The morphogenesis-related NDR kinase (MOR) pathway regulates morphogenesis in fungi. In spite of the high conservation of its components, impairing their functions results in highly divergent cellular responses depending on the fungal species. The reasons for such differences are unclear. Here we propose that the species-specific connections between cell cycle regulation and the MOR pathway could be partly responsible for these divergences. We based our conclusion on the characterization of the MOR pathway in the fungus Ustilago maydis. Each gene that encodes proteins of this pathway in U. maydis was deleted. All mutants exhibited a constitutive hyperpolarized growth, contrasting with the loss of polarity observed in other fungi. Using a conditional allele of the central NDR kinase Ukc1, we found that impairing MOR function resulted in a prolonged G2 phase. This cell cycle delay appears to be the consequence of an increase in Cdk1 inhibitory phosphorylation. Strikingly, prevention of the inhibitory Cdk1 phosphorylation abolished the hyperpolarized growth associated with MOR pathway depletion. We found that the prolonged G2 phase resulted in higher levels of expression of crk1, a conserved kinase that promotes polar growth in U. maydis. Deletion of crk1 also abolished the dramatic activation of polar growth in cells lacking the MOR pathway. Taken together, our results suggest that Cdk1 inhibitory phosphorylation may act as an integrator of signaling cascades regulating fungal morphogenesis and that the distinct morphological response observed in U. maydis upon impairment of the MOR pathway could be due to a cell cycle deregulation.
Fungal cells exhibit a diverse array of shapes and sizes. Since fungal morphogenesis is driven by localized membrane expansion and cell wall deposition, the wide variation of fungal morphology observed could be explained by the fact that different fungi use different elements to produce and regulate cell wall formation. However, although the inventory of gene products involved in hyphal morphogenesis continues to expand, it seems clear that regulatory and structural elements are shared by the majority of fungi. A second hypothesis to explain the wide morphological variation observed could be that fungi use similar elements but ‘wiring’ between these elements are unique for each fungus (Harris, 2011). The morphogenesis-related (MOR) pathway is a good example supporting this hypothesis. The MOR pathway, also named the RAM (regulation of Ace2 and morphogenesis) pathway, is conserved among different fungi (Maerz and Seiler, 2010). This pathway includes a nuclear Dbf2-related (NDR) kinase termed Cbk1 in Saccharomyces cerevisiae (Racki et al., 2000), Orb6 in Schizosaccharomyces pombe (Verde et al., 1998), and COT1 in Neurospora crassa (Yarden et al., 1992), associated with a regulatory subunit, Mob2 (Hou et al., 2003; Maerz et al., 2009; Nelson et al., 2003). The NDR-Mob2 complex is activated by a germinal center kinase called Kic1 in S. cerevisiae (Nelson et al., 2003), Nak1 in S. pombe (Huang et al., 2003), and Pod6 in N. crassa (Seiler et al., 2006). Additional elements are Hym1/Mo25-related (Karos and Fischer, 1999; Nelson et al., 2003) and Tao3/Mor2/Fry-related scaffold-like proteins (Du and Novick, 2002; Kanai et al., 2005). What made this pathway extremely interesting is the fact that in spite of such a broad conservation in their elements, abrogation of this pathway has different morphological outcome depending on the fungi. While in S. cerevisiae loss-of-function mutations in RAM components lead to defects in cell separation after mitosis, loss of polarity, and defect in cell integrity (Bidlingmaier et al., 2001; Nelson et al., 2003), a similar range of mutations in Cryptococcus neoformans (another budding yeast) resulted in hyperpolarized growth (Walton et al., 2006). In contrast, in the filamentous fungi N. crassa (Yarden et al., 1992), Aspergillus nidulans (Johns et al., 2006), Colleototrichum trifolii (Buhr et al., 1996) and Claviceps purpurea (Scheffer et al., 2005), this pathway is not required for the establishment of polarity per se, but mutants display hyperbranched growth, indicating that, in these organisms, this pathway is required to restrict excessive branch formation in subapical regions of the cell.
The reasons for these divergences seem to be unrelated to the core signaling components as all of these core elements as well as their interactions are conserved among these various fungi (Maerz and Seiler, 2010). However, the inputs and outputs to and from this central core are most likely unique. In other words, a plausible explanation for the morphological differences in MOR mutants may be explained by an organism-specific integration of the MOR pathway in a cellular signaling context. Very few upstream pathway activators as well as downstream targets of this signaling cascade are known (Maerz and Seiler, 2010) and the study of the cellular processes targeted by the MOR pathway could help understand the wide morphological defect observed in these different fungi.
