Pathogenesis, morphogenesis and cell cycle are connected in the fungal pathogen Ustilago maydis. Here we report the characterization of the catalytic subunit of the cyclin-dependent kinase, encoded by the gene cdk1, and the two B-type cyclins present in this organism, encoded by the genes clb1 and clb2. These cyclins are not redundant and appears to be essential for cell cycle. The analysis of conditional mutants in cyclin genes indicates that Clb1 is required for G1 to S and G2 to M transitions, while Clb2 is specifically required for the onset of mitosis. Both Clb1 and Clb2 carry functional destruction boxes, and expression of derivatives lacking D-boxes arrested cell cycle at a post-replicative stage. High levels of Clb1 generated cells with anomalous DNA content that were hypersensitive to microtubule-destabilizing drugs. In contrast, high levels of Clb2 induce premature entry into mitosis, suggesting that Clb2 is a mitotic inducer in U. maydis. In addition, Clb2 affects morphogenesis, and overexpression of clb2 induces filamentous growth. Furthermore, we have found that appropriate levels of Clb2 cyclin are critical for a successful infection. Mutant strains with half a dose of clb2 or high level of clb2 expression are impaired at distinct stages in the infection process. These data reinforces the connections between cell cycle, morphogenesis and virulence in this smut fungus.
Ustilago maydis, a basidiomycete fungus, causes smut disease of maize. In this fungus, pathogenesis and sexual development are intricately interconnected, to the extent that U. maydis is completely dependent on the plant to accomplish a complete sexual cycle (Banuett, 1995). Haploid cells grow in yeast-like unicellular form, dividing by budding, and induction of the pathogenic phase requires the mating of two compatible haploid cells and the generation, after cell fusion, of an infective dikaryotic filament, which invades the plant (Kahmann et al., 2000). The transition from budding growth to mating is a response to environmental factors and the presence of compatible pheromone. In previous work, we demonstrated that mating is linked to cell cycle, and that cells of U. maydis, when exposed to pheromone, undergo a cell cycle arrest in G2 phase prior to mating (García-Muse et al., 2003). This cell cycle arrest contrasts with pheromone-induced cell cycle arrest in ascomycete yeasts such as Saccharomyces cerevisiae and S. pombe, which takes place in G1 phase (Sprague and Thorner, 1992; Davey, 1998), and suggests alternative mechanisms linking pheromone-response and cell cycle arrest in this basidiomycete fungus. To determine the mechanisms involved in pheromone-induced cell cycle arrest in U. maydis, we first found it necessary to make a careful study of the basic molecular components of cell cycle regulation, because so little information is available for this organism in this respect.
Previous studies related to cell cycle in Ustilago maydis have been focused on morphological descriptions of asynchronous cultures (O'Donnell and McLaughlin, 1984; Jacobs et al., 1994; Snetselaar and McCann, 1997; Steinberg et al., 2001; Banuett and Herskowitz, 2002). Vegetatively growing U. maydis cells normally produce one polar bud per cell cycle (Jacobs et al., 1994). Studies correlating nuclear density and cell morphology showed that cells complete DNA synthesis before beginning to form buds. In other words, the bud formation takes place in G2 phase (Snetselaar and McCann, 1997). This is in contrast to bud formation in S. cerevisiae, for example, which is reported to occur during S phase (Pringle and Hartwell, 1981). Once the bud is nearly mature, the nucleus migrates to the bud, where it divides (Holliday, 1974; O'Donnell and McLaughlin, 1984; Snetselaar, 1993; Banuett and Herskowitz, 2002). In rapidly growing cells, because cell division produces daughter cells with a mass similar to mother cells, G1 phase is very brief and the S phase begins shortly after cytokinesis. Therefore, cells have a 2C DNA content for most of the cell cycle (Snetselaar and McCann, 1997). Morphological studies of the cytoskeleton in dividing U. maydis cells have clearly defined the microtubule and actin organization during cell cycle, proving excellent markers of different cell cycle stages. For instance, U. maydis cells undergoing G2 phase assemble long microtubules towards the growth region while at the onset of mitosis, this network disassembles and is replaced by a spindle and prominent astral microtubules (Steinberg et al., 2001; Banuett and Herskowitz, 2002).
In eukaryotes, major cell cycle controls regulate the onset of S phase and mitosis and ensure that these events occur in the correct sequence. Central to these controls are the cyclin-dependent kinases (CDKs), which are composed of a catalytic subunit and a regulatory subunit called cyclin. In fungi, a single catalytic subunit, encoded by cdc2 in fission yeast and by CDC28 in budding yeast, is required for both these cell cycle transitions (Krylov et al., 2003). Fungal Cdc2 homologues become associated with different cyclins that function during G1 for the onset of S phase and later in the cell cycle for the onset of mitosis (Nasmyth, 1993). In multicellular eukaryotes several catalytic subunits are present, which form a variety of complexes with different cyclins, regulating progression through the cell cycle (Nigg, 1995).
In an effort to characterize the molecular basis of the cell cycle regulation in Ustilago maydis, we have isolated the gene encoding the CDK catalytic subunit as well as the genes encoding the two B-cyclins of U. maydis. We have characterized the phenotypes of diverse gain-of-function and loss-of-function mutations in the cyclin genes in order to determine their function in mitosis and to distinguish any individual roles they may have. We show that these cyclins play a primary non-redundant role in different cell cycle events from S phase to progression through mitosis. Furthermore, we have found that the B-type cyclin Clb2 acts as a mitotic inducer in U. maydis and plays a fundamental role in morphogenesis and pathogenesis. We show that strains carrying anomalous clb2 gene doses are affected in their cellular shape, and in the ability to successfully infect. To our knowledge this is the first report linking a cell cycle defect to virulence in a phytopathogenic fungus. The data reported in this work reinforce the postulated connections between cell cycle, morphogenesis and pathogenicity in Ustilago maydis.
Materials and Methods
Strains and growth conditions
For cloning purposes the E. coli K12 derivative DH5α (Bethesda Research Laboratories) was used. All U. maydis used in this study are listed in Table 1. Strains were grown at 28°C in yeast extract, peptone, dextrose medium (YPD), yeast extract, peptone, sucrose medium (YEPS) [modified after Tsukuda et al. (Tsukuda et al., 1988)], complete medium (CM) or minimal medium (MM) (Holliday, 1974). Conditional strains were grown in MM with nitrate (MM-NO3) as the sole nitrogen source as described previously (Banks et al., 1993). To shut-off the Pnar1 promoter, strains to be tested were grown in MM-NO3 until OD600 of 0.2, pelleted by centrifugation, washed twice with minimal medium without nitrogen, and incubated in MM with ammonium as the nitrogen source (MM-NH4) or YPD (Brachmann et al., 2001). For induction of the Pcrg1 promoter (Bottin et al., 1996; Brachmann et al., 2001), strains to be tested were grown in CM medium with 2% glucose as a carbon source (CMD) or rich medium (YPD) until OD600 of 0.2, pelleted by centrifugation, washed twice with water and incubated in CM with 2% arabinose (CMA) or rich medium with 2% arabinose (YPA). All chemicals used were of analytical grade and were obtained from Sigma or Merck.
