mg/ml) was incubated with 20 l of p12 cks1 , p13 suc1 , or control agarose beads (previously preblocked with 10 mg/ml bovine serum albumin) for 3 h at 4°C. Unbound proteins were then removed by extensive washing, as described previously (18), and bound proteins were eluted by boiling the beads in the presence of SDS and ␤-mercaptoethanol. Samples of the eluted proteins were separated on a 12% SDS-polyacrylamide gel and transferred to PVDF (polyvinylidene difluoride) membrane. Western blots were performed, as described previously (18), with crude anti-CRK3 antiserum, used at 1:100 dilution (plus or minus 2 g/ml blocking peptide). Antigens were detected using a chemiluminescence detection system (Pierce SuperSignal).
Kinase Assays-The S-100 lysates were diluted to 0.4 mg/ml and incubated with 20 l of p12 cks1 , p13 suc1 , or control beads as for immunoblotting above. Removal of nonspecifically bound protein and the histone H1 kinase assay were performed as described previously (15,18). Phosphorylated histone H1 was quantified by scintillation counting or by phosphorimaging.
Complementation of S. cerevisiae cdc28 Mutants-The crk3 gene was expressed in three Saccharomyces cerevisiae cdc28 temperature-sensitive alleles using the yeast vector pRS416-met (22). pRS416-met is a 2-m based plasmid containing the URA3 gene as a selectable marker and a methionine repressible promoter to regulate expression levels. A 1.3-kb EcoRI/XhoI fragment, containing the full-length crk3 gene, was excised from pGL89, and cloned into the EcoRI/XhoI site of pRS416met. Yeast were transformed with the resultant vector (pGL120) or a positive control plasmid, pBP120, containing the Caenorhabditis elegans ncc-1 gene (23) using the lithium acetate method and clones tested for their ability to complement the temperature-sensitive mutant strains. Colonies were resuspended in 50 l of distilled water, and 5 l were added to YNB plates supplemented with appropriate amino acids, including methionine at 0, 50, 100, 200, 500, or 1000 M, but lacking in uracil. One set of plates was incubated at the permissive temperature of 23°C, the other at 37°C. The temperature-sensitive mutant strains used were cdc28 -13, cdc28 -4, and cdc28-IN (24). To confirm that the crk3 gene was expressed in each of the yeast strains, Western blots were performed on yeast cell extracts with the anti-CRK3 peptide antisera.
Transfection of L. mexicana-The crk3 gene was fused to a 6-histidine tag using the pQE60 vector (QIAGEN Inc.) by constructing unique NcoI and BglII sites on the 5Ј and 3Ј ends of the coding region using PCR with the plasmid clone pGL89 and the following primers (with engineered restriction sites underlined): 5Ј-GCCCATGGCTTCGTTTG-GCCGTG-3Ј and 5Ј-CGCAGATCTCCAACGAAGGTCGCTGAACC-3Ј. PCR was performed with Pfu polymerase (Stratagene). The PCR product was digested with NcoI and BglII and cloned into NcoI/BglIIdigested pQE60 to give the plasmid pGL91. This encodes CRK3 protein with a 6-histidine tag at the C-terminal end (WRSHHHHHH-COOH, where the underlined W is the last authentic CRK3 amino acid). The PCR-based cloning into pQE60 introduced a change in the CRK3 sequence at the second amino acid position, changing the authentic serine to an alanine residue (resulting in an inactive kinase). To subclone the crk3his gene into the pX shuttle vector (25) and repair this amino acid change, the following protocol was used. An EcoRI/SstII fragment, corresponding to the 5Ј end of the crk3 gene, was excised from the plasmid clone pGL89, and an SstII/HindIII fragment, corresponding to the 3Ј end of the crk3his gene, was excised from the pGL91 plasmid. These two fragments were ligated together and cloned into BlueScript (pGL92). The repaired crk3his gene was then excised with SmaI and XhoI, the XhoI site was filled-in using Klenow, and the fragment was cloned into the shuttle vector, pXneo (25), restricted with SmaI. A plasmid with the insert in the correct orientation was designated pGL95. Transfections were carried out as described previously (17,26). The cell line expressing CRK3his from the pX vector in wild-type L. mexicana genetic background was named wt[pXCRK3his].
Depletion of CRK3his from Cell Extracts-1.5 ml of S-100 cell lysates were prepared from 1 ϫ 10 8 L. mexicana or wt[pXCRK3his]. The extracts were made up to 50 mM imidazole and 500 l incubated with 200 l of Ni-NTA agarose for 1 h at 4°C. Both the original lysate and the eluate from the column (CRK3-depleted) were then incubated with p13 suc1 or control beads, as described above, and the bound protein was assayed for histone H1 kinase activity together with the washed Ni-NTA agarose beads.
