Hrk1 Plays Both Hog1-Dependent and -Independent Roles in Controlling Stress Response and Antifungal Drug Resistance in Cryptococcus neoformans

The HOG (High Osmolarity Glycerol response) pathway plays a central role in controlling stress response, ergosterol biosynthesis, virulence factor production, and differentiation of Cryptococcus neoformans, which causes fatal fungal meningoencephalitis. Recent transcriptome analysis of the HOG pathway discovered a Hog1-regulated gene (CNAG_00130.2), encoding a putative protein kinase orthologous to Rck1/2 in Saccharomyces cerevisiae and Srk1 in Schizosaccharomyces pombe. Its function is not known in C. neoformans. The present study functionally characterized the role of Hrk1 in C. neoformans. Northern blot analysis confirmed that HRK1 expression depends on the Hog1 MAPK. Similar to the hog1Δ mutant, the hrk1Δ mutant exhibited almost complete resistance to fludioxonil, which triggers glycerol biosynthesis via the HOG pathway. Supporting this, the hrk1Δ mutant showed reduced intracellular glycerol accumulation and swollen cell morphology in response to fludioxonil, further suggesting that Hrk1 works downstream of the HOG pathway. However, Hrk1 also appeared to have Hog1-independent functions. Mutation of HRK1 not only further increased osmosensitivity of the hog1Δ mutant, but also suppressed increased azole-resistance of the hog1Δ mutant in an Erg11-independent manner. Furthermore, unlike the hog1Δ mutant, Hrk1 was not involved in capsule biosynthesis. Hrk1 was slightly involved in melanin production but dispensable for virulence of C. neoformans. These findings suggest that Hrk1 plays both Hog1-dependent and –independent roles in stress and antifungal drug susceptibility and virulence factor production in C. neoformans. Particularly, the finding that inhibition of Hrk1 substantially increases azole drug susceptibility provides a novel strategy for combination antifungal therapy.


Introduction
The HOG (High Osmolarity Glycerol response) pathway plays central roles in several cellular functions in filamentous and nonfilamentous fungi, encompassing both pathogens and nonpathogens. This pathway is not only important in maintaining normal cellular homeostasis under a plethora of environmental stresses, but is also required for development and differentiation of fungi (see reviews, [1,2]). Particularly for pathogenic fungi, the HOG pathway is one of the major signaling cascades for controlling production and regulation of virulence factors [2]. The mammalian counterpart for the fungal HOG pathway is the p38 MAPK pathway, which also plays a pivotal role in stress response, immunity modulation, and cell death [3,4].
The basidiomycetous fungus, Cryptococcus neoformans, causes fatal lung and brain infections if left untreated. The HOG pathway in C. neoformans is one of the major signaling pathways to control stress response, differentiation, and virulence factor production and is structurally similar to those found in other fungi. External signals are sensed and transferred by the two-component-like phosphorelay system, which consists of hybrid sensor histidine kinases, named Tco (Two-component-like) proteins, the histidinecontaining phosphotransfer (HPt) protein, Ypd1, and two response regulators, Ssk1 and Skn7. Downstream of the phosphorelay system, Ssk1 further activates a mitogen-activated protein kinase (MAPK) module, which comprises a Hog1 MAPK, Pbs2 MAPK kinase (MAPKK), and Ssk2 MAPKK kinase (MAPKKK). This signaling cascade controls cellular responses against a variety of environmental cues, including osmotic shock, high temperature, oxidative stress, UV irradiation, heavy metal, antifungal drugs, and toxic metabolites [2,5,6,7,8,9]. The Skn7 response regulator is also involved in Na + stress response, oxidative stress response, and antifungal drug resistance, but mostly in a Hog1-independent manner [6].
Regardless of the common features of the Cryptococcus HOG pathway, the pathway has undergone unique specialization in terms of regulation and cellular functions. Most notably, the Hog1 MAPK is constitutively phosphorylated under unstressed conditions in a number of clinical and environmental C. neoformans strains, including strain H99 (serotype A genome sequencing platform strain) and B-3501A (serotype D genome sequencing platform strain), which is in stark contrast to most fungi, where Hog1 is unphosphorylated under normal conditions but phosphorylated in response to environmental stresses [5,8]. The constitutive phosphorylation of Hog1 appears to provide unique phenotypes to the HOG mutants in C. neoformans, including increased production of two important virulence factors of C. neoformans, antiphagocytic capsule and antioxidant melanin, and enhanced sexual differentiation [5,6,8]. The Ssk2 MAPKKK has been identified as a major upstream signaling component that is responsible for constitutive phosphorylation of the Hog1 MAPK through comparative analysis of meiotic maps between the serotype D f1 sibling strains B-3501 and B-3502, which show differential Hog1 phosphorylation patterns [8]. Two nonsynonymous coding sequence changes (L240F and M738V) in the Nterminal region of Ssk2 between strains B-3501 and B-3502 appear to define the different phosphorylation patterns of Hog1 [8].
