Cytoplasmic retention and degradation of a mitotic inducer enable plant infection by a pathogenic fungus

*For correspondence: [email protected] Present address: † Department of Biology, New York University, New York, United States; ‡ Centro Nacional de Biotecnologı ´a (CSIC), Madrid, Spain Competing interests: The authors declare that no competing interests exist. Funding: See page 29 Received: 31 May 2019 Accepted: 16 October 2019 Published: 17 October 2019 Reviewing editor: Christian S Hardtke, University of Lausanne, Switzerland Copyright Bardetti et al. This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited. Cytoplasmic retention and degradation of a mitotic inducer enable plant infection by a pathogenic fungus Paola Bardetti 1† , So ´nia Marisa Castanheira 1‡ , Oliver Valerius 2 , Gerhard H Braus 2 , Jose ´Pe ´rez-Martı´n 1 * 1 Instituto de Biologı ´a Funcional y Geno´ mica (CSIC), Salamanca, Spain; 2 Department of Molecular Microbiology and Genetics, Institute for Microbiology and Genetics, Georg-August-University, Go ¨ttingen, Germany Abstract In the fungus Ustilago maydis, sexual pheromones elicit mating resulting in an infective filament able to infect corn plants. Along this process a G2 cell cycle arrest is mandatory. Such as cell cycle arrest is initiated upon the pheromone recognition in each mating partner, and sustained once cell fusion occurred until the fungus enter the plant tissue. We describe that the initial cell cycle arrest resulted from inhibition of the nuclear transport of the mitotic inducer Cdc25 by targeting its importin, Kap123. Near cell fusion to take place, the increase on pheromone signaling promotes Cdc25 degradation, which seems to be important to ensure the maintenance of the G2 cell cycle arrest to lead the formation of the infective filament. This way, premating cell cycle arrest is linked to the subsequent steps required for establishment of the infection. Disabling this connection resulted in the inability of fungal cells to infect plants. DOI: https://doi.org/10.7554/eLife.48943.001 Introduction Sexual reproduction is widely conserved within the eukaryotic life tree. The final outcome of this process is the fusion of two distinct haploid nuclei into a single diploid nucleus. To avoid an imbalance in the nuclear genetic information provided by each mating partner, the cell cycle status of both nuclei should be the same before karyogamy. In metazoans, this synchronization often occurs once the two partner nuclei are in the same zygotic cytoplasm (Austin, 1978). However, in simple eukaryotes such as fungi or unicellular algae, cell cycle synchronization occurs before cell fusion (Hartwell, 1973). In these organisms, cell cycle synchronization is mediated by the recognition of signals, often pheromones secreted by distinct mating partners. The paradigmatic and best studied case is the budding yeast Saccharomyces cerevisiae. In this fungus, pheromone recognition is mediated by plasma membrane-located receptors, which transmit the signal toward the cell cycle machinery using a widely conserved MAP kinase cascade (Bardwell, 2005). Pheromone recognition in budding yeast results in G1 cell cycle arrest, which is maintained throughout all cell fusion process, resulting in a diploid zygote that is able to resume the cell cycle or enter into meiosis (Elion, 2000). Premating cell cycle synchronization at G1 phase seems to be the rule, as other fungi such as Schizosaccharomyces pombe (Davey, 1998), diatoms (Moeys et al., 2016), and -most likelyalgae such as Chlamydomonas reinhardtii (Joo et al., 2017) and the slime mold Dictyostelium discoideum (Ishida et al., 2005) apply the same principle. However, there is one exception in the fungal maize smut pathogen Ustilago maydis, where premating cell cycle synchronization in response to secreted sexual pheromones occurs at G2 phase (Garcı´a-Muse et al., 2003). Paradoxically, in spite of the distinct cell cycle stage for arrest, the elements involved in the transmission process (pheromone and receptors, MAPK cascade and transcription factors) are similar to those described in fungi that undergo arrest at G1 Bardetti et al. eLife 2019;8:e48943. DOI: https://doi.org/10.7554/eLife.48943 1 of 33 RESEARCH ARTICLE phase (Mu ¨ller et al., 2003;Vollmeister et al., 2012). This result strongly suggested that the differences were linked to alternative wiring of the U. maydis pheromone MAPK cascade with cell cycle regulators, although these connections were largely unknown. The reasons for the distinct cell cycle response to pheromone in U. maydis are likely related to the unusual developmental steps that mating triggers in this fungal system. In U. maydis, virulence and sexual development are intricately interconnected because the mating of two compatible budding haploid cells is the prerequisite to induce the infectious stage (Brefort et al., 2009; Vollmeister et al., 2012). Pathogenic development is mediated by two independent loci: the alocus, which encodes a pheromone-receptor system, and the b-locus, which encodes a pair of homeoproteins (bW and bE). On the plant surface, infection is initiated upon the recognition of mating pheromone secreted by haploid cells of the opposite mating type (Bo ¨lker et al., 1992). This recognition induces G2 cell cycle arrest as well as the formation of long conjugation tubes (Garcı´aMuse et al., 2003;Spellig et al., 1994), which grow toward each other and fuse at their tips (Snetselaar et al., 1996). Cytoplasmic fusion is not followed by karyogamy, resulting in a dikaryotic cell. After cell fusion on the plant surface, the G2 cell cycle arrest is sustained, and the single dikaryotic cell grows in a polar manner, producing the infective filament. This hypha expands, accumulating the cytoplasm at the tip of the filament, whereas the distal parts of the hypha become vacuolated and are sealed off by the insertion of regularly spaced septa, resulting in the formation of characteristic empty sections (Steinberg et al., 1998). This growth mode enables the fungus to progress along the plant surface, most likely to find an appropriate point of entry. Eventually, the hyphae stop polar growth in response to an as yet unidentified signal, and their tips swell to form appressoria and penetrate the cuticle (Snetselaar and Mims, 1992;Snetselaar and Mims, 1993). eLife digest Many fungi that cause diseases in plants need specialized structures to penetrate the plant’s tissues. To form these structures, the fungus must carefully control when and where its cells divide. As in other organisms, the sequence of events that lead to a fungal cell dividing in two are known as the cell cycle. Progress through the distinct steps in the cell cycle is regulated by enzymes including many that add or remove phosphate groups on other proteins. It remains unclear which regulatory enzymes allow any plant-infecting fungus to control its cell cycle when it forms an infection structure, but one fungus that could help answer this question is Ustilago maydis, the cause of a disease known as corn smut. The corn smut fungus forms infection structures after two different mating strains meet on the surface of the plant, stop dividing and then fuse. This implies that the cell cycles of both strains need to be coordinated to allow the fungus to infect the plant. The two strains recognize each other via chemical signals known as pheromones, and Bardetti et al. now show that pheromone recognition in the corn smut fungus results in an enzyme called Cdc25 being disabled, which in turn causes cell division to stop. Cdc25 is a phosphatase, meaning it removes phosphate groups from cell cycle regulators that are found in the nucleus of the cell. Specifically, Cdc25 targets phosphate groups that would otherwise inhibit the activity of these proteins. Further experiments showed that, following pheromone recognition, Cdc25 is disabled via a two-step process: first it is prevented from entering the nucleus which keeps it away from its targets, and then it is degraded. Bardetti et al. went on to show that this last step was required for the fungus to infect corn plants, since interfering with the breakdown of Cdc25 impairs its ability to stop the cell cycle and form an infection structure. Entry into plant tissue is a critical step for any parasites looking to invade a plant. Since it is difficult to reach the interior of plants with pesticides, most antimicrobial treatments in plants aim at prevention rather than cure. This means that increasing the delay between a fungus recognizing the surface of a plant and penetrating its tissues could give more time to prevent infections. These new findings represent a step towards achieving that goal, though more research is needed to better understand the molecular mechanisms required for the formation of infection structures in plantinfecting fungi. DOI: https://doi.org/10.7554/eLife.48943.002 Bardetti et al. eLife 2019;8:e48943. DOI: https://doi.org/10.7554/eLife.48943 2 of 33 Research article Developmental Biology Microbiology and Infectious Disease Once the filament enters the plant, the cell cycle is reactivated, and the fungus proliferates inside the plant. During the differentiation process resulting in plant penetration, the presence of a sustained G2 cell cycle arrest is mandatory and the impairment of this cell cycle arrest resulted in the inability of the fungus to infect plants (Castanheira and Pe ´rez-Martı´n, 2015). This cell cycle arrest is imposed first by the activation of the pheromone cascade and then maintained during the growth of the infective filament by a transcriptional