The NDR/LATS Kinase Cbk1 Controls the Activity of the Transcriptional Regulator Bcr1 during Biofilm Formation in Candida albicans

In nature, many microorganisms form specialized complex, multicellular, surface-attached communities called biofilms. These communities play critical roles in microbial pathogenesis. The fungal pathogen Candida albicans is associated with catheter-based infections due to its ability to establish biofilms. The transcription factor Bcr1 is a master regulator of C. albicans biofilm development, although the full extent of its regulation remains unknown. Here, we report that Bcr1 is a phosphoprotein that physically interacts with the NDR kinase Cbk1 and undergoes Cbk1-dependent phosphorylation. Mutating the two putative Cbk1 phosphoacceptor residues in Bcr1 to alanine markedly impaired Bcr1 function during biofilm formation and virulence in a mouse model of disseminated candidiasis. Cells lacking Cbk1, or any of its upstream activators, also had reduced biofilm development. Notably, mutating the two putative Cbk1 phosphoacceptor residues in Bcr1 to glutamate in cbk1Δ cells upregulated the transcription of Bcr1-dependent genes and partially rescued the biofilm defects of a cbk1Δ strain. Therefore, our data uncovered a novel role of the NDR/LATS kinase Cbk1 in the regulation of biofilm development through the control of Bcr1.


Introduction
Biofilms are surface-attached microbial communities embedded in an extracellular matrix. Cells in a biofilm exhibit phenotypic properties different from those of their planktonic counterparts, including an increased resistance to the host immune system and to antimicrobial agents [1][2][3]. Several tissues, as genitourinary or oral epithelia, and biomedical devices can serve as substrates for biofilm development. In this context, biofilm formation is a key feature in microbial pathogenesis.
Among the pathogenic fungi, C. albicans is one of the organisms most commonly associated with implant-related infections [2,4,5]. C. albicans is a polymorphic fungus that can change between three different forms: yeast, pseudohyphae and hyphae. Morphogenetic transitions are critical for the acquisition of proper biofilm architecture: initially, a basal layer of cells is formed when yeast cells attach to a surface followed by cell division and proliferation. In a second phase, cells differentiate into hyphal and pseudohyphal forms and produce extracellular material; the development of these forms and the increase in extracellular matrix deposition would finally arise in a dense and mature biofilm structure. Genes required for hyphal morphogenesis, cell wall remodeling, amino acid and lipid metabolism and glycolytic processes have been involved in the progression of biofilm formation in C. albicans (for a review, see [6]). Notably, biofilm development requires the activation of specific transcription programs different from those of free-living planktonic cells [7]. Tec1, a hypha-specific gene regulator; Bcr1, required for the expression of different cell wall proteins and Zap1, which governs matrix production, are examples of C. albicans biofilm transcriptional regulators [8][9][10][11]. In particular, Bcr1 has been shown to regulate the expression of a subset of genes encoding cell wall-anchored proteins including members of the agglutinin-like protein family, such as Als1 and Als3, and the hyphal wall protein Hwp1 [9,12]. Deletion of BCR1 results in defective biofilm formation in vivo and in vitro because of altered cell-to-cell interactions mediated by Als1, Als3 and Hwp1 [9].
The RAM signaling network is a conserved pathway that controls cell separation, polarized growth and cell integrity in yeast [13][14][15][16][17]. In Saccharomyces cerevisiae, the central core of the pathway consists of the Cbk1 kinase, a member of the NDR/LATS kinase family, its binding partner Mob2, the scaffolding protein Tao3, and the Ste20-like kinase Kic1. The activity of Cbk1, the main effector of the RAM pathway, is regulated by phosphorylation in a Kic1-and Tao3-dependent manner [18]. While Cbk1 polarity targets still remain largely unknown, the control of cell separation depends on the regulation of the transcription factor Ace2 [14,19,20]. It has been shown that Cbk1 phosphorylates Ace2 at the end of mitosis, triggering its accumulation at the daughter cell nucleus and leading to the asymmetric expression of genes involved in septum degradation [20]. Orthologues of the RAM pathway have been recently reported in C. albicans [16,17]. As in S. cerevisiae, RAM mutants exhibited cell separation defects and loss of cell polarity. In addition, the phenotypic analysis of a library of protein kinase mutant strains has shown that a cbk1D mutant is defective in biofilm formation although the function of Cbk1 in this process has not been further characterized [21].
The protein Cbk1 is a basophilic S/T protein kinase that exhibits a high specificity for motifs with His at the 25 position and the basic amino acids Lys or Arg at either the 23 or 22 positions [20]. This motif also seems to be the substrate of other kinases of the NDR/LATS family, such as the Drosophila melanogaster Warts/Lats and human LATS1 [22][23][24]. A search for this motif in the C. albicans proteome revealed that 0.6% of the proteins contain two or more putative Cbk1 consensus sites. Consistent with the conserved role of NDR kinases in the control of transcriptional activity, transcription factors were enriched among the putative targets, including several involved in biofilm development such Ace2 (3 sites), Bcr1 (2 sites), Nrg1 (3 sites) and Zap1 (3 sites) [25,8,10,26]. Therefore, this observation suggested that the NDR/LATS kinase Cbk1 could regulate biofilm formation by acting on transcription factors that trigger the biofilm developmental program in C. albicans. In the present study, we focus on the role of Cbk1 in the control of Bcr1, a master regulator of biofilm formation [27]. We and others previously established that Cbk1 is essential for hyphal development [16,17]. Here, we demonstrate that Cbk1 is also a regulator of biofilm formation through phosphorylation of Bcr1 at T191 and S566 residues. Thus, the Cbk1 kinase is emerging as a pivotal regulator of several developmental programs that are essential for the biology and pathogenesis of C. albicans.