Isolation of cdk1 gene
Two sets of degenerate oligonucleotides were synthesized according to the nucleotide sequences that encode two conserved regions in CDK from different fungi: KLADFGLA (CDK1: 5′AARYTNGCNGAYTTYGGNYTNGCN3′) and WYRAPE (CDK2a: 5′YTCNGGNGCNCGRTACCA3′ and CDK2b: 5′YTCNGGNGCYCTRTACCA3′). The CDK1/CDK2a and CDK1/CDK2b pairs were used for amplification with 50 ng FBD11 DNA as template in a volume of 50 μl. PCR products were generated in the following reaction mixture: 10 mM Tris-HCl pH 8.0, 50 mM KCl, 1.2 mM MgCl2, 100 μM dNTP, 50 μM of each primer and 2.5 units Taq polymerase. Conditions for PCR cycling included denaturation at 94°C for 1 minute, annealing at 45°C for 1 minute and extension at 72°C for 2 minutes. Selected fragments were isolated and cloned into pGEM-T Easy (Promega). Positive clones containing inserts were chosen and the nucleotide sequence of each plasmid insert was determined in both directions by the ABI model 373A Auto Sequence System (Perkin Elmer/Applied Biosystems). Three different kinds of inserts of the same size (93 bp) were obtained. The conceptual translation of these fragments generated amino acidic sequences with similarity to the sequence of CDK proteins from other fungi. Sequences flanking these fragments were obtained with a PCR-walking strategy (Siebert et al., 1995) using the Genome Walker system (Clontech) as directed by the manufacturer. The analysis of the sequence revealed three different ORFs: cdk1, which shows the highest similarity to S. pombe Cdc2, the main CDK in fission yeast (Beach et al., 1982); cdk3, which shows the highest sequence similarity to S. cerevisiae Srb10, a CDK involved in transcriptional regulation (Liao et al., 1995); and crk1, with the highest similarity to S. cerevisiae Ime2 (Garrido and Pérez-Martín, 2003).
Isolation of clb1 and clb2 genes
Two sets of degenerate oligonucleotides were synthesized according to the nucleotide sequences that encode two conserved regions in B-cyclins from different fungi: MVA/SEY (G2-1a: 5′ATGGTNKCNGARTAY3′ and G2-1b: 5′ATGGTNAGYGARTAY3′) and FIAA/SKY (G2-2a: 5′TTYATHGCNKCNAARTAY3′ and G2-2b: 5′TTYATHGCNAGYAARTAY3′). Pairs of oligonucleotides were used for amplification with 50 ng FBD11 DNA as template as above. Two different kinds of inserts of the same size (282 bp) were obtained. The conceptual translation of these fragments generated amino acidic sequences with similarity to the sequence of B-type cyclins proteins from other fungi. Sequences flanking one of these fragments (the one encoding Clb1) were obtained with a PCR-walking strategy (Siebert et al., 1995) as above. Flanking sequences to the fragment encoding Clb2, were generously provided by Peter Schreier (Bayer CropScience AG, Monheim, Germany).
Plasmids utilized in this study are listed in Table 2. Sequence analysis of fragments generated by PCR was performed with an automated sequencer (ABI 373A) and standard bioinformatic tools. Integration of the plasmids into the corresponding loci was verified in each case by diagnostic PCR and subsequent Southern blot analysis.
A fragment carrying the entire cdk1 ORF sequence without stop codon and flanked by SpeI and EcoRI sites was obtained by amplification of genomic DNA with the primers CDK1 (5′GGACTAGTCATATGGACAAGTATCAAAGGATCGAA3′) and CDK4 (5′CGGAATTCTGTGAGGAGCCTCCTGAAGTACGG3′). This fragment was cloned in the pBS-FLAG-HYG plasmid digested with SpeI and EcoRI. The pBS-FLAG-HYG plasmid carries a copy of the FLAG epitope, and a hygromycin B resistance marker (J. P.-M., unpublished). After digestion with SacII the pCDK-FLAG plasmid was integrated by homologous recombination into the cdk1 locus.
This plasmid was produced by ligation of a pair of DNA fragments flanking the cdk1 ORF into pSMUT, a U. maydis integration vector containing a hygromycin B resistance cassette (Bölker et al., 1995) digested with BamHI and XhoI. The 5′ fragment (flanked by BamHI and NotI sites) spans from nucleotide –426 to nucleotide –18 (considering the adenine in the ATG as nucleotide +1) and it was produced by PCR amplification using the primers PCDK-A (5′ATTTGCGGCCGCTGGACAAACATCTTGGTCGGA3′) and PCDK-B (5′CGGGATCCAAGGGCAGGCGAAGAGCGACC3′). The 3′ fragment (flanked by NotI and XhoI sites) spans from nucleotide +740 to nucleotide +1094 and it was produced by PCR amplification using the primers TCDK-A (5′CCGCTCGAGCCCCACTGACGACGTTTGGCC3′) and TCDK-B (5′TAAAGCGGCCGCGACGAGGCCAGCACGAAAAAA3′). After digestion with NotI the pKOCDK plasmid was integrated by homologous recombination into the cdk1 locus.
A fragment carrying the entire clb1 ORF sequence without stop codon and flanked by SpeI and EcoRI sites was obtained by amplification of genomic DNA with the primers ATGclb1 (5′GGACTAGTCATATGTCTCAGAACATCGTAAGCGCTC3′) and CLB-tag (5′CGGAATTCCTCCGTCGCTTCGTACGCGTT3′). This fragment was cloned in the plasmid pBS-VSV-HYG digested with SpeI and EcoRI. The plasmid pBS-VSV-HYG carries a copy of the VSV epitope and a hygromycin B resistance marker (J.P.-M., unpublished). After digestion with KpnI the pCLB1-VSV plasmid was integrated by homologous recombination into the clb1 locus.
A fragment carrying the entire clb2 ORF sequence without stop codon and flanked by SpeI and EcoRI sites was obtained by amplification of genomic DNA with the primers CLB2-1 (5′GGACTAGTCATATGCCACAACGCGCTGCCCTT3′) and CLB2-9 (5′CCGAATTCAGGCTGACCTGCTTGAGTCGA3′). This fragment was cloned in the plasmid pBS-MYC-HYG digested with SpeI and EcoRI The plasmid pBS-MYC-HYG carries three copies of MYC epitope and a hygromycin B resistance marker (J.P.-M., unpublished). After digestion with BlpI the pCLB2-MYC plasmid was integrated by homologous recombination into the clb2 locus.
This plasmid was produced by ligation of a pair of DNA fragments flanking the clb1 ORF into pNEBHyg(+), a U. maydis integration vector containing a hygromycin B resistance cassette (Brachmann et al., 2001) digested with EcoRI and SacII. The 5′ fragment (flanked by EcoRI and KpnI) was produced by PCR using the primers CYC1-2 (5′GGGGTACCTCGCGGTCATCGTACTAGACAG3′) and CYC1-3 (5′CAGCGCAAGCTGAGAATTCAACTTCCACTC3′). This fragment spans from nucleotide –1668 to nucleotide –894 (considering the adenine in the ATG as nucleotide +1). The 3′ fragment (flanked by KpnI and SacII) spans from nucleotide +2113 to nucleotide +2539 and it was produced by PCR amplification using the primers CYC1-7 (5′TCCCCGCGGAGACCTTCTAGATATCTTCCC3′) and CYC1-8 (5′CGGGGTACCCCAATCCGCTTATGATTC3′). After digestion with KpnI the pKOCLB1 plasmid was integrated by homologous recombination into the clb1 locus.
This plasmid was produced by ligation of a pair of DNA fragments flanking the clb2 ORF into pNEBCbx(+), a U. maydis integration vector containing a carboxine resistance cassette (Brachmann et al., 2001) digested with EcoRI and SacII. The 5′ fragment (flanked by EcoRI and KpnI) was produced by PCR using the primers CYC2-2 (5′CGGGGTACCTTTCTGCTATTGGCTTCAGCA3′) and CYC2-3 (5′CGGAATTCGGCTCGGAGTTTACTGCGGTAG3′). This fragment spans from nucleotide –498 to nucleotide –14 (considering the adenine in the ATG as nucleotide +1). The 3′ fragment (flanked by KpnI and SacII) spans from nucleotide +1753 to nucleotide +2278 and it was produced by PCR amplification using the primers CYC2-7 (5′TCCCCGCGGTCAATCTGAGAGGCACAGTTC3′) and CYC2-8 (5′CGGGGTACCTGGCTTTCGCACATCACATGG3′). After digestion with KpnI the pKOCLB2 plasmid was integrated by homologous recombination into the clb2 locus.
