Stage-specific Differences in Cell Cycle Control in Trypanosoma brucei Revealed by RNA Interference of a Mitotic Cyclin*

African trypanosomes have a tightly coordinated cell cycle to effect efficient segregation of their single or-ganelles, the nucleus, flagellum, and kinetoplast. To investigate cell cycle control in trypanosomes, a mitotic cyclin gene ( CYC6 ) has been identified in Trypanosoma brucei . We show that CYC6 forms an active kinase complex with CRK3, the trypanosome CDK1 homologue, in vivo . Using RNA interference, we demonstrate that absence of CYC6 mRNA results in a mitotic block and growth arrest in both the insect procyclic and mammalian bloodstream forms. In the procyclic form, CYC6 RNA interference generates anucleate cells with a single kinetoplast, whereas in bloodstream form trypanosomes, cells with one nucleus and multiple kinetoplasts are observed. Fluorescence-activated cell sorting analysis shows that bloodstream but not procyclic trypanosomes are able to reinitiate nuclear S phase in the absence of mitosis. Taken together, these data show that procyclic trypanosomes can undergo cytokinesis without completion of mitosis, whereas a mitotic block in bloodstream form trypanosomes inhibits cytokinesis but not kinetoplast replication and segregation nor an additional round of nuclear DNA synthesis. This

The African trypanosome, Trypanosoma brucei, is a unicellular parasitic protozoan that causes sleeping sickness in humans. T. brucei has a complex biphasic life cycle with different developmental forms playing specific roles within each of two hosts (1). The procyclic form in the tsetse fly gut and the long slender form in the mammalian bloodstream establish infection in each host, whereas the metacyclic form in the tsetse salivary glands and the short stumpy form in the mammalian bloodstream are arrested in G 0 /G 1 phase of the cell cycle and are preadapted to life in the new host. Upon transmission, concomitant with differentiation to the new developmental form, the cell cycle arrest is released. Thus, there is an integral link between the cell cycle and parasite differentiation during the life cycle of T. brucei (2). In addition, there is a requirement for the trypanosome to replicate and segregate its single organelles (the nucleus, the kinetoplast (which contains the DNA of the single mitochondrion), flagellum, and basal body) (3,4), which implies added complexity in cell cycle control not present in relatively simple eukaryotes such as yeast.
Ultrastructural studies of the procyclic form have described a number of markers of cell cycle position and identified discrete phases within the trypanosome cell cycle (3). During G 1 phase, the probasal body (lying adjacent to the flagellar basal body) matures. Daughter flagellum outgrowth follows, and new probasal bodies for each mature basal body are formed. Kinetoplast S phase is much shorter than nuclear S phase and commences just prior to nuclear S phase (4), suggesting that (as yet unidentified) interorganelle control mechanisms may coordinate DNA synthesis. Early in G 2 phase, the replicated kinetoplast segregates, and this is followed by nuclear mitosis and finally cytokinesis. It has been postulated that entry to cytokinesis may be more dependent on kinetoplast division and segregation than on mitosis (5).
Tightly controlled regulation of DNA replication and chromosome segregation is essential for maintenance of genome stability. Eukaryotes have therefore evolved a complex ordered series of checkpoints in order to bring about accurate DNA replication, mitosis, and cytokinesis. Most checkpoints and cell cycle control proteins are highly conserved from yeast to humans. However, at least one cell cycle checkpoint present in mammalian cells has been suggested to be absent from trypanosomes, since treatment of procyclic trypanosomes with the anti-microtubule agent rhizoxin results in cytokinesis occurring in the absence of mitosis (5,6). In contrast, treatment of procyclic trypanosomes with aphidicolin prevented cells from entering mitosis, indicating that a mitotic entry checkpoint is present (5). It also remains possible that additional checkpoints may operate to cope with requirements specific to trypanosomes (e.g. in order to coordinate the replication and segregation of the kinetoplast with that of the nucleus).
The involvement of cyclin-dependent kinases in the control of the cell cycle is well established in yeast and higher eukaryotes (7,8). In T. brucei, a number of putative CDKs 1 (known as Cdc2-related kinases (CRKs)) and cyclins have been described (1,9,10). The cyclins can be grouped into different classes according to sequence homology to other known cyclins. CYC2, CYC4, CYC5, and CYC7 exhibit homology to the PREG1/ PHO80 class of cyclins (1,10) and may play roles in nutrient sensing; CYC3, CYC6, and CYC8 share homology with mitotic * This work was supported by the Medical Research Council (MRC) and by the British Council. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EBI Data Bank with accession number(s) AJ496538, AJ496539, and AJ496540.
§ An MRC Senior Research Fellow. To whom correspondence should be addressed. E-mail: j.mottram@udcf.gla.ac.uk. cyclins; and CYC9 is a cyclin C homologue and thus may play a role in transcriptional regulation in T. brucei (1). Recently, CRK3 was shown to form an active kinase complex with CYC2 in vivo, the first example of a trypanosome CDK-cyclin complex; however, the role of this complex in vivo is not known (10). Studies on T. brucei CRK3 and its orthologue from Leishmania suggest that CRK3 is a functional trypanosomatid CDK1 (1). By a combination of yeast two-hybrid analysis and co-immunoprecipitations with histone H1 kinase assays, we extend previous studies to show that CRK3 also forms an active kinase complex with CYC6 in vivo.