To identify downstream target genes and to clarify the complex signaling network of the Cryptococcus HOG pathway, comparative transcriptome analysis has recently been performed, revealing a number of novel characteristics of the pathway [9]. First, the HOG pathway negatively controls expression levels of ergosterol biosynthesis genes, including ERG11, resulting in increased cellular ergosterol content when the HOG1, PBS2, SSK2, or SSK1 genes are mutated. Therefore, mutation of the HOG pathway components confers dramatic synergistic antifungal effects with amphotericin B, which binds to the ergosterol of fungal cell membranes and forms a transmembrane channel, causing a leakage of essential ions and eventually cell death. Increased cellular ergosterol content by inhibition of the HOG pathway is likely to provide more binding targets for the polyene drug and thereby enhance its antifungal efficacy. Furthermore, a number of cadmium stressresponsive genes are differentially regulated and indeed the HOG mutants show increased resistance to cadmium treatment. These features of the Cryptococcus HOG pathway have not been described previously in other fungi.
This prior transcriptome analysis discovered several hitherto uncharacterized protein kinases, whose expression appears to be differentially regulated by the Cryptococcus HOG pathway [9]. Among these we have identified a gene, (CNAG_00130.2), which has greatly decreased basal expression in both ssk1D and hog1D mutants. CNAG_00130.2 shows significant homology to Hog1dependent MAPK-activated protein kinases (MAPKAPKs) in other fungi. In S. pombe, two MAPKAPKs, Srk1 (also known as Mkp1) and Cmk2 (also known as Mkp2), have been reported to physically interact with and be activated via phosphorylation by the Sty1 MAPK (Hog1 homologue) [10,11]. Expression of SRK1 is induced by a range of environmental stimuli, including oxidative or osmotic stress, and heat shock, in a Sty1-dependent manner [11]. Srk1 is involved in conjugation, meiosis, and sporulation [11] whereas Cmk2 modulates the oxidative stress response [10]. Srk1 is also required for regulation of S. pombe cell cycle progression [12]. In S. cerevisiae, two MAPKAPKs, Rck1 and Rck2, have been discovered and between these Rck2 is activated by phosphorylation via the Hog1 MAPK [13,14,15]. Although the rck1D, rck2D, and rck1D rck2D mutants do not show any distinctive phenotypes, both Rck1 and Rck2 are involved in oxidative stress response and meiosis [16,17].
In this study we confirmed that expression of CNAG_00130.2 is controlled by the Hog1 MAPK in C. neoformans and thereby named it Hrk1 (Hog1-regulated kinase 1). To functionally characterize Hrk1, we constructed targeted deletion mutants for HRK1 in both wild-type and hog1D mutant strain backgrounds and performed an array of phenotypic analyses in comparison with the hog1D mutant. Our data suggest Hrk1 plays both Hog1-dependent and -independent roles in controlling stress and antifungal drug response and production of virulence factors, such as melanin, in C. neoformans. However, Hrk1 is dispensable for virulence of C. neoformans. Notably, mutation of HRK1 not only suppressed azoleresistance of the hog1D mutant, but also further increased azolesensitivity compared to the wild-type, strongly suggesting a novel drug target for treatment of cryptococcosis.

Results
Identification of Hog1-regulated kinase 1 (Hrk1) in C. neoformans The prior transcriptome analysis for elucidation of the downstream signaling network governed by the HOG pathway revealed that a C. neoformans gene (CNAG_00130.2), which encodes a putative member of the Ca 2+ /calmodulin-dependent protein kinase superfamily, is significantly downregulated even under unstressed condition in both ssk1D and hog1D mutants, but not in the skn7D mutant [9] (Fig. 1A). To further confirm whether this gene is truly regulated by the HOG pathway, we performed Northern blot analysis. In agreement with the microarray data, basal expression levels of CNAG_00130.2 were significantly lower in both hog1D and ssk1D mutants than the wild-type and skn7D mutant strains (Fig. 1B), strongly indicating that it is indeed differentially regulated by the HOG pathway in C. neoformans. Therefore the CNAG_00130.2 was named Hrk1 (Hog1-regulated kinase 1).
