Cdk5 kinase regulates the association between adaptor protein Bem1 and GEF Cdc24 in the fungus Ustilago maydis

Polarized cell growth is fundamental for morphogenesis and development of both unicellular and multicellular organisms. In many fungal pathogens, host invasion frequently correlates with the ability to switch between isotropic and polar growth or with changes in the direction of polarity axes (Gow et al., 2002). It is clear that the ability to form polarized hyphae may represent an `Achilles' heel' that can be exploited to limit fungal invasion (Harris, 2006). How these processes are regulated in conjunction with the induction of the virulence program in pathogenic fungi has been the focus of much research recently. We use the corn smut fungus, Ustilago maydis, as a model system for analysis of the connection between polarity regulation and pathogenicity. This dimorphic phytopathogen undergoes strong polar growth as part of its pathogenic development (Bölker, 2001; García-Pedrajas and Gold, 2004; Kahmann and Kämper, 2004). Recently, we described the identification and characterization of a Cdk5/Pho85 homolog in U. maydis, and provided evidence that this kinase has a major role in the maintenance of cell polarity and virulence in this phytopathogen (Castillo-Lluva et al., 2007). Cyclin-dependent kinases from the Cdk5/Pho85 family are thought to play important roles in morphogenesis. Although originally described in yeast (Carroll and O'Shea, 2002; Huang et al., 2007), in mammals Cdk5 represents an important example of a master element, which allows efficient cytoskeleton remodeling in migrating neurons in response to external signals (Xie et al., 2006).

The establishment of polarized growth in fungi is a complex process in which cortical landmarks lead to the local activation of Rho-like GTPases, which then direct the polarization of the actin cytoskeleton as well as the spatial regulation of the secretory apparatus (Pruyne and Brestcher, 2000; Park and Bi, 2007). Work conducted in Saccharomyces cerevisiae showed that cytoskeletal polarization is guided by the distribution of the small GTPase Cdc42 and its guanine-nucleotide exchange factor (GEF) Cdc24 at the plasma membrane (Pruyne and Brestcher, 2000; Wedlich-Soldner and Li, 2004). This complex is maintained at sites of polarized growth by binding to the adaptor protein Bem1, which seems to act as a scaffold protein (Gulli et al., 2000; Butty et al., 2002). Activated Cdc42 transduces signals to multiple downstream effectors, which in turn signal to the actin cytoskeleton (Park and Bi, 2007). A similar scheme has been proposed in U. maydis (Leveleki et al., 2004), although a distinctive feature compared with budding yeast is the presence of a Rac1 homolog in addition to a Cdc42 homolog (Weinzierl et al., 2002). In U. maydis, Rac1 and its activator Cdc24 are involved in the regulation of polar growth (Mahlert et al., 2006). By contrast, Cdc42 together with its activator Don1, is involved in the regulation of cell separation, a complex process that requires the formation of primary and secondary septa (Weinzierl et al., 2002; Mahlert et al., 2006). We reported that in U. maydis the Cdk5 kinase is required for the activation of Rac1 via the accumulation of the GEF Cdc24 at the cell pole, which therefore explains the loss-of-polarity defect in cdk5^ts cells (Castillo-Lluva et al., 2007). Nevertheless, the mechanisms behind the Cdk5-dependent stabilization of Cdc24 at the cell pole in U. maydis are not well understood. Here, in order to investigate the role of Cdk5 during polarized growth, we characterized the U. maydis Bem1 homolog and its interaction with Cdc24. Our results suggest that Bem1 acts as an adaptor protein that helps to stabilize Cdc24 at the cell poles. Interestingly, this stabilization depends on the presence of an active Cdk5 kinase. Our data led us to propose a working model for the involvement of Cdk5 in the regulation of polar growth in U. maydis.

We described previously that cdk5^ts mutant cells growing at restrictive temperature were unable to polarize growth. As this inability correlated with the delocalization of the U. maydis Cdc24 homolog from the cell tip (Castillo-Lluva et al., 2007), we argued that Cdk5 was responsible of the polar localization of Cdc24. However, we found no proof of direct interaction between Cdk5 and Cdc24, indicating either that this delocalization could be a consequence and not the cause of the lack of polar growth; or that the Cdk5 acts on other elements involved in the apical localization of Cdc24.