Morphogenesis and cell cycle are intricately connected in fungi. Moreover, progression through the cell cycle is critically dependent on signals induced by environmental factors or stress stimuli, as well as intracellular conditions. Therefore, cell cycle is a primary target of signaling cascades in all organisms and it makes sense to think that the MOR pathway also could play a function in cell cycle regulation. In metazoan, it is well known that cascades including NDR kinases such as the Hippo pathway controls cell proliferation and some of its elements are considered as tumor suppressors (Pan, 2010). In fungi, one of the few characterized downstream targets of this pathway is the transcriptional factor Ace2 from S. cerevisiae, which controls cell separation after mitosis (Colman-Lerner et al., 2001; Nelson et al., 2003). In addition, Ace2 is phosphorylated by Cdc28, the mitotic CDK in budding yeast (Brace et al., 2011; Mazanka and Weiss, 2010). However, supporting our idea about species-specific targets, Ace2 is not conserved through the fungal kingdom and seems to be specific to S. cerevisiae and closed relatives such as Candida albicans (Mulhern et al., 2006). The intricacies between cell cycle regulation and morphogenesis in fungi as well as the apparent non-conservation of the MOR pathway targets led us to think that some of the differences observed in MOR pathway regulation in fungi could be related to species-specific connections between this pathway and cell cycle regulation. We also base this assumption on our previous experiments addressing the connections between the cell cycle and MAPK-based signaling pathways in the phytopathogenic fungus Ustilago maydis (Pérez-Martín et al., 2006). This fungus has been used as a model system in fungal development for many years (Steinberg and Pérez-Martín, 2008). Using this model organism, we studied the connections between cell cycle regulation and two different MAPK cascades: the pheromone-dependent cascade (García-Muse et al., 2003) and the Cell Wall Integrity (CWI) pathway (Carbo and Pérez-Martín, 2010). We found that for both pathways, the distinct cell cycle connections with otherwise conserved signaling pathways resulted in dramatic morphological differences when compared with the well-known responses in other fungi such as S. cerevisiae. For instance, upon treatment with cell wall stressors, S. cerevisiae and U. maydis cells showed different morphological responses that are a consequence of the specific regulation of cell cycle in response to these stress: in U. maydis, the CWI cascade activation accelerates the G2/M transition whereas it induces a G2 arrest in S. cerevisiae (Carbo and Pérez-Martín, 2010; Harrison et al., 2001).
The idea that the species-specific differences in MOR pathway regulation could be the consequence of species-specific wiring with cell cycle, as we observed in other signaling pathways, prompted us to analyze the MOR pathway in U. maydis. A previous report (Dürrenberger and Kronstad, 1999) established that mutations in ukc1, the gene encoding the NDR kinase in U. maydis, resulted in hyperpolarized growth, which contrasted with the defects in polarity observed in cell deleted for CBK1 in S. cerevisiae.
In this work, we characterized the components of the MOR pathway along with the central NDR kinase Ukc1, as well as the connections between Ukc1 and cell cycle regulation, focusing on the morphological consequences of these interactions.
Characterization of Mob2 in U. maydis
The fungal NDR kinases described to date belong to two major groups. One group of NDR kinases is crucial for coordination between mitotic exit and septum formation as well as cell separation (Dbf2 group) while the other group controls polar cell morphology (Cot1 group) (Maerz et al., 2009). In U. maydis three NDR kinases can be identified by sequence analysis (supplementary material Fig. S1A). One of them, the Ukc1 kinase, belongs to the Cot1 group. Moreover, analysis of the morphological consequences of loss-of-function mutations in ukc1 supported the inclusion of this kinase in the Cot1 group. Deletion of ukc1 resulted in cells that show defects in cell separation and grow in a hyperpolarized way (supplementary material Fig. S1B). These defects are similar to those observed in Cryptococcus neoformans, another basidiomycete yeast (Walton et al., 2006), and contrast – at least with respect to sustained polar growth – with the effects observed in other yeasts like S. cerevisiae (Nelson et al., 2003) or S. pombe (Verde et al., 1998), where a problem in the maintenance of polar growth occurs.
NDR kinases require association with Mob co-activators, and this interaction is specific for each kinase group (Hergovich, 2011). In silico search in the manually annotated MIPS U. maydis database (see http://mips.helmholtz-muenchen.de/genre/proj/ustilago/), identified three protein entries in the genome of U. maydis that were most similar to fungal Mob1 and Mob2 and to the more distantly related Mob family member MOB3/phocein (Fig. 1A). Based on these sequence similarities, we assigned the names Mob1 to UM04352, Mob2 to UM12135, and Mob3 to UM00493. It is worth noting that in fungi, Mob2 homologs appeared to form two distinct clusters (Maerz et al., 2009). Homologs from filamentous fungi (N. crassa and A. nidulans) were clustered together, while homologs from yeast-like fungi were clustered in another group (Fig. 1A). NDR kinases from the Cot1 group specifically interact with Mob2-like proteins. To address whether that was the case in U. maydis, we first created a strain harboring a deletion of the mob2 gene and morphological defect were analyzed. We observed that the morphology of cells lacking Mob2 was indistinguishable from the morphology of cells carrying a deletion of the ukc1 gene (Fig. 1B).
Given the phenotypic similarity, we constructed a double Δukc1 Δmob2 mutant and found that loss of mob2 did not exacerbate the morphological defects of ukc1 cells, supporting a genetic interaction between ukc1 and mob2 (Fig. 1B). We also analyzed their ability to physically interact by co-immunoprecipitation. For this, we generated strains harboring tagged versions of Ukc1 and Mob2, the protein Ukc1-myc and t7-Mob2. Tagged alleles were as functional as wild-type versions and they were able to complement the defective morphology of mutant strains (supplementary material Fig. S2). The results from co-immunoprecipitation assays indicated that Mob2 and Ukc1 tagged version were co-immunoprecipitated with antibodies against both epitopes (Fig. 1C). All together these data support the idea that Ukc1-Mob2 are the orthologs in U. maydis of Cot1-Mob2.