We demonstrate using RNA interference (RNAi) that CYC6 is essential for mitosis in the African trypanosome, but that a mitotic block produces different phenotypes in the procyclic and bloodstream life cycle stages. In the procyclic form, anucleate cells with a single kinetoplast (termed zoids) (5, 6) are generated, whereas in bloodstream form trypanosomes, cells with a single nucleus but multiple kinetoplasts are observed. Bloodstream form but not procyclic trypanosomes reinitiate nuclear S phase in the absence of mitosis. From these data, we conclude that procyclic trypanosomes can undergo cytokinesis without completion of mitosis, but inhibiting mitosis in bloodstream form trypanosomes prevents cytokinesis but neither kinetoplast replication and segregation nor an additional round of nuclear DNA synthesis.

EXPERIMENTAL PROCEDURES
Recombinant DNA Techniques-Standard DNA techniques were performed as described (11). Transformation of yeast was achieved using the lithium acetate method (12). Plasmid isolation from Escherichia coli, and DNA extraction from agarose gels used Qiaprep and Qiaex kits from Qiagen, respectively. DNA sequencing was performed by the DNA sequencing service (University of Glasgow) using an ABI automatic sequencer.
Cloning CYC6 and TbCKS1-The T. brucei genome databases (available on the World Wide Web at www.tigr.org/tdb/mdb/tbdb/ and www. sanger.ac.uk/Projects/T_brucei) were searched using the sequences of known cyclins to identify potential T. brucei homologues. Using sequence information derived from these databases, CYC6 was cloned using a PCR-based approach. The sequences of CYC6 have been submitted to EMBL under accession numbers AJ496539 and AJ496540. The genome databases were also searched for the T. brucei homologue of SUC1/CKS1 using the leishmanial CKS1 sequence (13). PCR with oligonucleotides OL861 and OL862 (Table I) was used to amplify the TbCKS1 sequence from EATRO 795 genomic DNA. The gene was cloned, sequenced, and submitted to EMBL under accession number AJ496538.
BLAST searches were carried out using the search engine at NCBI (available on the World Wide Web at www.ncbi.nlm.nih.gov/BLAST/), which also incorporated a conserved protein domain search.
Yeast Two-hybrid Assay-The Hybrid Hunter system (Invitrogen) was used. Plasmids pGL176 (LexA:CRK1) and pGL177 (LexA:CRK3) have been described previously (10). CRK4 (accession number AJ413200) was cloned into the SacI/SalI sites of pHybLex/Zeo following PCR with oligonucleotides OL372 and OL373, generating plasmid pGL395. The open reading frame of CRK5 was previously cloned into expression vector pHD675 (14), generating pGL220. 2 To clone CRK5 into pHybLex/Zeo, a 1.6-kb EcoRI-XhoI fragment from pGL220 was subcloned into the EcoRI-XhoI sites of pHybLex/Zeo, generating pGL412. This fragment contained a 5Ј-truncated CRK5 gene, since CRK5 contains an internal EcoRI site. PCR with oligonucleotides OL490 and OL491 was used to amplify the remainder of the gene and clone it into the EcoRI site of pGL412, generating pGL442 (LexA: CRK5). CRK6 (accession number AJ505556) was cloned into the EcoRI/ SalI sites of pHybLex/Zeo following PCR with oligonucleotides OL849 and OL850, generating plasmid pGL627.
A fragment of CYC6 containing the cyclin box homology region was amplified by PCR using oligonucleotides OL895 and OL896 and cloned into BamHI/HindIII-cut pYESTrp to generate pGL680 (B42:CYC6). TbCKS1 was cloned into the HindIII and XhoI sites of pYESTrp following PCR with oligonucleotides OL861 and OL862, generating plasmid pGL656 (B42:CKS1). Plasmid inserts were sequenced to show that they were in-frame with either LexA or B42 and contained no mutations introduced by PCR.
Bait and prey plasmids were transformed separately into Saccharomyces cerevisiae strain L-40, and ␤-galactosidase filter lift assays (as described by the manufacturer) were performed to confirm that no plasmid alone could activate transcription of the reporter gene lacZ. L-40 strains expressing appropriate pairs of plasmids were then generated, and further ␤-galactosidase assays were performed to assay for protein-protein interactions. Control plasmids expressing LexA:Fos, B42:Jun, and B42:lamin (Invitrogen) were also transformed into L-40 in the appropriate combinations to provide positive and negative controls for the assays. Expression of fusion proteins was confirmed by Western blotting of cell lysates with antibodies (Invitrogen) against either LexA or the V5 epitope, which is fused to the B42 domain, according to the manufacturer's protocol.