Cdc24 is an essential protein in U. maydis and localizes at the cell tips; its absence results in cells that are unable to grow in a polar manner (Castillo-Lluva et al., 2007). Besides the domains typically found in Rho-GEFs (Hoffman and Cerione, 2002), U. maydis Cdc24 contains a protein C (PC) domain at its C-terminus (Fig. 1A). Importantly, the PC domain is conserved in S. cerevisiae Cdc24 and S. pombe Scd1 GEFs, and it is required to interact directly with Bem1 and Scd2, respectively, proteins that help to the polar localization of the corresponding GEF (Ito et al., 2001; Butty et al., 2002; Endo et al., 2003). Bearing this sequence conservation in mind, we reasoned that putative adaptor proteins, probably a Bem1 homolog, should also help to localize Cdc24 at the cell tip in U. maydis. To address this issue, a BLAST homology search of the U. maydis genome database (http://www.broad.mit.edu/annotation/fungi/ustilago_maydis/index.html) using the amino acid sequence of S. cerevisiae Bem1 (Bender and Pringle, 1991; Chenevert et al., 1992) and S. pombe Scd2 (Endo et al., 2003) identified an open reading frame (UM05711.1) encoding a protein of 714 amino acids. This protein was the only one in the U. maydis genome that shared significant sequence homology with Bem1 and Scd2 (P values of 3e^–38 and 6e^–73, respectively) and therefore we designated it UmBem1. Strikingly, in spite of the presence of two tandem SH3 domains, the recently described CI domain (Yamaguchi et al., 2007) and a PX (phox homology) domain, UmBem1 lacks a PB1 (phox and Bem1p 1) domain (Fig. 1B). Importantly, Bem1-Cdc24 interaction in S. cerevisiae is mediated by the PB1 domain of Bem1 and the PC1 domain of Cdc24 (Ito et al., 2001; Butty et al., 2002). This kind of interaction is also conserved between the S. pombe scaffold protein Scd2 and the Cdc24-like protein Scd1 (Endo et al., 2003). However, the absence of this domain in UmBem1 does not necessarily preclude the putative interaction with Cdc24, because it could be either cryptic (i.e. not conserved at the sequence level but conserved at the structural level) or simply that Cdc24 interacts with Bem1 through a noncanonical PB1 interaction (Sumimoto et al., 2007) as has been described in the mammalian MEK5-ERK5 interaction (Nakamura et al., 2006).

To accurately compare the effects of the absence of Bem1 with those observed in the absence of Cdc24 (which is essential for growth), we constructed the conditional alleles bem1^nar1 or cdc24^nar1. In these alleles, the coding regions of bem1 or cdc24 were fused to the nar1 promoter, which is induced by growing the cells in nitrate as the nitrogen source, and strongly repressed in medium with ammonium or amino acids as nitrogen source (Brachmann et al., 2001). We constructed strains in which the native alleles were replaced by these conditional alleles. In accordance with the essential role of Cdc24, conditional cdc24^nar1 cells were unable to form colonies when shifted to repressive medium (Fig. 1C; YPD). However, bem1^nar1 cells were able to form colonies in these conditions, indicating that Bem1 was not essential for growth, although the smaller size of the colonies suggest some impairment on growth with respect to wild-type cells (Fig. 1C).

We compared in liquid culture, cells carrying the bem1^nar1 allele with cells carrying the cdc24^nar1 conditional allele. In both cases, incubation under restrictive conditions for nar1 expression resulted in round cells. We performed double DAPI and Calcofluor White staining to detect nuclei and septa in these cells (Fig. 1D). We found that cdc24^nar1 cells arrested as cell doublets, showing a single nucleus per cell compartment. For bem1^nar1 cells, 93% of the cells lost their polarity. First, they accumulated as cell doublets, which then became cell chains owing to impairment in cell separation. We frequently observed cell compartments carrying two nuclei (Fig. 1D, arrowheads). We believe that in these cells nuclear division occurs but the cytokinesis is somehow delayed.