RAM mutants show severe defects in biofilm formation in vitro
Although a role for Cbk1 in biofilm formation has been reported [21], no quantification of biofilm-formation defect of cells lacking Cbk1 has been published. To characterize the severity of the phenotype of cbk1D cells in biofilm formation, cells were grown under biofilm inducing conditions using a microfermentor model [7] and compared to wild-type and bcr1D reference strains. In the wild-type strain, cells thoroughly colonized the Thermanox plastic slide producing a thick and cottony biofilm. In contrast, cbk1D mutant developed very poor haze-like biofilm structures in which colonization of the plastic slides and biomass production were dramatically reduced, similar to that of bcr1D cells ( Figure 1A and 1B). The function of Cbk1 in biofilm formation depends on the RAM pathway since mutants defective in different components of the RAM network (kic1D, tao3D, and mob2D) phenocopied cbk1D cells. The same strains were tested using a biofilm formation model on silicon squares [8] to perform confocal scanning laser microscopy (CSLM). As expected, wild-type reference strain produced a complex biofilm structure of interconnected yeasts, pseudohyphae and hyphae whereas bcr1D mutant developed a rudimentary biofilm that included yeast cells with few pseudophyphae and hyphae. However, RAM mutants produced biofilms comprised exclusively of isolated clumps of yeast cells ( Figure 1D). Thus, our data indicates that the absence of RAM signaling causes a severe defect in biofilm development.
The structural defects of RAM mutant biofilms could be due to their inability to develop hyphae [16,17]. Nevertheless, in contrast to other mutants defective in hyphal formation such as cph1D efg1D [7], RAM mutants were even unable to form a compact primary layer of yeast cells on the slides, which might suggest a defect in adhesion. The RAM pathway regulates the activity of the transcription factor Ace2, which controls the expression of cell wall genes [19,28,15]. As shown in Figure 1, no significant differences were found in biomass production in microfermentor experiments between ace2D and RAM mutants. It has been reported that the ace2D mutant has a moderate defect in adherence to plastic surfaces [25]. In order to assess whether the defects in attachment to plastic or silicone surfaces of the RAM mutants were due to an Ace2 misregulation, the adherence of RAM and ace2D mutants was quantified and compared to wildtype and bcr1D reference strains ( Figure 1C; see Materials and Methods). Whereas ace2D cells showed a 50% reduction in adherence compared to the wild-type control, the defect exhibited by RAM mutants was more dramatic, with a 90% decrease in the number of adherent cells similar to that of bcr1D cells. Therefore, these results indicate that adherence impairments of RAM mutants could not be solely due to defects in Ace2 transcriptional regulation.

