Systematic analysis of the Candida albicans kinome reveals environmentally contingent protein kinase-mediated regulation of filamentation and biofilm formation in vitro and in vivo

ABSTRACT Protein kinases are critical regulatory proteins in both prokaryotes and eukaryotes. Accordingly, protein kinases represent a common drug target for a wide range of human diseases. Therefore, understanding protein kinase function in human pathogens such as the fungus Candida albicans is likely to extend our knowledge of its pathobiology and identify new potential therapies. To facilitate the study of C. albicans protein kinases, we constructed a library of 99 non-essential protein kinase homozygous deletion mutants marked with barcodes in the widely used SN genetic background. Here, we describe the construction of this library and the characterization of the competitive fitness of the protein kinase mutants under 11 different growth and stress conditions. We also screened the library for protein kinase mutants with altered filamentation and biofilm formation, two critical virulence traits of C. albicans. An extensive network of protein kinases governs these virulence traits in a manner highly dependent on the specific environmental conditions. Studies on specific protein kinases revealed that (i) the cell wall integrity MAPK pathway plays a condition-dependent role in filament initiation and elongation; (ii) the hyper-osmolar glycerol MAPK pathway is required for both filamentation and biofilm formation, particularly in the setting of in vivo catheter infection; and (iii) Sok1 is dispensable for filamentation in hypoxic environments at the basal level of a biofilm but is required for filamentation in normoxia. In addition to providing a new genetic resource for the community, these observations emphasize the environmentally contingent function of C. albicans protein kinases. IMPORTANCE Candida albicans is one of the most common causes of fungal disease in humans for which new therapies are needed. Protein kinases are key regulatory proteins and are increasingly targeted by drugs for the treatment of a wide range of diseases. Understanding protein kinase function in C. albicans pathogenesis may facilitate the development of new antifungal drugs. Here, we describe a new library of 99 protein kinase deletion mutants to facilitate the study of protein kinases. Furthermore, we show that the function of protein kinases in two virulence-related processes, filamentation and biofilm formation, is dependent on the specific environmental conditions.

is due, at least in part, to its relatively tractable genetics compared to some other fungal pathogens, particularly in the pre-CRISPR-Cas9 era (4).Indeed, C. albicans was one of the first human fungal pathogens for which large-scale mutant collections were available (4)(5)(6).Since that time, collections of deletion mutants involving defined functional classes of proteins or near genome-wide coverage of proteins have been developed for C. glabrata (7), Cryptococcus neoformans (8,9), and Aspergillus fumigatus (10) using a variety of genetic technologies.In C. albicans, large-scale collections of strains containing alleles under the control of a tetracycline-regulated promoter (11) or comprised of bar-coded heterozygous deletion mutans (12) are also available.These genetic tools have been invaluable for systematic gene discovery and functional analysis.
One of the most widely employed C. albicans mutant collections is the bar-coded transcription factor (TF) deletion set constructed by the Johnson Laboratory (5).It has been used for many single-mutant analyses as well as pooled-competition experiments using both in vitro and in vivo models of C. albicans biology and pathogenesis (13)(14)(15)(16).Deletion mutants of regulatory proteins such as TFs are powerful tools for the analysis of complex phenotypes, particularly when combined with network and expression analysis (16,17).A second broad category of information-rich regulatory proteins is protein kinases (PKs).PKs frequently serve as upstream activators of TFs and function within discrete signaling pathways to relay information within the cell.In addition to their value in the molecular analysis of complex biological processes, PKs are among the most widely drugged targets in modern pharmacology.Because the antifungal pharmacopeia is limited to only three classes of drugs for the treatment of life-threatening fungal infections, PKs represent a reasonable set of proteins to target for the development of new antifungal drugs.
Based on the above considerations, we felt that a systematic set of PK deletion mutants would be a useful addition to the current palate of C. albicans genetic resources.Prior to the initiation of this project, Blankenship et al. had reported and characterized a large set of PK transposon insertion mutants (18).While the project was underway, another group reported the use of a PK deletion set (86 PKs) generated in the SC5314 background to identify PKs involved in iron utilization (19).In addition, Lee et al. reported the use of a set of tetracycline-regulated PKs to characterize the kinome relating to in vitro filamentation (11).We adopted an approach modeled on that used by Homann et al. to generate the TF deletion set mentioned above (5).Specifically, the PKs were generated in the same genetic background (SN) to allow a more direct comparison of TF and PK phenotypes.We also used the same set of barcode sequences to mark the PK deletion mutants to facilitate the combined use of these libraries in the analysis of a given phenotype without having to re-optimize primer sequences.Finally, the library will be made publicly available through the Fungal Genetics Stock Center.
Here, we report the construction of a non-essential PK deletion set and its charac terization under a variety of growth conditions.We have also screened the library to identify PKs required for two related C. albicans virulence traits: (i) filamentation and (ii) biofilm formation.We find that each of these traits is dependent upon an extensive network of PKs and identify multiple PKs not previously associated with filamentation or biofilm formation.We also show that two MAPK pathways, the cell wall integrity (CWI) and hyper-osomotic-glycerol (HOG) pathways, play previously underappreciated or unrecognized roles in filamentation with the Mkc1 MAPK of CWI pathway required for the maintenance phase of filamentation in vitro.Finally, we demonstrate that the HOG is much more important during in vivo biofilm formation than would be predicted from its effect on in vitro biofilm formation.

Construction of protein kinase deletion set in the SN genetic background
A set of 119 PKs and PK-related genes were collated from databases and previous publications.The parental strain SN152 (Arg − , His − , Leu − ) was used to generate the deletion mutants through a CRISPR-Cas9 approach (Fig. 1A).The ORFs were replaced using a LEU2 auxotrophic marker containing 1 of 30 barcode sequences in the 5′ flanking region.The barcode sequences in the PK mutants are those previously used by Homann et al. in the construction of the TF deletion library (5).Therefore, a common set of barcodes can be used to interrogate both TF and PK deletion sets in the same genetic background.A p5 Illumina adaptor sequence was also placed upstream of the barcodes so that the region could be amplified using a forward primer complementary to the adaptor and a reverse primer homologous to a region of the LEU2 marker (see Fig. S1A for the schematic of the full construct).The homozygous PK deletion strains match the auxotrophic markers of the SN95 reference strain (Arg − , His − ).Previous studies by Noble and Johnson have shown that the virulence properties of SN95 are comparable to a prototrophic strain in a mouse model of disseminated candidiasis (20).
The open reading frame for the targeted PKs was replaced by a LEU2-marked deletion construct using a split-marker approach that results in a pair of inverted repeats on the 5′ and 3′ ends of the LEU2 gene.The inverted repeats can, therefore, be used to recycle the LEU2 marker using the CRISPR-Cas9-induced-excision (CRIME) approach developed by Huang and Mitchell while leaving the barcode in place (21).As a result, the PK mutants have a wide variety of genetic markers available for additional strain construction and modification.In addition, they can be readily returned to prototrophy using plasmids available in the community.The deletion set will be available from the Fungal Genetic Stock Center following publication.
Of the 119 ORFs selected, homozygous deletion mutants were obtained for 99 PK genes (Fig. 1B); we obtained heterozygous but not homozygous deletions for 20 of the targeted PKs.Table S1 contains lists all PK homozygous mutants along with the barcode for each mutant as well as the oligonucleotides and guide RNAs used to construct the strain.Because our goal was to construct a large deletion set, we did not exhaus tively optimize transformation conditions for PK genes that did not yield homozygous mutants after 2-3 independent attempts.In other organisms, many PKs are essential and, therefore, we compared our set of 20 genes to the literature of essential genes and the specific locus through the Candida Genome Database.As shown in Fig. 1C and Table S2, 15 of the 20 mutants were classified as essential in the literature.Three of the five mutants that were not classified as essential [PKC1 (22), YCK2 (23), and VPS34, (24)] have been constructed.It is not clear why we could not obtain mutants for these PKs.The literature strains were obtained through sequential deletion of the alleles, whereas we used on one-step deletion approach.It is possible that the heterozygous mutant is able to adapt to the reduced function of the PK and, therefore, is better positioned to tolerate the loss of the gene altogether.