We analyzed the subcellular localization of Bem1 and Cdc24. For this, we constructed a strain carrying chromosomal functional CFP and YFP C-terminal fusions of Bem1 and Cdc24, respectively. Both proteins showed the same pattern of subcellular distribution. They colocalized at the cell tips and at the bud neck during cytokinesis (Fig. 2A). We also tested whether Bem1 was important for Cdc24 localization, and vice versa (Fig. 2B). Firstly, we analyzed Bem1-GFP localization in cdc24^nar1 cells growing under restrictive conditions. Interestingly, we found that Bem1-GFP localization (Fig. 2B, arrowheads) was independent of Cdc24. By contrast, when the localization of Cdc24-GFP was analyzed in bem1^nar1 cells growing under restrictive conditions, we found that Cdc24-GFP localization at the cell tips was dependent on Bem1 (Fig. 2B).

Overexpression of cdc24 resulted in the formation of elongated cells (Castillo-Lluva et al., 2007). Encouragingly, we found that overexpression of bem1 also produced this kind of filamentous growth (Fig. 2C). Moreover, we found that this filamentous growth was dependent on cdc24 (Fig. 2D). By contrast, we observed that the filamentous phenotype observed after overexpression of cdc24 was independent of bem1 (Fig. 2D). We observed that after overexpression, Cdc24-GFP fusion protein accumulated at the tip even in a Bem1-defective background (supplementary material Fig. S1).

Altogether, these observations suggest that at physiological levels Bem1 is helping Cdc24 to localize at the cell tip, although high levels of Cdc24 can bypass the Bem1 requirement indicating that additional secondary elements are involved in the polar localization of Cdc24.

Since the genetic evidence suggests that Bem1 and Cdc24 could be involved in the promotion of polar growth in U. maydis, we sought to analyze whether Cdc24 and Bem1 physically interact. For this, we constructed a strain carrying chromosomal functional versions of tagged Bem1-3MYC and Cdc24-3HA. Lysates prepared from the corresponding strains were immunoprecipitated with anti-MYC and anti-HA antibodies, and the immunoprecipitates were analyzed for the presence of Bem1 (using anti-MYC antibodies) and Cdc24 (using anti-HA antibodies) by western blot assay. The results indicated that, as expected, Bem1 and Cdc24 were able to physically interact (Fig. 3A).

We sought to add further support to the idea that Bem1 interacts with Cdc24 in U. maydis. For this, we took advantage of the Bimolecular Fluorescence Complementation (BiFC) technique (for a recent review, see Kerppola, 2006). In this approach, the two non-fluorescent halves of the GFP variant yellow fluorescent protein (YFP) are fused separately to two potentially interacting proteins. The interaction of the two fusion proteins leads to the reconstitution of the YFP variant and hence fluorescence. We fused the N-terminal fragment of YFP (NYFP, residues 1-157) to the N-terminus of Bem1 and the C-terminal fragment of YFP (CYFP, residues 158-239) to the C-terminus of Cdc24. Cells expressing individual fusion proteins (i.e. NYFP-Bem1 or Cdc24-CYFP) as well as the pair combination (NYFP-Bem1 and Cdc24-CYFP) were examined by fluorescence microscopy (Fig. 3B). Expression of either fusion alone did not produce detectable fluorescence. Interestingly, cells carrying the NYFP-Bem1 Cdc24-CYFP pair produced a fluorescent signal that accumulated preferentially at the cell tips, confirming the physical interaction between these two proteins that was demonstrated by immunoprecipitation.

We previously reported that Cdc24 localization at the cell tips was dependent on Cdk5 (Castillo-Lluva et al., 2007; Fig. 4A). As Cdc24-GFP localization at the tips was also dependent on Bem1, we wondered whether Bem1-GFP localization was affected by defective cdk5. Strikingly, we found that Bem1-GFP was still present at the cell tips in cdk5^ts cells growing at restrictive conditions (Fig. 4B). We also examined whether overexpression of bem1 bypassed the requirement of cdk5 to induce polar growth, as shown for cdc24 (Castillo-Lluva et al., 2007). Interestingly, we found that bem1 overexpression was also able to induce polar growth in a cdk5^ts genetic background (Fig. 4C,D).