This plasmid was constructed by ligation of a pair of DNA fragments into pRU2, a U. maydis integration vector containing the promoter of the nar1 gene (Brachmann et al., 2001) digested with NdeI and EcoRI. The 5′ fragment (flanked by EcoRI and KpnI) was produced by PCR using the primers CYC1-2 and CYC1-3. The 3′ fragment (flanked by NdeI and KpnI) was obtained by PCR amplification with primers ATGclb1 and CLB-tag and spans from nucleotide +1 to nucleotide +2114. After digestion with KpnI the pCLB1nar plasmid was integrated by homologous recombination into the clb1 locus.
This plasmid was constructed by ligation of a pair of DNA fragments into pRU2 digested with NdeI and EcoRI. The 5′ fragment (flanked by EcoRI and KpnI) was produced by PCR using the primers CYC2-2 and CYC2-3. The 3′ fragment (flanked by NdeI and KpnI) was obtained by PCR amplification with primers CLB2-1 and CYC2-6 (5′CGGGTACCTCCTCGGGATCGAGCCCCATGA3′) and spans from nucleotide +1 to nucleotide +773. After digestion with KpnI the pCLB2nar plasmid was integrated by homologous recombination into the clb2 locus.
Plasmids overexpressing clb1
The pRU11-CLB1 plasmid carries the VSV-tagged allele clb1-1 cloned under the control of the Pcrg1 promoter. It was constructed by cloning a NdeI-AflII fragment from pCLB1-VSV into the same sites of pRU11 (Brachmann et al., 2001). Plasmid pRU12-CLB1 was constructed in a similar way, but the pRU12 plasmid was used instead of pRU11. This plasmid carries a mutant version of the Pcrg1 promoter (that we called Pcrg1*), which is around five times less active in the presence of arabinose than the Pcrg1 promoter present in pRU11 (Brachmann et al., 2001). The plasmids carrying Clb1 versions lacking the destruction boxes were constructed by exchanging the wild-type clb1 ORF as a NdeI-EcoRI fragment from pRU12-CLB1 with fragments carrying the deleted versions obtained after PCR amplification with the following oligonucleotides: (i) pRU12-CLB1ΔDB1: ATG-Clb1 and CLB-dbTAG (5′CGGAATTCCTTGACGCTGGGCAGCAGCCTT3′); (ii) pRU12-CLB1ΔDB2: CLB3 (5′GGACTAGTACTTTATGGAGATTTGCTCG3′) and CLB-TAG; (iii) pRU12-CLB1ΔDB1-2: CLB3 and CLB-dbATG. All of them were integrated after SspI linearization into the succinate dehydrogenase (cbx) locus as described previously (Brachmann et al., 2001).
Plasmids overexpressing clb2
The plasmid pRU11-CLB2 carries the MYC-tagged allele clb2-1 cloned under the control of Pcrg1 promoter. It was constructed by cloning of a NdeI-AflII fragment from pCLB2-MYC into the same sites of pRU11 (Brachmann et al., 2001). A similar construction was made by using the pRU12 plasmid, that carries the weaker Pcrg1* promoter, giving the plasmid pRU12-CLB2. The derivative lacking the destruction box, pRU12-CLB2ΔDB was constructed by exchanging the wild-type clb2 ORF as a NdeI-EcoRI fragment from pRU12-CLB2 with a fragment carrying the deleted version obtained after PCR amplification with the CLB2-2 (5′CGGGATCCATATGGTCGCCAGAGCCAATGCA3′) and CLB2-9 primers.
The pCU1-CLB2 plasmid expressing the clb2-1 allele under the control of the constitutive Phsp70 promoter was constructing by cloning a NdeI-AflII fragment from pCLB2-MYC into the same sites of pCU1 (Brachmann et al., 2001). All of them were integrated after linearization with SspI into the succinate dehydrogenase (cbx) locus.
Specific cell cycle arrests
Cell cycle arrests were carried out as described previously (García-Muse et al., 2003). To enrich the population in cells in G1 phase, a feed-starve regimen was followed as described previously (Holliday, 1965).
Protein analysis and kinase assay
For the preparation of crude protein extracts, cells were harvested by centrifugation at 4°C and washed twice with ice-cold water. The cell pellet was resuspended in ice-cold HB buffer (25 mM Mops pH 7.2, 15 mM MgCl2, 15 mM EGTA, 1% Triton X-100, 20 μg/ml leupeptin, 40 μg/ml aprotinin, 0.1 mM sodium orthovanadate and 15 mM p-nitrophenyl phosphate). An equal volume of glass beads was added to this suspension and cells were broken by two bursts of vigorous vortexing for 3 minutes at 4°C. The glass beads and cell debris were removed by centrifugation for 5 minutes in a microfuge. The supernatant was used for the different assays.
The affinity precipitation of Cdk1 was done by incubating 20 μl of Suc1-Sepharose beads (Calbiochem) with total protein extracts (50-150 μg) in HB buffer for 2 hours at 4°C with gentle agitation as described previously (Surana et al., 1993). The beads were then collected by centrifugation and washed six times with 1 ml of HB buffer.
For co-immunoprecipitation analysis, approximately 3.5 mg of total protein extracts were incubated with 1 μg of the monoclonal anti-myc 9E10 or anti-VSV-G (Roche Diagnostics Gmb) for 1 hour on ice, and then protein G-Sepharose beads (Pharmacia-Biotech) were added and incubated for 30 minutes at 4°C with agitation. Immunoprecipitates were washed six times with 1 ml of HB buffer.
For the kinase reaction, protein precipitates were incubated at 25°C for 10 minutes in KIN buffer (2 mg/ml histone H1, 200 μM [γ-32P]ATP, 100 cpm/pmol in 25 mM Mops pH 7.2). The reaction was terminated by adding 5 μl of 5× Laemmli buffer, and then boiled for 3 minutes and loaded onto a 12.5% SDS-PAGE gel. Phosphorylated histone H1 was visualized by autoradiography.
Western analysis was performed on total protein extracts or the protein precipitates (50-100 μg), separated on 8-10% SDS-PAGE. Anti-PSTAIRE (Santa Cruz Biotechnology), anti-FLAG (M2; Kodak), anti-myc 9E10 (Roche Diagnostics Gmb) and anti-VSV-G (Roche Diagnostics GmbH) antibodies were used at 1:10000 dilution in phosphate-buffered saline +0.1% Tween + 10% dry milk. Anti-mouse-Ig-horseradish peroxidase and anti-rabbit-Ig-horseradish peroxidase (Roche Diagnostics Gmb) were used as a secondary antibody, at 1:10000 dilution. All western analyses were visualized using enhanced chemiluminescence (Renaissance®, Perkin Elmer).
Total RNA was isolated as described previously (Garrido and Pérez-Martín, 2003). A 750 bp fragment spanning nucleotides +271 to +1021 (considering the adenine in the ATG as nucleotide +1) of the clb1 gene and an 800 bp fragment that spans nucleotides +699 to +1499 of the clb2 gene were used as a probe. A 5′-end labeled oligonucleotide complementary to the U. maydis 18S rRNA (Bottin et al., 1996) was used as loading control in northern analyses. A phosphorimager (Molecular Imager FX, Bio-Rad) and the program Quantity One (Bio-Rad) were used for visualization and quantification of radioactive signals.
Flow cytometry, light microscopy and image processing
Flow cytometry was performed as described previously (Garrido and Pérez-Martín, 2003). Nuclear staining was done using DAPI as described previously (Garrido and Pérez-Martín, 2003). WGA and calcofluor staining was performed as described previously (Wedlich-Söldner et al., 2000). Microscopy analysis was performed using a Zeiss Axiophot microscope. Frames were taken with a cooled CCD camera (Hamamatsu C4742-95). Epifluorescence was observed using standard FITC and DAPI filter sets. Cfp (cyan fluorescent protein) was analyzed with specific filter set (BP436, FT455, BP480-500). Image processing was performed with Image Pro Plus (Media Cibernetics) and Photoshop (Adobe).