Quantitative phenotyping of PK mutants under diverse growth conditions using competitive fitness assays
PKs have a broad range of functions.To begin to characterize these functions, the growth of each PK deletion mutant was compared to the auxotrophic marker-matched SN95 strain using a flow cytometry-based competitive fitness assay (Fig. 2A).SN95 was fluorescently labeled with mNEON using a previously reported plasmid.A 1:1 mixture of mNEON-SN95 and an unlabeled PK mutant was then incubated in microtiter plates under each condition.The ratio of SN95 to PK mutant in the culture at time 0 and 24 hours was determined by gating on the mNEON signal (SN95) and forward scatter (total cells).The PK mutants were first compared to SN95 in rich medium (YPD) at 30°C to assess general growth effects.A representative example of raw data for the ckb1∆∆ mutant at 30°C in YPD is shown in Fig. 2B.Seven PK mutants (ckb1∆∆, frk1∆∆, crk1∆∆, rad53∆∆, sok1∆∆, cla4∆∆, and mss2∆∆) were less fit than SN95 under these conditions (log 2 fitness score [FS] ≤ −1, P value < 0.05, Student's t-test; Table S3).
To compare the fitness scores to more typical plate-based growth assays, we first examined the discordant phenotype of the regulatory (CKB1) and catalytic subunits (CKB2) of the B component of the casein kinase complex (Fig. 2C).Consistent with the competitive fitness data, the ckb1∆∆ mutant showed a growth defect at 30°C on solid agar YPD media while the ckb2∆∆ mutant did not.Three additional mutants with reduced FSs in YPD along with one mutant with no phenotype (bub1∆∆) and one with an increased FS (kin3∆∆) were also tested.The mutants with reduced FS and WT fitness showed similar phenotypes while the kin3∆∆ mutant did not show evidence of an increased growth rate (Fig. 2D); however, it is not clear that this assay would be the most sensitive for such a phenotype.The frk1∆∆ and crk1∆∆ mutants had very similar FSs (0.26 and 0.30, respectively; Table S3) and showed similar growth phenotypes on solid agar.Based on these data, it appears that the FS provides a reasonably sensitive and quantitative measure of mutant fitness.FSs under all other conditions were normalized to YPD at 30°C to control for intrinsic growth defects.
The fitness of each PK mutant was determined in 10 different conditions including temperature, salt, membrane, osmotic, and pH stress.In addition, three non-glucose carbon sources were examined.The results of this screening along with the hierarchi cal clustering of the mutants are summarized in the heat map shown in Fig. 3A.The components of the cell wall integrity (CWI) and hyperosmotic glycerol (HOG) pathways cluster together as expected (25,26).Similarly, Sak1 functions upstream of the key metabolic kinase Snf1 and, consistent with those functions, the paired cluster together as do the Sak1-Snf1 pair (27).In addition, two relatively unstudied but paralogous kinases, Sky1 and Sky2 (28), cluster closely as well.By contrast, the two catalytic subunits of the cAMP-PKA kinase (TPK1 and TPK2) and the individual components of the CK2 casein kinase complex (Cka1, Cka2, Ckb1, and Ckb2; Fig. 2C) do not cluster together.The two catalytic subunits of PKA show distinct filamentation phenotypes in C. albicans (29).In the case of the latter, the discordance may be because the loss of the regulatory subunits of the CK2 complex leads to distinct phenotypes compared to the loss of function of the catalytic subunits.
Table S4 summarizes the sets of mutants with a log 2 FS ≤ −1 with a P value < 0.05 (Student's t-test) in each of the 10 conditions.Bud32 and Mec1 were the only two mutants with reduced fitness in all conditions.Although neither PK has been extensively studied in C. albicans, both Bud32 and Mec1 are conserved across fungi and have been characterized in other systems.Bud32 is part of the kinase, putative endopeptidease, and other proteins of small size (KEOPS) complex that regulates multiple critical cellular functions such as telomere regulation, cell cycle progression, and chromatin/transcrip tion (30).Mechanistically, at least part of the pleomorphoric functions of the KEOPS complex is due to its role in tRNA modification.Although C. albicans Bud32 has not been previously studied in detail, the extensive phenotypes observed for the bud32∆∆ mutant are consistent with reported functions in other fungi such S. cerevisiae and Cryptococcus neoformans.Mec1 is part of the Mec1-Rad53 DNA damage signaling pathway and, therefore, the pleomorphic phenotypes are not surprising (31).
Because C. albicans SNF1 was long thought to be essential, its effect on non-carbon metabolism phenotypes has not been investigated (27).In S. cerevisiae, Snf1 is required for cellular responses to other environmental stresses in addition to glucose limitation and is regulated by SAK1 (32).Consistent with their known functions during growth on non-glucose carbon sources, both the snf1∆∆ and sak1∆∆ mutant grew poorly on media containing galactose, glycerol, and acetate (Fig. 3B).Both the sak1∆∆ and snf1∆∆ mutants have reduced fitness under osmotic, alkaline, and acidic stresses (Fig. 3B).However, unlike the S. cerevisiae snf1∆ mutant, the C. albicans snf1∆∆ mutant is not sensitive to elevated temperature (Fig. 3B).Although loss of Snf1 affects many cellular processes, it does not lead to a global loss of fitness.
Finally, the utility of phenotypic hierarchical clustering has significant limitations, particularly in cases where kinases thought to function together do not cluster as expected.This is most likely due to the relatively small set of conditions we used for our analysis and the broad functions of PKs.PKs that cluster together are likely to have related functions; however, the lack of clustering does not mean that they do not have related functions under other conditions.For example, despite functioning in the same DNA damage pathway, the mec1∆∆ and rad53∆∆ mutants do not have overlapping phenotypes; however, we did not examine DNA damage conditions and, therefore, this discordance is likely due to differences in compensatory responses or other indirect processes.As such, the clustering results have high specificity in the sense that clustered PKs are likely to have similar functions under those specific conditions, but lacks sensitivity and cannot exclude the likelihood that PKs have related functions in other conditions.

An extensive network of PKs regulates C. albicans hyphal morphogenesis in vitro
The ability to transition between yeast and hyphal morphologies is one of the most well-studied C. albicans virulence traits.Many genes including PKs have been identified as contributors to this phenotype through gene-specific studies as well as large-scale studies.For example, Lee et al. recently reported a screen of strains containing tetracy cline-regulated PK mutants for those that regulate filamentation (11).To systematically characterize the filamentation phenotypes of our PK deletion mutant set, we used a plate-based primary screen in which the PK mutants were spotted on YPD, Spider medium, RPMI, and RPMI + 10% bovine calf serum (BCS) agar plates.The plates were incubated at either 30°C or 37°C with an auxotrophic marker-matched reference strain (SN95) control on each plate; the screen was performed in biological duplicate.Except for the YPD experiments, plates were photographed on incubation days 3, 4, and 5.Each strain was scored from −3 to 0 for peripheral colony filamentation using the system represented in Fig. S1B.Negative scores were based on filamentation at day 5 while mutants were classified as having +1 or +2 if they showed an increased filamentation score prior to day 5.The YPD plates were incubated at either 30°C or 37°C and photo graphed on days 1, 2, and 3 to identify colonies with evidence of filamentation when SN95 showed none.We recognize that central wrinkling is also a widely used filamentation phenotype but scoring this phenotype was poorly reproducible in our hands while filamentation was consistent between replicates (5,6).Therefore, we focused on peripheral filamentation.
Remarkably, 88 of 99 PK mutants tested had altered filamentation on at least one of the six filament-inducing conditions (the filamentation phenotypes for each strain are summarized in Table S5).In comparison, 40/155 transcription factor mutants had altered filamentation according to the screen reported by Homann et al. (5).Of this set, 18 mutants showed increased filamentation and 69 had a reduced filamentation score under at least one condition.All of the mutants with growth defects at either 30°C or 37°C had reduced filamentation.However, all mutants with growth defects formed colonies on the solid agar plates so it is not clear how much the tempera ture sensitivity contributed to the reduced filamentation.Although no PK mutant was hyper-filamentous on all conditions, 30 PK mutants had negative filamentation scores on all six hypha-inducing media and temperature combinations (Fig. 4A).Of the 30 PK mutants that showed reduced filamentation on all six conditions, 12 PK mutants (we excluded MEC1, RAD53, and BUD32 due to their pleomorphic growth defects) had not been previously reported to show decreased filamentation through a search of the CGD database and PubMed literature.
In this set of PK mutants, only two mutants showed evidence of filamentation on non-inducing conditions (YPD medium): cak1∆∆ and swe1∆∆.Both of these mutants showed dramatically increased colony wrinkling that was evident after 24 hours at 37°C but not at 30°C (Fig. S1C).We did not observe peripheral invasion for these strains up to 72-hour incubation.Cak1 has previously been shown to repress filamentation (11,33) while swe1∆∆ has been reported to have no filamentation phenotypes (34).As noted in Fig. 4A, the swe1∆∆ mutant has decreased filamentation under inducing conditions and increased filamentation under non-inducing conditions.Swe1 is involved in cell cycle processes and elongated cell morphology is a well-established phenotype of cell cycle mutants, providing a possible explanation for the phenotype.
Representative filamentation phenotypes on RPMI at 37°C are shown in Fig. 4B.The CKA2 deletion was reported to have no phenotype in another background but the filamentation conditions tested were not reported (35).No previous studies on the filamentation of the ckb2∆∆ mutant have been reported.Our data indicate that the function of the CK2 complex extends to filamentation and is likely to contribute to the reduced virulence of CK2 complex mutants (35).The Hyperosmotic Glycerol pathway (HOG) is annotated as suppressing filamentation in the CGD database.However, the literature indicates that the role of HOG1 in filamentation is more complex.In their initial characterization of the hog1∆∆ mutant, the Pla laboratory found that its phenotype was dependent on conditions.On solid Spider medium, the hog1∆∆ mutant shows no filamentation while the addition of low concentrations of serum to liquid YPD led to early filamentation of the hog1∆∆ mutant relative to its parental strain (36); the latter phenotype has led, in part, to its classification as a repressor of filamentation.We found that three kinases in the HOG pathway (SSK2, PBS2, and HOG1) all showed reduced filamentation with HOG1 and SSK1 mutants deficient under all conditions tested (Fig. 4A).The reduced filamentation phenotypes for hog1∆∆ and ssk1∆∆ mutants are shown in Fig. 4B and, therefore, suggest that the HOG pathway plays an under-appreciated role in the positive regulation of filamentation and a more limited role as a repressor of filamentation.
The literature is also somewhat mixed with respect to the role of the Cell Wall Integrity pathway (CWI) during filamentation.In their initial report on the mkc1∆∆ mutant, Navarro-Garcia found that it had reduced filamentation on solid Spider medium (37).Kumamoto demonstrated that Mkc1 plays a key role in contact-induced filamentation (38).By contrast, Xie et al. found that, among the kinases of the CWI pathway (22), the pkc1∆∆ mutant but neither the bck1∆∆ nor mkc1∆∆ had reduced filamentation in YPD + 10% BCS.On solid medium, we found that mutants of each CWI component MAP-family kinase (BCK1, MKK2, and MKC1) had reduced filamentation (Table S5) on at least one condition with the bck1∆∆ mutant deficient in each condition (Fig. 4A and  B).The fact that the bck1∆∆ mutant has consistently reduced filamentation on multiple inducing conditions but mutants of the other CWI pathway components do not suggest that they play roles other than simply regulating the next PK in the cascade.
Lastly, it is also important to note that our screening results recapitulated phenotypes for many of the previously reported PK mutants.For example, the venerable cAMP-pro tein kinase A pathway mutants matched the previously reported pattern in which the TPK1 mutant has reduced filamentation under most conditions while the TPK2 mutant filaments normally (Fig. 4B) (29).The RAM pathway component kinases KIC1 and CBK1 (Table S5; Fig. 4B) are also both required for filamentation (39).Overall, the role of PKs in filamentation appears to be highly dependent upon the specific conditions under which C. albicans filamentation is induced.