C. neoformans Hrk1 is highly homologous to S. pombe Srk1/Mkp1, S. cerevisiae Rck2, and mammalian MAPKAPK-2, all of which are known to be p38/Hog1 MAPK-dependent protein kinases [11,15,16,18] (Fig. 1C). Hrk1 is more homologous to Srk1 (52% identity in amino acid sequence) than Rck2 (45% identity) and MAPKAPK-2. Hrk1 shares several structural features with the yeast Hog1/Sty1-dependent MAPKAPKs. First, the glycine-rich domain (also known as the glycine-loop) and the highly conserved DFG box in the kinase catalytic domain were also observed in Hrk1 (Fig. 1D). Second, a putative MAPK-binding motif with a consensus RR sequence observed in other MAPKAPKs was also found in Hrk1 (Fig. 1D). The serine residue (Ser519) of Rck2, which is known to be activated by phosphorylation, was also conserved in Hrk1, although the Ser residue appears not to be conserved in other MAPKAPKs. Indeed the Thr411 in Cmk2 was found to be in the position equivalent to Ser519 residue in Rck2 and to be essential for phosphorylation by Sty1 for its activation [10].
HRK1 expression is differentially regulated in response to environmental cues via the Hog1 MAPK We examined whether HRK1 expression is modulated in response to environmental stress, such as osmotic/salt shock, oxidative stress, and fludioxonil treatment since the HOG pathway is one of major stress-activated signaling pathways in C. neoformans. The quantitative RT-PCR revealed that HRK1 expression was downregulated in the wild-type strain in response to 1 M NaCl ( Fig. 2A). In the hog1D mutant, expression levels of HRK1 were generally low under both unstressed and osmotically stressed conditions ( Fig. 2A). Interestingly, basal expression levels of HRK1 were also very low in the ssk1D mutant, but its expression increased at the later time point (60 min) during NaCl treatment ( Fig. 2A), indicating that HRK1 expression needs to be induced in response to osmotic shock in the absence of the Ssk1 response regulator, but requires the Hog1 MAPK.
In response to hydrogen peroxide (2.5 mM H 2 O 2 ), HRK1 expression was upregulated in the wild-type strain. Both basal and 2) from microarray data using total RNA isolated from the WT (H99) strain and ssk1D, skn7D, and hog1D mutants grown to the middle logarithmic phase at 30uC in YPD medium [9]. (B) Northern blot analysis for measuring basal expression levels of HRK1 using total RNA of (A). (C) The phylogenetic analysis of Hrk1 orthologs between C. neoformans and other eukaryotes. The following proteins were compared for the analysis: S. cerevisiae Rck1 (YGL158W, SGDID: S000003126), S. cerevisiae Rck2 (YLR248W, SGDID: S000004238), Homo sapiens MAPKAPK-2, N. crassa Hrk1 induced levels of HRK1 were also generally low in the hog1D mutant (Fig. 2B). Similar to the results seen during the osmotic stress response, however, HRK1 expression levels were strongly induced after 30 min of exposure to H 2 O 2 in the ssk1D mutant and became equivalent to those of the wild-type strain (Fig. 2B). These data indicate that Hrk1 expression can be induced by H 2 O 2 in an Ssk1-independent but Hog1-dependent manner. In response to fludioxonil, however, both basal and induced levels of HRK1 were significantly lower in both the hog1D and ssk1D mutants than in the wild-type strain (Fig. 2C). Northern blot analysis also displayed similar results (Fig. S1). Interestingly, HRK1 expression levels appear to be roughly proportional to the level of phosphorylated Hog1. In the H99 strain background, Hog1 is highly, constitutively phosphorylated under unstressed conditions [5], which may reflect high HRK1 basal expression levels. In response to NaCl or fludioxonil, Hog1 phosphorylation levels decrease [5,6], which is similar to NaCl or fludioxonil-induced reduction of HRK1 expression. Also the Ssk1-independent induction of HRK1 presumably reflects the Ssk1-independent phosphorylation of Hog1 in response to NaCl [6]. In conclusion, basal HRK1 expression appears to be mainly controlled by the Ssk1 response regulator and the Hog1 MAPK. In response to osmotic and oxidative stresses, however, HRK1 can be induced via Hog1 by an unknown signaling component, other than Ssk1.