Bcr1 is a phosphoprotein that interacts with Cbk1 in vivo
The above results suggested that the Cbk1/Mob2 complex, the main effector of the RAM pathway, might target key regulators of biofilm formation other than Ace2. The transcription factor Bcr1 controls the expression of cell-surface genes and it is required for biofilm formation in vitro and in vivo [8,9]. Interestingly, Bcr1 contains two putative Cbk1 consensus phosphorylation sites at T191 and S556, one of which is located at the end of the first Zn finger (Figure 2A), suggesting that it could be regulated by the RAM pathway. Western blot analysis of cells expressing a functional Bcr1 tagged allele (BCR1-HA) grown at 37uC in Spider medium showed that Bcr1 is a Author Summary C. albicans infections frequently involve the formation of biofilms on implanted devices such as indwelling catheters. These complex communities of surface-associated fungal cells embedded in a matrix of extracellular polysaccharides protect C. albicans from host defences and antifungal agents. In recent years, several genes involved in the development of biofilms of C. albicans have been identified. These studies have uncovered complex regulatory networks that control the activity of several transcription factors during different steps of biofilm development. Bcr1 is a transcription factor that plays a major role in this process and yet, its regulation has not been studied extensively. Here, we show that Bcr1 function in biofilm formation and virulence requires phosphorylation of threonine 191 and serine 556 by the NDR/LATS kinase Cbk1. Moreover, given that Cbk1 is also required for the onset and maintenance of hyphal growth, our study highlights this kinase as a pivotal regulator of several developmental programs that are essential for the biology and pathogenesis of C. albicans. To analyze the phosphorylation pattern in greater detail, we used two-dimensional gel electrophoresis followed by Western blotting (2D-WB). On the basis of this approach, Bcr1 showed a complex phosphorylation pattern with multiple spots along the pH gradient ( Figure 2C). Treatment with l-phosphatase produced a single dot that ran at the middle of the pH gradient. When cell extracts from the cbk1D strain were analyzed by 2D-WB, the relative abundance of the different Bcr1 isoforms was shifted to intermediate spots, corresponding to less phosphorylated isoforms ( Figure 2C). Thus, the phosphorylation state of Bcr1 appears to be Cbk1-dependent.
To obtain further evidence that Bcr1 interacts with Cbk1 in vivo, we performed co-immunoprecipitation experiments from extracts of cells expressing Cbk1-myc and Bcr1-HA from their native promoters. This approach clearly showed an in vivo interaction between these two proteins ( Figure 2D). Taken together, these data suggest that Bcr1 could be a direct target of the NDR/LATS kinase Cbk1 in vivo.

The Cbk1 consensus phosphorylation sites in Bcr1 are required for adherence and biofilm formation in vitro
To determine whether the putative Cbk1 phosphorylation sites are important for Bcr1 function in vivo, we mutated them and analyzed the effect of these mutations in biofilm formation. We constructed phosphomimetic and phosphodefective versions of Bcr1 by replacing the T191 and S556 putative Cbk1 phosphoacceptor residues with Glu or Ala, respectively. The mutant alleles were used to replace the wild-type copy of a heterozygous BCR1/ bcr1D strain, resulting in strains containing the mutant alleles under the control of the native promoter as the sole source of Bcr1 in the cell. The phenotypes of these bcr1 mutants were compared to those of the BCR1 heterozygous and bcr1D null mutant strains. It has been reported that bcr1D mutants exhibit severe defects in adherence and biofilm formation [8,9]. We did not observe any significant difference in adherence between the bcr1 T191E , bcr1 S556E and bcr1 T191E/S556E (referred to as bcr1EE) mutant strains and the BCR1/bcr1D reference strain ( Figure 3A). Furthermore, biofilm biomass production by the three phosphomimetic mutants was similar to that of the reference wild-type control and no significant statistical differences were found between them (bcr1 T191E , p = 0.0826; bcr1 S556E , p = 0.7472 and bcr1 T191E/S556E , p = 0.1282) ( Figure 3B). In addition, CSLM imaging revealed that biofilm architecture of these mutants was similar to that of the wild type ( Figure 3C). In contrast, all strains carrying mutations to alanine (bcr1 T191A , bcr1 S556A and bcr1 T191A/S556A referred to as bcr1AA) showed a significant decrease in the number of adherent cells and in biofilm development. These defects were more severe in the bcr1 T191A and bcr1AA mutants, which showed values closer to the bcr1D mutant, than in the bcr1 S556A mutant. Finally, CSLM imaging revealed that the biofilm structure of alanine mutants was very rudimentary and mainly composed of yeast cells, with few pseudohyphae and hyphae. However, the bcr1 S556A mutant exhibited a more organized and compact primary cell layer than that formed by the bcr1 T191A and bcr1AA strains, suggesting a different contribution of the two phosphorylation sites to Bcr1 regulation. Thus, these data indicate that the Cbk1 phosphorylation sites in Bcr1 are physiologically relevant for Bcr1 function during biofilm formation.