Mating and plant infection
To test for mating, compatible strains were co-spotted on charcoal-containing PD plates (Holliday, 1974), which were sealed with Parafilm and incubated at 21°C for 48 hours.
Plant infections were performed as described previously (Gillissen et al., 1992) with the maize cultivar Early Golden bantam (Old Seeds, Madison, WI, USA). Filaments inside the plant tissue were stained with chlorazole black E as described previously (Brachmann et al., 2003).
Protein sequences of fungal CDKs and cyclins were downloaded from PubMed (http://www.ncbi.nlm.nih.gov/entrez/query.fcgi) and aligned in ClustalX (Thompson et al., 1997). Phylogenetic dendrograms were constructed using the minimal evolution method with a nearest neighbor joining tree as starting point and 500 Bootstrap replicates.
The cdk1 gene encodes the mitotic cyclin-dependent kinase in U. maydis
To identify genes encoding Cdc2 homologues in U. maydis, a PCR-based approach was adopted using degenerate primers that had been designed on the basis of conserved regions of CDK catalytic subunits from other fungi, as described in Materials and Methods. Three different PCR products were isolated and their sequences used to identify the full-length ORF by a PCR walking approach. After conceptual translation of the isolated ORFs, only one of the derived amino acid sequences showed the so-called PSTAIRE motif (Pines and Hunter, 1991) characteristic of functional Cdc2 homologues. The genomic sequence of this ORF was designated cdk1 on the basis of the phenotypic characterization show below. It consisted of 1115 bp, which encoded a putative protein of 298 amino acids (accession number AY260971) with a calculated molecular mass of 34 kDa. Comparison of genomic and cDNA indicated that cdk1 contains two introns of 121 bp and 100 bp length that are located 37 and 324 bp downstream of the start codon. Cdk1 shares 67% amino acid identity with Cdc2 from S. pombe, 66% with S. cerevisiae Cdc28, and 69% with Candida albicans Cdc28 (Fig. 1B). The Cdk1 protein contains all of the motifs characteristic of the Cdc2 homologues (Fig. 1A): (i) the above mentioned, PSTAIRE motif which has been implicated in the binding of the regulatory subunit, i.e. cyclin, to the CDK (Jeffrey et al., 1995); (ii) the specific threonine residue within the T-loop (at position 167 in U. maydis Cdk1) that is recognized by CDK-activating kinase (Fleig and Gould, 1991); (iii) the conserved tyrosine residue (Tyr 15) to be phosphorylated by the inhibitory Wee1 kinase (Fleig and Gould, 1991).
A cell lysate of a U. maydis wild-type strain separated by SDS-PAGE and probed with an anti-PSTAIRE antibody revealed the presence of two bands that migrated around 34-kDa (Fig. 1C). To determine which one of these cross-reactive band represented the Cdk1 protein, the chromosomal cdk1 gene was replaced by the cdk1-1 allele, which produces a Cdk1 protein that carries a copy of the FLAG epitope at the C terminus. An extract of the strain carrying the cdk1-1 allele probed with an anti-PSTAIRE antibody showed that one of the bands shifted to a lower electrophoretic mobility (because of the size increase caused by the FLAG tag) with concomitant staining with the anti-FLAG antibody (Fig. 1C).
The S. pombe protein Suc1 is known to bind specifically to mitotic CDKs with high affinity (Ducommun and Beach, 1990). Consequently, Suc1 binding would provide a convenient assay for characterizing the U. maydis Cdk1 protein. Cell lysates from both a wild-type strain and the strain carrying the cdk1-1 allele were then incubated with Suc1-conjugated Sepharose beads, and the precipitates were separated by SDS-PAGE followed by immunoblotting with anti-FLAG or anti-PSTAIRE antibodies (Fig. 1D). From total extracts, Suc1 precipitated only one of the two polypeptides recognized by the anti-PSTAIRE antibody, and in the tagged strain, the anti-FLAG antibody detected the Suc1-bound protein. Controls were made in which protein extracts of wild-type and tagged strains were incubated with Sepharose beads, in order to detect non-specific interaction between Cdk1 and the matrix. No Cdk1 was recovered in the bound fraction in these assays (not shown).
To determine the specific kinase activity associated with Cdk1, cell lysates of the tagged strain were then incubated with an anti-FLAG M2 affinity gel and the immunoprecipitates were analyzed by western blot assays with anti-PSTAIRE antibodies (Fig. 1E) or assayed for histone H1 kinase activity (Fig. 1F). Cell lysates from the wild-type strain were used as a negative control. The precipitated complex was able to phosphorylate histone H1 quite efficiently, while the negative control showed no kinase activity (Fig. 1F). The kinase activity correlates with the presence of a band detected by the anti-PSTAIRE antibody in the immunoprecipitate (Fig. 1E).
These lines of evidence strongly suggest that the Suc1-associated, and anti-PSTAIRE-recognized 34-kDa protein is likely to be the catalytic subunit of the U. maydis mitotic CDK. To prove that indeed the cdk1 gene encodes a protein essential to growth, we inactivated one of the two cdk1 alleles by replacement with a hygromycin B resistance cassette (giving the cdk1Δ::hph null allele) in the diploid strain FBD11, and after sporulation we analyzed the meiotic progeny. No viable haploid hygromycin B resistant cells were obtained in the meiotic products indicating that the cdk1 gene encodes an essential mitotic CDK in U. maydis.
The clb1 and clb2 genes encode B-type cyclins
To isolate genes encoding B-type cyclins from U. maydis, a degenerate PCR approach using oligonucleotide primers corresponding to conserved regions of fungal B-type cyclins was utilized. Two different fragments were recovered, and the respective full-length genes were isolated as described in Materials and Methods. The genomic sequences of these ORFs were designated clb1 and clb2, and they encoded putative proteins of 707 and 539 amino acids respectively (accession numbers AY260969 and AY260970). Comparison of genomic DNA and cDNA indicated that both genes were intronless (not shown). The conceptual translation of clb1 and clb2, indicated that the predicted proteins contained typical motives shared among all members of the B-type cyclin family, such as the FLRRXSK motif-less conserved in Clb2- (Fitch et al., 1992) and the `destruction box' motif, which is involved in the degradation of cyclins mediated by the APC/cyclosome (Glotzer et al., 1991) (Fig. 2A). Sequence comparison indicates an overall 30% sequence identity between U. maydis cyclins and B-type cyclins from other fungi (not shown). A dendrogram analysis (Fig. 2B) indicates that the two U. maydis proteins fall in two different branches. In one branch, U. maydis Clb1 is grouped along with the S. pombe cyclins Cig2 and Cdc13, which are involved in the G1 to S and G2 to M transitions, respectively (Fisher and Nurse, 1995). The second branch grouped U. maydis Clb2 protein with the S. cerevisiae Clb3, the S. pombe Cig1 and C. albicans Cyb4 proteins. From these cyclins only a clear role has been assigned to Clb3, which is a G2/M phase cyclin (Fitch et al., 1992; Richardson et al., 1992).