U. maydis orthologs of MOR network genes
Observing the dramatic phenotype of Δukc1 and Δmob2 mutants, we wondered whether it was related to defects in the MOR pathway or to a specific role of this kinase-regulator pair, distinct from the MOR pathway in U. maydis. To address this question we searched in the U. maydis genome database at MIPS for ORFs displaying amino acid sequence similarity to MOR signaling network components from S. cerevisiae and S. pombe.
The NDR/Mob2 complex is activated by a germinal center (GC) kinase and coordinated by Hym1/Mo25-related and Tao3/Mor2/Fry-related scaffold proteins (Maerz and Seiler, 2010). We found an entry (UM11396) encoding a member of GCK-III subfamily of eukaryotic Ste20 kinases, which shows a high sequence similarity with S. cerevisiae Kic1 (E-value 9e−95) and S. pombe Nak1 (E-value 3e−104). We also found single entries for the scaffold Hym1 (UM10613, E-values of 2e−58 and 3e−102 with S. cerevisiae Hym1 and S. pombe Pmo25, respectively) and Tao3 (UM10098, E-values of 6e−157 and 3e−95 with S. cerevisiae Tao3 and S. pombe Mor2, respectively). In addition to these core elements, we also identified the homolog of the modulator protein Sog2 (UM02656, E-values of 2e−16 and 4e−31 with the S. cerevisiae and S. pombe counterparts, respectively). A putative homolog to the downstream effector Ace2 was also found although the similarity region was restricted to the Zinc finger domain (UM10181, E-value of 3e−21 with S. cerevisiae Ace2).
We deleted each of the putative U. maydis MOR genes identified above. All mutated cells but Δace2 mutants showed defects in growth and cell morphology that were indistinguishable from the defects observed in Δukc1 and Δmob2 mutants (Fig. 2), corroborating the idea that impairment of the MOR pathway in U. maydis resulted in a constitutive hyperpolarization that differs from the loss of polarity seen in S. cerevisiae. The absence of morphological consequences in cells lacking the putative ace2 as well as the low similarity value strongly suggested that this protein was not conserved in U. maydis.
Hyperpolarized growth in U. maydis correlates with increased actin structures at the growth tip
The reasons for the divergent response among fungal MOR mutants regarding to polar growth control were not addressed yet. The actin cytoskeleton has been suggested to be one of the elements involved in the morphological response upon impairment of the MOR pathway (Maerz and Seiler, 2010). In fact, data from several fungal MOR mutants analyzed to date indicated anomalies in actin cytoskeleton, although the consequences of MOR mutations in actin cytoskeleton varied in each fungus. For instance, in MOR mutant cells from S. pombe, C. albicans and N. crassa, the actin cytoskeleton appears dispersed around the cell cortex (Hou et al., 2003; Song et al., 2008; Verde et al., 1998; Ziv et al., 2009). In contrast, in C. neoformans it was described that the hyperpolarized growth associated to RAM mutants correlated with abnormal actin localization at the growing tips (Walton et al., 2006). Observing these divergences in actin localization, we tested whether in U. maydis hyperpolar growth was also associated to changes in actin cytoskeleton in MOR mutants. To this end, we introduced into Δukc1 cells either a fimbrin-GFP fusion to detect actin patches or a Myo5-GFP fusion as a surrogated marker of actin cable activity (Castillo-Lluva et al., 2007). We found that fimbrin-GFP showed much higher fluorescent signal at the growing tips in Δukc1 cells than in control cells (Fig. 3A; supplementary material Fig. S3), while the fluorescent signal produced by Myo5-GFP was not significantly different between mutant and control cells (Fig. 3B; supplementary material Fig. S3) suggesting that actin patch localization, but not actin cables were affected in Δukc1 cells.
These data were in accordance with the described RAM mutants from C. neoformans (Walton et al., 2006). To delineate functional similarities between U. maydis and C. neoformans MOR/RAM pathways, we analyzed the subcellular localization of Ukc1. A C-terminal Ukc1-GFP fusion protein was created and inserted at the ukc1 native locus of a wild-type strain. This fusion allele was as functional as a wild-type allele since cells carrying this mutant allele had a wild-type appearance. Representative images of cells at different stages of the cell cycle were assembled to show that Ukc1-GFP fusion protein localized in different cell regions during cell cycle progression (Fig. 3C). In unbudded cells (G1 phase), Ukc1-GFP was found within the nucleus (Fig. 3C, arrowheads) as well as at the tip where a new bud will appear (Fig. 3C, arrows). At the bud emergence stage, it localized profusely at the bud tip as well as in the nucleus. When the daughter cell was ready to separate from the mother cell, Ukc1-GFP localized at the cell separation site and in both mother and daughter nuclei. These results contrast with those showed in C. neoformans, where Cbk1-DsRed fusions were localized diffusely throughout the cytoplasm and to punctuated spots resembling vesicles (Walton et al., 2006). While this distinct localization would suggest differences in cellular function between these two fungi, the subcellular localization of Ukc1-GFP fusion in U. maydis was in accordance with the cellular defects observed in cells lacking ukc1, which ranged from alteration in growth polarity to cell separation. The localization of Ukc1 at the tip of the cell seems to be a general trend in these kinases as it was described both in yeast, including S. cerevisiae (Bidlingmaier et al., 2001; Weiss et al., 2002) and C. albicans (Gutiérrez-Escribano et al., 2011) as well as in filamentous fungi such as N. crassa (Maerz et al., 2012; Seiler et al., 2006). In addition, we could observe the Ukc1-GFP protein within the nucleus as described for Cbk1 in S. cerevisiae (Colman-Lerner et al., 2001; Weiss et al., 2002). However, while Cbk1p localizes within the nucleus only during the G1-M transition, Ukc1 was found associated with the nucleus during all cell cycle phases, suggesting additional or different functions for Ukc1 protein kinase in U. maydis compared to S. cerevisiae.