regulator called b-factor, which is encoded in the b-locus and composed of two subunits (bW and bE), provided by each mating partner. The mechanisms involved in this sustained cell cycle arrest have been described only partially. While the manner by which the presence of b-factor arrests the cell cycle is comprehended in detail, the molecular intricacies associated with the cell cycle arrest induced in response to pheromone are still unknown. G2/M transition in U. maydis is regulated by the presence of two distinct cyclin-dependent kinase (CDK) complexes: Cdk1-Clb1 and Cdk1-Clb2 (Garcia-Muse, 2004). Of these, the limiting step is provided by the activity of the Cdk1-Clb2 complex, which is controlled by the inhibitory phosphorylation of Cdk1. The level of this phosphorylation depends on the relative activity of the Wee1 kinase (which inhibits Cdk1) and the Cdc25 phosphatase (which activates Cdk1) (Sgarlata and Pe ´rez-Martı´n, 2005a;Sgarlata and Pe ´rez-Martı´n, 2005b). Not surprisingly, the mechanism by which the b-factor arrests the cell cycle at G2 during the growth of the dikaryotic infective filament relies on the increase of Cdk1 inhibitory phosphorylation: The b-factor activates the DNA damage response (de Sena-Toma ´s et al., 2011;Mielnichuk et al., 2009) in the absence of DNA damage (TenorioGo ´mez et al., 2015), resulting in the phosphorylation of Cdc25, promoting thereby its interaction with 14-3-3 proteins, which in turn inactivates the phosphatase by its retention in the cytoplasm (Mielnichuk and Pe ´rez-Martı´n, 2008); at the same time, the b-factor represses the transcription of hsl1, which encodes a kinase that downregulates Wee1 kinase, increasing as a consequence the level of inhibitory phosphorylation of Cdk1 (Castanheira et al., 2014). Moreover, a second cell cycle brake is added during the formation of the infective filament since the b-factor also activates the transcription of biz1, a transcriptional regulator that represses the transcription of clb1, encoding the b cyclin required for the second Cdk1-cyclin complex involved in G2/M transition (FlorParra et al., 2006;Garcia-Muse, 2004). Here we tried to uncover the elements required for the pheromone-induced cell cycle arrest. Although it was possible that the pheromone response shared the same regulatory scheme as the b-induced cell cycle arrest, we show that this is not entirely the case and that some of the elements involved are different. We report that the pheromone response MAPK cascade is distinctly wired to cell cycle regulation, resulting first in the inhibition of the nuclear localization of Cdc25, via importin phosphorylation, and second, once the MAPK signaling reaches a threshold, in the degradation of the accumulated Cdc25. These steps were found to be required to ensure the maintenance of cell cycle arrest during the infection process. Inability to do so resulted in a strong defect in virulence. Results Activation of the pheromone cascade promotes G2 cell cycle arrest via inhibitory phosphorylation of Cdk1 We sought to address the molecular mechanisms behind the G2 cell cycle arrest observed upon pheromone response in U. maydis. This response requires, in each mating partner, the recognition of the compatible pheromone by its cognate receptor and the transmission of the signal through a conserved MAPK cascade (Mu ¨ller et al., 2003;Vollmeister et al., 2012). However, the expression of pheromone and pheromone receptor genes requires poor nutritional conditions, which enables the activity of the transcriptional regulator Prf1 that activates the promoters from the a-locus (Hartmann et al., 1999;Kaffarnik et al., 2003) (Figure 1—figure supplement 1A). Since changes in nutritional conditions could alter the cell cycle pattern in this fungus (Pe ´rez-Martı´n et al., 2006), we took advantage of the previous description of an activated allele of the pheromone cascade MAPKK Fuz7 (fuz7 DD ,Figure 1—figure supplement 1B), whose expression faithfully recapitulates the pheromone response in U. maydis (Mu ¨ller et al., 2003;Zarnack et al., 2008). In this way, we make the activation of the pheromone MAPK cascade independent of the elements located Bardetti et al. eLife 2019;8:e48943. DOI: https://doi.org/10.7554/eLife.48943 3 of 33 Research article Developmental Biology Microbiology and Infectious Disease upstream of this cascade (i. e. receptors and pheromones) allowing us to focus on the connections between the pheromone response MAPK cascade and cell cycle regulators. When an ectopic copy of the fuz7 DD allele was expressed under the control of the crg1 promoter (induced by arabinose and repressed by glucose) (Figure 1—figure supplement 1C and D), it mimicked the G2 cell cycle arrest observed when pheromone is sensed by U. maydis (Garcı´aMuse et al., 2003): cells accumulate 2C DNA content, carrying a single nucleus with an intact nuclear membrane (U. maydis breaks down its nuclear envelope at mitosis; Straube et al., 2005) (Figure 1A and B). Furthermore, this cell cycle arrest was dependent on Kpp2, the downstream MAPK, but independent of Prf1 (Figure 1—figure supplement 1E). We observed that the expression of the fuz7 DD allele was correlated with an increase in the level of inhibitory phosphorylation of Cdk1, which has been reported to be associated with G2 cell cycle arrest in U. maydis (Sgarlata and Pe ´rez-Martı´n, 2005a;Sgarlata and Pe ´rez-Martı´n, 2005b) (Figure 1C and Figure 1—figure supplement 2A). Moreover, the impairment of Cdk1 inhibitory phosphorylation, either by the expression of cdk1 AF , an allele refractory to inhibitory phosphorylation, or by the downregulated expression of wee1, the cognate kinase responsible for inhibitory phosphorylation (Sgarlata and Pe ´rez-Martı´n, 2005b), abrogated the fuz7 DD -dependent cell cycle arrest (Figure 1D and Figure 1—figure supplement 2B–F). These results supported the notion that the G2 cell cycle arrest associated with the activation of the pheromone cascade was dependent on the inhibitory phosphorylation of Cdk1. The molecular mechanisms for pheromone cascade-mediated cell cycle arrest are likely to be different from those described for b-dependent cell cycle arrest The mechanism of cell cycle arrest induced by the b-factor, which is responsible for the sustained G2 arrest during the growth of the infective filament, also involves an increase in the level of Cdk1 inhibitory phosphorylation (Mielnichuk et al., 2009). Moreover, in agreement with a previous report (Zarnack et al., 2008), we observed that the transcription of hsl1, a negative regulator of the Wee1 kinase, was strongly downregulated upon expression of the fuz7 DD allele (Figure 1—figure supplement 3A), as it occurs in b-dependent cell cycle arrest (Castanheira et al., 2014;Heimel et al., 2010). These findings prompted us to think that pheromone-dependent and b-dependent cell cycle arrests might share the same molecular mechanisms. To add further support to this idea, we analyzed the involvement in the fuz7 DD -dependent cell cycle arrest of other elements required for b-dependent cell cycle arrest, like the Chk1 kinase. However, our results indicated that Chk1 was not involved in fuz7 DD -dependent cell cycle arrest (Figure 1—figure supplement 3B). In the same way, it was already reported that expression of the fuz7 DD allele does not upregulate biz1 expression (Flor-Parra et al., 2006;Zarnack et al., 2008), suggesting different mechanisms for pheromoneand b-dependent cell cycle arrest. The target of Cdk1 inhibitory phosphorylation in U. maydis is the Cdk1-Clb2 complex, which is the main regulatory control for G2/M transition (Garcia-Muse, 2004;Sgarlata and Pe ´rez-Martı´n, 2005b). The amount of Cdk1 inhibitory phosphorylation depends on the relative activity levels of the Wee1 kinase and the Cdc25 phosphatase (Sgarlata and Pe ´rez-Martı´n, 2005a). We analyzed the effects of expression of the fuz7 DD allele on the levels of these regulators (Figure 1E). We observed that the protein levels of Cdc25 dropped abruptly upon expression of the fuz7 DD allele, although this decrease in Cdc25 levels cannot be attributed to a decrease in the mRNA levels of its gene (Figure 1—figure supplement 3A). Since the observed downregulation of Cdc25 levels could account for G2 cell cycle arrest (Sgarlata and Pe ´rez-Martı´n, 2005a), we aimed to bypass the fuz7 DD -dependent cell cycle arrest by overexpression of cdc25 at the same time as expression of the fuz7 DD allele. However, contrary to our expectations, overexpression of an ectopic copy of cdc25 does not abrogate cell cycle arrest, in spite of the presence of high protein levels of Cdc25 in these conditions (Figure 1F and Figure 1— figure supplement 4). This result not only strongly suggested that the decrease of Cdc25 levels was not the cause of the cell cycle arrest, but it also supported our view of distinct molecular mechanisms between pheromoneand b-factor-induced cell cycle arrest: While high levels of Cdc25 do not affect the ability to arrest the cell cycle by the pheromone cascade activation, in the b-dependent cell cycle arrest, Cdc25 is retained at the cytoplasm by Bmh1 (14-3-3 protein), and this repression can be overwhelmed by high levels of Cdc25 (i. e. overexpressing Cdc25) (Mielnichuk et al., 2009). Bardetti et al. eLife 2019;8:e48943. DOI: https://doi.org/10.7554/eLife.48943 4 of 33 Research article Developmental Biology Microbiology and Infectious Disease Figure 1. Expression of fuz7 DD allele promotes a G2 cell cycle arrest that depends on Cdk1 inhibitory phosphorylation. (A) Cells expressing the fuz7 DD allele accumulated with a 2C