Inhibition by Olomoucine-SBCRK was bound to p13 suc1 beads, as described above, and eluted with 2 mg/ml free p13 suc1 . CRK3his was purified on Ni-NTA agarose, as described above, and eluted by incubation with 100 mM EDTA in lysis buffer for 30 min at 4°C. CRK3his was then bound to p13 suc1 beads and assayed in situ. Samples of enzyme were incubated with increasing concentrations of olomoucine (Alexis Corporation UK (Ltd)), and kinase activity was assayed as described. Recombinant human cdc2/cyclin B (New England Biolabs) was used as a positive control.

RESULTS
Characterization of the crk3 Gene from L. mexicana-The L. mexicana crk3 gene was isolated using the same PCR strategy and degenerate oligonucleotide primers that proved successful in the cloning of the T. brucei crk3 gene (16). The full-length gene (Fig. 1) was isolated from a Dash II genomic library (19) and subcloned into Bluescript on a 2-kb HindIII DNA fragment. Mapping of the clone revealed that a mitochondrial elongation factor G gene homologue was located about 3.5-kb downstream of crk3. Interestingly, a homologous gene is located immediately downstream of the crk3 gene in T. brucei (16). The predicted protein encoded by the leishmanial crk3 gene shows the greatest degree of homology to the cdc2 family of serine/threonine protein kinases when compared with protein sequence data bases. The predicted amino acid sequence is compared in Fig. 2 with that of human cdc2 and of the homologous proteins from T. brucei (TbCRK3) and from T. cruzi (TcCRK3). The three trypanosomatid CRK3s each have an unusual 19-amino acid N-terminal extension, when compared with human cdc2, that is not highly conserved in sequence and has only one conserved residue, an arginine at position 10. Overall, the leishmanial CRK3 has 78% identity with TbCRK3, 77% identity with TcCRK3, and 54% identity with human cdc2. The leishmanial CRK3 contains all the domains and residues characteristic of the serine/threonine protein kinase family. In addition, CRK3 also contains the conserved residues and domains which are important for the regulation of cdc2 activity. This includes equivalent residues to human cdc2 at Thr-14 and Tyr-15, in the ATP-binding domain, and Thr-161 (Fig. 2). These three residues are highly conserved both within the CDK family and between species. This implies that CRK3 activity may be controlled by similar post-translational mechanisms as exist in other eukaryotes (through a positively regulating kinase, wee1, and a negatively regulating phosphatase, cdc25), although no homologues of these regulatory enzymes have been identified in Leishmania to date.
The "PSTAIRE" box, consisting of the 16-amino acid sequence EGVPSTAIREISLLKE, is a highly conserved domain thought to be involved in the recognition and binding of the cyclin partner (27)(28)(29). The corresponding domain in the leishmanial CRK3 (EGIPQTALREVSILQE) has six substitutions in comparison with Hscdc2 (see Fig. 2) and is highly conserved with the trypanosome CRK3 homologues. The presence of this domain in the leishmanial kinase suggests that cyclin binding may play an important regulatory role, although again no cyclin homologues have been identified in Leishmania. Cyclin genes have, however, been isolated from T. brucei (30), 2 implying that regulation of kinase activity through cyclin binding is conserved in trypanosomatids.
The L. mexicana crk3 Gene Is Single Copy and Is Expressed throughout the Life Cycle-Southern blot analysis, using as a probe a 1.2-kb EcoRI/HindIII genomic fragment which covers the complete open reading frame of the crk3 gene, detected a single hybridizing band in four out of five restriction digests (Fig. 3A, lanes 2-5). The SalI restriction digest (lane 1) gave one major hybridizing fragment of approximately 3.5-kb and one weakly hybridizing fragment of approximately 6 kb. The presence of two hybridizing bands is due to the internal SalI site within the crk3 gene (Fig. 1). Together with mapping data obtained from the clones, these results indicate that crk3 is single copy in the leishmanial genome.
When the same probe was hybridized to a Northern blot of poly(A) ϩ mRNA derived from the three life cycle stages of L. mexicana, a 1.7-kb mRNA was detected (Fig. 3B). Relative hybridization intensity was quantified by phosphoimaging using ␣-tubulin as a control. This gave a ratio of 5:1:1 for replicative promastigote:stationary phase metacyclic:replicative amastigote.