Based on the data presented above, we hypothesized that Cdk5 may regulate the Bem1 and Cdc24 interaction. Therefore, we wondered whether Cdk5 would affect the ability of Bem1 to interact with Cdc24 and whether Cdk5 would associate with Bem1 and/or Cdc24. To test this, we introduced chromosomal functional versions of tagged Bem1-VSV and Cdc24-3HA in 3MYC-tagged cdk5 and cdk5^ts allele strains. We incubated both strains at restrictive conditions for cdk5^ts cells (4 hours at 34°C) and analyzed the putative interactions amongst these proteins by immunoprecipitation assays. It is worth mentioning that during this period of incubation cdk5^ts mutant cells do not lose viability (Castillo-Lluva et al., 2007). We observed that in the control strain it was possible to obtain coimmunoprecipitation of Bem1 and Cdc24, as well as Bem1 and Cdk5, but not Cdc24 and Cdk5 (Fig. 5A). By contrast, the Cdc24-Bem1 complex was not present in coimmunoprecipitates from cdk5^ts cells, although the interaction between Cdk5 and Bem1 was still present. These data support the notion that Cdc24 and Bem1 interaction is dependent on an active Cdk5 protein and that Cdk5 and Bem1 interact.

We also repeated the BiFC assay as above using NYFP-Bem1 and Cdc24-CYFP fusions in a cdk5^ts genetic background. In agreement with the coimmunoprecipitation results, we were unable to obtain a fluorescent signal when cells were grown under restrictive conditions (Fig. 5B).

We described previously that Cdk5-GFP accumulates in the nucleus (Castillo-Lluva et al., 2007). Interestingly, Cdk5-GFP also appeared at the cell tips, where it colocalized with Bem1-RFP (Fig. 6A). To determine whether the interaction between Cdk5 and Bem1 occurs at the tip of the cell, we carried out a BiFC assay between Cdk5 and Bem1. Since we found no interaction between Cdc24 and Cdk5, we also included this interaction pair as a control. We fused the N-terminal fragment of YFP (NYFP, residues 1-157) to the N-terminal end of Bem1 and Cdc24, and the C-terminal fragment of YFP (CYFP, residues 158-239) to the C-terminal end of Cdk5. As before, in each fusion, the two protein fragments were joined by a linker of 8 Gly-Ala residues. Cells expressing individual fusion proteins (i.e. NYFP-Bem1, NYFP-Cdc24 and Cdk5-CYFP) as well as pair combinations (i.e. NYFP-Bem1 Cdk5-CYFP and NYFP-Cdc24 Cdk5-CYFP) were examined by fluorescence microscopy (Fig. 6B). Expression of either fusion alone did not produce detectable fluorescence. Interestingly, cells carrying the NYFP-Bem1 Cdk5-CYFP pair produced a fluorescence signal that appeared at the cell tip. Strikingly, we also found that a fluorescence signal also accumulated within the nucleus.

The above results gave the surprising result that Bem1 was able to interact with Cdk5 within the cell nucleus. One possibility is that once the fluorescent protein was reconstituted, the interaction between the two halves is so strong that will keep the interacting partners together. Thus, fluorescence in the nucleus will indicate past interaction, for instance at the cell tip, and that later on, the complex aberrantly accumulates in the nucleus. Alternatively, it could be possible that a fraction of Bem1 could be located at the cell nucleus, where it could be able to interact with Cdk5, as well as other proteins. In this respect, it is worth mentioning that when bem1-GFP cells were mounted onto the microscope slide for some time (i.e. 10 minutes) occasionally Bem1-GFP signal was observed within the nucleus (see supplementary material Fig. S2). This effect, which we believe was a consequence of oxygen depletion on the slide, was never found when observing cdc24-GFP cells. Therefore, we reasoned that Bem1 could be transiently located in the cell nucleus and actively (i.e. energy dependently) exported to the cytoplasm. A conserved factor, also present in U. maydis, involved in nuclear protein export is the protein Crm1 (Hutten and Kehlenbach, 2007). To determine whether Bem1 was exported from the nucleus in a Crm1-activity-dependent manner, we investigated Bem1-GFP localization in cells treated with leptomycin B (LMB), a well-known and specific inhibitor of Crm1-mediated nuclear export (Yashiroda and Yoshida, 2003). A parallel experiment was carried out with cdc24-GFP cells. Bem1-GFP as well as cdc24-GFP cells were treated with LMB (100 ng/ml) or its solvent ethanol. At various times following the addition of the drug, a portion of each culture was photographed and counted to determine the number of cells with nuclear accumulation of Bem1-GFP and Cdc24-GFP. We found that following LMB treatment for less than 10 minutes, Bem1-GFP accumulated in the nucleus of most cells (89%) (Fig. 7A,C). By contrast, Cdc24-GFP localization was not modified after LMB treatment (Fig. 7B,C).