Cbk1-mediated phosphorylation of Bcr1 is required for full ALS3 and ALS1 expression
To determine whether Cbk1 phosphorylation sites in Bcr1 were important to regulate its transcriptional activity, we analyzed the expression levels of Bcr1-dependent genes in cells containing the phosphomutant Bcr1 proteins. Figure 4A shows ALS3 and ALS1 transcript levels measured by RT-PCR in the wild-type, bcr1D, bcr1EE and bcr1AA strains. Both ALS3 and ALS1 expression in bcr1EE cells was similar to that of the wild-type strain, whereas a significant reduction was observed in the bcr1AA mutant. Nevertheless, expression was still significantly higher in the bcr1AA mutant compared to the bcr1D null mutant, in which the transcripts were dramatically reduced, as previously described [8,9]. The decrease in the expression of these Bcr1-dependent genes observed in bcr1AA cells was not due to changes in Bcr1 abundance, since Bcr1 protein levels were similar in the bcr1AA-HA, bcr1EE-HA and BCR1-HA strains ( Figure 4B). Therefore, these findings suggest that phosphorylation of Bcr1 at Cbk1 phospho-acceptor sites is important for full expression of Bcr1-dependent genes ALS3 and ALS1.
The bcr1EE allele partially complements the cbk1D biofilm defect in vitro The hypothesis stated above predicts that the bcr1EE allele might suppress, at least partially, the biofilm defects observed in a cbk1D strain. We tested this prediction by integrating the bcr1EE or bcr1AA alleles in a cbk1D/cbk1D BCR1/bcr1D background. The adherence ability and biomass production of the double mutants were tested and compared to those of the parental strain. While the results obtained with the cbk1D bcr1AA mutant were similar to those found for the cbk1D reference strain, the introduction of bcr1EE in the cbk1D background resulted in a significant increase in the number of adherent cells and biomass ( Figure 5A and 5B). The cbk1D bcr1EE double mutant exhibited adherence values 3fold-higher than those of the cbk1D mutant and biofilm biomass production was doubled, although these values were still lower than those found in the wild type. CSLM imaging also revealed qualitative differences between cbk1D bcr1EE and cbk1D strains ( Figure 5C). Whereas the rudimentary biofilms formed by the cbk1D strain were composed of few groups of round yeast cells, the cbk1D bcr1EE strain produced a biofilm with higher cell density that included yeast and pseudohyphal cells. Therefore, the phosphomimetic bcr1EE mutant partially rescues the biofilm formation defects of the cbk1D strain.
Given the phenotypic differences between both strains, we next analyzed whether the introduction of bcr1EE allele had any effect on Bcr1 transcriptional activity. RT-PCR assays showed that ALS3 and ALS1 expression levels in the cbk1D mutant were dramatically decreased compared to wild-type levels ( Figure 5D). Notably, the cbk1D bcr1EE double mutant exhibited a significant increase in ALS3 and ALS1 transcripts versus the cbk1D reference strain (5.3fold increase for ALS3 and 3-fold increase for ALS1). Altogether, these results strongly suggest that the RAM pathway regulates Bcr1 transcriptional activity and, therefore, biofilm development, through Cbk1-dependent phosphorylation.