Downregulation of ukc1 resulted in prolonged G2 phase
To understand the basis of the enhanced polar growth observed when MOR pathway was disabled in U. maydis, we constructed a strain carrying the conditional allele ukc1nar1, in which the nar1 promoter replaced the promoter of ukc1 (Fig. 4A). Using this allele, the expression of ukc1 is induced by growing cells in nitrate as the nitrogen source, and strongly repressed in medium with ammonium or amino acids as nitrogen source (Brachmann et al., 2001). Cells carrying the ukc1nar1 allele were incubated at repressing conditions during 24 h. Samples were obtained at different time points and used for microscopic observations: cell walls were visualized by staining the cells with FITC-labeled wheat germ agglutinin (WGA), a lectin that binds to oligomeric chitin (Nagata and Burger, 1974), and the number of nuclei was counted after DAPI staining (Fig. 4B,C). In growing conditions that allowed expression from the nar1 promoter (minimal medium with nitrate, time 0 h), cells carrying ukc1nar1 were morphologically indistinguishable from control strains. However, in repressing conditions, the downregulation of ukc1 correlated with an increased growth at both cell poles as well as a dramatic delay in nuclear division (doubling time for nuclei was around 6.3 hours while in control cells, mitosis occurs every 2.1 hours) resulting in enlarged cells that were strikingly polarized (Fig. 4B; supplementary material Fig. S4). In cells where nuclear division was apparent, a septum between dividing cells was eventually formed, although no cell separation occurred (Fig. 4B).
In fungi, cell size and cell division are connected in a way that cell divides only when it reaches a certain critical size (Rupes, 2002). However, we could observe that in ukc1nar1 cells growing in repressing conditions, this coupling of size/division was abolished, since nuclear division was actively delayed independently of the cell size. To address whether this delay occurs in a specific cell cycle phase, we analyzed the DNA content of these cell aggregates by FACS analysis, and correlated it with the nuclei number at different times. Assuming an equal partition of DNA per nuclei, our results indicated that these nuclei mostly carried a 2C DNA content, suggesting that nuclear division was delayed during G2 phase (Fig. 4D).
Hyperpolar growth in MOR-defective cells depends on Wee1 kinase
In U. maydis, cell size in daughter cells is dependent on the length of G2 phase: the longer the G2 phase, the larger the daughter cell and vice versa (Pérez-Martín et al., 2006). Therefore a delay in G2 phase is compatible with the enlarged cell compartments we observed in ukc1-defective cells. Previous reports (Sgarlata and Pérez-Martín, 2005a; Sgarlata and Pérez-Martín, 2005b) indicated that in U. maydis, length of G2 phase depends on the levels of Cdk1 inhibitory phosphorylation at Tyr15, which is modulated by two opposed cell cycle regulators: the Wee1 kinase and the Cdc25 phosphatase. The alteration of the level of one of these regulators results in a morphological change: the overexpression of Wee1 kinase or downregulation of Cdc25 phosphatase induced the formation of prolonged cells arrested in G2 phase (Sgarlata and Pérez-Martín, 2005a; Sgarlata and Pérez-Martín, 2005b). To know whether the G2 delay observed in MOR deleted cells could be the consequence of a change in Cdk1 phosphorylation, we analyzed the level of Cdk1 phosphorylation at Tyr15 upon downregulation of ukc1nar1. For this, we used a specific antibody raised against the Tyr15 phosphorylated human Cdc2 peptide (Y15P), which also recognizes the Cdk1 Tyr15-phosphorylated form of U. maydis (Sgarlata and Pérez-Martín, 2005b). We found an increase of around 1.5 fold in the levels of Tyr15P-Cdk1 when the ukc1nar1 allele was repressed (Fig. 5A,B). Such an increase in the level of inhibitory phosphorylation in cells where ukc1 expression was repressed could account for a longer G2 phase.
Since levels of Cdk1 inhibitory phosphorylation in U. maydis are dependent on the protein level of Wee1 kinase (Sgarlata and Pérez-Martín, 2005b), we wondered whether the levels of Wee1 were affected when MOR pathway was disabled. Using a previously described myc-tagged Wee1 allele (Sgarlata and Pérez-Martín, 2005b), we found that the levels of Wee1 protein increased around 2.5 times upon downregulation of ukc1nar1 expression (Fig. 5A). This increase in protein level was not related to the wee1 mRNA level, which showed no significant change upon downregulation of ukc1 (Fig. 5C).
To establish a cause-effect relationship between the increase in Wee1 levels and the apparent G2 cell cycle delay in response to ukc1 knock-down, we constructed a strain in which both ukc1 and wee1 were under the control of the nar1 promoter, so expression of both genes can be downregulated at the same time (wee1 is essential in U. maydis). After 12 hours of incubation in restrictive conditions (complete medium), the apparent doubling time for nuclei in the double mutant was shorter than 3 hours (Fig. 5D), supporting the idea that the increase in Wee1 levels was responsible, at least partially, for the cell cycle delay when MOR pathway was disabled. Strikingly, we also observed that double mutant cells do not display the elongated morphology observed in MOR mutant cells (Fig. 5E).