Hrk1 regulates osmotic stress response and fludioxonil susceptibility by controlling intracellular glycerol accumulation via the Hog1 MAPK To explore the function of Hrk1, the HRK1 gene was disrupted by targeted deletion and comparative mutant analysis with other HOG mutants was performed. The hrk1D mutant allele was generated by overlap PCR and biolistically transformed into the wild-type C. neoformans serotype A H99 (MATa) and KN99a (MATa) strains as described in Materials and Methods. Three independent hrk1D mutants (YSB270, YSB271, and YSB272) were generated in the MATa background and confirmed by Southern blot analysis (Fig. S2). For verification of any phenotypes of the hrk1D mutant, a complemented strain (hrk1D+HRK1) was constructed by re-integrating the wild-type HRK1 into its native locus. Furthermore, to address whether Hog1 is the only upstream MAPK for regulating expression and activity of Hrk1 and which Hog1 processes are likely mediated through Hrk1, we also constructed the hog1D hrk1D double mutant (Fig. S3). If Hog1 is the only upstream regulator of Hrk1, the hrk1D hog1D mutant should exhibit phenotypes similar to the hog1D mutant.
Next, we monitored fludioxonil sensitivity of the hrk1D mutant. Fludioxonil hyperactivates the HOG pathway, triggering overaccumulation of intracellular glycerol content, and causes defective cytokinesis and cell swelling, which eventually results in cell cycle arrest [7]. Therefore, deletion of HOG1 confers complete resistance to fludioxonil. Similar to the hog1D mutant, the hrk1D mutant was highly resistant to fludioxonil, although the hog1D mutant was more resistant than the hrk1D mutant ( Fig. 4A). Reintegration of the HRK1 gene restored fludioxonil sensitivity of the hrk1D mutant. Under exposure to fludioxonil the wild-type strain showed marked cell swelling, indicative of intracellular glycerol over-accumulation ( Fig. 4B and 4C). The hrk1D mutant showed less cell swelling in response to fludioxonil than the wild-type strain, although more cell swelling than the hog1D mutant. To further support these findings, measurement of intracellular glycerol content upon fludioxonil treatment revealed that the hrk1D mutant had much lower levels of intracellular glycerol content than the wild-type strain, albeit to a greater extent than the hog1D mutant. Collectively, these data indicate that Hrk1 is the major downstream component of the Hog1 MAPK pathway to relay fludioxonil-responsive signaling, but other minor signaling component(s) may exist downstream of the Hog1 MAPK to confer full fludioxonil sensitivity through control of glycerol synthesis.
Because mutants with defects in cell wall integrity, such as cna1D, mpk1D, and ras1D, are hypersensitive to fludioxonil [7,19], we also examined the role of Hrk1 in maintaining cell wall integrity. In contrast to previously described mutants, the hrk1D mutant showed wild-type levels of sensitivity to SDS, Congo red, DTT, and high temperature (Fig. S5). Furthermore, hrk1 mutation did not exacerbate the increased sensitivity of the hog1D mutant to high temperature and SDS (Fig. S5). Taken together, these data indicate that Hrk1 is not required for maintenance of cell wall integrity.

Hrk1 is involved in oxidative stress response in a Hog1-independent manner
The finding that HRK1 expression is induced in response to H 2 O 2 led us to examine the role of Hrk1 in the oxidative stress response of C. neoformans. S. pombe Srk1 and S. cerevisiae Rck2 are also known to be involved in oxidative stress response in a Sty1 or Hog1-dependent manner [16,20]. In C. neoformans, the HOG pathway plays a critical role in defending against diverse oxidative damaging agents, such as H 2 O 2 , menadione, tert-butyl hydroperoxide (tBOOH), and diamide [5,9]. The hrk1D mutant was clearly distinct from the hog1D mutant in oxidative stress response (Fig. 5). The hrk1D mutant showed wild-type levels of sensitivity to H 2 O 2 , menadione, and tBOOH, unlike the hog1D mutant that is hypersensitive to these agents ( Fig. 5 and data not shown). However, the hrk1D hog1D mutant exhibited higher H 2 O 2 sensitivity than the hog1D mutant, indicating that Hrk1 plays a minor role in oxidative stress response.
Diamide, which is not only an exogenous oxidant oxidizing thiol groups of proteins and glutathione but also a radiosensitizer perturbing DNA repair system, is known to confer oxidative stress response in a manner distinct from H 2 O 2 [21]. As reported previously [19], the hog1D mutant showed increased resistance to diamide treatment (3 mM) compared to WT, which is in stark contrast to the results from H 2 O 2 treatment (Fig. 5). Conversely, the hrk1D mutant was more sensitive to diamide than WT, which is in opposition to the pattern found in the HOG mutants (Fig. 5). Complementation with the WT HRK1 gene restores normal diamide susceptibility to the hrk1D mutant, further confirming the involvement of Hrk1 in diamide resistance. Furthermore, mutation of HRK1 decreased diamide-resistance of the hog1D mutant (Fig. 5). Taken together, these data suggest that C. neoformans differentially senses oxidative stress conferred by H 2 O 2 and diamide, and that the HOG pathway and Hrk1 protein kinase appear to act on the oxidative stress response independently.