Some proteins that are actively exported from the nucleus contain a hydrophobic nuclear export signal (NES) that binds Crm1 (Hutten and Kehlenbach, 2007). Examination of the protein sequence of Bem1 revealed the presence of a short hydrophobic region with the potential to act as a Crm1-dependent NES (Fig. 8A). To determine whether this sequence acts as a NES we generated a mutant (Bem1-NESΔ) in which three of the hydrophobic residues were substituted with alanine. In addition, we also fused GFP at the C-terminus to localize the protein. In comparison with a strain carrying a non-mutated version of Bem1 fused to GFP (Bem1-GFP) (Fig. 8B, top panel), we found that Bem1-NESΔ-GFP strongly accumulated at the cell nucleus, although it also appeared dispersed within the cytoplasm (Fig. 8B, middle panel). Interestingly, these mutant cells showed a rounder morphology than wild-type cells suggesting some polarity defects. Since the mutated NES sequence is located inside the conserved domain PX (which is predicted to be involved in the interaction with other factors), we cannot unequivocally ascribe the lack of polarity observed in the bem1-NESΔ mutant cells to the nuclear accumulation of Bem1, because other functions could potentially be affected. To test whether forced nuclear extrusion would restore the wild-type morphology to the bem1-NESΔ mutant cells, we cloned the NES from protein kinase A inhibitor (PKI) (Hutten and Kehlenbach, 2007) onto the C-terminus of the Bem1-NESΔ mutant. This protein, Bem1-NESΔ+NES-GFP, did not localize on the nucleus, indicating that the extra PKI NES was functional, and interestingly bem1-NESΔ+NES-GFP cells grew in a polar manner, although a complete wild-type morphology was not obtained (Fig. 8B, bottom panel).

In this work we provide evidence that the adaptor protein Bem1 in U. maydis helps to localize the GEF Cdc24 at the cell tips where polar growth occurs. Several results support this conclusion. Bem1 and Cdc24 colocalized at the cell tip and physically interacted, which was demonstrated using two different approaches: co-immunoprecipitation analysis and BiFC assays. We also found that, although the localization of Bem1 was independent of Cdc24, at physiological levels Cdc24 requires Bem1 for its localization. This dependence agrees with the similar morphological defects obtained when the levels of Bem1 or Cdc24 were reduced using conditional mutants. However, it seems that the Bem1 requirement is not strict and that additional elements would help Cdc24 to localize at the cell tips. We based this conclusion on two results: Bem1 was not essential for growth, whereas Cdc24 was, and cdc24 overexpression bypassed the Bem1 requirement for polar growth and for its localization at the cell tip (supplementary material Fig. S1). Accordingly, the roles of Bem1 during polar growth regulation in U. maydis mirrored those described in other yeast such as S. cerevisiae and S. pombe (Butty et al., 2002; Endo et al., 2003). Here, we do not address whether additional elements (such as Rac1 or Cla4 proteins) could interact with Bem1, as happens in S. cerevisiae or S. pombe, but our future research will explore this. Nevertheless, what sets apart the U. maydis Bem1-Cdc24 interaction from the other yeasts is its dependence on Cdk5, a cyclin-dependent kinase that regulates polar growth in the smut fungus (Castillo-Lluva et al., 2007). We determined that Cdk5 regulates the ability of Cdc24 and Bem1 to form a complex. Both coimmunoprecipitation and BiFC assays support this conclusion. However, the nature of this regulatory step remains unknown. We proved that Bem1 and Cdk5 interact; however, we found no evidence of Cdk5-dependent protein modification (i.e. phosphorylation) of Bem1 or Cdc24 (J.P.-M., unpublished observations). The proposed Cdk5-dependent regulatory step could be rather less direct. For instance, Cdk5 may downregulate GTPase-activating proteins (GAPs), leading to Rac1-GTP, which may promote Bem1-Cdc24 complex formation. A recent report on S. cerevisiae indicates that GAPs, which are located at the cell tip, are targets of the Pho85 kinase, the S. cerevisiae Cdk5 homolog (Sopko et al., 2007). Alternatively, it is possible that Cdk5 promotes the interaction between Bem1 and Cdc24 independently of its kinase activity. Recently, it has been shown in S. pombe that the Pom1 kinase affects the functionality of the GAP Rga4 by physically interacting with it, although the kinase does not phosphorylate it (Tatebe et al., 2008). Additional efforts will be required to discern all these possibilities in U. maydis.