Altogether, these results uncover an interesting connection between MOR pathway and cell cycle, as the MOR pathway seems to exert a positive control during G2/M transition. Furthermore, the unexpected finding that downregulation of wee1 expression abolishes the strong hyperpolar growth observed in MOR mutant cells, provides important insight in the understanding of the morphological outcome observed when MOR pathway is disabled (Fig. 5D).
Crk1, a regulator of polar growth in U. maydis, is responsible for the morphology of MOR mutants
Wee1 is a tyrosine kinase which main target described so far is the catalytic subunit of mitotic CDK. Recently, additional targets of Wee1, like kinesin, were described in Drosophila (Garcia et al., 2009). An appealing possibility was that the apparent dependence on Wee1 for the strong polar growth observed in U. maydis MOR mutants was related to uncharacterized Wee1 target(s) involved in morphogenesis. However, we believe that this Wee1-dependency is related to the ability of Wee1 to phosphorylate its cell cycle target Cdk1 and in consequence to delay the G2/M transition. We based this assumption on the observation that expression of cdk1AF, a Cdk1 mutant allele refractory to inhibitory phosphorylation and therefore mimicking the downregulation of wee1 with respect to the cell cycle (Sgarlata and Pérez-Martín, 2005b), phenocopied the effect of decreasing the Wee1 levels in MOR mutant cells (supplementary material Fig. S5). Expression of cdk1AF decreases the hyperpolar growth usually observed in MOR mutant cells.
In U. maydis there is a strong correlation between G2 phase and polar growth. Bud formation takes place during G2 phase and relies almost entirely on polar growth (Steinberg et al., 2001). Therefore, any condition resulting in a G2 cell cycle delay or arrest in U. maydis also correlates with a strong polar growth (Mielnichuk et al., 2009; Pérez-Martín, 2009; Pérez-Martín and Castillo-Lluva, 2008). Although the details are largely unknown, it is likely that several elements could be involved in the communication between cell cycle and polar growth during G2 phase in U. maydis. One of these elements appears to be the kinase Crk1. This MAPK-like kinase, which has roles during the mating process in U. maydis (Garrido and Pérez-Martín, 2003; Garrido et al., 2004), had strong effects on the ability of cells to sustain polar growth: cells defective in this kinase had a rounder morphology that contrasted with the elongated bud morphology of wild-type U. maydis cells, while overexpression of crk1 resulted in a dramatic elongated growth (supplementary material Fig. S6). In cells arrested in G2 phase after impairment of mitotic CDK activity, crk1 expression level dramatically raised (supplementary material Fig. S7A,B). Moreover, in the absence of Crk1 function, the strong polar growth associated to G2 cell cycle arrest was abolished (supplementary material Fig. S7C), suggesting that enhancement of polar growth after G2 cell cycle arrest could be related to the upregulation of crk1 expression.
Because of the connections between G2 phase, polar growth and Crk1, we wondered whether, as it happens in G2-arrested cells, the strong polar growth observed when MOR pathway was disabled showed any dependence on Crk1 function. We deleted the crk1 gene in ukc1nar1 cells and found to be the case: double mutant cells did not show the dramatic polar growth observed in ukc1-defective cells (Fig. 6A). Concurrently, crk1 mRNA as well as protein levels were also higher in cells where ukc1 expression was downregulated (Fig. 6B,C). These results provide a framework to understand the dramatic polar growth observed when MOR pathway is disabled: absence of MOR activity resulted in an increase of Crk1 levels, promoting the strong polar growth associated with mutations of MOR components in U. maydis. However, although the upregulation observed in the levels of crk1 mRNA can be explained by a negative regulation (direct or indirect) exerted by MOR pathway components on crk1, it also could be explained as a side effect of the observed G2 enlargement due to MOR pathway inactivation. To distinguish between these two hypotheses, we analyzed crk1 levels in cells with disabled MOR pathway (ukc1nar1 cells) in which wee1 expression was knocked down (avoiding the observed G2 enlargement consequence of MOR downregulation), and found that in these conditions the crk1 levels do not increase (Fig. 6B).
Taken together, these results indicate that part of the dramatic morphological outcome with respect to polar growth observed in U. maydis cells upon disabling the MOR pathway was directly related to the wiring between cell cycle regulation and the MOR pathway in this fungus.
Cdk1 inhibitory phosphorylation may act as an integrator of signaling cascades regulating polar growth in U. maydis
The activation/inhibition of the master regulators of Cdk1 inhibitory phosphorylation, the kinase Wee1 and the phosphatase Cdc25, resulted in the decision to enter mitosis or not. Thereby, this Cdk1 inhibitory phosphorylation has been used by several organisms (including U. maydis) as an integrator of signals related to cell size (Daga and Jimenez, 1999; Jorgensen and Tyers, 2004; Pérez-Martín et al., 2006; Rupes, 2002). The results shown in the previous paragraph strongly suggest the idea that cell cycle regulation could also act as an integrator of signals related to the ability to sustain or not polar growth in U. maydis. In such a scenario, different signals – some promoting polar growth, others opposing it – could be involved in the regulation of Cdk1 inhibitory phosphorylation, which may be acting as an integrator of the final morphological outcome.