Antibodies and Immunoblotting-Antiserum specific for T. brucei CRK3 was generated by immunizing a sheep with recombinant CRK3 that was tagged with 6 histidines at the C-terminal end, generating antibody AB24. The mouse monoclonal antibody BB2 (15) was used as the anti-TY antibody. Proteins were separated by SDS-PAGE and transferred to polyvinylidene difluoride membrane. Western blots were performed as described previously (13) with either a 1:20 dilution of anti-TY antibody or a 1:200 dilution of AB24 followed by a 1:20,000 CYC6ty sense TCAGCGGGAGGTCCATACTAACCAGGATCCACTTGACGC OL953 CYC6ty antisense TGAGCGTCAAGTGGATCCTGGTTAGTATGGACCTCCCGC OL1155 CYC6 sense CCCAAAGCTTATGAATCCCACGGCACTTCG OL1156 CYC6 antisense GAGTGGGCCCTTACACAGCAGAGTGTAACG OL1079 CYC2 antisense TGAGTATTTGCCTATGCGACGGATG dilution of the appropriate horseradish peroxidase-conjugated secondary antibody. The West-Dura chemiluminescence detection system (Pierce) was used to visualize antigens. Transfection of T. brucei-Culture and transfection of T. brucei was carried out essentially as described previously (16). Where necessary, hygromycin was added at a concentration of 50 g ml Ϫ1 (procyclic cells) or 5 g ml Ϫ1 (bloodstream form cells), G418 at 10 g ml Ϫ1 (procyclic cells) or 2.5 g ml Ϫ1 (bloodstream form cells), zeocin at 10 g ml Ϫ1 (procyclic cells), and phleomycin at 2.5 g ml Ϫ1 (bloodstream form cells).
Generation of a T. brucei CYC6ty-overexpressing Cell Line-To generate TY-tagged CYC6, two annealed oligonucleotides OL952 and OL953, which composed the TY epitope flanked by Bpu1102I-compatible overhangs, were ligated into the unique Bpu1102I site at the 5Ј-end of CYC6. Correct integration of the TY tag was confirmed by sequence analysis. CYC6ty was amplified by PCR using the oligonucleotides OL1155 and OL1156, sequenced to confirm no mutations had been introduced by PCR and ligated into HindIII/ApaI-cut pHD675 (14) to generate pGL840. Plasmid pGL840 was linearized by digestion with NotI and transfected into the procyclic cell line EATRO 795 pHD449 (10). Overexpression of CYC6ty was induced in midlog phase cultures by growing cells in medium containing 50 ng ml Ϫ1 tetracycline for 48 h. Cells were harvested by centrifugation, washed in phosphate-buffered saline, and either snap-frozen in a dry ice/ethanol bath (for immunoprecipitations and kinase assays) or resuspended in Laemmli buffer (for SDS-PAGE and Western blot analysis).
Immunoprecipitations and Histone H1 Kinase Assays-To immunoprecipitate CYC6ty or CYC2ty kinase complexes from appropriate cell lines, trypanosome S100 lysates (10 7 cell equivalents) (prepared as described previously (10)) were incubated with 20 l of antibody for 1 h at 4°C. 50 l of protein G-Sepharose (0.5 mg ml Ϫ1 ) was added, and samples were incubated for a further 30 min before being washed and assayed for histone H1 kinase activity as described previously (10).
To show CRK3 co-precipitates with CYC6ty, anti-TY monoclonal antibody was first cross-linked to protein G beads using disuccinimidyl suberate using the Seize X protein G immunoprecipitation kit (Pierce). 500 l of S100 lysate from cells overexpressing CYC6ty were diluted 1:1 with binding buffer 1 (0.14 M NaCl, 8 mM Na 2 PO 4 , 2 mM potassium phosphate, 10 mM KCl, pH 7.4) and incubated with either protein G beads alone or protein G-anti-TY beads overnight at 4°C. Following extensive washing of the beads with 25 mM Tris, 0.15 M NaCl, pH 7.2, beads were resuspended in Laemmli buffer and boiled to release immunoprecipitated proteins from the antibody. Samples were analyzed by Western blotting with anti-TY and anti-CRK3 antibodies.
p12 cks1 and p13 suc1 Binding Assays-Leishmanial p12 cks1 or yeast p13 suc1 protein was coupled to Amino-link beads at a concentration of 5 mg ml Ϫ1 as described previously (13). Amino-link beads coupled to Tris-HCl, pH 7.4, to block the reactive sites were used as a negative control. S100 lysates were incubated with 50 l of control, p12 cks1 , or p13 suc1 beads for 1 h at 4°C. Beads were extensively washed with lysis buffer and used for either immunoblotting or histone H1 kinase assays.
Generation and Induction of RNAi Cell Lines-For RNAi, the procyclic 427 pLew13 pLew29 and the bloodstream form 427 pLew13 pLew90-6 cell lines (17) were transfected with pGL622 or p2T7ti/green fluorescent protein (GFP) (18). Plasmid pGL622 was generated by subcloning a 408-bp BamHI-XhoI fragment of CYC6 into vector p2T7ti (18). Plasmid p2T7ti/GFP contains the full open reading frame for GFP (758 bp) between the two opposing T7 promoters. Independent clones were generated by limited dilution cloning. To induce RNAi in procyclic forms, the cell line was grown to a density of 0.5-1 ϫ 10 7 cells ml Ϫ1 , diluted to give a density of 5 ϫ 10 5 cells ml Ϫ1 , and cultured in the presence or absence of 50 ng ml Ϫ1 tetracycline. Cells were counted daily using a Neubauer improved hemocytometer, and cultures were diluted to 10 6 cells ml Ϫ1 when the density approached 10 7 cells ml Ϫ1 . Bloodstream form cells were induced at a density of 10 5 cells ml Ϫ1 with 1 g ml Ϫ1 tetracycline and were diluted back to 10 5 cells ml Ϫ1 whenever the cell density approached 10 6 cells ml Ϫ1 .