Analysis of Bcr1 phosphomutants in disseminated murine candidiasis
In addition to its role in biofilm formation, Bcr1 has been shown to contribute to C. albicans virulence. Indeed, inactivation of BCR1 results in a fitness defect in a mouse model of disseminated candidiasis when compared to the wild-type and other C. albicans knock-out mutant strains [29]. Hence, we tested whether the Cbk1dependent phosphorylation of Bcr1 could contribute to its function in virulence. Immuno-competent BALB-c mice were infected intravenously with C. albicans BCR1/BCR1, BCR1/bcr1D, bcr1D/ bcr1D, bcr1AA and bcr1EE strains and mortality was recorded over a 14-day period. As shown in Figure 6, inactivation of one copy of BCR1 did not impair C. albicans virulence while inactivation of the two BCR1 alleles resulted in significant reduction of C. albicans virulence (p,0.005), as anticipated from previous observations [29]. Similarly, the bcr1AA mutant showed reduced virulence, suggesting that Cbk1-mediated phosphorylation of Bcr1 is required for C. albicans virulence. The most striking result was obtained with the bcr1EE mutant that showed reduced virulence to a level similar to that observed with the bcr1D and bcr1AA mutants ( Figure 6). These results suggest that phosphorylation and dephosphorylation of Bcr1 is required for full virulence of C. albicans.