A surprising result that we found is that Bem1 seems to be transiently located at the cell nucleus. Results presented here demonstrate that U. maydis Bem1 is subject to nuclear export. Nuclear accumulation of Bem1 could be induced by treating cells with Leptomycin B, an inhibitor of the nuclear export factor Crm1. Moreover, Bem1 contains a sequence that resembles a NES, and mutation in hydrophobic residues, which is predicted to be important for NES activity, resulted in the accumulation of Bem1 in the nucleus. Since we did not detect nuclear Bem1 in normal conditions, we believe that this nuclear localization is transient. We have no idea about the reasons for such a localization. Although we detected a strong interaction in the BiFC assay between Cdk5 and Bem1 in the cell nucleus, this does not necessarily imply that the physiological interaction occurs in this cell compartment. Retaining Bem1 at the cell nucleus affects the ability of the cell to grow in a polar manner. Mutations in the NES sequence produced rounder cells than in the wild type, and the addition of a heterologous NES sequence alleviates this defect. One interesting possibility could be that by regulating the subcellular localization of Bem1, the cell will be able to determine the amount of polar growth. Alternatively, because Bem1 has been linked to other processes, such as vacuole homeostasis, in other fungi (Han et al., 2005), it could be possible that Bem1 has additional unknown targets in the cell nucleus. There are other examples of proteins involved in the regulation of polar growth that transiently localize within the cell nucleus. During polar growth in S. cerevisiae, Cdc24 is transiently located at the cell nucleus where it interacts with the adaptor protein Far1 (Toenjes et al., 1999; Shimada et al., 2000). In response to pheromone, which induces polar growth in budding yeast, Far1 helps Cdc24 to be recruited to the site where polar growth will be initiated allowing the interaction with Bem1 (Butty et al., 1998; Nern and Arkowitz, 2000). It will be challenging to determine whether a similar regulatory scheme also works for Bem1 in U. maydis.

All strains have the genetic background of FB1 (a1b1) (Banuett and Herskowitz, 1989) and are summarized in supplementary material Table S1. U. maydis strains were grown at 28°C unless otherwise specified. The permissive temperature for temperature sensitive (ts) strains was 22°C and the restrictive temperature was 34°C. U. maydis cells were grown in complete medium (CM) amended with the corresponding carbon source (Holliday, 1974) unless otherwise specified. The expression of genes under the control of the crg1 and nar1 promoters, were all performed as described previously (Brachmann et al., 2001).

U. maydis strains were constructed by transformation of progenitor strains with linearized plasmids (Tsukuda et al., 1988). The integration of the plasmids into the corresponding loci was verified by diagnostic PCR and subsequently by Southern blotting.

Standard molecular techniques were followed (Sambrook et al., 1989). Protein extraction and immunoprecipitation were performed as previously described (Garrido et al., 2004; Butty et al., 2002).

To produce the conditional bem1^nar1 allele we followed a strategy consisting of ligation of a pair of fragments into pRU2 (Brachmann et al., 2001) digested with NdeI and EcoRI. The 5′ fragment (0.74 kbp, flanked by EcoRI and EcoRV) was produced by PCR amplification using the primers Bem1-3 (5′-AGTGATATCGGCTGGCAAGACTCGAAACCG-3′) and Bem1-4 (5′-CCGAATTCGTTAGGAGATCGTGATTGTCGTCAAG-3′) and spans nucleotide–869 to–122 with respect to the translation start. The 3′ fragment (1.58 kbp, flanked by NdeI and EcoRV) was obtained by PCR amplification using the primers Bem1-1 (5′-CATATGAAGGGTCTCAAAGATTTTCGCAGG-3′) and Bem1-4 (5′-TCTCTGGTTGCTGATATCTAGCCCTCC-3′) spans nucleotide +1 to nucleotide +1587. The resulting plasmid, pBem1nar was integrated, after digestion with EcoRV by homologous recombination into the bem1 locus.