To test this idea we took advantage of our previous work describing the connection between the cell wall integrity (CWI) pathway and cell cycle in U. maydis (Carbo and Pérez-Martín, 2010). CWI pathway over-activation in U. maydis accelerates the G2/M transition by activating the mitotic inducer Cdc25, a phosphatase that opposes the kinase activity of Wee1. Shorter G2 phase results in the inability to support polar growth, which has to be avoided in conditions of cell wall damage. We analyzed the consequences on polar growth of a simultaneous activation of CWI pathway (i.e. accelerating G2/M transition by promoting Cdk1 dephosphorylation and opposing polar growth) and inactivation of MOR pathway (i.e. prolonging G2 phase by increasing Cdk1 inhibitory phosphorylation and promoting polar growth). For that, we introduced into ukc1nar1 cells a transgene that ectopically expressed mkk1DD encoding an activated version of the CWI pathway MAPK activating kinase (Carbo and Pérez-Martín, 2010). Cells expressing the mkk1DD allele showed a clear isotropic growth, while cells in which ukc1 expression was downregulated showed hyperpolarized growth (Fig. 7A). Strikingly, MOR downregulation in conditions of CWI pathway over-activation resulted in a dramatic attenuation of the strong polar growth associated with disabling of the MOR pathway (Fig. 7A). Moreover, when the level of Cdk1 inhibitory phosphorylation in these strains was assessed, a clear correlation between the degree of Cdk1 Tyr15 phosphorylation and the extent of polar growth in the cells was observed (Fig. 7B,C).
These data allow us to propose that in U. maydis Cdk1 inhibitory phosphorylation can be used as an integrator of multiple incoming signals to determine a response regarding polar growth regulation.
The MOR pathway includes a network of interacting proteins that controls several aspects of fungal life including growth, differentiation and virulence (Maerz and Seiler, 2010). Ukc1, the central NDR kinase from the MOR pathway in U. maydis, was described previously (Dürrenberger and Kronstad, 1999), although no details of additional elements were provided at that time. In this study, we expanded the analysis of this pathway, describing Mob2, the NDR activator, as well as the rest of the common elements, including the upstream activating kinase Nak1, the scaffold-like proteins Tao3, Hym1 and the Sog2 modulator. All U. maydis MOR mutants exhibited the same class of defects indicating that they were components of the same pathway.
One of the major open questions that remain from the analysis of the fungal MOR pathway is why, in spite of MOR proteins being conserved, mutations in various fungi lead to defects ranging from loss of polarity to hyperpolarization and increased branch initiation. One of the most dramatic examples is the difference in cellular morphology between S. pombe and U. maydis when the MOR pathway is disabled. Although their cell division mode is different (one is a fission yeast while the other is a budding yeast), these two yeasts share a similar growth pattern that depends mainly on polar growth. Furthermore, both yeast showed the maximum polar growth during G2 phase, which length is controlled by Cdc2/Cdk1 inhibitory phosphorylation levels. However, while in U. maydis ukc1 mutants the polar growth is enhanced, loss-of-function mutation of orb6 in S. pombe resulted in spherical cells that lost their polarized cell shape (Verde et al., 1998). Here, we wish to propose that the distinct response observed in U. maydis MOR mutants with respect to the ability to control polar growth might be a consequence of the specific wiring between MOR pathway and cell cycle in this fungus. We found that downregulation of ukc1 expression using a conditional allele resulted in an increase of the level of Cdk1 inhibitory phosphorylation, and as result an enlargement of the G2 phase. Inhibitory phosphorylation of Cdk1 in U. maydis is dependent on the kinase Wee1, and we found higher levels of Wee1 upon downregulation of ukc1. In contrast to what we observed in U. maydis, the Orb6 kinase from S. pombe has been reported to act as a dose-dependent inhibitor of mitosis: a decrease of Orb6 levels leads to mitotic advance (shortening of G2 phase) and Orb6 overexpression can delay the onset of mitosis (enlarging the G2 phase) (Verde et al., 1998). Genetic data indicated that, as we found in U. maydis, Wee1 was also involved in this regulatory connection (Hou et al., 2003; Verde et al., 1998). It is worth noting that the distinct effect of MOR pathway in cell cycle mirrors the different morphological outcome observed upon disabling MOR pathway in these yeasts. In U. maydis, the Cdk1-Clb2 complex is the main target of Wee1-mediated inhibitory phosphorylation (García -Muse et al., 2004; Sgarlata and Pérez-Martín, 2005b). In addition to its role inducing mitosis, the Cdk1-Clb2 complex negatively controls polar growth in U. maydis. Any condition resulting in inhibition of Cdk1-Clb2 complex produces both a G2 delay/arrest as well as the induction of strong polar growth. These conditions include mutations affecting either the Cdk1-Clb2 complex directly (García-Muse et al., 2004); the balance of inhibitory phosphorylation via Wee1 or Cdc25 activities (Sgarlata and Pérez-Martín, 2005a; Sgarlata and Pérez-Martín, 2005b); the response to checkpoint activation such as the DNA damage response (Pérez-Martín, 2009); or the response to developmental cues, such as the virulence program (Mielnichuk et al., 2009). We observed that conditional removal of wee1 or expression of a cdk1 allele refractory to inhibitory phosphorylation prevented the hyperpolar growth observed in ukc1 mutants. These data strongly suggest that the extent of polar growth observed in MOR mutant could be a side consequence of the level of inhibitory phosphorylation of Cdk1. Whether this explanation could also explain the results from S. pombe orb6 mutants is unclear. In S. pombe it was concluded that Orb6 had two roles during cell cycle: to maintain polarized cell growth and to delay the onset of mitosis. For these two roles, two independent pathways were proposed: one acting on Cdc42 activity while the other acting most likely upstream Wee1 kinase (Das et al., 2009; Verde et al., 1998). However, some connections linking cell cycle and polar growth in S. pombe could also be proposed to explain the observed morphological effects of mutations in MOR pathway, since overexpression of wee1 alleviated the lack of polarity of orb6 mutants (Hou et al., 2003; Verde et al., 1998).