Discussion
Control of gene expression is a conserved function of NDR/ LATS kinases in eukaryotic cells [30,31]. The aim of this work was to assess whether the NDR/LATS kinase Cbk1 was involved in transcriptional control during C. albicans biofilm development. A survey for Cbk1 consensus sites in the C. albicans proteome allowed the identification of a subset of C 2 H 2 Zn finger transcriptional regulators of biofilm development as putative Cbk1 targets (Bcr1, 2 sites; Ace2, 3 sites; Nrg1, 3 sites; and Zap1, 3 sites). Since these proteins are required at different stages of biofilm formation [25,8,10,26], this observation might suggest that Cbk1 regulates transcriptional activity at different steps during C. albicans biofilm development. Here, we have studied the function of Cbk1 at early stages of biofilm formation, and our findings suggest that full activation of Bcr1 is dependent on Cbk1 phosphorylation.
Recently, a phenotypic analysis of a collection of protein kinase mutants identified Cbk1 as a kinase required for biofilm formation [21]. In agreement with this observation, we have found that RAM mutants have severe biofilm formation defects similar to that of the cbk1D strain. The impaired biofilm development of RAM mutants could be due to their inability to develop hyphae [32,16,17], since hyphal formation is important for biofilm development in C. albicans [33]. However, a cph1D efg1D strain defective in hyphae development is able to produce rudimentary biofilms and express a set of biofilm-related genes in a microfermentor model [7]. In contrast to cph1D efg1D cells, RAM mutants did not produce a compact layer of yeast cells in the same biofilm-formation model, suggesting a severe defect in cellsurface adherence. Given the central role of Bcr1 in this process, we rationalized that Cbk1 phosphorylation could be required for Bcr1 activation.
The results presented in this report indicate that Cbk1 regulates Bcr1 function during biofilm development, and this is based on the following evidences. First, biofilm mass production and adherence of RAM mutants were similar to that of bcr1D cells. Second, using 2D gels we have shown that the Bcr1 phosphorylation status was partially Cbk1-dependent. In addition, coimmunoprecipitation experiments revealed a physical interaction between Cbk1 and Bcr1. Third, the bcr1AA allele lacking Cbk1 phosphorylation sites phenocopied bcr1D cells, whereas the bcr1EE allele behaved as the wild-type, indicating that phosphorylation of the two Cbk1consensus sites is essential for Bcr1 function. In addition, the expression of Bcr1-dependent genes were reduced in bcr1AA cells as compared to the BCR1/bcr1D strain, suggesting that phosphorylation of Cbk1 sites in Bcr1 is required for full activation of the transcriptional program required for biofilm formation. In support of this idea, an upregulation in ALS3 and ALS1 expression was observed in the bcr1EE cbk1D strain in comparison to the cbk1D mutant (5.3-fold increase for ALS3 and 3-fold increase for ALS1). Given that overexpression of these adhesins in a bcr1D background restores biofilm formation in vitro and in vivo [9,34], the downregulation of ALS3 and ALS1 observed in bcr1AA cells could account for their phenotypic defects. Finally, the phosphomimetic bcr1EE allele partially rescued the biofilm defects of cbk1D cells. This partial rescue could be taken as an indication that Cbk1 has additional functions in biofilm development that are independent of the two Bcr1 phosphorylation sites (T191 and S556). One possibility could be that Cbk1 also plays a role in regulating the Ace2 transcription factor. In S. cerevisiae, Cbk1-dependent phosphorylation of Ace2 is required for its biological function [20]. In C. albicans, expression of Ace2 target genes depends on Cbk1 [17] and we and others have shown that Ace2 is required for adhesion to surfaces [25]. Alternatively, it is also possible that Cbk1 controls other regulatory factors required for biofilm development. This is based on the fact that ALS3 expression levels in bcr1EE cbk1D cells were still 2.6-fold lower than those of bcr1EE or BCR1/bcr1D cells ( Figure 5D), suggesting that full expression of ALS3 also requires Cbk1 in a Bcr1-independent manner. One possible scenario could be that Cbk1 also negatively regulates a repressor of ALS3 expression, such as Nrg1 that contains three Cbk1 consensus sites (S200, T251 and T281). The transcriptional repressor Nrg1 downregulates ALS3 transcription by binding to two regions in the ALS3 promoter [35]. Thus, it is possible that full activation of ALS3 expression during biofilm formation would require the inactivation of the Nrg1 repressor and the activation of the Bcr1 transcription factor, and that both events could be dependent on the phosphorylation of their Cbk1-consensus sites.
Another likely target of Cbk1 is the RNA-binding protein Ssd1 which contains multiple Cbk1-phosphorylation sites. In S. cerevisiae, the absence of Cbk1 in cells expressing functional Ssd1 severely impairs cell growth [36]. Recently, it has been shown that Ssd1 associates with specific mRNAs which encode proteins involved in bud growth and cell wall remodeling [37]. Cbk1-dependent phosphorylation of Ssd1 is required for translation of these mRNAs whereas Cbk1 inhibition promotes Ssd1 association with P bodies and thereby translational repression of Ssd1-associated mRNAs [37,38]. Given the growth defect of S. cerevisiae cbk1D cells is suppressed by deletion of SSD1, a functional Ssd1 is likely to cause a negative effect on cell growth that is inactivated by Cbk1. In C. albicans, the same genetic interaction between CBK1 and SSD1 has been described [16]. Therefore, the activation of Ssd1 could account for the biofilm defects shown by the RAM mutants. However, since expression of Bcr1EE is able to alleviate the biofilm defects of cbk1D cells, it is likely that Cbk1 controls biofilm formation, at least partially, through the regulation of Bcr1 transcriptional activity. Further work on the functional interaction between Cbk1 and Ssd1 is an interesting area to determine the role of translational control during biofilm formation in C. albicans.
Regulation through phosphorylation is an important mechanism for controlling the activity of transcription factors [39,40]. How Cbk1-dependent phosphorylation of Bcr1 regulates the transcriptional activity of the protein is unclear. In metazoans NDR/LATS kinase-dependent phosphorylation of YAP/TAZ coactivators results in their cytoplasmic retention and subsequent ubiquitination and degradation [23,24,41,42]. In S. cerevisiae, the   NDR/LATS kinase Cbk1 has the opposite functional output, since it promotes nuclear accumulation of the transcription factor Ace2 by phosphorylating two sites within its nuclear export sequence (NES) [20]. In C. albicans, our findings indicate that Cbk1 is not involved in regulating Bcr1 stability. Since Bcr1 does not contain an obvious NES close to the Cbk1 phosphorylation sites, we speculate that nuclear-cytoplasmic shuttling may not be a major output of Cbk1 phosphorylation. Therefore, a possibility is that the Cbk1-dependent phosphorylation might regulate the interaction of Bcr1 with other components of the transcription machinery.
Another significant conclusion of this work is that Cbk1dependent regulation of Bcr1 is important for virulence in C. albicans. As previously described [29], we found that cells lacking Bcr1 were less virulent in disseminated murine candidiasis. Supporting a role for Cbk1 in Bcr1 activation, the bcr1AA strain showed a similar reduced virulence. Strikingly, the phosphomimetic bcr1EE mutant also showed the same reduced virulence. Since C. albicans mutants that are locked in the yeast form are less virulent in disseminated murine candidiasis [43][44][45], one possibility is that the virulence defect of these mutants might be a consequence of a defect in hyphal formation. However, bcr1D cells [8] and BCR1 phosphomutants produced wild-type hyphae in hyphae-inducing conditions (not shown). Therefore, our results indicate that not only Cbk1-dependent phosphorylation of Bcr1, but also its dephosphorylation at specific moments of the infection process, are required for full virulence in in vivo models. Given that the addition of phosphates to specific residues of proteins can modify their interaction with other proteins [46,47], a dynamic Bcr1 phosphorylation state could be required to modify its interaction with other regulatory factors required for Bcr1dependent gene expression. This could aid in the colonization of different niches in response to environmental cues within the host. In C. albicans, developmentally regulated genes appear to be controlled by complex interactions between several transcription factors at their promoters [35,48]. The ALS3 gene is a good example of such complexity, since its expression is regulated by multiple transcription factors, including Efg1, Cph1, Bcr1, Nrg1, Rfg1 and Tup1 [35]. However, we could not exclude that the virulence defects observed in phosphomimetic bcr1 mutants were due to conformational changes that might reduce its function during infection.
In sum, our results indicate that the RAM signaling pathway plays an important role in regulating Bcr1 function through its phosphorylation at two specific residues mediated by the Cbk1 kinase and that this phosphorylation is required for proper biofilm formation and virulence. Interestingly, the two phosphorylation sites are also conserved in other species, such as Candida dubliniensis or Candida tropicalis. In the Candida parapsilosis Bcr1 ortholog, which is also required for biofilm formation [49], the putative Cbk1 phosphorylation site at the end of the first Zn finger domain (T214) is also present. Therefore, the role of the RAM pathway in the control of Bcr1 function might be conserved in other biofilmforming species. Table 1. Strains used in this study.