DNA content. Fluorescence/Activated Cell Sorter (FACS) analysis of the DNA content of a control strain and a strain carrying an ectopic copy of the fuz7 DD allele under the control of the crg1 promoter growing in inducing (Complete Medium Arabinose, CMA) and non-inducing (Complete Medium Glucose, CMD) conditions (Figure 1—figure supplement 1). The period of incubation in testing media is indicated Figure 1 continued on next page Bardetti et al. eLife 2019;8:e48943. DOI: https://doi.org/10.7554/eLife.48943 5 of 33 Research article Developmental Biology Microbiology and Infectious Disease In summary, these previous results indicated that some elements, like the downregulation of hsl1 (and thereby the upregulation of Wee1) seemed to be shared by band pheromone cascadeinduced cell cycle arrest. However, other elements, like the downregulation of biz1 and the Chk1mediated retention of Cdc25 at cytoplasm seemed to be unique for b-dependent cell cycle arrest. Pheromone cascade-induced cell cycle arrest depends on an alternative cyclin interacting with a Cdk-like kinase In our search for elements connected to the pheromone cascade that could be involved in the induction of the cell cycle arrest, we recalled the gene pcl12, which has been reported to be strongly induced by the pheromone MAPK (Flor-Parra et al., 2007). This gene encodes a cyclin from the Pcl family (Measday et al., 1997), and its ectopic expression under a regulatable promoter (such as the crg1 promoter) was sufficient to induce the formation of a cell structure resembling a conjugation tube, which was also arrested at G2 phase (Flor-Parra et al., 2007). For that reason, we were curious about the involvement of this protein in the pheromone cascade-induced cell cycle arrest. Indeed, we found that Pcl12 was required for the induction of cell cycle arrest as well as for the observed decrease in Cdc25 levels upon the expression of the fuz7 DD allele (Figure 2A and B). Pcl12 forms a complex with the essential cyclin-dependent kinase Cdk5 (Castillo-Lluva et al., 2007). The Pcl12-Cdk5 complex is required for sustained polar growth during the formation of the conjugation tubes in response to pheromone treatment (Flor-Parra et al., 2007). However, we found that Cdk5 was not required for fuz7 DD -induced cell cycle arrest (Figure 2—figure supplement 1). Since Pcl12 is a cyclin, this result suggested the existence of alternative partners (most likely kinases) for Pcl12 during pheromone cascade-induced cell cycle arrest. To identify these putative partners, we performed coimmunoprecipitation coupled with liquid chromatography-mass spectrometry (LC/MS) analysis of a GFP-tagged Pcl12 version in the presence of the expression of the fuz7 DD allele. We found the kinase Crk1 among the major peptides copurifying with Pcl12 (Figure 2—figure supplement 2,Figure 2—source data 1). Crk1 (Cdk-Related Kinase 1) was previously described as Figure 1 continued (hours). (B) Cells expressing the fuz7 DD allele induce conjugative hyphae that are arrested in G2 phase. Representative image of cells expressing the fuz7 DD allele and carrying NLS-GFP and Cut11-Cherry fusions to detect the nucleus and the nuclear envelope, growing in CMA for 6 hr. This image was a composition from various images to show different stages during the production of the conjugation hyphae. Bar: 15 mm. (C) Cells expressing the fuz7 DD showed increased levels of Cdk1 inhibitory phosphorylation (Cdk1Y15P). Data acquisition is described in Figure 1—figure supplement 2A and. Means are shown (Figure 1—source data 1). (D) Interfering with the Cdk1 inhibitory phosphorylation resulted in inability to arrest cell cycle upon fuz7 DD allele expression. Fuz7 DD -derived strains carrying the NLS-GFP reporter as well as the indicated mutations were incubated in inducing conditions (CMA) for 6 hr. Filaments were sorted as carrying 1, 2 or 3 and more nuclei. The graph shows the result from three independent experiments, counting more than 100 filaments each. Means and SDs are shown (Figure 1—figure supplement 2 and Figure 1—source data 2). (E) Protein levels of G2/M regulators upon fuz7 DD allele expression. Strains carrying HA-tagged versions of Clb2, Cdc25 and Wee1 and carrying the fuz7 DD allele or not (control) were incubated for the indicated time in induction conditions (CMA). Similar amount of protein extracts was separated by SDSPAGE. Immunoblots were incubated with an antibody against HA. As loading control, we used the Cdk1 protein, which can be detected using antiPSTAIRE (which recognizes both Cdk1 and Cdk5). (F) Overexpression of cdc25 does not abrogate the fuz7 DD -dependent cell cycle arrest. Representative images of cultures growing in inducing conditions (CMA) for 6 hr, from a control strain expressing the fuz7 DD allele, and a strain coexpressing both the fuz7 DD allele and an ectopic copy of cdc25 (Figure 1—figure supplement 4). DOI: https://doi.org/10.7554/eLife.48943.003 The following source data and figure supplements are available for figure 1: Source data 1. Data for Figure 1C. DOI: https://doi.org/10.7554/eLife.48943.008 Source data 2. Data for Figure 1D. DOI: https://doi.org/10.7554/eLife.48943.009 Figure supplement 1. Activation of the pheromone response cascade upon expression of the fuz7 DD allele. DOI: https://doi.org/10.7554/eLife.48943.004 Figure supplement 2. Fuz7 DD -dependent cell cycle arrest requires inhibitory phosphorylation of Cdk1. DOI: https://doi.org/10.7554/eLife.48943.005 Figure supplement 3. The mechanism of fuz7 DD -dependent cell cycle arrest is unrelated to b-dependent cell cycle arrest. DOI: https://doi.org/10.7554/eLife.48943.006 Figure supplement 4. Overexpression of cdc25 does not abrogate the fuz7 DD -dependent cell cycle arrest. DOI: https://doi.org/10.7554/eLife.48943.007 Bardetti et al. eLife 2019;8:e48943. DOI: https://doi.org/10.7554/eLife.48943 6 of 33 Research article Developmental Biology Microbiology and Infectious Disease Figure 2. Fuz7 DD -induced cell cycle arrest depends on an alternative cyclin interacting with a Cdk-like kinase. (A) The cyclin Pcl12 and the Cdk-like kinase Crk1 were required for Fuz7 DD -dependent cell cycle arrest. Representative images of cultures of strains carrying the fuz7 DD allele and the indicated mutations. Cultures were incubated for 12 hr in inducing conditions for fuz7 DD (CMA). Cells carried a constitutively expressed NLS-GFP reporter to detect nuclei and were stained with Calcofluor White (CFW) to detect septa. Note that filaments in the mutants were composed of cell compartments carrying one nucleus each and separated by septa. Bar: 20 mm. (B) The cyclin Pcl12 and the Cdk-like kinase Crk1 were required for Fuz7 DD -dependent decrease of Cdc25 levels. Western blot analysis to show the level of Cdc25 (upper blot) upon expression of fuz7 DD allele in cells growing in inducing conditions (CMA) for the indicated time. Levels of Cdk1 were used as loading control (bottom blot). (C) Ability to arrest the cell cycle upon expression of fuz7 DD or pcl12 in distinct mutant strains. The indicated strains, which also carried the NLS-GFP transgene, were incubated in inducing conditions (CMA) for 6 hr. Filaments from each culture were counted and sorted as carrying 1 (cell cycle arrested) or more than one nucleus (not arrested). The graph shows the result from three independent experiments, counting more than 100 filaments each. Means and SDs are shown (Figure 2—source data 2). Representative images corresponding to the respective cultures could be found at Figure 2E and F as well as at Figure 2—figure supplement 3. (D) Scheme of Crk1, showing the mutant alleles used in this work. These mutants were already described: crk1 KD , is a kinase-dead loss of function mutant; crk1 AEF is refractory to T-loop activation by Fuz7; crk1 AAA is refractory to phosphorylation by the MAPK Kpp2 (Garrido et al., 2004). (E) Crk1 is required for cell cycle arrest promoted upon expression of pcl12. Representative images of cultures of strains carrying an ectopic copy of pcl12 under crg1 promoter and the indicated mutations. Crk1 KD carried the K145A mutation that inactivates its kinase catalytic activity (Garrido et al., 2004). Cultures were incubated for 6 hr in inducing conditions for pcl12 (CMA). Cells carried a constitutively expressed NLS-GFP reporter to detect nuclei and were stained with Calcofluor White (CFW) to detect septa. Bar: 15 mm. (F) Representative images of cultures of strains carrying the fuz7 DD allele as well as the indicated mutations. Cultures were incubated for 6 hr in inducing conditions for fuz7 DD (CMA). Cells carried a constitutively expressed NLS-GFP reporter to detect nuclei and were stained with Calcofluor White (CFW). Bar: 15 mm. DOI: https://doi.org/10.7554/eLife.48943.010 The following source data and figure supplements are available for figure 2: Figure 2 continued on next page Bardetti et al. eLife 2019;8:e48943. DOI: https://doi.org/10.7554/eLife.48943 7 of 33 Research article Developmental Biology Microbiology and Infectious Disease a regulator of polar growth since its overexpression induces cell filamentation (Garrido and Pe ´rezMartı´n, 2003). In support of a functional role of the observed Pcl12 and Crk1 physical interaction, we found that loss of function of Crk1 abrogated fuz7 DD -dependent cell cycle arrest as well as the decrease in Cdc25 levels in a similar manner as the pcl12 mutant did (Figure 2A and B). Furthermore, we also observed that Crk1 was also required for cell cycle arrest upon ectopic expression of pcl12 (Figure 2C and E). These results sustained the idea that Crk1 and Pcl12 