We are aware that although our results suggest that in U. maydis the MOR pathway/cell cycle connection occurs via Wee1 kinase, we cannot ascertain whether this is a direct (e.g. Wee1 is a target of Ukc1) or indirect effect. It is interesting to note that previous research done in S. cerevisiae indicated that loss of Cbk1 function prevented bud emergence and caused a G2 arrest, which was the consequence of the activation of the morphogenesis checkpoint (Kurischko et al., 2008). In budding yeast, morphogenesis checkpoint operates by phosphorylation of Tyr15 of Cdc28 by Swe1, the S. cerevisiae homolog of Wee1 (Lew, 2003). In fact, the G2 arrest consequence of Cbk1 inhibition in S. cerevisiae was prevented by deletion of Swe1 (Kurischko et al., 2008). One interesting possibility is that defects in MOR pathway in U. maydis triggers some sort of morphogenesis checkpoint and as a consequence the cell cycle is affected as it happens in S. cerevisiae. No studies addressed so far whether such a checkpoint morphogenesis operates in U. maydis, although homologs to Hsl1 and Hsl7 proteins are present in the genome of U. maydis.
A second major unaddressed question concerns the mechanisms why the activity of mitotic CDK negatively controls polar growth in U. maydis. The relationship between mitotic CDK and polar growth is well known in S. cerevisiae, where Cdc28 associated to Clb cyclins inhibits apical growth and triggers isotropic bud growth (Lew and Reed, 1993). Recent work showed that this inhibition involves the Cdc28-Clb-dependent phosphorylation of Lte1, which in turn act in a hitherto unknown process that prevents untimely-polarized growth by interfering with the activity of small GTPases Ras and Bud1 during G2/M (Geymonat et al., 2010). Although this kind of explanation could be also applied to U. maydis, it is likely that more than one element could be involved in the connections between cell cycle regulation and polar growth. Here we provide data indicating that in U. maydis connections between Cdk1-Clb2 complex and polar growth are mediated via the transcriptional levels of the kinase Crk1. This kinase seems to be important for sustained polar growth in U. maydis: overexpression induces a strong polar growth while in cells lacking its function polar growth was attenuated. Moreover, in the absence of Crk1 protein, we observed that G2 arrest (upon downregulation of clb2 expression) does not result in hyperpolarized growth. We also noted a negative correlation between the activity of Cdk1-Clb2 complex and crk1 mRNA levels: a low activity of Cdk1-Clb2, either because of downregulation of clb2 or overexpression of wee1, resulted in high levels of crk1 expression, while a high activity of Cdk1-Clb2 such as downregulation of wee1 correlated to low crk1 mRNA levels. These data provide a framework that further supports our conclusion that hyperpolarized growth in U. maydis MOR mutants was a side consequence of the distinct cell cycle/MOR pathway connections. We found that crk1 was required for the hyperpolarized growth upon downregulation of ukc1 expression. Moreover, crk1 mRNA as well as protein levels were higher in these conditions establishing a cause-effect relationship. However both the hyperpolarized growth and the increase of Crk1 levels were dependent on the ability of cells to phosphorylate Cdk1 since concomitant downregulation of wee1 abolished both effects.
Besides the intricacies that explain the connections between MOR pathway, cell cycle and morphogenesis in U. maydis, one of the major conclusion of this work is that in this fungus the phosphorylation status of mitotic CDK appears to be acting as an integrator of signaling cascades with outcomes affecting, among other processes, the ability of the cell to support strong polar growth. Different signaling cascades seem to converge on the regulation of Cdk1 inhibitory phosphorylation in U. maydis. These cascades include the b-dependent pathogenic program (de Sena- Tomás et al., 2011; Mielnichuk et al., 2009), DNA damage response (Mielnichuk and Pérez-Martín, 2008; Pérez-Martín, 2009), the Cell Wall Integrity pathway (Carbo and Pérez-Martín, 2010) and the pheromone response pathway (S. Castanheira and J.P.-M., unpublished data). In general cell cycle has been considered as a final effector, which is accordingly modified upon previous integration of the various signaling cascades in the cell. In this work we challenge this view. Here we showed that in U. maydis the final outcome of two different signaling pathways (downregulation of MOR pathway and upregulation of CWI pathway) affecting the ability to sustain or not polar growth correlated with the level of inhibitory phosphorylation of Cdk1. Our results suggest that it is the integration of the activity of regulators promoting and inhibiting this phosphorylation that produces the final outcome with respect to polar growth. The ability to integrate distinct signals into the level of Cdk1 inhibitory phosphorylation is well established regarding the regulation of cell size in a number of organisms (including U. maydis). Here we propose that a similar mechanism of signal integration could account for the induction of hyperpolarized growth in U. maydis and we believe that this idea could be also applied to other fungi.