Name
Genotype Source

Ethics statement
All animal experiments adhered to the EU Directive 86/609 on the approximation of laws, regulations and administrative provisions of Member States regarding the protection of animals used for experimental and other scientific purposes, and to related national regulations. All experiments were performed according to the guidelines of the European Convention for the Protection of Vertebrate Animals Used for Experimental and Other Scientific Purposes (ETS No. 123). The protocol was approved by Institut Pasteur Health Center Animal Care Committee (Protocol number: 10.455). Recovery of organs was performed following euthanasia of animals, and all efforts were made to minimize suffering.

Media
C. albicans strains were grown in YPD (2% Bacto Peptone, 2% dextrose, 1% yeast extract) for transformation experiments. Transformants were selected on synthetic medium (2% dextrose, 6.7% YNB and auxotrophic supplements) or YPD supplemented with clonNAT at a final concentration of 200 mg ml 21 (Werner BioAgents) for Nat+ strains. For adherence and biofilm formation assays using microfermentors, overnight cultures were grown in synthetic medium with 0.4% dextrose supplemented with arginine, histidine and uridine (20 mg ml 21 ) and then diluted in the same medium supplemented with methionin (200 mg ml 21 ). For the biofilm formation on silicone squares, cells were grown in synthetic medium with 0.9% dextrose and supplemented with arginine, histidine, uridine and methionin at the concentrations mentioned above.

Plasmid and strain construction
The strains used in this study are listed in Table 1 and derived from the BWP17 strain [50]. Disruption and epitope-tagged strains were made according to the PCR-mediated system using pFA plasmids [51,52]. Strains were confirmed by PCR. Oligonucleotides were obtained from Biomers.net (Ulm, Germany) and are listed in Table 2. For site directed mutagenesis, fragments carrying bcr1 T191A/E or S556A/E modifications were generated by PCR from genomic DNA of the JC1084 strain. The PCR products were inserted into pSC-A vector (Strataclone PCR cloning kit) for sequencing and then digested with both EcoRV and ClaI for the T191 fragment or BstBI for the S556 fragment, giving rise to a DNA fragment of 798 bp or 666 bp, respectively. These fragments were used to swap the wild-type BCR1 regions of a pGEM-URA3based vector, that contained the complete wild type BCR1 ORF fragment cloned into the SacI site and a 39UTR-BCR1 region inserted into the NotI site (+194 to +704 from the stop codon) to direct integration of the plasmid to the BCR1 locus. The constructions were digested with SalI and AflII and transformed into a BCR1 heterozygous strain using a standard lithium acetate method [53]. The transformants were confirmed by PCR and sequencing. Strains used for biofilm formation assays were transformed to uracil and histidine prototrophy using StuIlinearized CIp10 [54] or SacI/SacII-digested ECC72 plasmids respectively. ECC72 contains 2.9 kb from the HIS1 locus (from 1000 bp upstream of the start codon to 1000 bp downstream of the stop codon) which were PCR amplified and cloned in the pGEM-T (Promega) EcoRV site.