Generation of cDNA and RT-PCR-Total RNA was prepared from cell pellets using Trizol (Invitrogen). 1 g of RNA was treated with RQ1 RNase-free DNase (Promega) for 3 h at 37°C before being heat-inactivated. For first strand cDNA synthesis, 0.325 g of RNA was mixed with 100 ng of random hexamers (Invitrogen) and denatured at 70°C for 10 min. To the RNA, dNTPs were added to 400 M, dithiothreitol was added to 10 M, and PCR buffer (5Ј-rapid amplification of cDNA ends kit; Invitrogen) was added to give a final concentration of 1ϫ. The reagents were mixed and incubated at 42°C for 1 min before 200 units of Superscript II RT (Invitrogen) was added, and the reactions were further incubated at 42°C for 50 min. Identical reactions were set up for each cDNA sample without RT to act as a control for genomic DNA contamination. The reactions were terminated by heating to 70°C for 15 min before treatment with 1 l of 10 mg ml Ϫ1 RNase A at 37°C for 30 min. Multiplex PCRs were set up, using 1 l of cDNA/reaction and PIGO primers (OL827 and OL828) plus either CYC6 (OL544 and OL746, or OL941 and OL545) or CYC2 (OL171 and OL1079) primers. PCRs contained 45 mM Tris-HCl, pH 8.8, 11 mM NH 4 SO 4 , 4.5 mM MgCl 2 , 6.7 mM ␤-mercaptoethanol, 4.4 M EDTA (pH 8.0), 100 mM dNTPs, 113 g ml Ϫ1 bovine serum albumin, 20 ng of each primer, 1 l of cDNA, and 1 unit of TaqDNA polymerase (ABI). The conditions for PCR were 1 cycle of 95°C for 30 s, 30 cycles of 95°C for 50 s, 50°C for 50 s, 72°C for 90 s, and 1 cycle of 72°C for 5 min.
4,6-Diamidino-2-phenylindole (DAPI) Staining and FACS Analysis-For DAPI staining, ϳ5 ϫ 10 5 cells were resuspended in 50 l of medium, spread on a glass microscope slide, air-dried, and fixed overnight in methanol at Ϫ20°C. Slides were removed, the methanol was allowed to evaporate, and 50 l of 1 g ml Ϫ1 DAPI in phosphatebuffered saline (PBS) with 0.5% 1,4-diazabicyclo[2.2.2]octane as antifading agent was added to the slide and spread by the addition of a coverslip. Slides were examined under UV light on a Zeiss Axioplan microscope, and images were processed using a Hamamatsu ORCA-ER digital camera and Openlab version 3.0.3 software.
For FACS analysis, cells were suspended at a density of 10 6 ml Ϫ1 in 70% methanol, 30% PBS and were incubated at 4°C overnight. Cells were washed in 10 ml of cold PBS, resuspended in 1 ml of PBS containing 10 g ml Ϫ1 propidium iodide and 10 g ml Ϫ1 RNase A, and incubated at 37°C for 45 min. FACS was performed with a Becton Dickinson FACSCalibur using detector FL2-A and an AmpGain value of 1.75. Data were analyzed using the software CellQuest version 3.3.

Identification of a T. brucei Mitotic Cyclin-We searched the
T. brucei genome databases and identified a sheared DNA clone containing a partial gene sequence displaying high similarity to mitotic A and B type cyclins. Using PCR, we were able to clone and sequence the remainder of the gene (Fig. 1A), which we designated CYC6. Sequencing of independent PCRs indicated that there are two alleles of CYC6. The first allele (CYC6-1) encodes a 436-amino acid protein with a leucine at position 309. The second allele (CYC6-2) contains an insertion of a G at position 827 of the open reading frame, resulting in a frameshift and a shorter protein of 426 amino acids (Fig. 1B). CYC6-2 also has a valine residue at position 309 in place of the leucine of CYC6-1. Evidence for the presence of two alleles rather than two genes encoding CYC6-1 and CYC6-2 was provided by Southern blot analysis of EATRO 795 procyclic genomic DNA using a fragment of CYC6 as probe (Fig. 1A), which indicated that CYC6 is most likely single copy (Fig. 1C). So far, we have no data to indicate whether the two forms of CYC6 play distinct roles in vivo. However, the presence of alleles encoding two distinct proteins is unusual and has not been reported for any other cell cycle-related gene in T. brucei to date.
BLAST searches of CYC6-1 revealed that it shares highest similarity with the mitotic B-type cyclins, cyclin 1 from Antirrhinum majus (19) (22% identity, rising to 57% identity over the cyclin box) and with cyclin S13-6 from Glycine max (20) (24% identity rising to 53% identity over the cyclin box). Additionally, a search for conserved domains in CYC6-1 revealed high homology to both N-and C-terminal cyclin fold domains, (accession numbers gn1CDD5902 and gn1CDD6896), thus showing that CYC6 has the structural characteristics of a cyclin. Fig. 1B shows an alignment of CYC6-1 and CYC6-2 with the cyclin boxes of other mitotic cyclins. Another feature characteristic of mitotic cyclins is the presence of a 9-amino acid destruction box, frequently present at the N-terminal end of the protein, which targets the cyclin for ubiquitin-mediated proteolysis via the anaphase-promoting complex at the end of mitosis (21). The consensus sequence for the destruction box in organisms studied to date is RXALGXIXN. In the absence of data on destruction box sequences in T. brucei, the best candidate is at positions 3-11 (Fig. 1B), but it diverges from the consensus sequence and, most notably, lacks the conserved arginine at position 1 (Fig. 1D).