PK-NRG1 genetic interactions provide insights into the timing of PK functions during filamentation
The conserved repressor Nrg1 plays a central role in the initiation of C. albicans filamentation (40,41).Upon transfer to filament-inducing conditions, the Nrg1 pro tein is rapidly degraded and the expression of its gene, NRG1, is reduced.This leads to activation of the hyphal transcriptional program and formation of hyphae (42).Constitutive expression of NRG1 from a heterologous promoter blocks filament formation in vitro and during infection (43).Conversely, deletion of NRG1 leads to initiation of filamentation (pseudohyphae) in the absence of typical hypha-inducing stimuli (40,41).Therefore, candidate genes could function either upstream of the relief of Nrg1 repression or downstream: the filamentation phenotypes of mutations in genes functioning upstream should be suppressed by deletion of NRG1 while those involved in positively regulating the hypha program after relief of Nrg1 repression should have altered filamentation relative to the nrg1∆∆ mutant.Therefore, we constructed a set of pk∆∆ nrg1∆∆ double mutants and characterized their effects on nrg1∆∆ mutant pseudohyphae formation in YPD and the effect of the nrg1∆∆ mutation on reduced hypha formation of PK mutants during hyphae induction (Fig. 5A).For these experi ments, we used liquid culture-based induction conditions to quantitatively characterize the distribution of cellular morphotypes.It is important to note that filamentation and invasion of solid medium, for instance, do not always correlate.
We chose 7 PK mutants with filamentation phenotypes and generated pk∆∆ nrg1∆∆ double mutants.The HOG (pbs2∆∆ and hog1∆∆ mutants) and CWI pathway (mkc1∆∆ and bck1∆∆ mutants) mutants were chosen because their roles have not been extensively studied.Similarly, at the time of our work, neither Ire1 nor Sky1 had been extensively characterized.Sok1 was also included because deletion of NRG1 has been shown to restore filamentation to the sok1∆∆ mutant (44).We first examined the effect of the seven PK mutations on pseudohyphae formation by the nrg1∆∆ mutant in YPD at 30°C and compared the results to their corresponding nrg1∆∆ double mutant along with SN95 and the nrg1∆∆ single mutant.
As expected, the nrg1∆∆ mutant formed primarily pseudohyphae along with a few hyphae and very few yeast under these conditions (Fig. 5B).The seven PK mutants grew almost exclusively as yeast with rare pseudohyphae.The hog1∆∆ and pbs2∆∆ mutants were the only mutants to form quantifiable hyphae to any extent (<10%).These observations may be related to previous reports indicating that HOG mutants form increased hyphae under specific conditions (36).Each PK nrg1∆∆ double mutant formed an increased proportion of yeast relative to the parental nrg1∆∆ mutant (Fig. 5B).Hog1, Bck1, and Sok1 had the strongest effects on pseudohyphae formation under these conditions.Ire1 and Mkc1, on the other hand, had minimal effects on pseudohyphae formation (Fig. 5B).These data indicate that Hog1, Bck1, and Sok1 modulate pseudo hyphae formation after relief of Nrg1 repression under non-inducing conditions.The phenotype of the sok1∆∆ nrg1∆∆ mutant is quite striking.Under inducing conditions, the Sok1 kinase is thought to phosphorylate Nrg1 leading to its degradation and, consequently, relief of hyphal expression (44).If that were the only role that it plays, then the sok1∆∆ nrg1∆∆ mutant should phenocopy the nrg1∆∆ mutant.The sok1∆∆ nrg1∆∆ mutant shows a dramatic increase in yeast formation relative to nrg1∆∆, indicating Sok1 must have at least some distinct roles during pseudohyphae formation.
Next, we asked whether the deletion of NRG1 suppressed the filamentation defects of PK mutants.To do so, we exposed the nrg1∆∆ double mutants to filament induction in RPMI + 1% BCS instead of 10% BCS to increase sensitivity for more subtle phenotypes.The sok1∆∆ nrg1∆∆ double mutant forms hyphae to a similar extent as the nrg1∆∆ mutant while the sok1∆∆ single mutant forms almost no hyphae (Fig. 5C).This is consistent with previous observations and the model that Sok1 mediates degradation of Nrg1 during hyphae initiation (44).Like the sok1∆∆ mutant and consistent with plate-based assays, the ire1∆∆ mutant forms essentially no hyphae or pseudohyphae under these inducing conditions (Fig. 5C).In this case, however, deletion of NRG1 does not restore hyphae formation as the ire1∆∆ nrg1∆∆ double mutant forms predominately yeast and pseudohyphae.Thus, it appears that Ire1 contributes to the initiation of filamentation after relief of Nrg1 repression.It is also clear that Ire1 is critically required for C. albicans filamentation but not pseudohyphae formation in the nrg1∆∆ mutant (Fig. 5B).
Deletion of the HOG pathway PKs Hog1 and Pbs2 caused increased pseudohyphae relative to hyphae with a very minor increase in yeast (Fig. 5C).In this case, the two components of this linear pathway had very similar phenotypes.The nrg1∆∆ double mutants of these PKs, however, had very different phenotypes.In the case of the hog1∆∆ mutant, deletion of NRG1 had a minor effect on the distribution of morphotypes relative to the hog1∆∆ single mutant (Fig. 5C).The inability of a nrg1∆∆ mutation to affect the hog1∆∆ mutant phenotype is consistent with Hog1 functioning mainly after relief of Nrg1 repression.In stark and surprising contrast, the pbs1∆∆ nrg1∆∆ mutant forms almost no hyphae and is predominantly in yeast form.This phenotype is consistent with Pbs2 also functioning downstream of the step relieving Nrg1.
The distinctions between the phenotypes of the hog1∆∆ nrg1∆∆ and pbs1∆∆ nrg1∆∆ mutants are striking.First, these observations indicate that Pbs2 must have functions in addition to activating Hog1 during filamentation and is a conclusion that has prece dent in other studies of the HOG pathway in C. albicans.Second, the pbs2∆∆ mutant forms hyphae and, consequently, the strong reduction in filamentation in the pbs2∆∆ nrg1∆∆ mutant indicates that Pbs2 is required for initiation of filamentation only in the absence of NRG1.As such, this is a phenotype unique to the double mutant and, therefore, is a synthetic genetic interaction.The synthetic phenotype between Nrg1 and Pbs2, therefore, implies that Nrg1 must have a role beyond repressing the initiation of filamentation.Nrg1 cannot simply be an on-off switch for filamentation; otherwise, all of the double mutant phenotypes would be explainable by an additive interaction between the phenotypes of the PK mutant and the nrg1∆∆ mutants.Supporting this model is the fact that Nrg1 protein is only transiently reduced at the initiation of filamentation but is readily detectable after 60 min to 90 min of hyphal induction in most media (45).Furthermore, the expression of NRG1 is reduced during filamentation but remains readily detectable throughout filamentation in vitro and in vivo.Thus, there is only a brief window during which Nrg1 levels are dramatically low and there is an opportunity for Nrg1 to play additional roles beyond its canonical function.
Further supporting the above assertion are the phenotypes of the nrg1∆∆ double mutants with CWI pathway PKs Bck1 and Mkc1.Consistent with previously reported data in liquid YPD + 10% BCS hyphal induction conditions (22), neither the bck1∆∆ nor the mkc1∆∆ mutant had a significant reduction in hyphae formation in RPMI + 10% BCS (Fig. 5C).The bck1∆∆ nrg1∆∆ and mkc1∆∆ nrg1∆∆ double mutants, on the other hand, formed significantly fewer hyphae relative to either single mutant with a corresponding increase in pseudohyphae (Fig. 5C).Once again, these are synthetic phenotypes and, as such, imply that the CWI pathway is critically important during the period of the hyphal program when Nrg1 repression is relieved.

The CWI pathway contributes to the maintenance phase of hyphal morpho genesis in vitro but not in vivo
To test the hypothesis that the CWI pathway functions primarily during hyphal elonga tion under liquid inducing conditions, we first measured the length of hyphae formed under these conditions.Supporting the hypothesis, the length of the filaments formed by SN95 is significantly longer than the filaments formed by the corresponding mkc1∆∆ mutant after 4 hours of in vitro induction (Fig. 5D).If the CWI pathway is important for the maintenance or elongation phase of hyphal morphogenesis then phosphorylation of Mkc1, the terminal kinase, should occur during filamentation.To test this hypothesis, we used an antibody that recognizes the phosphorylated form of Mkc1.After 4 hours of hyphae induction in RPMI +10% BCS, Mkc1 phosphorylation increased relative to yeast phase cells in rich medium (Fig. 5E); this is also consistent with previous data reported by Kumamoto in YPD medium (38).Because deletion of NRG1 leads to constitutive induction of the filamentation program and deletion of MKC1 in the nrg1∆∆ background increases the proportion of yeast (Fig. 5B), we asked whether Mkc1 was phosphorylated in the nrg1∆∆ mutant under non-inducing conditions.Indeed, Mkc1 is phosphorylated in the nrg1∆∆ mutant, suggesting it functions after relief of Nrg1 repression (Fig. 5E).These data support the conclusion that the CWI pathway is activated during filamentation and contributes to the maintenance phase of hyphal morphogenesis in the presence of external hypha-inducing stimuli.The synthetic genetic interactions of the HOG and CWI pathway mutants with Nrg1 also suggest that Nrg1 has roles in filamentation beyond repression of the hyphal program.
Lastly, we tested the role of Mkc1 during filamentation in vivo using an in vivo imaging assay of filamentation in the sub-epithelial stroma of mouse ears (46).The ear tissue is injected with a 1:1 mixture of mNEON labeled reference strain and a mutant strain labeled with iRFP; the ears are then imaged 24 hours post-infection and both % filamentous cells and length of the filaments are determined as described in Materials and Methods.In the mkc1∆∆ mutant, the extent and length of filament formation were not statistically different (Fig. S1D and E).Based on time course experiments, the 4-hour in vitro and 24-hour in vivo time points are points where the ratio of yeast to filaments and the lengths have reached a steady state (47).In contrast to the mkc1∆∆, the deletion of CBK1, a PK required for filamentation under all in vitro conditions tested, formed essentially no filaments in vivo (Fig. 5F; Fig. S2F).