To produce the conditional cdc24^nar1 allele we followed a strategy consisting of ligation of a pair of fragments into pRU2 (Brachmann et al., 2001) digested with NdeI and AflII. The 5′ fragment (0.5 kbp, flanked by BamHI and AflII) was produced by PCR amplification using the primers Cdc24-4 (5′-GGATCCACAACCACGACAGTCAATTGC-3′) and Cdc24-5 (5′-CTTAAGGTCGCGAGTCGCCAGACGAAT-3′) and spans nucleotide–712 to–212 to with respect to the translation start. The 3′ fragment (0.65 kbp, flanked by NdeI and BamHI) was obtained by PCR amplification using the primers cdc24-1 (5′-CATATGGCTGCTTCCACATCGCTTGCGTTC-3′) and cdc24-6 (5′-GTGTCATTGAGGTAGAGCTGTGAGACG-3′) and spans nucleotide +1 to nucleotide +649. The resulting plasmid, pCdc24nar was integrated, after digestion with BamHI by homologous recombination into the cdc24 locus.

To overexpress bem1, a 2.28 kbp fragment flanked by NdeI and AflII sites was amplified by PCR using the primers Bem1-1 and Bem1-2 (5′-CTTAAGACAATGACTCGGATGCATACCATT-3′). This fragment carried the bem1 ORF as well as around 100 bp of the 3′ downstream region. The fragment was cloned into the corresponding sites of pRU11 plasmid (Brachmann et al., 2001) under the control of the crg1 promoter. The resulting plasmid, pRU11-Bem1, was digested with XcmI to promote the insertion by homologous recombination into the cbx locus as described (Brachmann et al., 2001).

To overexpress cdc24, a 3.46 kbp fragment flanked by NdeI and AflII sites was amplified by PCR using the primers cdc24-1 and Cdc24-2 (5′-CTTAAGGGAAGTGGAGAAGCGAATCACGAA-3′). This fragment carried the cdc24 ORF as well as around 100 bp of the 3′ downstream region. The fragment was cloned into the corresponding sites of pRU11 plasmid (Brachmann et al., 2001) under the control of the crg1 promoter. The resulting plasmid, pRU11-Cdc24, was digested with SspI to promote the insertion by homologous recombination into the cbx locus as described (Brachmann et al., 2001).

C-terminal epitope tags were introduced in each gene as described previously (Castillo-Lluva et al., 2007). Briefly, the corresponding ORFs without a stop codon were flanked by BamHI and EcoRI sites by PCR amplification with the corresponding primers and the resulting fragment was cloned in the tagging vectors pBS-VSV, pBS-MYC or pBS-HA (Garrido et al., 2004) and p123GFP, p123RFP, p123CFP and p123YFPHyg (Flor-Parra et al., 2006) (J.P.-M., unpublished results). Different resistance markers (hygromicin and carboxine) were added to the vectors carrying the tagged ORFs to allow the selection after transformation of U. maydis cells. To amplify the bem1 ORF we used the primers Bem1-8 (5′-CGCGGATCCTACTCCACCACCCAAGGATGA-3′) and Bem1-9 (5′-CCGGAATTCGGAAGCATGGAGCATGAACTT-3′). The 1.21 kbp fragment was cloned into pBS-MYC (MYC-tag, hygromicin resistance), pBS-VSV (VSV-tag, hygromicin resistance), p123CFP (CFP tag, carboxine resistance), p123GFP (GFP tag, carboxine resistance), and p123RFP (RFP tag, hygromicin resistance) resulting in pBem1-MYC, pBem1-VSV, pBem1-CFP, pBem1-GFP and pBem1-RFP plasmids, respectively. To amplify the cdc24 ORF we used the primers Cdc24-7 (5′-CGCGGATCCAAGATGAGTAGTATCCTGGGT-3′) and Cdc24-8 (5′-CCGGAATTCGTTGACGACCAACTCGACTTC-3′). The 1.64 kbp fragment was cloned into pBS-HA (HA-tag, carboxine resistance), p123YFP (YFP tag, hygromicin resistance), and p123GFP (GFP tag, carboxine resistance), resulting in pCdc24-HA, pCdc24-YFP and pCdc24-GFP, respectively.