For clb1 and clb2 to qualify as the genes encoding true B-type cyclins, their protein products must fulfill at least two criteria: they must associate with the catalytic subunit Cdk1 to form the mitotic CDK, and their levels must fluctuate during the cell cycle, falling as cells enter G1 and rising again as cells enter S/G2/M. To examine these requisites, we exchanged the chromosomal locus of clb1 and clb2 with the functional clb1-1 and clb2-1 alleles, encoding VSV- and MYC-tagged versions of Clb1 and Clb2, respectively. To determine whether the levels of Clb1 and Clb2 fluctuate through the cell cycle, because no reproducible synchronization method is available so far in U. maydis, we used cell cultures arrested in different cell cycle stages. Protein extracts were prepared from cultures of the tagged strains enriched in G1 phase or arrested in the presence of hydroxyurea (S phase) and benomyl (G2/M phase). Clb1 and Clb2 protein levels were measured by western blotting with the use of anti-VSV and anti-MYC antibodies (Fig. 2C). Low amounts of Clb1 and Clb2 proteins were detected in cells enriched in G1 phase, while they were clearly present in cells arrested either in S phase or G2/M phase.
To show a physical association between the B-type cyclins and Cdk1, we used a double approach. Firstly, cell lysates of the tagged strains were incubated with Suc1-beads that bind specifically to CDKs, and the precipitates were separated by SDS-PAGE followed by immunoblotting with anti-VSV and anti-MYC antibodies. As shown in Fig. 2D,E (upper panels), both Clb1 and Clb2 were recovered in the fraction of proteins bound to Suc1-beads. Secondly, protein extracts from both tagged and wild-type cells were prepared, and VSV- and MYC-tagged proteins were immunoprecipitated and analysed for the presence of Cdk1 (Fig. 2D,E, lower panels). High levels of Cdk1 co-precipitated with Clb1-VSV and Clb2-MYC, while none was detected in precipitates of untagged extracts. Taken together, these results indicate that the clb1 and clb2 genes encode B-type cyclins in U. maydis.
Conditional removal of cyclin function
We tried unsuccessfully to delete the clb1 and clb2 genes in haploid cells, whilst we were successful in the removal of a single copy of each gene in diploid strains (not shown). Since these results strongly suggest that clb1 and clb2 represent essential genes, we constructed strains in which the cyclin function could be conditionally controlled. To achieve conditional expression of cyclin genes, the U. maydis Pnar1 promoter, which is induced by nitrate and repressed by ammonium (Banks et al., 1993; Brachmann et al., 2001), was used. Two chimeric alleles were constructed composed of the Pnar1 promoter fused to the coding region of clb1 (clb1nar) and the coding region of clb2 (clb2nar). In both cases, the respective native allele was replaced by the conditional allele.
In the clb1 conditional strain, TAU41, the clb1 mRNA was elevated around tenfold with respect to wild-type in minimal medium containing nitrate (MM-NO3), and it was not detectable in minimal medium supplemented with ammonium (MM-NH4) and rich medium YPD (Fig. 3A). TAU41 cells grew on solid minimal medium containing nitrate at a rate similar to control wild-type cells, but they were unable to form colonies when shifted to solid minimal medium containing ammonium or YPD (Fig. 3B), indicating an essential role of the Clb1 function.
In the clb2 conditional strain, TAU42, the clb2 mRNA was elevated around fourfold with respect to wild-type in permissive conditions (MM-NO3), whereas it was decreased fivefold in MM supplemented with ammonium (MM-NH4) and it was barely detectable in YPD medium (Fig. 3A). Consistent with this pattern of expression, the clb2 conditional strain showed no difference in colony growth with respect to a wild-type strain in minimal medium containing nitrate or ammonium, and was clearly defective in growth in YPD medium (Fig. 3B). The differences of transcriptional repression in non-permissive conditions between the clb1 and clb2 conditional alleles could be explained by the different tightness in the control of the Pnar promoter depending on the chromosomal context (our unpublished observations).
We analyzed the growth of the conditional strains in liquid medium by following the duplication time (by measuring the colony forming unit increase) and DNA content (by FACS analysis) of the cells during at least three generations after transfer from conditional medium (MM-NO3) to tester medium. U. maydis FB1 cells had a duplication time of 260, 210 and 90 minutes growing in MM-NO3, MM-NH4 and YPD, respectively. More than 80% of the wild-type cells growing in MM-NO3 had a 1C DNA content, indicating that they were passing through G1 phase. In MM-NH4 medium, this proportion decreased to less than 40% and in rich medium it was only around 10% of the cells (Fig. 3C).
For TAU41 strain, we observed that in permissive conditions (MM-NO3) the duplication time and DNA content profile of the population was similar to the wild-type strain. However, the cells stop dividing in approximately two generations after the transfer to restrictive conditions (MM-NH4 or YPD). This cell duplication arrest is concomitant with the accumulation of cells, of which approx. 50% have 1C and 50% have 2C DNA content (Fig. 3C).
The clb2 conditional cells growing in MM-NO3 had a slight decrease in the duplication time (around 240 minutes) and the proportion of cells passing through G1 decreases to 50% of the population. No differences were apparent in cells growing in MM-NH4, while after the transfer of cultures to rich medium, the cells stop dividing in three generations, with the cell population having a 2C DNA content (Fig. 3C).
Clb1 is required to enter in S and M phase
The DNA content analysis of the arrested clb1 conditional strain suggested that Clb1 is required at different cell cycle stages. To characterize the phenotype of conditional cells, we introduced the clb1 conditional allele in a FB1 strain carrying a Cfp-α tubulin fusion (Wedlich-Söldner et al., 2002), generating the UMP25 strain which showed all the previously described phenotypes of the conditional TAU41 strain (not shown). UMP25 cells incubated in permissive conditions for 9 hours were morphologically wild type (not shown). However, after 6 hours of incubation in restrictive conditions (rich medium), the conditional strain arrested with the cells showing two alternative morphologies (Fig. 4A): around 47% of the cell population was composed of unbudded cells, while the rest of the cells showed a developed bud that frequently (25% of total cells) was larger than the mother cell. In contrast, in wild-type cells growing in the same conditions, different stages of bud formation can be found (Fig. 4A, left panel). All arrested conditional cells contained a single nucleus (Fig. 4A). Measurement of the relative fluorescence of the nuclei by microphotometric methods (Snetselaar and McCann, 1997) indicated that fluorescence in budded cells were usually about twice as intense as that in unbudded cells (not shown). These data strongly suggest that Clb1 depletion resulted in a cell cycle arrest at two different points: a pre-replicative arrest (unbudded cells with 1C DNA content) and a post-replicative arrest (budded cells with 2C DNA content). Of note, in budded arrested cells, the nucleus could be found either in the mother, neck or bud compartments (Fig. 4A). Because in U. maydis there is a migration of the nucleus to the bud compartment prior to mitosis (Holliday, 1974; O'Donnell and McLaughlin, 1984; Snetselaar, 1993; Banuett and Herskowitz, 2002), we investigated the organization of microtubules to define which kind of post-replicative arrest is taking place after Clb1 depletion. In U. maydis the microtubule cytoskeleton changes from a G2 cytoplasmic array, which is characterized by three to four bundles of microtubules that point towards the growth region, to a mitotic spindle that elongates after the metaphase-anaphase transition (Steinberg et al., 2001; Banuett and Herskowitz, 2002). After shift to restrictive conditions, conditional cells that made large buds contained long microtubules bundles that extend to the apical growth region (Fig. 4B), supporting a G2 arrest. Taking together all these lines of evidence, we conclude that Clb1 is required both for G1 to S and G2 to M transitions.
In the course of these experiments, we observed that more than 24 hours incubation of conditional cells in rich medium resulted in a bizarre phenotype consisting of large polarized cells (more than 60 μm in length), with an altered MT cytoskeleton, which formed basal vacuoles and had, posteriorly, empty sections that were separated by septa (Fig. 4C). Although we have no clear explanation for this behavior, we suspect that this could be the usual response in U. maydis to a continuous polar growth in a situation of arrested cell division. For instance, these cells are reminiscent of the growth mode of the dikaryotic hyphae produced after a successful mating, conditions in which a cell division arrest and a permanent polarized growth are present (Steinberg et al., 1998).