Adherence
Surface adherence was determined incubating plastic slides (Thermanox, Nunc) with OD 600 = 1 cell cultures during 1 hour at room temperature and washing three times with PBS to remove non-adherent cells. Cells attached to the plastic surface were recovered by vortexing and resuspended in 15 ml of PBS, using 5 ml for counting the adherent cells with an optical microscope. For each strain the average of cells/field is the result of counting a total of 30 fields from three independent experiments.

Biofilm formation
Microfermentor assays were done as described previously [7] while the method of the silicone squares used for microscopy analysis was described in [12]. For dry mass measurements, after the incubation period in the microfermentor, the plastic slides were removed from the glass spatula and transferred to a 50 ml tube with 10 ml of water. Cells were detached from the plastic surface by vortexing and then recovered by filtration using Millipore 1.2 mm filters. Filters were dried during 48 hours at 60uC prior mass measurement. Total biomass of each biofilm sample was calculated by subtracting the mass of a blank filter subjected to the same washing and drying treatment. For each strain the experiment was repeated twice in duplicate.

Microscopy and image analysis
Biofilm development in microfermentors was recorded with a Nikon Coolpix digital camera. For the silicone square experiments, biofilms were observed by CSLM, after staining with calcofluor white 0.01% (vol/vol) and 10 mg ml 21 concanavaline A Alexa fluor 594 conjugate (Invitrogen) for 1 h in the dark at 37uC with 150 rpm agitation. CSLM was performed with an upright Zeiss Axioskop2 FS MOT LSM 510 multiphoton microscope using a Zeiss Achroplan 640/0.8 W objective. All CSLM stacks were assembled into projections using the Image J software.

RNA isolation and expression analysis
Cultures for RNA extraction were done as previously described [9]. For RNA extraction mid-logarithmic cell cultures were harvested, frozen in liquid nitrogen and stored at 280uC. 50 mg of cells were lysed in a FastPrep cell disruptor in the presence of 50 ml TRIzol Reagent (Invitrogen). Total RNA was isolated according to manufacturer's instructions. The quantity, quality and integrity of the RNA were analyzed in an Agilent's 2100 Bioanalyzer system. cDNA synthesis was performed with the SuperScript II First-Strand Synthesys System (Invitrogen) using oligo(dT), from 3 mg total RNA previously treated with DNAase I (Invitrogen). qPCR assays were done using SYBR Premix Ex Taq (TaKaRa). For each reaction, 1 ml cDNA was used. No-template and no-reverse transcription controls were included. The assays were carried out in duplicates at generic cycle conditions (95uC for 45 s and 40 cycles of 95uC for 5 s and 60uC for 31 s, followed by a dissociation step at 95uC for 15 s, 60uC for 1 min and 95uC for 15 s) in an Applied Biosystems 7300 Real-Time PCR System (Applied Biosystems). Relative gene expression quantification was achieved using the Pfaff1 method.

Disseminated murine candidiasis assays
Nine week-old female Balb c/J mice were infected intravenously with 5610 5 CFU/mouse. Ten mice per strain were tested per experiment and mice were kept in groups of 5 per cage. 2 days after infection 2 mice/strain were sacrificed, the kidneys taken, homogenized and plated on SD medium containing 50 mg/ml ticarcillin and 10 mg/ml gentamycin. The fungal burden was determined by counting CFUs after 2 days at 30uC. Mice were checked 2 times per day for survival over a period of 14 days. At day 14 all remaining mice were sacrificed. Results were analyzed using the software GraphPad Prism5.