were working together during the pheromone cascade-induced cell cycle arrest, most likely as a complex. Interestingly, Crk1 activation was previously described to be linked to the pheromone cascade: In order to promote hyperpolarized growth, Crk1 must be activated via phosphorylation of its T-loop by the MAPKK Fuz7, as well as by phosphorylation of the C-terminal end by the MAPK Kpp2 (Garrido et al., 2004). Therefore, we wondered whether these described connections between Crk1 and the pheromone cascade during the promotion of filamentous growth were also required for fuz7 DD -induced cell cycle arrest. To address this question, we took advantage of previously described mutant alleles of crk1 refractory to phosphorylation by Fuz7 (crk1 AEF ) or by Kpp2 (crk1 AAA ) (Figure 2D) (Garrido et al., 2004). Much to our surprise, we found that cells carrying crk1 AEF or crk1 AAA mutant alleles were able to arrest the cell cycle upon the expression of the fuz7 DD allele (Figure 2F). Moreover, we also found that the ectopic expression of pcl12 was still able to arrest the cell cycle in cells carrying different mutations affecting the phosphorylation of Crk1 by the MAPK cascade, such as the loss-of-function in fuz7 or kpp2, or crk1 alleles refractory to phosphorylation (Figure 2C and Figure 2—figure supplement 3). These results indicated that the described activation of Crk1 via the pheromone cascade during filamentation played no role in the cell cycle arrest induced by the same MAPK cascade. To explain this apparent paradox, we propose that Crk1, which is a kinase that have features of both CDK-like and MAPK-like kinases (see discussion section for a detailed description of this hypothesis), can be activated by two distinct manners: either as a MAPK, by T-loop phosphorylation through the pheromone MAPK cascade, to support polar growth; or as a CDK, by interaction with Pcl12 cyclin, to support cell cycle arrest. Distinct activation mechanisms probably also involve distinct targets explaining the distinct outcomes. Both Crk1 and Pcl12 were previously described as regulators involved in morphogenesis of the conjugative tube (Flor-Parra et al., 2007;Garrido et al., 2004), although the relationships between these factors at this level were not studied. We observed that, in contrast to the cell cycle arrest, which is abrogated by a single mutation in either pcl12 or crk1, the effects of each gene mutation on the morphology of the resulting filament upon expression of fuz7 DD allele were distinct, and the double mutant was more affected than single mutants (Figure 2—figure supplement 4A). On the basis of these genetic interactions, we propose that Pcl12 and Crk1 work together, most likely forming a complex, and acting on the pheromone cascade-induced cell cycle arrest. However, it seems that these proteins also work in parallel pathways during the control of conjugation tube morphogenesis, most likely through the Pcl12-Cdk5 complex in one pathway, and through Crk1 receiving MAPK cascade signals in the other (Figure 2—figure supplement 4B). Figure 2 continued Source data 1. Data from LC/MS. DOI: https://doi.org/10.7554/eLife.48943.015 Source data 2. Data for Figure 2C. DOI: https://doi.org/10.7554/eLife.48943.016 Figure supplement 1. Cdk5 is not required for cell cycle arrest upon expression of the fuz7 DD allele. DOI: https://doi.org/10.7554/eLife.48943.011 Figure supplement 2. Protein interacting with Pcl12-GFP. DOI: https://doi.org/10.7554/eLife.48943.012 Figure supplement 3. Cell cycle induced by the ectopic expression of pcl12 is independent on MAPK-mediated phosphorylation of Crk1. DOI: https://doi.org/10.7554/eLife.48943.013 Figure supplement 4. Loss-of-function mutations in crk1 and pcl12 showed additive defects in morphology. DOI: https://doi.org/10.7554/eLife.48943.014 Bardetti et al. eLife 2019;8:e48943. DOI: https://doi.org/10.7554/eLife.48943 8 of 33 Research article Developmental Biology Microbiology and Infectious Disease Kap123, the importin for Cdc25, seems to be phosphorylated upon activation of the pheromone cascade The results shown above can be included in a working model in which the pheromone cascadedependent induction of the expression of pcl12 enables the formation of a Crk1-Pcl12 complex, which is responsible for the cell cycle arrest at G2. In support of this view, we observed that the cell cycle arrest induced by ectopic expression of pcl12 was dependent on the inhibitory phosphorylation of Cdk1 (Figure 3—figure supplement 1A and B), similarly to the fuz7 DD -induced cell cycle arrest. However, in clear contrast with the observed drop in Cdc25 levels upon fuz7 DD expression, the levels of Cdc25 did not decrease upon the ectopic expression of pcl12 (Figure 3—figure supplement 1C). This result reinforced our conclusion that the decrease in Cdc25 levels was not responsible for pheromone cascade-induced cell cycle arrest at G2. Strikingly, we also observed that in the filaments produced upon ectopic pcl12 expression, a GFP-Cdc25 fusion was excluded from the nucleus. Moreover, this exclusion was abrogated, as it was the cell cycle arrest, if the activity of Crk1 was eliminated (Figure 3A and B). Since the phosphatase Cdc25 must be transported to the nucleus to activate mitosis entry (Mielnichuk and Pe ´rez-Martı´n, 2008), we hypothesized that retaining Cdc25 in the cytoplasm could explain the observed G2 cell cycle arrest upon ectopic pcl12 expression and, most likely, also upon fuz7 DD expression. This hypothesis was reminiscent of the b-induced retention of Cdc25 at the cytoplasm via interaction with 14-3-3 proteins. However, we considered unlikely to be the same mechanism, because for b-dependent cell cycle arrest such cytoplasmic retention can be saturated by overexpression of Cdc25, and we observed that it was not the case for the pheromone cascade-induced cell cycle arrest (Figure 1F). One of the interactors of Pcl12 obtained from the coimmunoprecipitation coupled with LC/MS analysis was the uncharacterized protein UMAG_15014. Sequence phylogeny analysis indicated that this protein (renamed Kap123) belongs to the family of bimportins of class 3 and 4 (Figure 3—figure supplement 2A). Since Pcl12 is a cytoplasmic protein (Flor-Parra et al., 2007), we considered Kap123 unlikely to be involved in the transport of Pcl12 to the nucleus. A previous report showed that Sal3 from S. pombe, which is one of the members of the class 3 bimportin family, was involved in the nuclear import of Cdc25 in this fungus (Chua et al., 2002). Therefore, we decided to analyze whether U. maydis Kap123 was required for the nuclear localization of Cdc25, and we found that it was (Figure 3—figure supplement 2B and C). This result prompted us to hypothesize that Kap123 could be a target of the Crk1-Pcl12 complex and that the action of this complex could disable the interaction between Kap123 and Cdc25 upon pheromone signaling. In this way, interference with the nuclear localization of Cdc25 could explain pheromone cascade-induced G2 cell cycle arrest. Using GFP-trap beads, we analyzed the ability of a Kap123-GFP fusion to interact with Cdc25 under conditions for the ectopic expression of pcl12 in the presence or absence of functional Crk1 kinase. Additionally, to discard any effect of the expected interaction between Kap123 and Cdc25 as a consequence of the induced cell cycle arrest, we used the overexpression of Wee1 as a control for G2 cell cycle arrest (Sgarlata and Pe ´rez-Martı´n, 2005b). In support of our hypothesis, we observed that the presence of a Crk1-Pcl12 complex, but not G2 cell cycle arrest alone, disables the ability of Kap123 to interact with Cdc25 (Figure 3C). We also observed that upon fuz7 DD expression, the mobility of a Kap123-HA allele in SDS-acrylamide gels was reduced (Figure 3D), and this decrease in mobility can be eliminated by lphosphatase treatment of the protein immunoprecipitates (Figure 3E). Moreover, we found similar electrophoretic mobility reduction upon ectopic expression of pcl12 (Figure 3F). The decrease in the electrophoretic mobility of Kap123 upon expression of fuz7 DD was dependent on the presence of functional alleles of crk1,pcl12, and kpp2 (Figure 3G). In the case of ectopic expression of pcl12, Crk1 was required for the decreased electrophoretic mobility of Kap123, but neither fuz7 nor kpp2 were required (Figure 3H). These results mirrored the same genetic requirements as the cell cycle arrest, suggesting a causal relationship between phosphorylation of Kap123 and cell cycle arrest. Bardetti et al. eLife 2019;8:e48943. DOI: https://doi.org/10.7554/eLife.48943 9 of 33 Research article Developmental Biology Microbiology and Infectious Disease Figure 6. Effects of the absence of cell cycle arrest in the ability to mate and to infect plants. (A)kap123 T867A mutant strains are able to infect plants. Disease symptoms caused by crosses of wild-type and kap123 T867A mutant strains. The symptoms were scored 14 days after infection. Two independent experiments were carried out and the average values are expressed as percentage of the total number of infected plants (n: 30 plants in each experiment) (Figure 6—source data 1). (B)kap123 T867A mutant is able to mate. Crosses of control as well as kap123 T867A mutant strains carrying compatible mating types (a1 b1 and a2 b2) in charcoal-containing agar plates. Positive fuzzy phenotype can be detected as a white-appearance mycelial growth. Note that mutant combinations were slightly affected in the ability to produce fuzzy phenotype. Plates were incubated at 22˚C for two days. (C)kap123 T867A mutant filaments were affected