Materials and Methods
Strains and growth conditions
Ustilago maydis strains are listed in supplementary material Table S1 and are derived from FB1 background (Banuett and Herskowitz, 1989). Media were prepared as described (Holliday, 1974). Controlled expression of genes under the crg1 and nar1 promoter was performed as described previously (Brachmann et al., 2001; García-Muse et al., 2004). FACS analyses were described previously (García-Muse et al., 2003).
DNA, RNA and protein analysis
U. maydis DNA isolation was performed as previously described (Tsukuda et al., 1988).
RNA isolation and northern analysis were performed as described previously (Castillo-Lluva et al., 2004). Protein extracts and western blotting were performed as described previously (Garrido et al., 2004; Sgarlata and Pérez-Martín, 2005b). All quantification was done using a Phosphorimager (Molecular Dynamics). To detect the phosphorylated and non-phosphorylated forms of Cdk1 commercial antibodies were used as described (Sgarlata and Pérez-Martín, 2005b). Primary antibody was followed by a secondary antibody conjugated to horseradish peroxidase and immunoreactive proteins were visualized using a chemiluminescent substrate. The chemiluminescent signal was analyzed using ChemiDoc XCS+ (Molecular Imager, Bio-Rad).
Plasmid and strain construction
Plasmid pGEM-T easy (Promega) was used for subcloning and sequencing of genomic fragments and fragments generated by PCR. Sequence analysis of fragments generated by PCR was performed with an automated sequencer (ABI 373A) and standard bioinformatic tools. To construct the different strains, transformation of U. maydis protoplasts with the indicated constructions was performed as described previously (Tsukuda et al., 1988). Gene replacement into the corresponding loci was verified by diagnostic PCR and subsequent Southern blot analysis.
Deletion of each gene was performed using PCR-based gene targeting (Brachmann et al., 2004). Briefly, a pair of DNA fragments flanking the corresponding kinase ORF were amplified and ligated to antibiotic resistance cassettes via SfiI sites. The 5′ and 3′ fragments were amplified using specific oligonucleotide pairs (supplementary material Table S2). Each flanking fragment was about 1 kb in length.
To produce a conditional ukc1nar1 allele, we constructed a plasmid by ligation of a pair of fragments into pRU2-HYG, digested with NdeI and EcoRI. The ukc1 N-terminal region (flanked by EcoRI and KpnI) was produced by PCR using the primers ukc1-2 and ukc1-3. The ukc1 promoter region (flanked by NdeI and KpnI) was obtained by PCR amplification with primers ukc1-1 and ukc1-4. The resulting plasmid pUKC1nar1HYG was integrated, after digestion with KpnI, by homologous recombination into the U. maydis genome.
To construct the ukc1-GFP allele, the C-terminal part of the ukc1 ORF without the stop codon was amplified with the primer par ukc1-14/ukc1-17 and ligated to the p123 plasmid. The ukc1 fragment was placed in phase with the gfp gene. This plasmid (p123-ukc1) was digested by SacI and introduced by homologous recombination into the U. maydis genome.
To construct the ukc1-myc allele, the C-terminal part of the ukc1 ORF without the stop codon was amplified with the ukc1-11/ukc1-12 primer pair and ligated to the pBS-MYC plasmid in which it was cloned in phase with three copies of the myc epitope. This plasmid (pBS-ukc1-myc) was digested by EcoNI and introduced by homologous recombination into the U. maydis genome.
To construct the t7-mob2 allele, the mob2 ORF was amplified by PCR with mob2-15/mob2-5 primer pairs. The ATG codon start was exchanged with a BamHI site where a T7 epitope was inserted in the same translational phase. The resulting plasmid (pT7-mob2) was digested by SspI and introduced by homologous recombination into the U. maydis genome.
Samples were mounted on microscope slides and visualized in a Nikon Eclipse 90i microscope equipped with a Hamamatsu ORCA-ER CCD camera. All the images in this study are single plane. Standard DAPI, Rhodamine and GFP filter sets were used for epifluorescence analysis. The software used with the microscope was MetaMorph 7.1 (Universal Imaging, Downingtown, PA). Images were further processed with Adobe Photoshop 8.0.
Supplementary material available online at http://jcs.biologists.org/lookup/suppl/doi:10.1242/jcs.107862/-/DC1
We are indebted to Prof. Jimmy Correa-Bordes (UNEX, Spain) for critical reading and suggestions about this manuscript. We also thank members of Signalpath and Ariadne Consortium for helpful discussions.
↵*Present address: Department of Molecular and Cell Biology, University of California, Berkeley, CA 94720, USA
This work was supported by the Spanish government [grant number BIO2011-27773 to J.P.M.]; and the European Union [grant numbers MRTN-CT-2005-019277 to J.P.M., PITN-GA-2009-237936 to J.P.M.].
Supplementary material available online at http://jcs.biologists.org/lookup/suppl/doi:10.1242/jcs.107862/-/DC1
- Accepted June 6, 2012.
- © 2012. Published by The Company of Biologists Ltd