C. albicans biofilm formation requires a diverse set of protein kinases
C. albicans forms biofilms on inanimate surfaces such as intravascular catheters, medical devices and prosthetics, and dental appliances to cause a variety of human diseases (48).Biofilms also form on animate surfaces in the host, contributing to oropharyngeal candidiasis for example (49).To identify the network of PKs required for in vitro biofilm formation, we used a moderate throughput imaging assay in two commonly used media: Spider medium and RPMI + 10% BCS.Biofilm formation was assayed after 24-hour incubation at 37°C (50).Mutants were scored according to the thickness of the biofilm visualized by cross-section images of the well bottom based on two replicates per mutant and relative to SN95 strains imaged at the same time.We classified the mutants as having minor (−1), moderate (−2) or severe defects (−3).Figure 6A shows the reference strain along with representative examples of PK mutants with moderate or severe defects in RPMI + 10% BCS.Images for all strains tested in RPMI + 10% BCS are provided in Fig. S2A and B while scoring of strains in the two conditions is summarized in Table S5.
In RPMI + 10% BCS, 30 PK mutants had reduced biofilm formation with five having severe effects (IRE1, KIC1, SAK1, SNF1, and BUD32) while 22 PK mutants had reduced biofilm formation in Spider medium with two having severe defects (KIC1 and BUD32).Seventeen PK mutants had defects in both media (Fig. 6B).Of this set of mutants, four (sak1∆∆, snf1∆∆, bud32∆∆, and mec1∆∆) have growth defects under a variety of conditions (Fig. 3A; Table S3) which likely contribute to their poor biofilm formation.We did not identify PK mutants with increased biofilm density.Because filamentation is part and parcel of C. albicans biofilm formation (49), we compared the set of PKs required for filamentation on RPMI + 10% BCS with those required for biofilm formation in the same medium (Fig. 6C).The majority of PK mutants with reduced biofilm formation under these conditions also had reduced filamentation (22/27).However, less than 50% of the PK mutants with reduced filamentation in RPMI + 10% BCS also showed reduced biofilm formation.Interestingly, although only a handful of PK mutants have been tested in a mouse model of oropharyngeal candidiasis, an example of in vivo biofilm formation, the frk1∆∆ mutant has been shown by Naseem et al. to have a virulence defect in this setting, suggesting it has a general role in biofilm formation in multiple conditions (50).
We selected 12 mutants (Fig. 7A through C; Fig. S3) for testing using a third medium, RPMI at 37°C, and a quantitative microtiter plate-based biofilm density assay; the set included mutants from all three classes of phenotype severity.We characterized density after initial adhesion and at 24 hours and 48 hours.Regardless of the severity of initial screening phenotype, all of the tested mutants showed reduced biofilm densities at both 24 hours and 48 hours with none showing discordant phenotypes between the two time points.Seven of the 12 mutants had reduced initial adhesion with the RAM pathway kinase mutant kic1∆∆ having the most severe adhesion phenotype (Fig. 8A); mutation of the other RAM pathway kinase CBK1 also led to reduced adhesion (Fig. 7B).Deletion of BUB1, one of the PKs with no filamentation defect (RPMI or RPMI + 10% BCS, 37°C), increased initial adhesion relative to SN95.However, the density of the biofilm was reduced at both 24 hours and 48 hours (Fig. 7C).The thickness of the biofilm is reduced in the bub1∆∆ mutant relative to SN95 (Fig. 7D) but hyphae extending from the basal layer of the biofilm are clearly visible.Thus, Bub1 is required for biofilm formation without dramatically affecting filamentation.In addition, we did not identify any significant growth defects for the bub1∆∆ mutant in any of the conditions we examined (Table S3).These data indicate that a large set of PKs contribute to C. albicans biofilm formation and, although many likely do so through their effect on filamentation, PKs must also regulate other processes that are required for this complex process to occur.

Sok1 is dispensable for the establishment of the basal, filamentous layer of C. albicans biofilms in vitro but required for vertical extension of the biofilm
One of the PKs that has a very well-defined role in in vitro filamentation is Sok1 (44).At the initiation of filamentation, Sok1 phosphorylates the conserved transcriptional repressor Nrg1, leading to its degradation.This combined with reduced expression of NRG1 is required for initiation of filamentation.Nrg1 represses the hypha transcriptional program and constitutive expression of NRG1 locks the cells in yeast form.Consistent with previous literature, the sok1∆∆ mutant displayed a severe phenotype in all three conditions of our filamentation screen, and deletion of NRG1 in the sok1∆∆ mutant restored filamentation to near wild-type levels, supporting the model that Sok1 is required for relief of Nrg1-mediated suppression of hyphae formation (44) (Fig. 5C; Fig. S4A).
As expected from its filamentation phenotype, the sok1∆∆ mutant showed reduced biofilm formation in RPMI + 10% BCS (Table S5).Imaging of the biofilm formed by the sok1∆∆ mutant shows a dense, narrow layer of cells with little of the vertical extension of hyphae seen in the WT biofilm (Fig. 8A).Surprisingly, the apical view of the biofilm formed by the sok1∆∆ mutant shows extensive hypha formation within this basal layer (Fig. 8A).These data suggest Sok1 is dispensable for initial hypha formation at the basal level of the biofilm.Previous studies by Upplapari et al. have shown that constitu tive expression of NRG1 prevents hyphae formation during biofilm-forming conditions, indicating the necessity for relief of Nrg1 repression for filamentation under these conditions (51).During filamentation, Sok1 is thought to contribute to the relief of Nrg1 repression (44).
These observations suggest that the specific environment encountered by C. albicans at the basal level of an in vitro biofilm triggers a Sok1-independent pathway leading to relief of Nrg1 hyphal repression.Our data and that reported by others indicate that the sok1∆∆ mutant fails to filament on semi-solid agar and in liquid conditions (Fig. 5C; Fig. S4A) (44).However, the in vitro biofilm is formed on a solid plastic substrate rather than on semi-solid agar or in liquid culture.To test whether adherence to plastic bypasses the requirement for Sok1 during filamentation, we incubated WT and sok1∆∆ strains in RPMIS medium at 37°C in 2 mL dishes containing a microscope coverslip.The slides were removed after 30 hours and the morphology of adhered cells was examined.The sok1∆∆ mutant cells adhered to the plastic of the coverslip but did not filament while WT cells formed robust hyphae and biofilm-like structures (Fig. S4B).Therefore, adherence to plastic does not appear to bypass the need for Sok1 during filamentation in the biofilm.
The environment of the C. albicans biofilm shows features of hypoxia ( 52) and we, therefore, hypothesized that SOK1 may be dispensable for filamentation under hypoxic conditions.To test this hypothesis, we spotted the sok1∆∆ mutant on RPMI + 10% BCS agar plates and incubated them under hypha-inducing conditions (37°C) in ambient air or a hypoxic environment.As expected, the sok1∆∆ mutant did not show peripheral invasion in ambient air, and its invasion defect was suppressed in the nrg1∆∆ sok1∆∆ double mutant (Fig. 8B).Under hypoxic conditions, however, the sok1∆∆ mutant showed peripheral invasion that was similar to both WT and nrg1∆∆ mutants, indicating that SOK1 is not required for filamentation under hypoxic conditions (Fig. 8B).Kowalski et al. have shown that oxygen gradients are established as fungal biofilms mature and that the oxygen content of the medium decreases with increasing distance from the air-liquid interface (53).We hypothesized that the low oxygen levels at the base of the biofilm may render SOK1 dispensable for this early stage of biofilm formation.
To further characterize the effect of Sok1 on biofilm development, we compared the cross-sectional thickness of the WT and sok1∆∆ mutant biofilms at three time points (2 hours, 6 hours, and 24 hours; Fig. 8C).There was no difference in cross-sectional biofilm thickness between the WT strain and sok1∆∆ mutant at 2 hours or 6 hours but a large difference was present at 24 hours (Fig. 8D).Apical views of the biofilm at 2 hours also clearly show that the sok1∆∆ mutant forms hyphae at each time point (Fig. 8E).Consequently, it appears that the hyphae from the sok1∆∆ mutant are limited to regions relatively close to the base of the biofilm.Thus, Sok1 is not required for filamentation during the initial establishment of the basal layer of the biofilm but is required for the development of the vertical extensions beyond the basal layer.However, it seems unlikely that hypoxia is the primary reason for the ability of the sok1∆∆ mutant to filament at the basal layer of the biofilm because it seems unlikely that oxygen levels would be reduced fast enough to account for robust filamentation at 2 hours.Although in the A. fumigatus system, Kowalski et al. did not observe basal hypoxia until 12 hours of incubation (53).
Additional experiments will be required to definitely establish the molecular basis of this observation.However, our data strongly support a model in which two mecha nisms mediate relief of Nrg1 repression during biofilm formation: a Sok1-independent mechanism operates at the initial basal stage of biofilm formation, and a Sok1-depend ent mechanism that drives filament extension in the higher oxygen concentrations above the basal layer.We also have found that the sok1∆∆ mutant can filament under hypoxic conditions.However, it appears that these two features of the effect of Sok1 on filamentation are not directly related.