Clb2 is required to enter in mitosis
The clb2 conditional cells arrested in rich medium with a 2C DNA content, suggesting a post-replicative arrest. As before, to characterize the phenotype of arrested cells, we introduced the clb2 conditional allele in cells carrying a Cfp-α tubulin fusion. We utilized this new strain, called UMP26, to study the differences between the conditional mutant and wild-type cells in the various growth conditions. The most obvious difference between the conditional strain and wild-type cells growing in liquid medium was the appearance in the mutant culture of a subpopulation of cells in which the bud was elongated to the extent that it was larger than the mother cell. This subpopulation was around 25% of the conditional cells growing in MM-NH4, and more than 60% in rich medium (Fig. 5A). A detailed analysis of the elongated cells arrested in rich medium, indicated that they carried a single nucleus (Fig. 5C) with a MT cytoskeleton that was consistent with G2 phase (Fig. 5B2).
One plausible interpretation of the above results could be that Clb2 levels are responsible for marking the end of G2 phase and the beginning of mitosis. In the conditional strain growing in restrictive medium, the absence of clb2 expression resulted in the inability of the cells to enter mitosis and the unlimited growth of the bud. Should this interpretation be correct, an increase in the levels of Clb2 would result in a shorter G2 phase. To examine this hypothesis, we inserted in a wild-type strain an ectopic copy of the clb2 coding sequence under the control of the Pcrg1 promoter (Brachmann et al., 2001). The expression of this transgene was repressed in glucose-containing medium and induced in arabinose-containing medium, leading to more than 30-fold increase in the levels of Clb2 after 3 hours incubation in complete medium containing arabinose (not shown). In these conditions, overexpression of clb2 resulted in a dramatic phenotype. The cells lost their typical budding pattern and acquired an elongated shape. Each cell was divided by septation into compartments that each contained a single nucleus (Fig. 6A,B). Each compartment was smaller than a normal U. maydis cell (average of less than 10 μm in length, versus 15-20 μm in wild-type cells). FACS analysis indicated that these cells accumulated DNA in genome sized multiples, consistent with a normal DNA replication (Fig. 6B). Our interpretation of such as hyphal-like growth is that the high levels of Clb2 promoted a premature entry into mitosis resulting in the inability to produce the bud and then a division by septation. Because of the dramatic change in morphology observed, we wondered whether the cells were able to support a constitutive high level of clb2 expression. We have introduced in wild-type cells an ectopic copy of the clb2 gene cloned under the control of the strong constitutive hsp70 promoter (which produces around a 40-fold increase in the levels of Clb2; not shown). The resulting strain, TAU26, was viable, but it grew in a filamentous way in liquid medium (Fig. 6C). Taken together, these results indicate a correlation between the bud formation, the length of the G2 phase and the levels of Clb2, supporting the conclusion that the levels of Clb2 could mark the length of G2 and the onset of mitosis.
Overexpression of Clb1 alters chromosomal segregation
In addition to Clb2, the transition from G2 to M also requires the presence of Clb1. To evaluate if Clb1 could also be involved in the control of the length of G2 in U. maydis, we examined the consequences of the overexpression of clb1. To do this, we introduced into a wild-type strain an ectopic copy of clb1 under the control of the Pcrg promoter. In this strain, growth in arabinose-containing medium gave a 50-fold increase in the levels of Clb1 with respect to a wild-type strain (not shown). We observed that such high levels of Clb1 correlated, in liquid cultures, with cells of different sizes that showed altered morphology (Fig. 7A). There were differences in intensity of DAPI staining between cells in these cultures, suggesting differences in DNA content per cell (Fig. 7A). Consistently, FACS analysis of the clb1-overexpressing cultures indicated the presence of cells with a DNA content lower than 1C as well as a population of cells with a DNA content higher than 2C (Fig. 7B). Furthermore, clb1-overexpressing cells growing in solid medium accumulated phloxine B (a vital stain excluded by living cells) (Fig. 7C), indicating that high levels of Clb1 protein resulted in a loss of viability.
S. cerevisiae cells overexpressing cyclin Clb5 showed an altered chromosome segregation (Sarafan-Vasseur et al., 2002) and mammalian cells overexpressing cyclin E and cyclin B1 also show defects in the segregation of the chromosomes and aneuploidy (Spruck et al., 1999; Yin et al., 2001). Taking into account these reports, a plausible interpretation for our results is that high levels of Clb1 interfere with a normal chromosomal segregation, resulting in aneuplody. In S. pombe, it has been reported that high levels of Cdc13 resulted in disassembly of microtubules (Yamano et al., 1996). In line with this argument, we found that U. maydis cells overexpressing clb1 were hypersensitive to the microtubule-destabilizing drug benomyl (Fig. 7D). It is plausible to assume that, in U. maydis, high levels of Clb1 interfere with the microtubule assembly, providing an explanation for the observed DNA segregation defects.
Expression of indestructible B-cyclins interferes with progression through mitosis in U. maydis
The previous results highlight the importance of a strict control in the level of mitotic cyclins in U. maydis. One of the ways that eukaryotic cells control the levels of mitotic cyclins is regulated proteolysis mediated by APC/C (Zachariae and Nasmyth, 1999). Cyclin destruction is thought to be mediated by a conserved motif, the destruction box (D-box). U. maydis Clb1 has two putative D-boxes as judged by inspection of its sequence (Fig. 8A). One of them is located at the N terminus while the second one is located at the very C terminus. Clb2 has a single putative D-box in its N terminus (Fig. 8A). We constructed truncated versions of both Clb1 and Clb2, in which the various D-boxes were removed (Fig. 8A). These mutant, as well as the full-length versions, were tagged at the C terminus either with the VSV epitope (Clb1 derivatives) or the MYC epitope (Clb2 derivatives) and they were cloned under the control of the Pcrg1* promoter (Brachmann et al., 2001). This promoter is a less-active version of the Pcrg promoter that in induction conditions gives only a 10-fold and a 7-fold increase in the levels of Clb1 and Clb2 proteins, respectively, compared to wild-type levels (not shown). The use of this promoter avoids the harmful effects of hyperexpression of clb1 or clb2, reported in previous sections (not shown). We observed that expression of Clb1 derivatives lacking a single D-box had no effect on the ability of the cells to form colonies. However, the expression of the truncated clb1 allele lacking the two D-boxes, clb1Δdb1-2, or the expression of the mutant clb2 allele lacking the D-box, clb2Δdb, prevented colony formation (Fig. 8C). This effect is consistent with previous reports of expression of cyclins lacking D-boxes in other yeast systems and it has been attributed to the inability of the mutant cyclins to be degraded by the proteasome, which results in a cell cycle arrest (Ghiara et al., 1991; Surana et al., 1993; Yamano et al., 1996; Yamano et al., 2000). To examine this notion further, first we sought to demonstrate that the U. maydis cyclin variants devoid of D-boxes were more stable than their wild-type counterparts. For this, cells carrying either the wild-type or the mutant alleles under the control of the Pcrg1* promoter were incubated in arabinose-containing medium to allow the expression of the proteins, and after 2 hours of incubation, glucose and cycloheximide were added to shut-off the production of further protein. Samples were removed at different times for western analysis (Fig. 8D). We observed that while Clb1 and Clb2 protein levels declined, the levels of the mutant versions lacking the D-boxes, Clb1Δdb1-2 and Clb2Δdb remain high and almost constant, consistent with more stable proteins.