To generate the BiFC partners, two halves of YFP were produced as described previously (Wilson et al., 2004). The NYFP half (spanning the first 157 amino acids of YFP plus an additional eight residues Ala-Gly linker) was obtained as a NdeI-flanked PCR fragment after amplification with primers FPNdeI (5′-CATATGGTGAGCAAGGGCGAGGAGCTGTTCA-3′) and BFIC1 (5′-CATATGGGCGCCTGCGCCTGCGCCGGCCCCCTTGTCGGCCATGATATAGACGTT-3′). The CYFP half (starting with an eight residue Gly-Ala linker and followed by the 158-239 residues from YFP) was obtained as an EcoRI fragment after amplification by PCR with primers BFIC2 (5′-GAATTCGGGGCCGGCGCAGGCGCAGGCGCCCAGAAGAACGGCATCAAGGTGAAC-3′) and FPEcoRI (5′-GAATTCACATGTTAATTATTACATGCTTAA-3′). To produce the NYFP-Bem1 and NYFP-Cdc24 fusions, the NdeI NYFP cassette was cloned into the corresponding NdeI site of pBem1nar and pCdc24nar, respectively. The resulting plasmids were introduced by homologous recombination into the corresponding native loci. To produce the Cdk5-CYFP fusion, the EcoRI CYFP cassette was cloned into the corresponding EcoRI site of a plasmid carrying the cdk5 ORF without the stop codon (Castillo-Lluva et al., 2007) and introduced by homologous recombination into the cdk5 locus. A similar strategy was utilized to produce the Cdc24-CYFP construction.

To construct the bem1-NESΔ mutant allele, a two-step mutagenesis protocol using PCR was utilized. Leu447, Met450, and Leu454 residues were exchanged with Ala residues. This change created a new restriction site (EagI) that was used to track the mutation by PCR amplification of genomic DNA. The amplifications were performed with the primer pairs BEM1-DIR (5′-TTCTAGATACTCCACCACCCAAGGATG-3′) and BEM1-NES2 (5′-TCAACGACATCGGCCGGACCGGGCGCGAGTGGAGCGATGCGTTCGCT-3′); BEM1-REV (5′-GCCATGGGAAGCATGGAGCATGAACTT-3′) and BEM1-NES1 (5′-AGCGAACGCATCGCTCCACTCGCGCCCGGTCCGGCCGATGTCGTTGA-3′). The fragment carrying the mutations was fused to p123-GFP or p123-PKINES-GFP and it was introduced into the corresponding locus by homologous integration. p123NESGFP carries a fusion of GFP to the protein kinase A inhibitor nuclear export sequence (LALKLAGLDI) (Kutay and Guttinger, 2005).

Samples were mounted on microscope slides and visualized in a Nikon eclipse 90i microscope equipped with a Hamamatsu ORCA-ER CCD camera. Some images were also acquired with a Hamamatsu Electron Multiplier CCD digital camera. Images were taken using the appropriate filter sets, a Nikon Plan Apo VC 100× NA 1.40 lens and Nikon oil-immersion Type A nd=1.515. The software used with the microscope was MetaMorph 6.1 (Universal Imaging, Downingtown, PA). Images were further processed using Adobe Photoshop 7.0. All the images in this study are in a single plane.

For DAPI/calcofluor (D/CF) staining, cells from 900 μl of cell culture were fixed by quickly and thoroughly mixing with 100 μl formalin (37% stock solution) and incubating on ice for at least 10 minutes. The fixed cells were washed twice with ice-cold PBS and resuspended in 10-40 μl PBS. To visualize DNA as well as the septa, both DAPI and calcofluor were included in the mounting medium (10 μl of 10 mg/ml paraphenylene diamine, 2 μl of 100 μg/ml DAPI, 47 μl H₂O, 3.5 μl saturated calcofluor, 38 μl glycerol). To mount the cells, 4 μl of cell suspension were delicately spread on the coverslip and any excess of suspension was removed. The coverslip was then left to dry in the dark. A small drop of mounting medium was placed on a slide onto which the coverslip was then dropped.

The authors acknowledge the valuable suggestions of anonymous reviewers. This work was supported by Grants from the Spanish government (BIO2005-02998) and EU (MRTN-CT-2005-019277). The authors declare no conflict of interest.