The RAM pathway governs biofilm formation in an Ace2-independent manner that is partially bypassed by overexpression of ALS3 and HGC1 in vitro and in vivo
The RAM pathway PKs (Cbk1 and Kic1) are established regulators of C. albicans biofilm formation and, accordingly, were among the most biofilm-deficient PK mutants in our screen (39).Gutierrez-Escribano et al. showed that Cbk1 phosphorylation of Bcr1 is required for biofilm formation in vitro and that this due, in part, to reduced expression of key biofilm-associated cell surface proteins such as Als3 (54).The transcription factor Ace2 is another bona fide Cbk1 substrate whose deletion mutant also has reduced biofilm formation (39).To determine whether Cbk1 phosphorylation of Ace2 was also involved in biofilm formation, we examined the density of biofilms formed by strains containing alleles of Ace2 lacking two or three of the Cbk1 consensus phosphorylation sites (55).Previous studies of these ace2 phospho-site mutants showed that they had reduced cell separation and increased lateral yeast formation during filamentation (56).As shown in Fig. S4C, the adhesion, 24-hour, and 48-hour biofilms formed by these strains were not statistically different than WT, indicating that the contribution of Ace2 to biofilm formation is independent of Cbk1.This is consistent with our findings that Cbk1-dependent Ace2 functions appear to be mainly related to cell separation (55).
Heterologous expression of ALS3 partially restores biofilm competency to the bcr1∆∆ mutant (56).Therefore, if the biofilm defects displayed by the cbk1∆∆ and kic1∆∆ mutants are primarily due to reduced Bcr1 activation, then the expression of ALS3 from a heterologous promoter in the cbk1∆∆ and kic1∆∆ mutants should also restore biofilm formation to those mutants.An important distinction between bcr1∆∆ and cbk1∆∆/kic1∆∆ mutants, however, is that the former can filament in SC5314-derived strains (57) while the latter are unable to filament under conditions examined to date (39).As discussed above, filamentation is a central step in the formation of mature C. albicans biofilms.Therefore, the important roles that Cbk1 and Kic1 play in filamentation may be the feature of their function most important for biofilm formation.We, therefore, attempted to drive filamentation in the kic1∆∆ and cbk1∆∆ mutants by overexpression of the cyclin HGC1.Overexpression of HGC1 is sufficient to drive filamentation in non-induc ing conditions and can bypass the filamentation defects of mutants such as the efg1∆∆ mutant (58).
We used the RBT5 promoter (P RBT5 ) to drive the expression of ALS3 and HGC1 in the cbk1∆∆ and kic1∆∆ mutants using methods described by Mao et al. (59).Expression of the RBT5 gene is strongly induced during biofilm formation and hypha formation in RPMI medium (59).Heterologous expression of ALS3 and HGC1 increased in vitro biofilm volume in both the cbk1∆∆ and kic1∆∆ mutants (Fig. 9A) although not to WT levels (Fig. 9B and C).Similar results were obtained for the ALS3and HGC1-overexpressing cbk1∆∆ mutants using the optical density assay (Fig. S4D).Importantly, overexpression of ALS3 and HGC1 does not improve biofilm formation in the ire1∆∆ mutant (Fig. S4E), indicating that there is specificity to the heterologous ALS3/HGC1 overexpression phenotypes.
Next, we imaged the biofilms at higher resolution to determine whether the overexpressing strains showed changes in cellular morphology.Indeed, HGC1 OE strains of both the cbk1∆∆ and kic1∆∆ mutants were elongated relative to the parental deletion mutants (Fig. 9D).By contrast, the ALS3 OE derivatives showed the round morphology of the parental strains.These data suggest that increased biofilm formation in the HGC1 OE strains may be due to increased filament-like cells while for ALS3 OE leads to morphology-independent partial rescue of biofilm formation in the RAM pathway.The ability of heterologous expression of ALS3 and HGC1 to rescue basal biofilm formation in the RAM PK mutants is not limited to in vitro conditions.Heterologous expression of both ALS3 and HGC1 in the cbk1∆∆ and kic1∆∆ mutants increased fungal burden in rat jugular catheters infected (60) with those strains relative to the parental strains (Fig. 9E and F).

The HOG pathway is required for biofilm formation in vivo
The Mitchell lab previously found that glycerol homeostasis plays an important role in biofilm formation (61).As implied by its name, a central function of the HOG pathway is to regulate cellular levels of glycerol (36).In our primary screen, we found that deletion of Pbs2, the MAPKK of the HOG pathway, but not of the terminal MAPK Hog1, led to a minor defect in RPMI + 10% BCS biofilm formation but not in Spider medium (Table S5).We, therefore, decided to test both the pbs2∆∆ and hog1∆∆ mutants using our secondary assay to further explore the apparent condition-dependent role of the HOG pathway during biofilm formation.In RPMI medium without BCS, the pbs2∆∆ and hog1∆∆ mutants both showed reduced biofilm density (Fig. 10A and B) at 24 hours and 48 hours.The hog1∆∆ mutant also had significantly reduced initial adhesion but the pbs2∆∆ mutant did not.Desai et al. (61) found that the addition of glycerol to the medium suppresses the biofilm defect in other mutants that contribute to intracellular glycerol synthesis (e.g., the rhr2∆∆ mutant).The addition of glycerol to RPMI restored the density of the biofilm formed by the pbs2∆∆ mutant to wild-type levels (Fig. S4E).These data indicate that, under some in vitro conditions, the HOG pathway contributes to biofilm formation to a modest extent due, in part, to its role in glycerol production.
Because the HOG pathway showed condition-dependent effects on biofilm formation, we wondered whether the pathway was important in a model of clinically relevant biofilms.To test this possibility, we again used the rat vascular catheter model (5).We tested the pbs2∆∆ mutant instead of the hog1∆∆ mutant in this experiment because the pbs2∆∆ mutant does not have a defect in initial adhesion.As shown in Fig. 10C, the fungal density of the catheter biofilm formed by the pbs2∆∆ mutant was reduced by >2log 10 CFU/catheter, a much stronger phenotype than expected based on the in vitro studies.We also compared the WT and pbs2∆∆ mutant biofilms using scanning electron microscopy (SEM).As expected, the WT biofilm was composed of an extensive network of hyphal structures with yeast phase cells and extracellular matrix scattered among it (Fig. 10D).By contrast, the pbs2∆∆ mutant biofilm is predominantly yeast-phase cells with extracellular matrix but few hyphal cells visible.The lack of filamentation of the pbs2∆∆ mutant in the rat catheter provides further evidence that the HOG pathway plays a condition-dependent role in C. albicans morphogenesis and biofilm formation.