in nuclear number. Crosses from compatible strains (wild-type or kap123 T867A mutant) carrying a NLS-GFP fusion under control of the b-factor-dependent dik6 promoter were scrapped from agar surface, mounted on microscopy slides and epifluorescence was observed. Representative images show DIC and fluorescence in GFP channel. Bar: 20 mm. (D) Quantification of the nuclear content of filaments obtained from charcoal plates. The graph shows the result from two independent experiments, counting more than 50 filaments each (Figure 6—source data 2). Means and SDs are shown. DOI: https://doi.org/10.7554/eLife.48943.028 The following source data and figure supplement are available for figure 6: Source data 1. Data for Figure 6A. DOI: https://doi.org/10.7554/eLife.48943.030 Source data 2. Data for Figure 6D. DOI: https://doi.org/10.7554/eLife.48943.031 Figure supplement 1. The presence of the kap123 T867A allele did not affect the ability of b-factor to arrest the cell cycle. Figure 6 continued on next page Bardetti et al. eLife 2019;8:e48943. DOI: https://doi.org/10.7554/eLife.48943 16 of 33 Research article Developmental Biology Microbiology and Infectious Disease cytoplasm (Figure 5C and D), but not its downregulation (Figure 7A). These observations suggested that although cytoplasmic retention was required, it was not sufficient to induce the downregulation of Cdc25. We hypothesized that Cdc25 could be a target (direct or indirect) of the pheromone MAPK cascade, which induces its degradation only when Cdc25 is retained at the cytoplasm and the activation of the pheromone cascade reaches some threshold level (as expected in the presence of high amounts of synthetic pheromone or expression of the activated fuz7 DD allele). To uncover the involvement of the MAPK cascade in the downregulation of Cdc25, we constructed a strain simultaneously expressing the fuz7 DD allele and pcl12 under the crg1 promoter. In this way, we can induce cell cycle arrest, as well as the retention of Cdc25 in the cytoplasm (as a consequence of the ectopic expression of pcl12), even in conditions that potentially disable the MAPK cascade, such as the disruption of the kpp2 gene. In agreement with our hypothesis, we found that under conditions of cell cycle arrest and the retention of Cdc25 in the cytoplasm, Kpp2 was required for the decrease in Cdc25 levels upon fuz7 DD expression (Figure 7C). Since we considered the possibility that Cdc25 was a direct target of Kpp2, we tried to detect physical interactions between Kpp2 and Cdc25. We used a strain that expressed both the fuz7 DD allele and pcl12 and carried a kinase-dead allele of Kpp2 fused to GFP (Kpp2 K50R ) (Mu ¨ller et al., 2003), as well as an HA-tagged Cdc25 allele. Using GFP-trap beads, we were able to detect interactions between Kpp2 and Cdc25, which were dependent on the expression of the fuz7 DD allele (i.e., high levels of MAPK signaling) and the presence of a wild-type Kap123 allele (i.e., retainment of Cdc25 in the cytoplasm) (Figure 7D). Taken together, our results were consistent with the idea that activation of the pheromone cascade resulted in the negative regulation of Cdc25 levels by a two-step mechanism: First, Cdc25 is retained in the cytoplasm as a result of the formation of the Pcl12-Crk1 complex, and then Cdc25 is potentially phosphorylated by the Kpp2 MAPK, which probably promoted its degradation. Compared to other fungal Cdc25-like phosphatases, U. maydis Cdc25 carries an unusually long N-terminal extension of approximately 270 amino acids with no similarity to orthologues in the database (Figure 7E). Previous attempts to ascribe a role to this domain were elusive because its deletion did not appear to affect the Cdc25 function in U. maydis growing under axenic conditions (Sgarlata and Pe ´rez-Martı´n, 2005a), although it affected the sensitivity of U. maydis cells to cell wall stressors (Carbo ´and Pe ´rez-Martı´n, 2010). We found that the removal of this N-terminal extension resulted in the stabilization of Cdc25 upon expression of the fuz7 DD allele (Figure 7F). Based on sequence, this region contains at least three putative MAPK phosphorylation sites (Ser 7 , Ser 23 and Ser 239 ,Figure 5E). With support from the observed physical interactions, we considered the possibility that Kpp2 phosphorylates Cdc25 in this region, promoting its degradation as described in mammalian Cdc25 in response to activation of the p38 MAPK (Uchida et al., 2009). To investigate this possibility, the three candidate Ser residues were mutated to Ala (cdc25 AAA allele). Encouragingly, we found that the triple mutant was resistant to the downregulation observed in the wild-type allele upon expression of the fuz7 DD allele (Figure 7G). Moreover, in accordance with our model explaining how the cell cycle is arrested in response to pheromone, neither the deletion of the N-terminal extension nor the alanine mutations affected the ability of cells undergo cell cycle arrest (Figure 7H) as well as the retention of Cdc25 at cytoplasm (Figure 7I). In summary, these data strongly suggested that the MAPK Kpp2 targets and downregulates the phosphatase Cdc25, most likely promoting its degradation. The pheromone-dependent decrease of Cdc25 levels is required for virulence We analyzed the capacity of sexually compatible strains (a1 b1 and a2 b2 mating type) carrying the cdc25 AAA allele to infect plants. Strikingly, we found that the presence of this mutant allele strongly affected the ability of fungal cells to produce disease to corn plants (Figure 8A). The defect in virulence could be attributed to problems related to the mating process itself or it might be related to Figure 6 continued DOI: https://doi.org/10.7554/eLife.48943.029 Bardetti et al. eLife 2019;8:e48943. DOI: https://doi.org/10.7554/eLife.48943 17 of 33 Research article Developmental Biology Microbiology and Infectious Disease Figure 7. Cdc25 is negatively regulated by Kpp2, the pheromone MAPK. (A) High levels of pheromone resulted in a decrease in Cdc25 levels. Western blot from extracts obtained from a1 mating-type cells carrying a HA-tagged Cdc25 endogenous allele that were incubated for 6 hr in the presence of the indicated synthetic a2 pheromone concentrations in CMD medium. Similar amount of protein extracts was separated by SDS-PAGE. Immunoblots were incubated with an antibody against HA. As loading control, we used the Cdk1 protein, which can be detected using anti-PSTAIRE (which Figure 7 continued on next page Bardetti et al. eLife 2019;8:e48943. DOI: https://doi.org/10.7554/eLife.48943 18 of 33 Research article Developmental Biology Microbiology and Infectious Disease subsequent steps, such as the ability of the fungus to proliferate within the plant, for example. To address this issue, we took advantage of the solopathogenic strain SG200, which is a haploid strain that carries the genetic information from the two different mating types and, as a consequence, does not require cell fusion to produce the infective hypha (Bo ¨lker et al., 1995). SG200-derived cells carrying the cdc25 AAA allele were as virulent as the control strain (Figure 8A), suggesting a problem in mating associated with the presence of the cdc25 AAA allele. Moreover, a cdc25 AAA mutant cross showed a weak fuzzy phenotype in charcoal plates, in comparison to wild-type control, indicating some kind of problem with the formation of the dikaryotic filament (Figure 8B). We wondered whether the impaired fuzzy phenotype in the mutant crosses was caused by defects in cell fusion. To address that question, we used the P dik6 -NLS-GFP reporter described above. Since this reporter will be only active in the presence of a functional b-factor, it can be used as an indirect readout of cell fusion between compatible sexual partners, observing the appearance of nuclear GFP fluorescence in the cell mixture resulting from cross. We crossed mutant haploid cells carrying the dik6 reporter, and we were able to observe filaments carrying nuclear GFP signal, indicating that cell fusion was not affected. Interestingly, these filaments were multinucleated (Figure 8C). In contrast to the multinucleated filaments found from crosses involving the Kap123 T867A mutant, which seemed not to be able to bypass the G2 barrier imposed by the b-factor (we never found more than four nuclei per filament, Figure 6D), a significant proportion of the filaments observed in cdc25 AAA crosses showed more than five nuclei (Figure 8D). Since cells carrying the cdc25 AAA allele were able to be arrested in response to activation of the pheromone cascade (Figure 7H), these observations prompted us to think that the presence of the cdc25 AAA allele could affect the G2 cell cycle arrest activated in the presence of a functional b-factor. However, we observed that the presence of the cdc25 AAA allele did not affect the ability of AB33 cells (see Figure 6—figure supplement 1for a description of this strain) to undergo cell cycle arrest upon the expression of compatible b proteins (Figure 8E). Figure 7 continued recognizes both Cdk1 and Cdk5). Cell response refers to percentage of cells showing conjugation tube in each culture. (B) Inability to retain Cdc25 at the cytoplasm resulted in stabilization of Cdc25 levels in response to expression of the fuz7 DD allele. Western blot analysis to show the level of Cdc25 (upper blot) in cells growing in inducing conditions (CMA) for the expression of fuz7 DD allele during the indicated time. Levels of Cdk1 were used as loading control (bottom blot). (C) The MAPK Kpp2 is required for down-regulation of Cdc25 levels. Strains expressing at the same