Recombinant p12 cks1 Binds to a 35-kDa Protein in Leishmanial Cell Lysates and to Recombinant CRK3his-The S. pombe
protein, p13 suc1 , interacts with cdc2 and cdk2 from many species and has been widely used as a highly specific affinity matrix for the purification of cdc2 when covalently cross-linked to a solid support, such as Sepharose beads (9,11,12,(31)(32)(33). This matrix, however, does not bind other CDK proteins such as cdk4, cdk5, and cdk6 (32). We have previously described a leishmanial homologue of p13 suc1 , p12 cks1 , which is also capable of binding p34 cdc2 from yeast and bovine cell extracts (18). Both p13 suc1 and p12 cks1 bind a histone H1 kinase activity from leishmanial cell extracts, SBCRK, which is thought to be the functional cdc2 homologue from this parasite (18). To investigate proteins that bind p12 cks1 and are components of the SBCRK complex, promastigote cell lysates were separated by SDS-PAGE and blotted onto a PVDF membrane, which was then incubated with recombinant p12 cks1 protein. Detection of bound p12 cks1 by Western blotting with an anti-p12 cks1 specific antibody (18) revealed that p12 cks1 binds specifically to a protein of approximately 35 kDa (Fig. 4, lane 1). A number of higher molecular mass proteins were also detected with this antiserum, but these were also present in the control (lane 2). In a parallel experiment, recombinant CRK3 protein was blotted onto a PVDF membrane and probed with p12 cks1 . Detection with the anti-p12 cks1 antiserum revealed that p12 cks1 also binds recombinant CRK3 under these conditions (lane 3).
CRK3 Binds p12 cks1 and p13 suc1 -Binding of CRK3 from leishmanial cell extracts to p13 suc1 and p12 cks1 beads was assessed by Western blot using an anti-CRK3 antibody, raised against the C-terminal 12 amino acids of CRK3. This antibody did not recognize reproducibly a band of the predicted molecular mass from leishmanial whole cell extracts. However, when CRK3 from promastigote cell extracts was concentrated by selective affinity binding to p13 suc1 and p12 cks1 beads, the anti-CRK3 antibody recognized a band of 35 kDa (Fig. 5A) which was abolished by the presence of competing peptide (not shown). When cell extracts of wild type L. mexicana were incubated with p13 suc1 and p12 cks1 beads, it was found that CRK3 bound to both proteins from all three life cycle stages (Fig. 5A). Less CRK3 protein bound to p13 suc1 and p12 cks1 from metacyclic cell extracts than from promastigotes or amastigotes (Fig. 5A) correlating with a reduced histone H1 kinase activity detected from the metacyclic form, in comparison to the 2 P. Neuville and J. C. Mottram, unpublished data.  1 and 4), stationary phase metacyclics (lanes 2 and 5), or axenic amastigotes (lanes 3 and 6) was separated on a 1.4% formaldehydeagarose gel, transferred onto nylon membrane, and hybridized with a crk3-specific probe (lanes 1-3) or an ␣-tubulin-specific probe (lanes 4 -6).
CRK3his Has Histone H1 Kinase Activity and Binds p13 suc1 -The crk3 gene was engineered with a 6-histidine tag on the C terminus and re-expressed in L. mexicana from the pX vector (25). Clonal cell lines expressing CRK3his (wt[pXCRK3his]) were selected by growth in 25 g/ml neomycin (G418). Cell extracts prepared from wild-type L. mexicana or wt[pXCRK3his] were incubated with excess Ni-NTA agarose in the presence of 50 mM imidazole, and the washed beads were assayed for the ability to phosphorylate histone HI (Fig. 6). Activity was detected in extracts from wt[pXCRK3his]) (lane 2) but not in wild type (lane 1) showing that the histidine-tagged kinase can be selectively affinity purified and is active. The same cell extracts were incubated with p13 suc1 or control beads (with saturating amounts of beads), and the bound histone H1 kinase activity was assessed (lanes 4 and 5, wild type; lanes 8 and 9, wt[pXCRK3his]). Similar p13 suc1 binding histone H1 kinase activity was observed in both cell extracts (Fig. 6, lanes  5 and 9), indicating that expression of the tagged CRK3 from the episome does not result in increased kinase activity. To assess the contribution of CRK3his to overall p13 suc1 binding kinase activity, the wild type and wt[pXCRK3his] flow-through from the Ni-NTA agarose column, in which CRK3his had been depleted, were incubated with p13 suc1 or control beads and assayed for histone H1 kinase activity (lanes 6 and 7, wild type; lanes 10 and 11, wt[pXCRK3his]). A reduction of over 50% for p13 suc1 binding histone H1 kinase activity was observed between wild type and wt[pXCRK3his] after Ni-NTA selection (compare lanes 7 and 11), indicating that CRK3his is the predominant kinase binding to p13 suc1 in this cell line and by inference that CRK3 is the kinase component of SBCRK.