DISCUSSION
PKs are critically important regulators of a wide variety of biological processes that have been extensively studied in a variety of organisms and are currently one of the most important drug targets in medicine (62).Historically, the model yeast, Saccharomyces cerevisiae, played a seminal role in the development of our understanding of PK biology (63).To further the study of PKs in the common human fungal pathogen C. albicans, we have constructed a set of 99 PK and PK-related gene homozygous deletion mutants.Prior to this effort, other large-scale collections of PK strains have been reported in C. albicans.The initial collection was a transposon-derived library in the BWP17 background which has been specifically interrogated for mutants affecting cell wall processes as well as biofilm formation and filamentation (18).A PK deletion mutant library has also been reported for the reference strain SC5314 and has been the basis for a study of iron homeostasis as well as the kinase Ypk1 (19,64).Finally, the characterization of a sub-library of PK mutants from the GRACE collection, a set of SC5314-derived strains in which ORFs are expressed from a doxycycline-regulated promoter (tetO) has been described recently and was used to study both filamentation and antifungal drug susceptibility (11,65).Our overarching goal for creating this deletion set was to provide a genetic resource for the community to complement the transcription factor deletion set and the PK deletion sets that are available in other important human fungal pathogens.
The PK deletion mutant library described here has four features.First, it is in the SN background and, therefore, is complementary to the widely used TF deletion set ( 5) and a large-scale collection of other C. albicans deletion mutants (6).Second, it is amenable to further genetic manipulation through both auxotrophic and dominant drug selection markers.Third, the mutations are barcoded and are suitable for competitive fitness experiments using pools of mutants (6,13,14).Fourth, we provide benchmark phenotyping for the mutants over a range of commonly employed growth conditions for future reference in selecting mutants for study.As discussed previously, there are a number of essential PKs and the majority of the PKs for which we only obtained heterozygous mutants have indications that they are likely to be either essential or have significant fitness defects (Fig. 1C).
Our systematic characterization of this collection of PK mutants supports a general conclusion regarding the function of PKs in C. albicans biology and pathobiology.Specifically, the function of non-essential PKs and the phenotypic consequences of their corresponding deletion mutants are highly dependent on the specific environmental conditions under which the mutants are examined.For example, only two mutants, mec1∆∆ and bud32∆∆, showed growth phenotypes across all 10 conditions examined; correspondingly, Mec1 and Bud32 are likely to be involved in the fundamental process of DNA repair and RNA metabolism, respectively.
The environmental contingency of PK function is, however, most apparent during filamentation and biofilm formation, two process that are critical contributors to human disease.While all but 11 PK deletion mutants had an effect on filamentation under one of the six conditions we examined, one-third of the filament-associated PKs (30/88) affected filamentation on all six conditions while 15 mutants altered filamentation under a single condition.An example of the latter observation is the tel1∆∆ mutant which does not filament on Spider medium at 37°C but is similar to WT on the other conditions (Table S5).A pathway-level example supporting this conclusion is the HOG pathway which is classified in most review literature as repressing filamentation.This is based on observations of HOG pathway mutant behavior in liquid medium growth under non-hypha-inducing conditions (36).Our results are consistent with these observations in that we saw small number of hyphae formed by the hog1∆∆ mutant in YPD at 30°C (Fig. 5B); however, the previous reports used slightly different conditions and did not quantify the numbers.In contrast to these results, all three of the HOG PK mutants showed significantly reduced filamentation on solid media induction conditions (Fig. 4A and B).Furthermore, the pbs2∆∆ mutant has a minor biofilm defect in vitro but a profound defect in an in vivo rat catheter biofilm model (Fig. 10B through D).The in vivo biofilm defect appears to be due to an inability of the pbs2∆∆ mutant to form filaments in the catheter.Our observations strongly support the conclusion that, under specific, infection-relevant conditions, the HOG pathway is involved in the positive regulation of filamentation while under other specific conditions it acts to suppress filamentation.
A second MAPK pathway for which we have found environmentally contingent functions during filamentation is the CWI.Previous work had demonstrated that the deletion mutant of the CWI pathway MAPK, MKC1, led to reduced filamentation and invasion on specific media when plated on solid agar (37).Furthermore, Kumamoto showed that contact with surfaces triggers the activation of the CWI pathway (38).In a liquid induction medium (YPD + 10% BCS), Xie et al. found that PKC1 was required for filamentation but that the CWI pathway MAP kinases BCK1, MKK2, and MKC1 were dispensable under these conditions.These observations were consistent with those of Kumamoto (38) who found that Mkc1 phosphorylation was reduced in the same liquid induction medium compared to conditions in which C. albicans was in contact with a solid or semi-solid substrate.We found that the CWI MAP kinase mutants had mediumdependent filamentation phenotypes of variable severity (Table S4) with the MAPKKK bck1∆∆ having the most severe phenotype on solid medium.In accordance with the results reported by Xie et al. (22), neither the mkc1∆∆ nor the bck1∆∆ mutants showed reduced filamentation in liquid RPMI + 1% BCS (Fig. 5C) but the filaments of the mkc1∆∆ mutant were significantly shorter than WT (Fig. 5D).It seems reasonable that cell wall remodeling during hyphal extension would require activation of the CWI pathway.This requirement for Mkc1 during hyphal elongation was not, however, observed in vivo (Fig. S1D and E).It is possible that the reduction in length observed in vitro may be due to a difference in the timing of the analysis and if earlier time points were used in vivo then a difference in the rate of filament lengthening might be observable.
The synthetic genetic interactions of CWI and HOG pathway PKs with NRG1 empha sized that repression of the hyphal transcriptional program cannot be the only function of Nrg1.Under most in vitro induction conditions, Nrg1 protein expression is restored as hyphal morphogenesis continues (44,45).Because Nrg1 regulates the expression of non-hyphal genes and other TFs, it seems likely that Nrg1 may play additional roles after initiation of hyphal morphogenesis (42).Our genetic interaction data with CWI pathway mutants suggest it may participate with CWI PKs in the maintenance of hyphal extension.These potential non-canonical roles for Nrg1 may be related to results from the Mitchell lab (66) as well as our own (67) describing Nrg1-related phenotypes and functions that are distinct from those expected based on its canonical function as an on-off switch for C. albicans filamentation.
Our analysis of the RAM pathway during biofilm formation also provides insights into the roles of filamentation during this process (39).Filamentation is not only a separate morphological state but is accompanied by a dramatic change in the expression of a variety of genes; indeed, many of these genes are typically expressed only in hyphae and play roles in the ability of C. albicans to form hyphae (41,42).Cbk1 phosphorylates Bcr1 to activate the expression of hypha-and biofilm-associated genes such as ALS3 (54).Overexpression of ALS3 in bcr1∆∆, cbk1∆∆, or kic1∆∆ mutants partially suppress their biofilm defects further supporting the important function of the Kic1/Cbk1-Bcr1-Als3 circuit during biofilm formation (Fig. 9A through D).This is likely due to increased intracellular or cell substrate adhesion mediated by the cell wall protein Als3 because there is no change in the morphology of the ALS3 OE derivatives of the cbk1∆∆ or kic1∆∆ mutants (Fig. 9D).
By contrast, overexpression of HGC1 in the cbk1∆∆ and kic1∆∆ mutants led to increased filament-like cells relative to the parental strains (Fig. 9D).Hgc1 is a cyclin that functions with Cdc28 to phosphorylate key proteins required for C. albicans polarized growth during filamentation (68).Cdc28/Hgc1 also phosphorylates the RAM network component Mob2 (69).Mob2, in turn, is required for Cbk1 function (39).To our knowledge, deletion of any RAM pathway mutant prevents initiation of filamentation under any condition reported in the literature.Therefore, the ability of HGC1 overexpres sion to increase filamentation in biofilms of the cbk1∆∆ and kic1∆∆ mutants cannot result from a linear Cdc28/Hgc1-Mob2-Cbk1 pathway.Instead, this result indicates that HGC1 overexpression bypasses this requirement for filamentation, presumably through a mechanism that functions downstream of the RAM pathway.These filamentous cells do not, however, extend above the basal layers as do WT cells, further emphasizing the conclusion that they are not completely normal hyphae.
We have also discovered that the development of the hyphal extensions from the basal layer of C. albicans biofilms involves an environmentally contingent function of Sok1.Previous genetic studies have linked Sok1 to the process of relieving Nrg1 repression and, consequent, initiation of filamentation (44).Although our results with the sok1∆∆ mutant under standard filamentation conditions were completely consistent with this model (Fig. 5C; Fig. S4A), we observed filamentation of the sok1∆∆ mutant during biofilm formation (Fig. 8).Further analysis indicated that the filamentation was restricted to the basal layer of the biofilm and that the profound filamentation defect of the sok1∆∆ mutant could be suppressed under hypoxic conditions (Fig. 8B).However, it is unlikely that hypoxia bypasses the requirement for Sok1 during establishment of the biofilm basal layer based on our time course experiments.Once again, we find that the regulation of C. albicans filamentation employs a variety of condition-dependent regulators.Further work will be required to understand the molecular mechanisms of this dramatic example of environmentally contingent regulation of C. albicans filamentation and biofilm formation.
As a resource, we hope that this collection will facilitate a more detailed understand ing for the role of PKs in the regulation of C. albicans biology.Considering the exten sive number of PKs that affect filamentation under at least one in vitro condition, the identification of specific targets or sets of targets that mediate the downstream functions of the PKs during filamentation will be required before such an understanding can be achieved.Integrating this genetic resource with existing, and future, phosphoproteomic analyses of the filamentation process as reported by Wilger et al. (70), Cao et al. (71), and Min et al. (72) are likely to lead to new insights into the specific targets required for filamentation.
As a brief illustration of the hypothesis-generating potential of such an approach, we consider the RAM PK Cbk1 (39), which is required for filamentation in all in vitro conditions as well as in our in vivo assay.The two most widely studied substrates of Cbk1, the TFs Ace2 and Bcr1, play condition-dependent roles in morphogenesis and, thus, cannot account for the profound and consistent filamentation defects displayed by cbk1∆∆ mutants (55,56,58).Interestingly, the deletion of NRG1 in the cbk1∆∆ background partially suppresses its lack of filamentation in a manner that appears to involve its canonical substrate Ssd1, an RNA-binding protein (73).However, neither the nrg1∆∆ cbk1∆∆ nor the ssd1∆∆ cbk1∆∆ double mutant forms true hyphae (73), indicating that Cbk1 has additional hyphae-associated substrates.The phosphoproteomic data set collected during C. albicans filamentation by Min et al. (72) showed that 58 proteins are phosphorylated at a Cbk1 consensus motif.GO term analysis revealed that this set of potential Cbk1 substrates is enriched for proteins involved in filamentous growth (16/58, FDR 0.01, Benjamini correction) and Swe1 and Bck1, two PKs required for filamentation.Thus, the combination of functional genetic data from the PK collection with phospho proteomic data derived from a specific C. albicans phenotype can rapidly generate hypotheses regarding the mechanisms of PK-mediated regulation.
The integration of phosphoproteomic data with functional data based on PK genetic experiments is also facilitated by multiple genetic resources for the study of C. albicans PKs now available.Our collection provides a set of PK mutants in a strain background amenable to further genetic manipulation and the ability to used barcode-facilitated pooled genetic screens.A similar deletion set has been generated by the Morschhauser lab in the widely used reference strain SC5314 (19).This set fits well with the extensive use of the SC5314 strain in virulence studies and avoids potential complications in the interaction of mutations with auxotrophic markers.Finally, strains containing tetracy cline-regulated PK alleles as the only source of the PKs are a subset of the larger GRACE collection and have been used to study PK function as well.The tetracycline-regulated alleles allow the study of essential PKs or low-fitness PK mutants and are also well validated in mouse models of infection.Two limitations of the tetracycline-regulated PK strains are that in the absence of doxycycline, the alleles tend to be over-expressed and complete suppression of expression may not be achieved in the presence of doxycycline.The latter limitation is most likely the reason for distinctions in some of the PK filamentation phenotypes found by Lee et al. and this work (11).
One of the themes that emerge from this work is that it is difficult to make general conclusions regarding the contribution of PKs to complex phenotypes such as filamentation and biofilm formation.These limitations apply to the non-screening experiments we present herein as well.For example, our conclusion that, in RPMI medium, phosphor ylation of Ace2 by Cbk1 is dispensable for biofilm formation may not apply to other conditions.Furthermore, it is possible that Ace2 is phosphorylated by Cbk1 but, that in RPMI medium, loss of Cbk1-dependent Ace2 function leads to a compensatory response that suppresses the phenotype.Similarly, the timing of a given PK function relative to relief of Nrg1 repression during filamentation is also likely to vary from one condition to another and our conclusions based on a single in vitro filament induction condition need to be viewed in light of such caveats.
In summary, the environmentally contingent nature of PK function suggests that the analysis of complex phenotypes such as filamentation and biofilm formation requires examination of multiple conditions and, as others have also emphasized, that care must be taken when generalizing the results to other niches and environments in the absence of direct experimentation.In this work, we examined only three media and two different temperatures; therefore, different results are to be expected if other conditions are employed.C. albicans adapts to a wide range of environmental niches and PKs mediate important sensing and signaling networks.Consequently, this complexity and variation in function should not be surprising.Indeed, the extensive network of PKs that affect C. albicans filamentation is likely due to the fact that filamentation and the related pathobiological process, biofilm formation, are an integral part of its physiology and, thereby, dependent upon the function of a wide variety of other physiologic processes and systems.