Evidence that CYC6 can function as a cyclin came from the observation that the cyclin box homology region of CYC6 could complement the S. cerevisiae G 1 cyclin conditional mutant, DL-1 (22) (data not shown). We were therefore interested to discover its kinase partner in T. brucei. The cyclin box homology region of CYC6 was tested in a two-hybrid interaction screen with CRK1 and CRK3-CRK6. We were unable to test for an interaction with CRK2, since LexA: CRK2 was autoactivatory in the absence of prey. CYC6 expressed in the host strain L-40 resulted in negligible ␤-galactosidase activity, whereas a positive reaction was detected for CYC6 with CRK3 ( Fig. 2) but not CRK1, CRK4, CRK5, or CRK6 (not shown). To confirm this interaction in vivo, a T. brucei cell line that expresses a tetracycline-inducible TY1-epitope-tagged CYC6 gene was generated (Fig. 3A). CYC6ty was immunoprecipitated from cell lysates with an anti-TY monoclonal antibody and demonstrated to be associated with a histone H1 kinase activity (Fig. 3B). Preincubation of the antibody with TY peptide abrogated the activity.
To determine whether CYC6ty interacts with CRK3 in vivo, CYC6ty immunoprecipitates were analyzed by Western blot- ting (Fig. 3C). CRK3 co-precipitated with CYC6ty, demonstrating that CRK3 can interact with CYC6ty in vivo. It has previously been shown that active CRK3 binds to leishmanial p12 cks1 and to yeast p13 suc1 beads, with CYC2ty associating with p12 cks1 but not p13 suc1 beads (10). To determine whether CRK3 complexed with CYC6ty bound to either p12 cks1 or to p13 suc1 , induced CYC6ty lysates were incubated with control, p12 cks1 , and p13 suc1 beads before being assayed for histone H1 kinase activity or being analyzed by Western blotting (Fig. 3, D and E). Histone H1 kinase activity was detected on p12 cks1 and p13 suc1 but not control beads (Fig. 3D), whereas CYC6ty bound to p13 suc1 but not to p12 cks1 beads. This contrasted with CYC2ty, which was shown to bind to p12 cks1 but not to p13 suc1 beads (Fig. 3E), consistent with previous observations (10). Thus, the binding of different cyclin partners to CRK3 appears to alter its affinity for leishmanial p12 cks1 versus yeast p13 suc1 .
We were also able to identify and clone T. brucei CKS1, a homologue of leishmanial CKS1 (13) and S. cerevisiae SUC1. The trypanosome p12 cks1 shares 67% identity with p12 cks1 of Leishmania mexicana. In the two-hybrid assay, a positive interaction, albeit weak, was detected for CRK3 and the T. brucei CKS1, but not when CKS1 was expressed alone (Fig. 2), indicating that trypanosome p12 cks1 can interact with CRK3. This finding is consistent with the ability of CRK3 to interact with leishmanial p12 cks1 (10).
CYC6 RNAi in Procyclic Trypanosomes-In order to further investigate the role of CYC6 in vivo, an RNAi approach was used. A fragment (see Fig. 1A) of CYC6 was cloned into vector p2T7ti (18) and transfected into the T. brucei procyclic 427 pLew13 pLew29 cell line (17). Two independent clones were selected, and growth curves were generated in the absence or presence of tetracycline, which induces expression of the CYC6 double-stranded RNA (Fig. 4A). Induction of CYC6 RNAi resulted in a growth arrest in procyclic cells ϳ48 -72 h postinduction. No such growth arrest was seen upon tetracycline induction of the RNAi cell line transfected with the p2T7ti construct containing the coding sequence for GFP (Fig. 4A), thus indicating that the growth arrest seen with CYC6 RNAi is not caused by the production of double-stranded RNA per se but is specifically associated with production of CYC6 doublestranded RNA.
To confirm that the growth arrest resulted from a specific down-regulation of CYC6 mRNA, an RT-PCR method was used. This approach was taken, since CYC6 could not be detected reliably by Northern analysis due to low abundance. cDNA generated from mRNA with or without induction was used in multiplex PCRs with primers for PIGO and either CYC6 or CYC2. One primer used to amplify CYC6 binds outside of the region used in the RNAi construct (Fig. 1A); hence, any products obtained should be derived solely from native RNA and not from RNA produced from the RNAi construct. The PCR products were then electrophoresed and subjected to Southern blotting with the relevant probes. T. brucei PIGO, a gene involved in lipid anchor biosynthesis, 3 was used as a loading control. The ratio of CYC6 or CYC2 product to PIGO product was assessed on a phosphor imager. Since each independent clone gave essentially the same phenotype, for this and subsequent analyses, only the data for one clone is shown. Induction of CYC6 RNAi in procyclic cells led to a dramatic decrease in CYC6 mRNA relative to PIGO after 28 h, whereas the CYC2 mRNA levels remained approximately constant (Fig.  4B). This indicates that the effects of the RNAi are due to a specific down-regulation of CYC6 mRNA.
To confirm that knockdown of CYC6 RNAi was associated with a down-regulation of CRK3-CYC6 kinase activity, cell lysates were assayed for p13-binding kinase activity. Following induction of CYC6 RNAi, the p13-binding kinase activity decreased to background levels (Fig. 4C), consistent with the notion that CYC6 RNAi leads to a reduction in CYC6 protein levels, resulting in reduced CRK3-CYC6 kinase activity.