CRK3his, purified on Ni-NTA beads and eluted with 100 mM EDTA, was incubated with p13 suc1 beads (lane 3). The detection of histone H1 kinase activity bound to the beads confirmed that CRK3his could bind p13 suc1 and retain histone H1 kinase activity.
Inhibition of SBCRK and CRK3his by Olomoucine-Olomoucine is a purine analogue that is a potent inhibitor of mammalian cdc2 with a narrow range of selectivity (34). Olomoucine was tested for its effects against SBCRK and CRK3his. Dose response curves gave IC 50 values for the inhibition of histone H1 kinase activity by olomoucine of 28 M for SBCRK and 42 M for CRK3his (Fig. 7).
crk3 Is Unable to Complement CDC28 Temperature-sensitive Mutants-To test if crk3 could complement an S. cerevisiae CDC28 ts mutant, the crk3 gene was introduced on a low copy plasmid into three strains containing different ts alleles. The plasmid contained a promoter repressible by methionine, allowing the expression levels of the CRK3 protein to be regulated (22). At the restrictive temperature, the cdc28 -13 mutant arrests at the G 1 /S boundary of the cell cycle while the other two mutants tested (cdc28 -4 and cdc28-IN) block at the G 2 /M transition (7,24). The crk3 gene was unable to restore growth at the restrictive temperature for any of the strains using a variety of expression levels. The expression of CRK3 was confirmed by Western blotting of yeast extracts with anti-CRK3 antibodies (not shown). The ncc-1 gene of C. elegans was used as a positive control for complementation of the ts phenotype (23). DISCUSSION We have previously described a histone H1 kinase that binds fission yeast p13 suc1 and whose activity is stage-regulated during the Leishmania life cycle (15,18). This cdc2-related kinase activity (SBCRK) is not encoded by the leishmanial crk1 gene as CRK1 does not bind p13 suc1 (18). Attempts to purify this activity and obtain amino acid sequence for the components of the kinase complex (predicted to be kinase and cyclin partners) proved problematic due to the low concentration of SBCRK in the cell. As an alternative approach, we used PCR and oligonucleotides designed to conserved regions of cdc2-related kinases to amplify crk genes. This approach had proved successful previously with the cloning of several crk genes from the closely related trypanosomatid T. brucei (16), and it also proved successful with L. mexicana as described in this paper. Several lines of evidence suggest that crk3 encodes a p13 suc1 binding kinase (SBCRK, (15)). (a) crk3 is predicted to encode a protein of 35 kDa, and p13 suc1 binds a similar size protein in leishmanial cell extracts (Fig. 4). (b) Antibodies can detect CRK3 from leishmanial cell extracts bound to p13 suc1 (Fig. 5). (c) CRK3 tagged with 6-histidine residues, expressed in L. mexicana and FIG. 5. CRK3 binding to p12 cks1 or p13 suc1 from the three different life cycle stages parallels histone H1 kinase activity. S-100 lysates prepared from promastigotes, metacyclics, or amastigotes were incubated at saturating concentrations with a limited number of p12 cks1 , p13 suc1 , or control beads. The proteins bound to the beads were assayed for the presence of CRK3 by Western blotting (panel A, lanes [1][2][3][4][5][6][7][8][9] or for histone H1 kinase activity (panel B). Dark gray bars, p12 cks1 ; black bars, p13 suc1 ; light gray bars, control.
FIG. 6. CRK3his has histone H1 kinase activity and binds p13 suc1 . S-100 lysates were prepared from promastigote L. mexicana wild type or wt[pXCRK3his] mutant and incubated with excess Ni-NTA agarose (lanes 1 and 2), p13 suc1 (lanes 5 and 9), or control beads (lanes  4 and 8). Histone H1 kinase activity bound to the beads was detected by SDS-PAGE and autoradiography. The supernatant from the Ni-NTA column was collected and incubated with p13 suc1 beads (lanes 7 and 11) or control beads (lanes 6 and 10), and histone H1 kinase activity bound to the beads was assessed. CRK3his bound to Ni-NTA agarose was eluted with 100 mM EDTA, bound to p13 suc1 beads, and assessed for histone H1 kinase activity (lane 3).