time fuz7 DD and pcl12 and carrying or not a loss of function allele of kpp2, were grown during the indicated time in inducing conditions (CMA). Extracts were analyzed by Western blot to detect Cdc25-3HA protein levels (upper blot) and the levels of Cdk1, which were used as loading control (bottom blot). (D) Kpp2 is able to interact with Cdc25 upon activation of the pheromone cascade. Soluble extracts from strains carrying Cdc25-3HA and a Kpp2 KD -GFP tagged in their corresponding endogenous loci, and carrying ectopic copies of the indicated genes under the control of crg1 promoter, as well as additional mutations (also indicated) were incubated with GFP-trap beads and the immunoprecipitates submitted to Western blot with anti-HA (Cdc25) and antiGFP (Kpp2 KD ) antibodies in succession. Cells were grown in inducing conditions (CMA, (A) or repressive conditions (CMD, (D) for crg1 promoter during 6 hr. (E) Scheme of U. maydis Cdc25, showing the N-terminal specific extension, and the postulated MAPK phosphorylation sites, with the Serine residue (in red) exchanged to Alanine in the mutant allele cdc25 AAA . (F) The N-terminal extension of Cdc25 determines its down-regulation in response to expression of the fuz7 DD allele. Western blot analysis to show the level of Cdc25 or Cdc25 D1–270 (upper blot) upon expression of fuz7 DD allele in cells growing in inducing conditions (CMA) for the indicated time. Levels of Cdk1 were used as loading control (bottom blot). (G) Putative MAPK phosphorylation sites were involved in the down-regulation of Cdc25 in response to expression of the fuz7 DD allele. Western blot analysis to show the level of Cdc25 or Cdc25 AAA (upper blot) upon expression of fuz7 DD allele in cells growing in inducing conditions (CMA) for the indicated time. Levels of Cdk1 were used as loading control (bottom blot). (H) Cell cycle arrest in response to expression of the fuz7 DD allele is unaffected by the absence of down-regulation of Cdc25. fuz7 DD -expressing strains carrying the indicated cdc25 alleles, which also carried the NLS-GFP transgene, were incubated in inducing conditions (CMA) for 6 hr. Filaments from each culture were counted and sorted as carrying 1 (cell cycle arrested) or more than one nucleus (not arrested). The graph shows the result from three independent experiments, counting more than 100 filaments each (Figure 7—source data 1). Means and SDs are shown. (I) Quantification of number of filaments showing GFP fluorescence associated with the nucleus in control and strains carrying the indicated cdc25 alleles, expressing pcl12 and incubated for 6 hr in inducing conditions (CMA). The graph shows the result from three independent experiments, counting 50 filaments each (Figure 7—source data 2). Means and SDs are shown. DOI: https://doi.org/10.7554/eLife.48943.032 The following source data is available for figure 7: Source data 1. Data for Figure 7H. DOI: https://doi.org/10.7554/eLife.48943.033 Source data 2. Data for Figure 7I. DOI: https://doi.org/10.7554/eLife.48943.034 Bardetti et al. eLife 2019;8:e48943. DOI: https://doi.org/10.7554/eLife.48943 19 of 33 Research article Developmental Biology Microbiology and Infectious Disease Figure 8. Pheromone-dependent down-regulation of Cdc25 is required for full virulence. (A)cdc25 AAA mutant strains are affected in the ability to infect plants. Disease symptoms caused by crosses of wild-type and cdc25 AAA mutant strains. Strikingly, solopathogenic SG200-derived strains were not affected. The symptoms were scored 14 days after infection. Two independent experiments were carried out and the average values are expressed as percentage of the total number of infected plants (n: 30 plants in each experiment) (Figure 8—source data 1). (B)cdc25 AAA mutant is affected in the Figure 8 continued on next page Bardetti et al. eLife 2019;8:e48943. DOI: https://doi.org/10.7554/eLife.48943 20 of 33 Research article Developmental Biology Microbiology and Infectious Disease In summary, we observed that the absence of downregulation of Cdc25 levels in response to pheromone does not affect neither pheromone-induced cell cycle arrest nor de novo (independent of previous pheromone-induced cell cycle arrest) b-induced cell cycle arrest. However, the absence of downregulation of Cdc25 levels seems to affect the cell cycle arrest in b-dependent filaments originating from crosses (i.e., resulting from a mating process). These unexpected results can be explained keeping in mind that, in U. maydis, during a regular cell cycle Cdc25 accumulates at G2 phase; and, that in contrast to pheromone-dependent cell cycle arrest, the b-dependent cell cycle arrest can be bypassed by high levels of Cdc25 (Mielnichuk et al., 2009). We believe that during the pheromone-dependent G2 cell cycle arrest, Cdc25 is accumulating in the cytoplasm of conjugation tubes, but that as the compatible conjugation tubes are getting close (and therefore encountering higher pheromone concentrations) the subsequent increase in the pheromone-cascade signaling, promotes a decrease in the Cdc25 levels that allows a proper b-dependent cell cycle arrest in the filament once the cells were fused. Inability to achieve this decrease (as the case of stable Cdc25 alleles) could result in impaired b-dependent cell cycle arrest and thereby defects in pathogenicity. Discussion Smut fungi are a widespread group of plant pathogens whose sexual development is coupled with virulence. All smut species investigated so far have in common the need to undergo a successful mating reaction to form an infective dikaryotic filament before being able to infect their host plant (Bakkeren et al., 2008). In the most studied member of this group, U. maydis, it has been described strong connections between the regulation of the cell cycle and the ability to infect plants: During the steps previous to plant infection, it is required a sustained G2 cell cycle arrest (Perez-Martin, 2012). The cell cycle arrest of the infective filament on plant surface observed in U. maydis seems to be more general, and it is also present in rust fungi like Uromyces phaseoli (Heath and Heath, 1978). The connections between cell cycle regulation and the virulence in plant fungal pathogens have been studied in detail in other systems, in addition to U. maydis, such as Magnaporthe oryzae and Colletotrichum orbiculare. In these systems, the infection depends on the formation of appressorium, a structure required for plant penetration. The appressoria from these fungi use a turgor-driven mechanical process to breach the plant cuticle. The functionality of this class of appressorium implies the increase of internal turgor pressure, which is linked to an elaborated program that includes from metabolic reprograming to morphological changes to produce a clearly defined structure with a thick, multilayered and highly melanized cell wall (Ryder and Talbot, 2015). The various steps required for the formation of appressorium in these phytopathogenic fungi are linked to the different cell cycle phases. There is a G1/S control to induce the formation of appressorium, first described in C. orbiculare (Fukada and Kubo, 2015), but also operating in M. oryzae (Fukada et al., Figure 8 continued formation of dikaryotic hyphae. Crosses of control as well as cdc25 AAA mutant strains carrying compatible mating types (a1 b1 and a2 b2) in charcoalcontaining agar plates. Positive fuzzy phenotype can be detected as a white-appearance mycelial growth. Note that mutant combinations were affected in the ability to produce fuzzy phenotype. Plates were incubated at 22˚C for two days. (B)cdc25 AAA mutant cells are able to fuse. Crosses from compatible strains (wild-type or cdc25 AAA mutant) carrying a NLS-GFP fusion under control of the b-factor-dependent dik6 promoter were scrapped from agar surface, mounted on microscopy slides and epifluorescence was observed. Representative images show DIC and fluorescence in GFP channel. Bar: 20 mm. (C)cdc25 AAA mutant filaments were affected in nuclear number. Graph shows the quantification of the nuclear content of filaments obtained from charcoal plates. The graph shows the result from two independent experiments, counting more than 50 filaments each (Figure 8— source data 2). Means and SDs are shown. (E) b-induced cell cycle arrest de novo is not affected by the presence of cdc25 AAA mutant allele. Representative images of cultures from AB33-derived strains that were incubated in inducing conditions (minimal medium with nitrate) for 8 hr. Bar: 20 mm. DOI: https://doi.org/10.7554/eLife.48943.035 The following source data is available for figure 8: Source data 1. Data for Figure 8A. DOI: https://doi.org/10.7554/eLife.48943.036 Source data 2. Data for Figure 8D. DOI: https://doi.org/10.7554/eLife.48943.037 Bardetti et al. eLife 2019;8:e48943. DOI: https://doi.org/10.7554/eLife.48943 21 of 33 Research article Developmental Biology Microbiology and Infectious Disease 2019); and there is a S-phase checkpoint required for appressorium differentiation in M. oryzae (Ose ´s-Ruiz et al., 2017;Ose ´s-Ruiz and Talbot, 2017). In contrast to appressoria from these phytopathogenic fungi, appressoria of U. maydis is not dependent on turgor pressure to penetrate the plant tissue. The appressorium directs the localized secretion of enzymes that weakens the plant cuticle, allowing the plant tissue penetration. For that reason, the morphogenesis process is less complex and appressoria are unmelanized, rather small swellings of the hyphal tip that form penetration structures that are less constricted (Snetselaar and Mims, 1993). However, in