General culture methods
All media was prepared using previously reported recipes (5).Unless otherwise indicated, C. albicans strains were pre-cultured prior to experiments by inoculation of a colony from an agar plate into yeast peptone dextrose (YPD) and incubation overnight at 30°C.Library PK mutants that had been genotypically confirmed were stored in YPD-glycerol stocks at −80°C and recovered by streaking on a YPD agar plate followed by incubation at 30°C for 2 days.

Protein kinase gene deletion library construction
All known/predicted protein kinase genes (Tables S1 and S2) were deleted using modified CRIME (CRISPR-Cas9-induced marker excision) approach enabling marker recycling through a direct repeat mediated crossover (21,74).The 3′ portion of the split cassette was amplified from the pMH03 plasmid using a universal LEU2 forward primer (OL48, Table S1) and a reverse primer with a pMH03 specific sequence followed by an 80nt portion of the 3′ UTR adjacent to the stop codon.5′ portion of the deletion cassette was modified to include a barcode at the 5′ end.To this end, a set of modified 5′ split cassette plasmids (pMH04) was created bearing the 48 barcodes used previously by Homann et al. (5) preceded by a common adapter sequence.The 5′ portion of the split cassette was then amplified from a pMH04-derived barcoding plasmid using a primer with 80nt target gene start codon proximal sequence followed by the adapter sequence and a universal LEU2 marker reverse primer (OL49, Table S2).For each protein kinase gene, two independent sgRNAs were designed using the ChopChop tool.The sgRNA and Cas9 expression cassettes were prepared as described in reference 68 and co-transformed into C. albicans SN152 strain together with the split deletion cassette (21).Transformants were selected on CSM-Leu media.Correct integration of the deletion cassette was verified by PCR and the absence of amplicon from the targeted ORF.

Construction of fluorescently labeled strains
Fluorescently labeled SN95 was generated with the pENO1-NEON-NAT1 and the mkc1∆∆ mutant was labeled with pENO1-iRFP-NAT1 plasmid as previously described ( 47) and transformants were selected on YPD containing 200 µg/mL nourseothricin.Fluorescent isolates were confirmed by flow cytometry.The resulting His − strains were converted to His + by integrating plasmid pGEM-HIS1 as previously described (75).The fluorescently NEON-labeled SN95 showed no difference in fitness relative to the SN95 parent in YPD at 30°C.

Construction of NRG1 deletion mutants
Strains lacking NRG1 in the SN95 and the PK mutant strains indicated in Fig. 5 were generated by CRISPR-Cas9-mediated disruption using gRNA and repair constructs reported previously (66).
The sgRNA cassettes for the 5′ regions of each gene were generated using splitjoint PCR using primers HGC1p sgRNA/F and HGC1p SNR52/R, ALS3p sgRNA/F and ALS3p SNR52/R.The constructs and guide RNAs were then transformed into cbk1Δ/Δ, kic1Δ/Δ and ire1∆∆ mutant strains.Transformants were selected on YPD + nourseothricin medium and correct integration was confirmed by PCR.

Flow cytometry competitive fitness assay
Competitive fitness assays were performed in 96-well plates containing 200 µL of medium per well.Wells were inoculated with individual strains to be tested and the reference strain SN95-mNeonGreen and grown statically overnight at 30°C.Dilutions were performed to standardize input of the unlabeled protein kinase mutant library strains and SN95-mNeonGreen in a 1:1 ratio, at 2 × 10 4 cells/mL final concentration in varying growth conditions, as indicated.Co-cultures were grown at 30°C or 37°C for 24 hours in ambient air before analyzing on an Attune NxT Flow Cytometer with CytKick autosampler.The gating strategy was optimized for single cells.mNeonGreen positive and negative populations were identified by histogram plot with 100,000 cells counted per sample.The percent of mNeonGreen negative cells in experimental growth medium conditions was divided by the percent of mNeonGreen negative cells at 30°C in YPD to determine competitive fitness.Three biological replicates of each condition were performed and the difference between mutant and SN95 was analyzed for statistical significance by two-sided Student's t-test (P < 0.05) in Microsoft Excel.Experimental growth media were as follows: YPD, YP + 2% galactose, YP + 2% glycerol, YP + 2% acetate, YPD + 150 mM HEPES (pH 4), YPD + 150 mM HEPES (pH 9), YPD + 1.5 M NaCl, YPD + 1.5 M sorbitol, YNB + arg + 1% glucose + AmSO 4 , YPD + 0.000407% SDS.All were screened at 30°C, with the exception of YPD at both 30°C and 37°C.Average competitive fitness scores were used to generate a heat map in Morpheus (https://software.broadinstitute.org/morpheus),hierarchically clustered by growth condition and strain with one minus Pearson correlation, average linkage.

Spot dilution assays
Cells from overnight cultures were washed twice with PBS prior to quantifying OD 600 .Strains were diluted to an OD 600 of 1, followed by 10-fold serial dilutions.Three microliters from each dilution was spotted on agar plates and grown inverted at 30°C or 37°C as indicated for 24 hours on YPD, YP + 2% galactose, YP + 2% glycerol, YP + 2% acetate, YPD + 1.5 M sorbitol, or YPD + 1.5 M NaCl.Images were acquired at 24 hours and are representative of triplicates.

Agar plate-based screening assay of filamentation phenotypes
Overnight cultures of PK mutants were diluted to 0.1 OD 600 and 5 µL was spotted on YPD, Spider medium, RPMI, or RPMI + 10% bovine calf serum plates; SN95 control strains were also spotted on each plate.The plates were incubated at 30°C or 37°C.Peripheral invasion was scored at day 5 using the scale shown in Fig. S1A for mutants with reduced filamentation.For mutants with increased filamentation, they were scored at the time point when the mutant showed increased filamentation relative to SN95.The screen was performed in biological duplicate.For experiments performed under hypoxia, the plates were placed in a GasPak and incubated as described above.

In vitro liquid culture-based filamentation assay
For in vitro hyphal induction in liquid media, C. albicans strains were incubated overnight at 30°C in YPD media, harvested, and diluted into RPMI supplemented with either 1% or 10% BCS at a 1:50 ratio and incubated at 37°C for 4 hours.Induced cells were fixed with 1% (vol/vol) formaldehyde.Fixed cells were then imaged using the Echo Rebel upright microscope with a 60× objective; pseudohyphae were defined as filamentous cells with non-parallel cell walls and hyphae were defined as filamentous cells with parallel cell walls.The assays were conducted in biological triplicates on different days.Differences in the distribution of morphotypes were analyzed using data pooled from independent replicates and Chi-squared test (P < 0.05).The length of the hyphae was determined as previously described.

Imaging assay of in vivo filamentation
The imaging of dBA/2 mice ears infected with fluorescently labeled SN95 (NEON) and mkc1∆∆ mutant (iRFP) was carried out as described previously (46).Each ear (n = 3) was inoculated with a 1:1 mixture of SN95:mkc1∆∆ mutant.The ears were imaged 24 hours post-inoculation.Multiple Z stacks (minimum 15) were used to score the yeast vs filamentous ratio.The cells were considered as a "yeast" if the cells were round and/or budded cells.Furthermore, yeast cells were required not to project through multiple Z stacks.The cells were considered "filamentous" if the cells contained an intact mother and filamentous which was at least twice the length of the mother body.A minimum of 100 cells from multiple fields were scored.Paired Student's t-test with Welch's correction (P > 0.05) was used to define the statistical significance which was carried out using GraphPad Prism software.The filament length of the in vivo samples was measured as described previously (46).Briefly, a Z stack image of the reference or mutant strain was opened in ImageJ software and the distance between the mother neck to the tip of the filament was measured.At least 50 cells per strain from multiple fields were measured.Statistical significance was determined by the Mann-Whitney U test (P > 0.05).

Imaging-based screening of PK mutant biofilm formation
As previously described (51), strains were inoculated into 96-well plates at an OD 600 of 0.05 in 100 µL Spider medium or RPMI + 10% BCS.The wells were incubated at 37°C for 90 min with shaking (60 rpm) to adhere cells.The wells were then carefully washed twice with PBS.Fresh medium (100 µL) was added to each well and returned to the incubator for 24 hours with shaking.The medium was carefully aspirated and the biofilms were fixed with 4% formaldehyde in PBS (100 µL) for 1 hour; washed with PBS and then treated with Calcofluor White (200 µg/mL) overnight with shaking (60 rpm).The wells were then sequentially treated with 50% and then 100% thiodiethanol solution for 1 hour.Images were collected using a Keyence BZ-X800E microscope.Processing of images was performed as described in Shama et al. (51).Each mutant was screened in biological duplicate and compared to SN95 biofilms formed contemporaneously.