The karyotype distribution of the cell populations was monitored through the time course by using DAPI staining to visualize nuclei and kinetoplasts (Fig. 5). Induction of CYC6 RNAi in procyclic form trypanosomes resulted in an absence of cells progressing through the cell cycle to a two-nuclei, twokinetoplast (2N2K) karyotype, while at the same time giving rise to significantly more 1N2K and 0N1K ("zoid") cells (Fig.  5A). Indeed, within the population, isolated cells of the 1N2K karotype could be seen dividing to give a 1N*1K daughter cell (1N* being where the nucleus has a 4C DNA content; see 3  CYC6ty-overexpressing cell line was grown in the absence or presence of tetracycline (tet) for 48 h before cell lysates were prepared and analyzed by SDS-PAGE followed by Western blotting with anti-TY antibody. B, histone H1 kinase assay. CYC6ty was immunoprecipitated from uninduced and induced CYC6ty cell lysates with anti-TY antibody that had been preincubated in the absence or presence of TY peptide. Immunoprecipitates were assayed for histone H1 kinase activity. C, co-immunoprecipitations. Lysates of cells overexpressing CYC6ty (3 ϫ 10 7 cell equivalents) were incubated with protein G or protein G coupled to anti-TY antibody. Immune complexes were dissociated from the beads by boiling and then analyzed by SDS-PAGE and Western blotting with anti-TY antibody or anti-CRK3 antiserum. D and E, p12 cks1 and p13 suc1 binding. Lysates of cells overexpressing CYC6ty were incubated with control, p12 cks1 , or p13 suc1 beads before being assayed for histone H1 kinase activity (D) or being analyzed by Western blotting with anti-TY antibody (E). As a control, anti-TY Western blots of CYC2ty lysates that had been incubated with control, p12 cks1 , or p13 suc1 beads were also carried out (E). tet, tetracycline; pep, peptide; C, control beads. below) and a zoid (Fig. 5B), suggesting that the absence of mitosis did not prevent cytokinesis.
To measure the DNA content of the cells through the time course, cells were analyzed by FACS (Fig. 6). RNAi of CYC6 resulted in a decrease in the proportion of cells with 2C DNA content, whereas the proportions of cells with either Ͻ1C DNA content or 4C DNA content increased. This is in keeping with the observed karyotype of the cells (i.e. an increase in 1N2K cells that have replicated their nuclear DNA (4C content, designated 1N*) and an increase in zoids (Ͻ1C DNA content)). The absence of cells with DNA contents greater than 4C indicates that, within the population as a whole, 1N*1K daughter cells do not undergo a further round of nuclear S phase.
CYC6 RNAi in Bloodstream Form Trypanosomes-To test whether RNAi of CYC6 would give the same phenotype in bloodstream form cells, the same construct was transfected into cell line 427 pLew13 pLew90-6 (17), and two independent clones were selected for further analysis. RNAi of CYC6 in bloodstream form cells resulted in a growth arrest after overnight induction (Fig. 7A). This growth arrest was associated with a specific reduction in CYC6 mRNA as determined by RT-PCR (Fig. 7B). After induction of CYC6 RNAi, a significant FIG. 4. RNAi of CYC6 in procyclic form cells. A, growth curves. RNAi cell lines were grown in the presence or absence of tetracycline, and cell number was monitored over time. Representative growth curves for each procyclic CYC6 RNAi clone are shown (upper graph). Filled triangles, clone A8 without tetracycline (tet); open triangles, A8 with tetracycline; filled squares, clone E11 without tetracycline; open squares, E11 with tetracycline. A representative growth curve for a GFP RNAi cell line is also shown (lower graph). Filled circles, clone G4 without tetracycline; open circle, G4 with tetracycline. B, Southern blots of RT-PCRs. RNA was prepared from cells after a 24-h induction. First strand cDNA was generated, and multiplex PCRs were carried out using PIGO primers with either CYC6 or CYC2 primers. Products were blotted onto membrane and visualized by probing with the relevant genes. RT, reverse transcriptase. C, kinase assays for CYC6 RNAi cell lines. Uninduced and induced cell lysates (T ϭ 48 h, 2 ϫ 10 6 cell equivalents) were incubated with control or p13 beads before being assayed for histone H1 kinase activity.

FIG. 5. Karyotype of procyclic form cells after CYC6 RNAi.
A, karyotype distribution. The karyotype distribution of procyclic CYC6 RNAi cells (clone A8) induced with tetracycline (tet) was monitored over time. Circles, 1N1K; squares, 1N2K; triangles, 2N2K; crosses, 0N1K. B, images of procyclic CYC6 RNAi cells. Cells (clone A8) were stained with DAPI (from left to right, phase contrast, DAPI-stained, and overlay). Top, uninduced. Bottom, tetracycline-induced (T ϭ 50 h). decrease in CYC6 mRNA relative to the PIGO mRNA occurred in both clones. A similar decrease in CYC2 mRNA was not seen, thus indicating that the effects of the RNAi are due to a specific down-regulation of CYC6 mRNA. Assaying the p13-binding kinase activity following induction of RNAi showed a decrease in CRK3-CYC6 activity down to background levels (Fig. 7C).