affinity purified on Ni-NTA agarose, has p13 suc1 binding histone H1 kinase activity (Fig. 6). (d) Depletion of CRK3his from leishmanial cell extracts reduces p13 suc1 binding histone H1 kinase activity (Fig. 6). (e) SBCRK and CRK3his have a similar IC 50 for histone H1 kinase inhibition with the ATP analogue, olomoucine (Fig. 7). Although these experiments provide strong evidence that CRK3 is the major cdc2-related kinase activity that binds p13 suc1 , and hence encodes the stage-regulated kinase previously described (15,18), it cannot be ruled out that other minor activities might bind p13 suc1 . For instance, CRK1 can bind p12 cks1 , the leishmanial homologue of p13 suc1 (18), but not p13 suc1 , and so a number of leishmanial CRKs might bind p12 cks1 in vivo, as has been demonstrated for human CDKs and p9 cks1 (32,33). Further evidence that CRK3 encodes the SB-CRK might have come from the generation of null mutants for crk3 by targeted gene disruption. However, attempts to create crk3 null mutants resulted in changes in cell ploidy, a phenomenon that has been used as a positive criterion for determining if a gene is essential in Leishmania (17,35). Thus it would appear that crk3 is essential. 3 Attempts to create a mutant cell line in which crk3his was expressed in a null mutant background, similar to that described previously for crk1 (17), also failed, possibly because the histidine tag on the C terminus of the protein interfered with some aspect of the function of the kinase.
Despite the high level of sequence identity between the leishmanial crk3 gene and CDC28, crk3 was unable to complement any of the three S. cerevisiae CDC28 mutants tested. The mutants were selected for alleles that caused conditional blocks at either the G 1 /S or the G 2 /M boundary; however, CRK3 could not function as a CDC28 replacement at either of these checkpoints. We had previously shown that the T. brucei crk3 and L. mexicana crk1 were unable to complement an S. pombe cdc2 ts mutant (15,16). It is possible that the large sequence divergence in the PSTAIR region of CRK3 (only 10/16 amino acid identity), which is a cyclin binding site (29), precludes the binding of yeast cyclins. Clearly the large phylogenetic distance between Leishmania and yeast (36) makes negative cross-species complementation tests such as this difficult to interpret. Northern analysis shows differential expression of crk3 mRNA in the three different life cycle stages of L. mexicana, with 5-fold more crk3 mRNA in promastigotes than in either amastigotes or metacyclics. This ratio does not reflect CRK3 activity (as assessed by SBCRK histone H1 kinase activity, (15,17)). The level of CRK3 activity does, however, parallel the amount of CRK3 protein which binds to p13 suc1 beads from cell extracts prepared from the three life cycle stages. When equivalent experimental conditions were used for SBCRK activity, and for the anti-CRK3 Western blot in which p13 suc1 was in excess, no CRK3 could be detected bound to p13 suc1 from the metacyclic extract (data not shown). Only when cell extract was in excess could CRK3 be detected by Western blot bound to p13 suc1 from metacyclics (Fig. 5A). There are a number of possible explanations for the reduced CRK3 binding to p13 suc1 from metacyclic extracts. (a) CRK3 could be translated at a low level in the metacyclic stage. crk3 mRNA was detected in metacyclics by Northern analysis, but it cannot be ruled out that it is either not recruited onto the ribosome or is translated with poor efficiency in metacyclics compared with the proliferative life cycle stages. (b) CRK3 protein is unstable and quickly degraded in metacyclic cells. (c) A putative inhibitor protein binds to the CRK3 kinase in metacyclic cells and interferes with the binding interaction between CRK3 and p13 suc1 . Our anti-CRK3 antisera cannot detect CRK3 on Western blots of wild-type lysates, presumably due to the low cellular levels of this kinase, and therefore it has proved difficult to determine the relative concentrations of CRK3 protein in the three lifecycle stages of wild-type L. mexicana. However, the low levels of CRK3 protein in the cell cycle-arrested metacyclic stage, but not a concomitant reduction in mRNA levels, might represent an unusual mechanism for controlling cdc2-related kinase activity. In mammalian cells, cdc2 is generally stable throughout the cell cycle (37,38), although both cdc2 mRNA levels and cdc2 protein decline to almost undetectable levels following cell growth arrest and then increase again when the cell re-enters the cell cycle (37,39). The mechanism by which the activity of leishmanial CRK3 is regulated during the life-cycle is currently under investigation.