Optical density-based assay of biofilm formation
Biofilm growth was assayed as described in Glazier et al. (76).Briefly, strains were grown overnight at 30°C in liquid YPD with shaking, washed twice in phosphate-buffered saline (PBS), diluted to an OD 600 of 0.5 in desired media.200 µL of the suspension was dispensed into six wells of a 96-well plate (Corning Incorporated 96-well plate, catalog no.3596).The cells were then incubated in a 37°C incubator for 90 minutes to allow adherence.Next, the media was removed, and the wells were washed once with PBS to remove non-adhered cells.Adherence was measured OD 600 using a SpectraMax plate reader.The biofilm-inducing media was replaced and cells were further incubated at 37°C, without shaking.After 24 hours or 48 hours, the media were aspirated; the cells were washed with PBS as above and the OD 600 was measured.Three biological replicates with six technical replicates per strain were analyzed for each strain and the experiment was repeated independently 2-3 times.Differences between strains were analyzed by ANOVA and multiple comparison correction with statistical significance set at an adjusted P value < 0.05.

Rat model of Candida albicans vascular catheter infection
In vivo C. albicans biofilms were studied using an external jugular-vein, rat-catheter infection model as previously described (59).Briefly, a 1 × 10 6 cells/mL inoculum for each strain or strain combination was allowed to grow on an internal jugular catheter placed in a pathogen-free female rat (16-week-old, 400 g) for 24 hours.The catheter volumes were then removed and the catheters were flushed with 0.9% NaCl.The biofilms were dislodged by sonication and vortexing.Viable cell counts were determined by dilution plating.Three replicates were performed for each strain.Differences in fungal burden were analyzed by ANOVA and correction for multiple comparisons with a significant difference defined as a P < 0.05.After a 24-hour biofilm formation phase, the devices were removed, sectioned to expose the intraluminal surface, and processed for SEM imaging.Briefly, 1 mL fixative (4% formaldehyde and 1% glutaraldehyde in PBS) was added to each catheter tube and tubes were fixed at 4°C overnight.Catheters were then washed with PBS prior to incubation in 1% OsO4 for 30 min.Samples were then serially dehydrated in ethanol (30%-100%).Critical point drying was used to completely dehydrate the samples prior to palladium-gold coating.Samples were imaged on an SEM LEO 1530, with Adobe Photoshop 2022 (v.23.2.2) used for image compilation.

Phospho-Mkc1 western blot
Western blotting was performed as described previously (77,78).Briefly, harvested cell pellets were resuspended in 300-500 µL of lysis buffer (50 mM Tris-HCl [pH 7.5], 150 mM NaCl, 1 mM EDTA, 1% Triton X-100) supplemented with protease and phosphatase inhibitor cocktail (Thermo Scientific).The suspension was then transferred to a tube with glass beads and lysed by 10 30-s bursts using a bead beater.The resulting lysate was cleared by centrifugation and total protein was determined by Bio-Rad protein assay dye concentrate according to the manufacturer's instructions.50 µg of the lysate was separated on a 10% Mini-Protean TGX gel (Bio-Rad) and transferred to a nitrocellulose membrane.Membranes were blocked for 60 minutes at room temperature in 5% BSA in 50 mM Tris (pH 7.5), 150 mM NaCl, 0.01% Tween 20 (TBST).The membranes were then probed with anti p44/p42 MAPK Thr202/Tyr204 antibody (Cell Signaling) followed by anti-rabbit horseradish peroxidase (HRP)-conjugated secondary antibody (Bio-Rad).Blots were imaged on a myECL imager (Thermo Scientific) using a SuperSignal West Femto ECL reagent (Thermo Scientific).

DIRECT CONTRIBUTION
This article is a direct contribution from Damian J. Krysan, a Fellow of the American Academy of Microbiology, who arranged for and secured reviews by James Konopka, Stony Brook University, and Michael Lorenz, The University of Texas Health Science Center at Houston.

ETHICS APPROVAL
The animal experiments described in this work were approved by the Institutional Animal Care and Use Committees at the University of Iowa and the University of Wisconsin.

FIG 1
FIG 1 Construction of protein kinase homozygous deletion collection in C. albicans.(A) Genotypes of parental and homozygous protein kinase deletion strain.(B) Venn diagram showing protein kinases targeted (yellow) and homozygous protein kinase mutants constructed (blue).(C) The 20 protein kinases for which only heterozygous deletion mutants were obtained.Red font indicates a protein kinase with previously reported to be potentially essential.Black font indicates a protein kinase for which a homozygous deletion mutant has been reported in the literature.

FIG 2 FIG 3
FIG 2 Competitive fitness assay of protein kinase deletion mutants.(A) Schematic of flow cytometry-based competitive fitness assay.(B) Representative competitive fitness data for ckb1∆∆ compared to SN95.(C) Spot dilution assay showing phenotype of ckb1∆∆ mutant relative to ckb2∆∆ and SN95.(D) Representative spot dilution growth phenotypes for strains without competitive fitness defects by flow cytometry (black font) and those with reduced fitness (red font).

FIG 4
FIG 4 Summary of filamentation phenotypes of protein kinase deletion mutants and representative phenotypes.(A) Table of protein kinase mutants with reduced filamentation on all filament-inducing conditions tested.Black font indicates protein kinases previously implicated in filamentation and blue font indicates protein kinases not previously reported to play a role in filamentation.(B) Representative phenotypes of the reference strain SN95 and protein kinase deletion mutants on RPMI medium incubated at 37°C.

FIG 5
FIG 5 Genetic interactions of protein kinase mutants with nrg1∆∆ mutant reveal role for cell wall integrity pathway during hyphal extension.(A) Strategy to assess the timing of protein kinase function during filamentation using genetic interactions with nrg1∆∆ mutant.(B and C) The indicated strains were incubated in YPD at 30°C (B) or RMPI + 1% bovine calf serum (C) and the distribution of yeast, pseudohyphae, and hyphae as determined by microscopy on fixed samples.The bars indicate the means of two independent experiments with greater than 100 cells counted per experiment.Chi-squared tests were used to evaluate the statistical significance of comparisons (P < 0.05): * indicates comparison to protein kinase single deletion mutant; # indicates comparison to nrg1∆∆ mutant; ∆ indicates comparison to SN95.(D) The length of hyphae formed after 4-hour induction in RPMI + 1% bovine calf serum was determined as described in Materials and Methods.Statistical significance was determined by Mann-Whitney; * indicates P < 0.05.(E) Western blot analysis of Mkc1 phosphorylation in the indicated strains using anti-p42/44 antibody in the indicated strains after incubation in YPD at 30°C or RMPI + 10% BCS at 37°C for 4 hours.Ponceau staining was used to assess loading.** indicates a non-specific band observed in serum-induced samples.The blots are representative of three independent experiments.(F) The percentage of filamentous cells in WT or cbk1∆∆ mutant strains 24 hours after inoculation of mouse ears.

FIG 7 FIG 8
FIG 7 Confirmation of biofilm phenotypes in optical density assay.(A-C) The indicated strains were incubated in microtiter plates in RPMI medium at 37°C.For adhesion, the plates were washed with PBS at 90 minutes and the 600 determined.Mature biofilms were measured at 24 hours and 48 hours.Error bars indicate the mean of 4-5 replicates with error bars standard deviation.**** indicates statistical significance (P < 0.05) by two-way ANOVA with correction for multiple comparisons.(D) Cross-sectional image of the bub1∆∆ and SN95 biofilms indicating that the bub1∆∆ mutant forms hyphae.

FIG 8 (
FIG 8 (Continued) thickness over time course.Bars are means of three independent experiments with error bars indicating standard deviation.* indicates statistical significant (P < 0.05) difference between SN95 and sok1∆∆ mutant by two-tailed Student's t-test.(E) Apical views of SN95 and sok1∆∆ mutant biofilms at 2 hours indicate filament formation in both samples.

FIG 9
FIG 9 The role of Cbk1 in biofilm formation is independent of Ace2 and partially suppressed by overexpression of ALS3 and HGC1.(A) Apical and cross-section images of 24-hour biofilms formed by the indicated strains in RPMI + 10% BCS at 37°C.Please note that the images of the SN95 control strain are duplicated to allow ease of comparison to both the cbk1∆∆ mutants (top panel) and kic1∆∆ mutants (bottom panel).(B and C) show a comparison of the biofilm volume for cbk1∆∆ (B) and kic1∆∆ (C) mutants and corresponding P RBT5 -ALS3/HGC1 overexpression strains.**** indicates a statistically significant difference between the indicated groups by two-way ANOVA with multiple comparisons test.(D) Fluorescence microscopy of Calcofluor-stained cells from biofilms for the indicated strains.(E and F) show the fungal burden of jugular venous catheters placed in rats that were infected with the indicated strains.The catheters were removed at 24 hours; the fungal cells were processed and plated as described in Materials and Methods.Bars indicate the mean of three replicates with standard deviation noted by error bars.**** indicates a statistically significant difference (P < 0.05) between groups by two-way ANOVA with multiple comparisons test.The fungal burden for cbk1∆∆ and kic1∆∆ mutants and corresponding P RBT5 -ALS3/HGC1 overexpression strains are shown in E and F, respectively.

FIG 10
FIG 10 The HOG pathway plays a condition-dependent role in biofilm formation.The biofilm density for the hog1∆∆ (A) and pbs2∆∆ (B) mutants was determined by optical density in RPMI medium at 37°C at the indicated time points.Error bars indicate the mean of 4-5 replicates with error bars standard deviation.**** indicates statistical significance (P < 0.05) by two-way ANOVA with correction for multiple comparisons.(C) The fungal burden of jugular venous catheters placed in rats that were infected with either SN95 or the pbs2∆∆ mutant.The catheters were removed at 24 hours; the fungal cells were processed and plated as described in Materials and Methods.Bars indicate the mean of three replicates with standard deviation noted by error bars.**** indicates a statistically significant difference (P < 0.05) between groups by Student's t-test.(D) Electron micrographs of biofilms formed in vascular catheters by SN95 and the pbs2∆∆ mutant.