An analysis of the karyotype distribution in the induced bloodstream form cells revealed a significant increase in 1N2K cells with a coincident decrease in 2N2K cells, together with cells with multiple kinetoplasts (1N3K, 1N4K increasing up to 1N11K) (Fig. 8). In the multikinetoplast cells, the nucleus was usually significantly enlarged and often bilobed. Cell morphology was frequently aberrant. These results suggest that the cells are unable to complete mitosis, but unlike the situation in procyclic cells, cytokinesis does not occur. Absence of cytokinesis does not, however, prevent kinetoplast replication and segregation or outgrowth of a new flagellum.
Since the nuclei in the multikinetoplast cells were significantly enlarged, it seemed possible that the arrested bloodstream form cells were undergoing repeated rounds of nuclear S phase in the absence of mitosis. FACS analysis (Fig. 9) showed over time an increase in cells with a DNA content of 4C, coincident with a decrease in cells with a 2C DNA content. However, unlike the case for the procyclic form, this was accompanied by an increase in cells with DNA contents of greater than 4C, suggesting that the single nucleated, multikinetoplast cells underwent further rounds of nuclear S phase. FACS analysis on the host strain 427 pLew13 pLew90-6 revealed the same profile as the uninduced RNAi cell line (data not shown), confirming that this strain was not tetraploid prior to transfection with the RNAi construct. DISCUSSION We report the identification and analysis of a novel mitotic cyclin gene CYC6 from T. brucei, thus providing the first analysis of a molecule crucial to mitosis in the African trypanosome. We base the classification of CYC6 as a mitotic cyclin on the high degree of sequence similarity to many B-type cyclins, the presence of a sequence motif similar to a destruction box, and its functional complementation of the S. cerevisiae G 1 cyclin conditional mutant DL-1. This strain has previously been shown to be promiscuous for the type of cyclin that can complement it (12,22). Further, we have shown that CYC6ty is able to interact with CRK3, the proposed functional Cdc2/CDK1 homologue in kinetoplastids (1,23), to form an active kinase complex. This is only the second cyclin-kinase complex in African trypanosomes to be identified. CRK3 interacts with the PREG1/ PHO80-like (24) trypanosome cyclin CYC2 (10) to form a p12 cks1 -binding complex and with the mitotic cyclin CYC6 to form a p13 suc1 -binding complex. Considering the proposed classifications of the cyclins, it is likely that CRK3 interacts with CYC2 in G 1 and has a G 1 /S-specific function, perhaps operating by sensing cell size or nutritional status, and with CYC6 at G 2 /M to bring about progression through mitosis. Recent data to support this is the finding that RNAi of CYC2 in the procyclic form of T. brucei leads to a G 1 cell cycle block. 4 Finally, using RNAi, we have shown that down-regulation of CYC6 mRNA results in a reproducible mitotic block phenotype in both procyclic and bloodstream form trypanosomes (Table  II). Induction of CYC6 RNAi caused a growth arrest after 48 -72 h in procyclic cells and after 10 -16 h in bloodstream form cells. The differences in rapidity of onset of the phenotype may, at least in part, be accounted for by the differences in generation time of the two life cycle forms (ϳ10 -12 h for the procyclic form and 5-6 h for the bloodstream form). Additionally, the fact that procyclic cells were able to undergo cytokinesis following the mitotic block means that the growth arrest RNAi cell lines were grown in the presence or absence of tetracycline (tet) and cell number was monitored over time. Representative growth curves for each bloodstream clone are shown. Filled triangles, clone P2A without tetracycline (tet); open triangles, P2A with tetracycline; filled squares, clone P4A without tetracycline; open squares, P4A with tetracycline. B, Southern blots of RT-PCRs. RNA was prepared from cells induced for 8 h. First strand cDNA was generated, and multiplex PCRs were carried out using PIGO primers with either CYC6 or CYC2 primers. Products were blotted onto membrane and visualized by probing with the relevant genes. C, kinase assays for CYC6 RNAi cell lines. Uninduced and induced cell lysates (T ϭ 21 h, 2 ϫ 10 6 cell equivalents) were incubated with control or p13 beads before being assayed for histone H1 kinase activity. did not become apparent until one cell cycle later.
In procyclic trypanosomes, knockdown of CYC6 mRNA resulted in an almost total absence of 2N2K cells, showing that cells were unable to undergo nuclear mitosis, although they were able to segregate their kinetoplasts. This indicates that CYC6 is required for nuclear mitosis but not for division of the replicated kinetoplast. FACS analysis confirmed that these cells did not reinitiate nuclear S phase but were able to undergo cytokinesis in the absence of mitosis to give a daughter cell with double nuclear DNA content (1N*1K) and a zoid (0N1K). This phenotype has previously been seen as a result of treating procyclic trypanosomes with the anti-microtubule agent rhizoxin (5,6). This work confirms, using ablation of a molecule essential for entry into mitosis, that procyclic form trypanosomes lack a cell cycle checkpoint preventing initiation of cytokinesis in the absence of completion of mitosis, which is present in other eukaryotes.