To assess whether the expression of the stabilized versions of mitotic cyclins induce some specific cell cycle arrest, we introduced the constructions expressing the mutant alleles in a strain carrying a Cfp-α tubulin fusion, and the cells growing in induction conditions were analyzed (Figs 9 and 10). Liquid cultures of the strain expressing the clb1Δdb1-2 allele were composed of budded cells in which three different nuclear morphologies were apparent. Around 60% of the cells carried a single condensed nucleus located near the neck; in 14% of the cells the nucleus appeared split in two halves and located in the neck, and 26% of the cells had two condensed nuclei located not far apart (Fig. 9A). The FACS analysis indicated that all cells had a 2C DNA content (Fig. 9B). Of note, in spite to the fact that cells growing in non-inducing conditions gave a clear microtubule pattern, the incubation in inducing conditions results in cells where no clear microtubule structures were apparent, indicating that the spindles are somewhat unstable in the presence of the mutant clb1 allele (Fig. 9C). A probable explanation of these phenotypes is that the stabilized mutant version of Clb1 interferes with microtubule assembly, impairing the proper execution of mitosis, and resulting in a cell cycle arrest.
The cells overexpressing the stable clb2Δdb allele also arrested with a 2C DNA content (Fig. 10A). However, the morphology was different from cells expressing indestructible Clb1. The arrested cells showed no buds, although they had two well separated nuclei (Fig. 10B). This morphology is similar to the first stages observed when clb2 is overexpressed (see Fig. 6A), and supports the notion that high levels of Clb2 impaired the budding process (most probably because it induces a premature entry into mitosis). However, in contrast to the overexpression of wild-type clb2, no septa were formed between nuclei (Fig. 10C), and microtubule spindles were present (Fig. 10D). Our interpretation of these results is that a failure in Clb2 degradation results in a cell cycle arrest after anaphase, with the cell unable to exit mitosis.
Progression through the infective process requires an accurate control of Clb2 levels
The growth of the fungus inside the plant involves distinct steps that have been described from microscopic studies (Snetselaar and Mims, 1992; Banuett and Herskowitz, 1996). After fusion of haploid fungal cells and penetration inside the plant, the fungal hypha continues to grow as an unbranched tip cell leaving highly vacuolated portions behind. As the infection progress, the hypha proliferates and septa partition cell compartments, containing a pair of nuclei each. Polar growth continues along the main hypha that begins to form clamp-like structures and branches. Finally, the hyphae undergo fragmentation to release individual cells that will produce the diploid spore (teliospore). During all this process, the host plant responds with the symptoms of infection. Induction of chlorosis is known to be one of the earliest detectable symptoms of U. maydis infection (Christenesen, 1963). It is followed by anthocyanin production, a known stress reaction in plants. Finally, the hyperproliferation of plant cells in response to the fungus produces the typical plant tumors. Given the connections between morphogenesis and pathogenesis in U. maydis, we were interested in determining whether TAU26 cells that produce constitutively high levels of Clb2 protein and had in a hyphal-like growth (Fig. 6C), were able to cause disease symptoms upon inoculation into corn seedlings. We mixed TAU26 cells with the wild-type compatible haploid strain FB2 and we inoculated plants to test the virulence of the dikaryon produced upon fusion. These inoculations included positive controls of crosses between wild-type strains (FB1 and FB2). As indicated in Table 3, disease symptoms were observed on more than 90% of the plants inoculated with the compatible wild-type strains FB1 and FB2. In contrast, of the 67 plants inoculated with the combination of TAU26 and FB2, 66% showed no observable symptoms and the remaining 44% demonstrated only mild symptoms. The latter plants showed mostly chlorosis emanating from the point of inoculation (Fig. 11A). Interestingly, anthocyanin production was never observed in these plants. To asses at which level TAU26 cells are defective, we first tested the ability of TAU26 cells to mate in charcoal mating plates (Holliday, 1974). The visible white filamentous growth indicative of the dikaryon was slightly reduced in the mutant mixture (Fig. 11B). We analyzed the dikaryotic hyphae and observed both in wild-type and mutant mixtures the typical filament where hyphal extension occurs in the absence of mitosis and a fixed volume of cytoplasm migrates forward with the expanding apex, leaving behind an extensively vacuolated and otherwise empty distal cell compartment (Steinberg et al., 1998) (Fig. 11C,D). The only apparent difference is the `spiky' aspect of the mutant hypha. These results indicate that high levels of Clb2 decreased but did not negate dikaryotic formation. The observation of chlorosis around the site of inoculation indeed suggested that the mutant filaments were capable of growing within the plant tissue. Consequently, symptomatic leaves were sampled, stained and examined microscopically for the presence of the fungus (Fig. 11E,F). Fungal growth was clearly present in maize tissue in plants inoculated with either wild-type or mutant strains. The most obvious difference between mutant and wild-type filaments was the absence of teliospores in mutant mixtures, indicating that mutants failed to complete sexual development. In addition, we noted that the frequency of incipient branches appeared to be higher for filaments in plants infected with the mutant mixture. These results indicate that an extra dose of Clb2 impairs the ability of fungal cells to cause symptoms and progress in the infective cycle.
In order to gain further evidences about the role of Clb2 in virulence, we analyzed the ability to cause disease symptoms in plants of a diploid strain carrying a single allele of the clb2 gene (UMP32). Previously, we have observed that the UMP32 cells were larger than wild-type cells, with elongated buds that sometimes carried more than one nucleus (Fig. 12A). We considered this phenotype as a consequence of the lower levels of Clb2 in these cells, which could result in a delay in mitosis – as we showed in haploid clb2 conditional strains incubated in restrictive conditions (Fig. 5B1). This haploinsufficiency was specific for clb2, since a diploid strain carrying a single copy of the clb1 gene (UMP21), had a wild-type phenotype (not shown). First, we have tested for colony morphology on charcoal mating plates. The filamentous growth phenotype that arises from mating between haploid strains is also displayed by diploid strains of U. maydis that are heterozygous for both the a and b mating-type loci. As shown in Fig. 12B, filamentous colonies are formed by the wild-type diploid strain FBD11 as well as by the mutant diploid strains UMP21 and UMP32, indicating that half a dose of clb1 or clb2 is not detrimental at this step. Second, we inoculated plants with cultures of FBD11, UMP21 and UMP32 cells. As shown in Table 3, the strain carrying a single clb1 allele was as virulent as the diploid wild-type control. In contrast, we found that diploid cells defective in one clb2 allele were capable of causing chlorosis and anthocyanin production but tumors were never observed. Interestingly, the anthocyanin production was even stronger in plants infected with the UMP32 strain that in plants infected with FBD11 strain. We also studied the fungal growth inside the plant tissue. The hyphae of the UMP21 cells were similar to those of wild-type cells. In contrast, several differences were observed in the UMP32 hyphae: they were wider, less septate and they showed fewer branches than wild-type hyphae. No teliospores were present in the UMP32-infected plants (not shown).
Taken together, these results strongly suggest that Clb2 levels in U. maydis must be accurately controlled during the infection process, and that inability to do so blocks the complete sexual development.
In the current study, we have characterized the genes encoding the CDK catalytic subunit, Cdk1 and two regulatory subunits, the B-type cyclins Clb1 and Clb2 in the phytopathogenic fungus U. maydis. The cdk1 gene is essential and encodes a protein with high sequence similarity to other members of the Cdc2 protein family. The clb1 and clb2 genes encode B-type cyclins. Analysis of their coding sequences indicates typical sequence signatures of B-type cyclins. Both proteins interact with Cdk1 and their levels change in cells arrested at different cell cycle stages. A search performed in the U. maydis genome indicated that no other B-type cyclins were present (http://www.broad.mit.edu/annotation/fungi/ustilago_maydis/index.html). This situation is in contrast to other well-know fungi such as S. cerevisiae or S. pombe, where there is a certain degree of redundancy in cyclin content and in the roles these cyclins play. However, the situation in U. maydis is likely to be more general. For instance, the human pathogen Candida albicans appears to have only two B-type cyclin (J. Berman, personal communication).