spite of this apparent lack of complexity, there is also a clear connection with cell cycle regulation. G2 cell cycle arrest is mandatory to promote the differentiation of the appressorium, and the impairment in cell cycle arrest resulted in a lack of virulence (Castanheira et al., 2014). This obligation for cell cycle arrest at G2 phase is most likely related to the fact that mitosis and the morphogenetic program responsible for appressorium formation compete for the same cytoskeletal components. Because of this competition, it makes sense that cellular controls exist to force these two processes to be incompatible and therefore to prevent infective filaments from entering mitosis (i.e., to arrest the progression of the cell cycle at G2) (Pe ´rezMartı´n et al., 2016). G2 cell cycle arrest is established before mating and is maintained once the dikaryotic infective filament is formed, until the fungus enters the plant. During this period, cell cycle arrest is sustained by two distinct regulatory networks: First, by the pheromone-recognition cascade in each mating partner and second, once cytoplasmic fusion occurs, by the presence of a homeodomain transcriptional regulator called b-factor (which, among other targets, represses the a-locus, encoding the pheromone and receptors; Urban et al., 1996). Our results showed that the two regulatory networks recruit both shared and specific elements to induce and sustain cell cycle arrest (Figure 9). The G2/ M transition in U. maydis is controlled by the level of inhibitory phosphorylation of the Cdk1-Clb2 complex, and both regulatory networks rely on an increase in Cdk1 inhibitory phosphorylation to arrest the cell cycle. To achieve this increase, both regulatory networks upregulate the kinase responsible for this inhibitory phosphorylation, the Wee1 kinase. This upregulation is indirect and is mediated by the transcriptional repression of hsl1, which encodes a kinase that represses Wee1 activity. However, the repression of hsl1 is not enough to arrest the cell cycle at G2: loss-of-function mutants in Hsl1 are able to progress through the cell cycle, although they have an extended G2 phase (Castanheira et al., 2014). To arrest the cell cycle, a second element is required, namely the downregulation of Cdc25, the phosphatase that removes the Cdk1 inhibitory phosphorylation. Is at this step where the two regulatory networks differ. The pheromone cascade promotes the formation of a kinase-cyclin complex, Crk1-Pcl12, which phosphorylates and inhibits the b-importin Kap123, which is required for the nuclear translocation of Cdc25. The b-factor, however, activates the cascade responsible for the DNA damage response (DDR, composed of Atr1 and Chk1 kinases), resulting in phosphorylation of Cdc25, which promotes its cytoplasmic retention via 14-3-3 proteins (de Sena-Toma ´s et al., 2011;Mielnichuk and Pe ´rez-Martı´n, 2008;Mielnichuk et al., 2009). This downregulation of Cdc25 also provides a time window to load a third element (not present in pheromone-induced cell cycle arrest) consisting of the transcriptional repression (via an intermediate regulator called Biz1) of clb1, which encodes a second b-type cyclin that is also required for G2/M transition (Flor-Parra et al., 2006). The reasons for using distinct mechanisms to retain Cdc25 in the cytoplasm are probably related to the degree of reversibility of each developmental step during the infection process. The pheromone recognition by each mating partner has to culminate with the fusion of the respective conjugation tubes. If for some reason this step does not occur, the unmated cells must be able to return to the previous vegetative status. In other words, the pheromone cell cycle arrest should be reversible if the stimulus (pheromone) disappears and mating was not successful. A control mediated by the phosphorylation of importin could probably be easily reversed by some phosphatase (specific or not) once the signaling through the pheromone cascade is abrogated. The formation of the dikaryotic filament, however, is a more terminal decision in the sense that once the mating partners fuse their respective cytoplasms, if they are compatible at the b-locus, the infective filament is committed to infecting the plant. The presence of two independent cell cycle brakes, provided by the retention of Cdc25 at cytoplasm by 14-3-3 proteins and by the transcriptional repression of clb1 could make this step less reversible. Bardetti et al. eLife 2019;8:e48943. DOI: https://doi.org/10.7554/eLife.48943 22 of 33 Research article Developmental Biology Microbiology and Infectious Disease Figure 9. Two regulatory networks recruit both shared and specific elements to induce and sustain a G2 cell cycle arrest during the production of the infective structures in U. maydis. Scheme showing the distinct steps at cell cycle level occurring during the formation of the infective filament of U. maydis. Arrows and bars denote positive and negative interactions. Yellow stars denote phosphorylation. For details, see discussion section. DOI: https://doi.org/10.7554/eLife.48943.038 Bardetti et al. eLife 2019;8:e48943. DOI: https://doi.org/10.7554/eLife.48943 23 of 33 Research article Developmental Biology Microbiology and Infectious Disease Our results also reveal the importance of the transition from pheromone-dependent to b-dependent cell cycle arrest once the respective partners fuse. Although the downregulation of Cdc25 activity consists of the retention of Cdc25 in the cytoplasm in both regulatory networks, there is one relevant difference: while the inhibition of nuclear translocation (pheromone-dependent) cannot be bypassed by high concentrations of Cdc25, the retention of Cdc25 by 14-3-3 (b-dependent) is sensitive to the amount of Cdc25 (probably because of saturation of the available retention proteins). This difference gains importance because Cdc25 most likely accumulates in the cytoplasm of conjugation tubes during G2 cell cycle arrest. Once the respective cytoplasms fuse, if the amount of Cdc25 has not been previously tuned, the 14-3-3 retention system will probably be overwhelmed, and the time window required for the establishment of permanent G2 cell cycle arrest will not be provided. For that reason, it seems logical that once the pheromone cascade reaches some signaling threshold (probably reflecting a close encounter between compatible conjugation tubes), it triggers the downregulation of Cdc25 to levels that are most likely easily assimilable by the 14-3-3 proteins. The importance of adjusting the levels of Cdc25 is reflected by the impaired ability of cells carrying the cdc25 AAA allele to infect plants. Central to the pheromone-induced cell cycle arrest in U. maydis was the complex formed by the kinase Crk1 and the cyclin Pcl12. Both proteins were previously reported to be required for the proper morphogenesis of the infective filament in this fungus. Interestingly, in this function, they were part of distinct complexes or regulatory networks: Pcl12 complexed with the cyclin-dependent kinase Cdk5 (Flor-Parra et al., 2007), while Crk1 was activated in a similar way to MAPK via T-loop phosphorylation by the upstream MAPK kinase (Garrido et al., 2004). In this work, we showed that Pcl12 and Crk1 seem to form a complex with specific roles in cell cycle arrest upon pheromone recognition and that neither Cdk5 nor MAPK cascade phosphorylation are involved. How Pcl12 controls the activity of Crk1 in this process is beyond the scope of this work. However, we can anticipate some possibilities. Crk1 belongs to the family of RD kinases, which have a conserved arginine residue preceding an aspartate residue in the catalytic loop (Johnson et al., 1996). For this reason, the members of this family require an activation segment (T-loop) that undergoes a conformational change to fold into an active catalytic site. This conformational change can be achieved either by the phosphorylation of specific residues (such as TxY in MAP kinases) or by interaction with other proteins (such as cyclins in cyclin-dependent kinases) (Nolen et al., 2004). Crk1 carries a TEY signature at its T-loop, similar to a MAP kinase, but it is worth mentioning that it was cloned from a genetic screening devoted to uncovering CDK-like proteins in U. maydis and that the predicted three-dimensional folding of the catalytic domain of Crk1 fits the structure of Cdk2 (Garrido and Pe ´rez-Martı´n, 2003). We propose that Crk1 can be activated either by T-loop phosphorylation (through the pheromone MAPK cascade) or by interaction with Pcl12 cyclin. The activation of Crk1 in the absence of Pcl12, which is dependent on T-loop phosphorylation, resulted in the promotion of polar growth without affecting cell cycle regulation (Garrido et al., 2004). However, cell cycle arrest by Crk1 was dependent on Pcl12 cyclin but independent of T-loop phosphorylation. The use of proteins unrelated to cyclins to activate CDKs is well described, including the so-called K-cyclins from viruses, p35 as a coactivator of mammalian Cdk5 and the family of RINGO/Speedy proteins (Nebreda, 2006). However, the use of a cyclin to activate a noncanonical cyclin-dependent kinase has not, to the best of our knowledge, been previously described. The importance of this alternative activation mechanism can be appreciated by considering that Crk1 belongs to a widely conserved family of fungal kinases, the founding member of which is Ime2, a central regulator of meiosis in S. cerevisiae. The Ime2-related kinases exhibit amazing variety in controlling sexual developmental programs in fungi, although the