In bloodstream form cells, RNAi knockdown of CYC6 resulted in a significant decrease in the formation of 2N2K cells, consistent with CYC6 being required for mitosis. In many cases, the nuclei of the aberrant bloodstream form cells were significantly enlarged and bilobed, often to such an extent that only a very thin thread of DNA connected the two lobes. In contrast, bilobed nuclei were not seen in the procyclic 1N*2K cells. It is unlikely that these differences are a result of differing levels of penetrance of the RNAi in the two life cycle stages, since the p13-binding kinase activity is reduced to background levels following RNAi in both stages, suggesting similar levels of CYC6 down-regulation. Instead, this difference in phenotype could indicate that CYC6 acts later in mitosis in bloodstream forms than in procyclic forms. In contrast to the situation in  procyclic cells, however, the absence of large numbers of zoids and the presence of cells with multiple kinetoplasts indicate that bloodstream form trypanosomes are unable to undergo cytokinesis in the absence of mitosis. The majority of aberrant cells had karyotypes of 1N3K or 1N4K, but cells with one nucleus and up to 11 kinetoplasts were occasionally seen. This may indicate that, unlike the situation in the procyclic form, the mitosis to cytokinesis checkpoint present in other eukaryotes is operational in bloodstream form trypanosomes. Alternatively, it can be considered that structural constraints of cell division in the bloodstream form render cytokinesis impossible in the absence of mitosis. In the procyclic form, it was suggested that entry into cytokinesis may be largely dependent on kinetoplast division and segregation (5). However, it is clear that kinetoplast division and segregation is not sufficient for initiation of cytokinesis in bloodstream form cells. Bloodstream form cells were also able to reinitiate nuclear S phase in the absence of mitosis or cytokinesis, thus providing an explanation for the enlarged nuclei seen by microscopy. These data indicate that in this form, the checkpoint preventing reentry into S phase until cytokinesis and G 1 have been completed is either lacking or has been inactivated by the knockdown of CYC6. The latter option would be consistent with the situation in yeast where reentry into S phase is inhibited by high CDK1 activity (25). Recently, RNAi of FLA1 (flagella adhesion glycoprotein gene) (26) and of GPI8, encoding an enzyme involved in attachment of glycosylphosphatidylinositol anchors to proteins (27) in bloodstream form T. brucei, was shown to result in a block in cytokinesis. In these cells, multiple rounds of nuclear mitosis and kinetoplast segregation as well as flagella duplication in the case of GPI8 RNAi occurred. Thus, it seems that in the absence of cytokinesis in bloodstream form cells, there is no mechanism to prevent multiple rounds of reentry into S phase and organelle segregation. Similarly, if cytokinesis is blocked in the procyclic form of T. brucei by RNAi of FLA1 or ␣-tubulin or in the epimastigote form of T. cruzi by treatment with vinca alkaloids, repeated rounds of mitosis and kinetoplast duplication can occur (26,28,29). It remains to be seen whether in these cells this is due to inactivation of CRK3-CYC6 having occurred prior to the block in cytokinesis or whether trypanosomes lack a traditional mitosis to S phase checkpoint altogether. In the latter case, this would imply that reentry into S phase in the same cell cycle is possible until cytokinesis is initiated. In any case, it will be important to bear this in mind when analyzing RNAi cell lines in the future, since proteins uninvolved in the cell cycle, may, by invoking a block in cytokinesis through a cell cycle-independent mechanism, appear to generate a cell cycle phenotype.
One final point on checkpoints is that, at least in the procyclic form, there must be additional controls operating at G 1 /S to prevent the 1N*1K (4C nuclear DNA content) daughter cells from undergoing a further round of S phase. In mammalian cells, a p53-dependent G 1 tetraploidy checkpoint exists to prevent cells that have failed to segregate their chromosomes during the previous mitosis from progressing through G 1 , through transactivation of the CDK inhibitor p21 (30,31). It will prove interesting to discover exactly how procyclic trypanosomes bring about a similar arrest.
This work shows that there is an absolute requirement for CYC6 for mitosis, and the phenotypes seen with RNAi (Table  II) of CYC6 argue against any redundancy of function between the putative mitotic cyclins in T. brucei. It is possible that one of the other mitotic-like cyclins, CYC3 or CYC8, could be involved in later mitotic events, such as exit from mitosis, or segregation of the two daughter nuclei following nuclear division. Further, it is now apparent that kinetoplast division and segregation are not dependent on the G 2 /M cyclin-kinase complex CRK3-CYC6, raising the possibility that another cyclinkinase complex, perhaps involving one of the other mitotic-like cyclins, CYC3 or CYC8, is required for this process in the trypanosome cell cycle. Although the cell cycles of the nucleus and kinetoplast occur approximately synchronously, suggesting that they may be linked, they are not interdependent, as shown by the treatment of procyclic cells with okadaic acid, which prevents kinetoplast replication and cytokinesis but not mitosis (32), and in this study where kinetoplast replication can repeatedly occur in the absence of mitosis. The involvement of different cyclin-kinase complexes in the two cell cycles would provide an explanation for these phenomena.
Finally, although the T. brucei cell cycle has features common to other eukaryotes, such as a single CDK interacting with more than one cyclin to carry out different functions, trypanosome-specific features also exist. Replication and division of its mitochondrion appears to be carefully controlled by as yet unidentified molecules that are not required for nuclear division. Structural constraints of cell division may also mean that different and possibly novel cell cycle control mechanisms operate during mitosis and cytokinesis. Indeed, the unusual mechanism for achieving chromosome segregation in the trypanosome, proposed as a lateral stacking model (33), may indicate that unique molecular pathways underlie mitosis in African trypanosomes. Further careful molecular dissections of the events occurring at mitosis and cytokinesis in both bloodstream and procyclic forms will be required to fully determine the operational cell cycle checkpoints in trypanosomes and to elucidate how such control mechanisms evolved in the different life cycle stages.