Clb1 plays an essential role in the cell and is required to perform the G1 to S and the G2 to M transitions. In S. pombe, the G1 to S and G2 to M transitions are performed by Cig2 and Cdc13, respectively (Booher and Beach, 1987; Hagan et al., 1988; Bueno and Russell, 1993; Connolly and Beach, 1994), although in S. pombe cells deleted for the cig2 gene, Cdc13 substitutes for Cig2 to bring about the G1 to S transition (Fisher and Nurse, 1996; Mondesert et al., 1996). Interestingly, in a dendrogram analysis of B-type cyclins from different fungi, U. maydis Clb1 was clustered along with the Cig2 and Cdc13 cyclins. How does Cdk1-Clb1 kinase first promote G1 to S transition early in the cell cycle and then prevent the reinitiation during G2? A hypothesis that seems reasonable is to assume that additional elements, specific for each transition, modify the scope of target molecules of this complex. Clb2 appears to be specific for G2/M transition and it seems to be a rate-limiting regulator for entry into mitosis. Cells carrying a conditional clb2 allele growing in restrictive conditions were arrested in G2 phase with an elongated bud displaying an active polarized growth. In contrast, high levels of Clb2 expression resulted in short cells that divided by septation, producing a hyphal-like growth. In U. maydis, the G2 phase is characterized by polar growth that results in the elongation of the bud. It could well be that the levels of Clb2 mark the length of G2 and then the size of the bud. This could be related with a G2/M size control operating through the levels of Cdk1-Clb2. The postulated role of Clb2 is reminiscent of the roles proposed for cyclin A in humans (Furuno et al., 1999) and cyclin B2 in plants (Weingartner et al., 2003) as promoters of mitosis. In human cells it has been demonstrated that cyclin A/CDK2 activity is a rate-limiting component for entry into mitosis because exogenous active cyclin A/CDK2 will drive cells prematurely into mitosis (Furuno et al., 1999). In plants, induction of ectopic cyclin B2 expression drives cells to enter mitosis earlier, causing developmental abnormalities in transgenic plants (Weingartner et al., 2003). We do not yet have a clear view of how Clb1 and Clb2 cooperate in the G2/M transition. For instance, Clb2 could regulate Clb1 activity in G2 or both cyclins could act in different targets to jointly induce mitosis. Because of the different phenotypes observed in cells overexpressing either clb1 or clb2, we favored the latter hypothesis.
The Clb1 and Clb2 mitotic cyclins contain typical destruction boxes. By deletion analysis we demonstrated that they are functional. The inability to down-regulate the cyclin levels properly resulted in cell cycle arrest. We have shown that cells expressing a mutant version of Clb2 lacking the destruction box are unable to exit mitosis and arrest after anaphase, as has been demonstrated with stabilized versions of mitotic cyclins in other organisms (Surana et al., 1993; Yamano et al., 2000). The expression of the clb1Δdb1-2 allele, in contrast, produces a cell cycle arrest at some point between metaphase and anaphase, with cells unable to assemble spindles. We cannot determine whether the inability to proceed through mitosis in these cells is a consequence of the microtubule disassembly or whether the cell cycle arrest promotes such disorganization. However, we favored the first explanation. In S. pombe it has been reported that high levels of Cdc13 resulted in disassembly of microtubules (Yamano et al., 1996). Furthermore, consistent with the idea that in U. maydis high levels of Clb1 interfere with microtubule assembly, we showed that overexpression of clb1 resulted in cells that were hypersensitive to the microtubule inhibitor benomyl and had anomalous DNA content. A general feature observed in basidiomycete yeast is the presence of an extensive microtubule network in G2 that disassembles prior to mitosis and then forms the mitotic spindle (Steinberg et al., 2001; Kopecká et al., 2001; Banuett and Herskowitz, 2002). We entertained the idea that an initial step to enter in mitosis could be the premitotic activation of Cdk1-Clb1 that could induce the disassembly of the cytoplasmic microtubule network. Such a role of Cdk1-Clb1 in the induction of mitosis could help to explain the fact that in clb1 conditional cells incubated in restrictive conditions, when the bud is present, the nucleus migrates to the daughter compartment (a pre-mitosis step) but the microtubule cytoskeleton still keeps a typical G2 appearance. In this line of argument, it could be possible that the posterior inhibition of Clb1-Cdk1 activity (i.e. by proteolysis of Clb1) allows the assembly of microtubules to form the mitotic spindles. If this is the case, the presence of an indestructible version of Clb1 could preclude the spindle assembly, as we observed.
Taking all the above conclusions into account we suggest a working model of how the cell cycle proceeds in U. maydis. In our model, the G1 to S transition requires Cdk1-Clb1 activity. Whether Cdk1-Clb1 is the only requirement or additional elements (i.e. G1 cyclins) are needed to induce the G1/S transition remains to be clarified. Once DNA replication ends, the formation of the bud marks the beginning of G2. Once, the proper bud size is reached, mitosis is induced. The onset of mitosis requires both Cdk1-Clb1 and Cdk1-Clb2, with a specific role of Cdk1-Clb2 determining the length of G2 phase. Once mitosis is induced, cyclins must be degraded, most probably by APC/C. Based upon the results with stabilized cyclins, we postulate that Clb1 should be removed earlier in mitosis, whilst Clb2 should be removed later. We believe that this working model will provide a basis where additional regulatory elements, like APC/C, CDK inhibitory elements and other additional cell cycle regulators could be added in future. Isolation of specific cell cycle termosensitive mutants and studies with synchronous cultures could help us come to a better understanding of these issues.
In Ustilago maydis, an accurate control of cell cycle must be required not only in cells growing in vegetative conditions, but also during the infection process. Hyphal growth within the plant is a dynamic process that requires several stages of differentiation and then, pathogenesis, morphogenesis and cell cycle are predicted to be connected in this fungus. Our findings strongly suggest that progression through the different steps in infection requires a careful control of cell cycle. We had found that infection of a plant by fungal cells with high levels of Clb2 produces only symptoms of chlorosis, while infection with fungal cells carrying half a dose of Clb2 induces the production of anthocyanin by the plant, but tumor production does not occur. We believe that the accurate control of cyclin levels could be required not only to adjust the fungal morphogenesis inside the plant, but also to properly induce the developmental program that allows other processes such as plant-fungus signaling or biosynthesis and release of tumor inducing chemicals, for instance. Interestingly, the phenotypes found in clb2 mutants were reminiscent of mutants defective in genes involved in signaling in U. maydis. For instance, in mutants defective in ubc1, which encodes the regulatory subunit of PKA, or in ukb1, which encodes a serine/threonine protein kinase with a role in budding and filamentous growth, the development of the fungus is somehow interrupted in the shift from colonization and ramification to tumor formation (Gold et al., 1997; Abramovitch et al., 2002). It has been proposed that the defect of these signaling mutants in tumor induction could be related to an inability to sense or respond to host stimuli that are required for the production of the signal or signals that U. maydis emits to trigger tumor formation (Abramovitch et al., 2002). It is tempting to speculate that inability to adapt the cell cycle in response to plant signals impairs the execution of the fungal developmental program that induces the tumor formation in the plant. It follows from this speculation that cell cycle control must be an important downstream target in the fungus-plant interaction. Future experiments will address these connections.
We wish to thank Dr Peter Schreier, for BLAST analysis of the U. maydis genome and critical reading of the manuscript. Prof. David W. Holloman is thanked for a critical reading and improvement of the manuscript. We thank Prof. Judith Berman for critical reading of the manuscript. The help of Prof. Jaime Correa-Bordes and Dr Miguel Angel Blanco with the use of Suc1 to isolate CDK is also thanked. The Bayer CropScience AG is acknowledged for providing the genomic sequence of Umclb2. We are grateful to Prof. R. Kahmann and laboratory members for generous supply of reagents and support. This work was supported by Grant BIO99-0906 and BIO2002-03503 from MCyT to J.P.-M. and Grant SP1111 from DFG to GS. This manuscript is dedicated to the memory of Ira Herskowitz.
- Accepted September 17, 2003.
- © The Company of Biologists Limited 2004