targets and physiological inputs seem to be very species-specific (Irniger, 2011). Alternative mechanisms of activation could help to explain this plethora of roles. Bardetti et al. eLife 2019;8:e48943. DOI: https://doi.org/10.7554/eLife.48943 24 of 33 Research article Developmental Biology Microbiology and Infectious Disease Materials and methods Key resources table Reagent type (species) or resource Designation Source or reference Identifiers Additional information Ustilago maydis Strains used in this study are listed in Supplementary file 1 Perez-Martin lab N/A E. coli strain DH5aCGSC 12384 Antibody Anti-PSTAIRE (Rabbit polyclonal) Santa Cruz Biotechnology, Inc Sc-53 RRID:AB_2074908 Antibody Anti-phospho-Cdc2 (Tyr15) (10A11, Rabbit monoclonal) Cell Signaling 4539 RRID:AB_560953 Antibody Anti-HA, High affinity (3F10, Rat monoclonal) Roche 1 867 423 RRID:AB_390919 Antibody Anti-HA-HRP conjugate, High affinity (3F10, Rat monoclonal) Roche 12 013 819 001 Antibody Anti-c-Myc-HRP conjugate (9E10, mouse monoclonal) Roche 1 814 150 Antibody Anti-GFP, Living Colors (JL-8, mouse monoclonal) Clontech 632380 RRID:AB_2314359 Antibody Anti-rabbit IgG-HRP conjugate (Donkey polyclonal) Amersham Biosciences NA934 RRID:AB_772206 Antibody Anti-rat IgG-HRP conjugate (Mouse, monoclonal) SIGMA R7636 RRID:AB_1840005 Antibody Anti-mouse IgG-HRP conjugate (Goat polyclonal) SIGMA A0168 RRID:AB_257867 Chemical compound, drug Hygromycin B Roche 834 555 Chemical compound, drug clonNAT (nourseothricin) Werner BioAgents CAS#96736-11-7 Chemical compound, drug G418 Formedium G418-1 Chemical compound, drug Carboxine SIGMA 45371 Chemical compound, drug Phleomycin InvivoGen ant-ph-1 Chemical compound, drug Roche Protease Inhibitor Cocktail Roche 11-697-498-001 Chemical compound, drug PhosSTOP Roche 04-906-837-001 Chemical compound, drug Synthetic U. maydis a2 pheromone Proteomic Services from National Center of Biotechnology, CSIC, Madrid N/A Commercial assay or kit High Pure RNA isolation kit Roche 11828665001 Commercial assay or kit GFP-Trap_MA Chromotek gtma-400 Continued on next page Bardetti et al. eLife 2019;8:e48943. DOI: https://doi.org/10.7554/eLife.48943 25 of 33 Research article Developmental Biology Microbiology and Infectious Disease W. 2016. A sex-inducing pheromone triggers cell cycle arrest and mate attraction in the diatom Seminavis robusta. Scientific Reports 6:19252. DOI: https://doi.org/10.1038/srep19252 Mu ¨ller P, Weinzierl G, Brachmann A, Feldbru ¨gge M, Kahmann R. 2003. Mating and pathogenic development of the smut fungus Ustilago maydis are regulated by one mitogen-activated protein kinase cascade. Eukaryotic Cell 2:1187–1199. DOI: https://doi.org/10.1128/EC.2.6.1187-1199.2003,PMID: 14665454 Nebreda AR. 2006. CDK activation by non-cyclin proteins. Current Opinion in Cell Biology 18:192–198. DOI: https://doi.org/10.1016/j.ceb.2006.01.001,PMID: 16488127 Nolen B, Taylor S, Ghosh G. 2004. Regulation of protein kinases; controlling activity through activation segment conformation. Molecular Cell 15:661–675. DOI: https://doi.org/10.1016/j.molcel.2004.08.024,PMID: 15350212 Ose ´s-Ruiz M, Sakulkoo W, Littlejohn GR, Martin-Urdiroz M, Talbot NJ. 2017. Two independent S-phase checkpoints regulate appressorium-mediated plant infection by the rice blast fungus Magnaporthe oryzae. PNAS 114:E237–E244. DOI: https://doi.org/10.1073/pnas.1611307114,PMID: 28028232 Ose ´s-Ruiz M, Talbot NJ. 2017. Cell cycle-dependent regulation of plant infection by the rice blast fungus Magnaporthe oryzae.Communicative & Integrative Biology 10:e1372067. DOI: https://doi.org/10.1080/ 19420889.2017.1372067,PMID: 29259729 Pe ´rez-Martı´n J, Castillo-Lluva S, Sgarlata C, Flor-Parra I, Mielnichuk N, Torreblanca J, Carbo´ N. 2006. Pathocycles: Ustilago maydis as a model to study the relationships between cell cycle and virulence in pathogenic fungi. Molecular Genetics and Genomics 276:211–229. DOI: https://doi.org/10.1007/s00438-0060152-6,PMID: 16896795 Pe ´rez-Martı´n J. 2009. DNA-damage response in the basidiomycete fungus Ustilago maydis relies in a sole Chk1like kinase. DNA Repair 8:720–731. DOI: https://doi.org/10.1016/j.dnarep.2009.01.023,PMID: 19269260 Perez-Martin J. 2012. Cell cycle and morphogenesis connections during the formation of the infective filament in Ustilago maydis. In: Perez-Martin J, di Pietro A (Eds). Morphogenesis and Pathogenicity In Fungi. Berlin Heidelberg: Springer-Verlag. p. 97–114. DOI: https://doi.org/10.1007/978-3-642-22916-9_6 Pe ´rez-Martı´n J, Bardetti P, Castanheira S, de la Torre A, Tenorio-Go´ mez M. 2016. Virulence-specific cell cycle and morphogenesis connections in pathogenic fungi. Seminars in Cell & Developmental Biology 57:93–99. DOI: https://doi.org/10.1016/j.semcdb.2016.03.017,PMID: 27032479 Pope PA, Bhaduri S, Pryciak PM. 2014. Regulation of cyclin-substrate docking by a G1 arrest signaling pathway and the cdk inhibitor Far1. Current Biology 24:1390–1396. DOI: https://doi.org/10.1016/j.cub.2014.05.002, PMID: 24909323 Ryder LS, Talbot NJ. 2015. Regulation of appressorium development in pathogenic fungi. Current Opinion in Plant Biology 26:8–13. DOI: https://doi.org/10.1016/j.pbi.2015.05.013,PMID: 26043436 Sgarlata C, Pe´ rez-Martı ´n J. 2005a. The cdc25 phosphatase is essential for the G2/M phase transition in the basidiomycete yeast Ustilago maydis.Molecular Microbiology 58:1482–1496. DOI: https://doi.org/10.1111/j. 1365-2958.2005.04925.x,PMID: 16313631 Sgarlata C, Pe´ rez-Martı ´n J. 2005b. Inhibitory phosphorylation of a mitotic cyclin-dependent kinase regulates the morphogenesis, cell size and virulence of the smut fungus Ustilago maydis.Journal of Cell Science 118:3607– 3622. DOI: https://doi.org/10.1242/jcs.02499,PMID: 16046476 Snetselaar KM, Bolker M, Kahmann R. 1996. Ustilago maydis mating hyphae orient their growth toward pheromone sources. Fungal Genetics and Biology 20:299–312. DOI: https://doi.org/10.1006/fgbi.1996.0044, PMID: 9045760 Snetselaar KM, Mims CW. 1992. Sporidial Fusion and Infection of Maize Seedlings by the Smut Fungus Ustilago Maydis .Mycologia 84:193–203. DOI: https://doi.org/10.1080/00275514.1992.12026126 Snetselaar KM, Mims CW. 1993. Infection of maize stigmas by Ustilago maydis : light and electron microscopy. Phytopathology 83:843–850. DOI: https://doi.org/10.1094/Phyto-83-843 Spellig T, Bo ¨lker M, Lottspeich F, Frank RW, Kahmann R. 1994. Pheromones trigger filamentous growth in Ustilago maydis.The EMBO Journal 13:1620–1627. DOI: https://doi.org/10.1002/j.1460-2075.1994.tb06425. x,PMID: 8157001 Steinberg G, Schliwa M, Lehmler C, Bo ¨lker M, Kahmann R, McIntosh JR. 1998. Kinesin from the plant pathogenic fungus Ustilago maydis is involved in vacuole formation and cytoplasmic migration. Journal of Cell Science 111 ( Pt 15:2235–2246. PMID: 9664045 Straube A, Weber I, Steinberg G. 2005. A novel mechanism of nuclear envelope break-down in a fungus: nuclear migration strips off the envelope. The EMBO Journal 24:1674–1685. DOI: https://doi.org/10.1038/sj.emboj. 7600644,PMID: 15861140 Szabo ´Z, To ¨nnis M, Kessler H, Feldbru ¨gge M. 2002. Structure-function analysis of lipopeptide pheromones from the plant pathogen Ustilago maydis.Molecular Genetics and Genomics 268:362–370. DOI: https://doi.org/10. 1007/s00438-002-0756-4,PMID: 12436258 Tenorio-Go ´mez M, de Sena-Toma´ s C, Pe´ rez-Martı ´n J. 2015. MRNand 9-1-1-Independent activation of the ATRChk1 pathway during the induction of the virulence program in the phytopathogen Ustilago maydis.PLOS ONE 10:e0137192. DOI: https://doi.org/10.1371/journal.pone.0137192,PMID: 26367864 Terfru ¨chte M, Joehnk B, Fajardo-Somera R, Braus GH, Riquelme M, Schipper K, Feldbru ¨gge M. 2014. Establishing a versatile golden gate cloning system for genetic engineering in fungi. Fungal Genetics and Biology 62:1–10. DOI: https://doi.org/10.1016/j.fgb.2013.10.012,PMID: 24211735 Tsukuda T, Carleton S, Fotheringham S, Holloman WK. 1988. Isolation and characterization of an autonomously replicating sequence from Ustilago maydis.Molecular and Cellular Biology 8:3703–3709. DOI: https://doi.org/ 10.1128/MCB.8.9.3703 Bardetti et al. eLife 2019;8:e48943. DOI: https://doi.org/10.7554/eLife.48943 32 of 33 Research article Developmental Biology Microbiology and Infectious Disease Uchida S, Yoshioka K, Kizu R, Nakagama H, Matsunaga T, Ishizaka Y, Poon RY, Yamashita K. 2009. Stressactivated mitogen-activated protein kinases c-Jun NH2-terminal kinase and p38 target Cdc25B for degradation. Cancer Research 69:6438–6444. DOI: https://doi.org/10.1158/0008-5472.CAN-09-0869,PMID: 1 9638579 Urban M, Kahmann R, Bo ¨lker M. 1996. Identification of the pheromone response element in Ustilago maydis. Molecular & General Genetics : MGG 251:31–37. DOI: https://doi.org/10.1007/bf02174341,PMID: 8628244 Vollmeister E, Schipper K, Baumann S, Haag C, Pohlmann T, Stock J, Feldbru ¨gge M. 2012. Fungal development of the plant pathogen Ustilago maydis.FEMS Microbiology Reviews 36:59–77. DOI: https://doi.org/10.1111/j. 1574-6976.2011.00296.x,PMID: 21729109 Xu D, Farmer A, Chook YM. 2010. Recognition of nuclear targeting signals by Karyopherin-bproteins. Current Opinion in Structural Biology 20:782–790. DOI: https://doi.org/10.1016/j.sbi.2010.09.008 Zarnack K, Eichhorn H, Kahmann R, Feldbru ¨gge M. 2008. Pheromone-regulated target genes respond differentially to MAPK phosphorylation of transcription factor Prf1. Molecular Microbiology 69:1041–1053. DOI: https://doi.org/10.1111/j.1365-2958.2008.06345.x,PMID: 18627457 Bardetti et al. eLife 2019;8:e48943. DOI: https://doi.org/10.7554/eLife.48943 33 of 33 Research article Developmental Biology Microbiology and Infectious Disease