Hrk1 controls azole resistance in a Hog1-independent manner
The HOG pathway controls expression of ergosterol biosynthesis genes, including ERG11, and governs susceptibility to polyene and azole drugs. Accordingly, the HOG mutants exhibited increased sensitivity to the polyene class of antifungal agents, such as amphotericin B, but decreased sensitivity to the azole drugs, such as fluconazole and ketoconazole [9]. Increased ergosterol in the cell membrane by inhibition of the HOG pathway provides more binding sites for amphotericin B and enhances its antifungal activity whereas increased expression of ERG11 requires higher concentration of azole drugs for its inhibition and thereby decreases their antifungal efficacy. To address whether Hrk1 controls sensitivity to polyene and azole drugs via the HOG pathway, we measured drug sensitivity of the hrk1D mutant. Compared to the hog1D mutant, the hrk1D mutant exhibited only a minor hypersensitivity to amphotericin B (Fig. 6A). However, the hrk1D mutant exhibited greater sensitivity to fluconazole than the wild-type strain whereas the hog1D mutant showed resistance (Fig. 6A). Notably, deletion of HRK1 suppressed azole drug resistance of the hog1D mutant (Fig. 6A). The hrk1D hog1D mutant was even more sensitive to fluconazole and ketoconazole than the wild-type strain, implying that Hrk1 is involved in controlling azole sensitivity in a Hog1-independent manner.
To further verify these data, we examined expression levels of ERG11, the target gene of azole drugs, in the hrk1D mutant as ERG11 expression is known to be induced in the hog1D mutant [9]. ERG11 expression was indeed increased in the hog1D mutant compared to the wild-type strain as reported before, but not in the hrk1D mutant ( Fig. 6B and 6C). Furthermore, hrk1D mutation did not further affect ERG11 expression in the hog1D mutant. These data suggest that Hrk1 controls polyene and azole drug sensitivity in a Hog1-independent manner.
The role of Hrk1 in virulence factor production and differentiation of C. neoformans Another key feature of the Cryptococcus HOG pathway is its involvement in production of two major virulence factors, capsule and melanin, and sexual differentiation [5,6,8]. Perturbation of the HOG pathway increases capsule and melanin production in the C.
neoformans H99 strain background. Our previous transcriptome analysis revealed that expression of the polysaccharide capsule biosynthesis genes, including CAP10, CAP59, CAP60 and CAP64, is induced by perturbation of the HOG pathway [9]. Therefore, we have monitored capsule and melanin production in the hrk1D mutant. The hrk1D mutant displayed normal levels of capsule production comparable to the wild-type strain, whereas the hog1D mutant showed increased capsule production as previously described [5,6,8]. Furthermore the hrk1D mutation did not affect increased capsule production of the hog1D mutant (Fig. 7A). Therefore, Hrk1 does not appear to be involved in the Hog1regulated capsule biosynthesis.
We have also monitored melanin biosynthesis levels in the hrk1D mutant. Previously we have reported that mutation of the HOG pathway increases LAC1 expression and subsequently enhances melanin production [9]. Unlike the hog1D mutant, deletion of HRK1 did not increase but instead slightly decreased melanin production (Fig. 7B), suggesting that Hrk1 plays a minor role in melanin biosynthesis.
The HOG pathway is negatively involved in sexual differentiation of C. neoformans by repressing pheromone production under normal conditions [5]. However, Hrk1 was not involved in sexual differentiation of C. neoformans. Even in a bilateral cross between MATa and MATa hrk1D mutants, normal filamentous growth and basidia and basidospore formation were observed similar to the cross between wild-type strains (Fig. S6). Thus, we conclude that Hrk1 is not involved in sexual differentiation of C. neoformans.

Hrk1 is dispensable for virulence of C. neoformans
The role of Hrk1 in osmoadaptation, oxidative stress response, and melanin biosynthesis prompted us to investigate its role in virulence of C. neoformans, given the importance of these factors for survival within a host organism. Mice infected with the hrk1D mutant showed symptoms of sickness similar to those infected with the wild-type strain (Fig. 7C). The hrk1D+HRK1 complemented strain exhibited slightly reduced virulence compared to the wildtype strain. During construction of the hrk1D+HRK1 mutant, repeated biolistic transformation may affect its normal virulence (Fig. 7C). These results suggest that Hrk1 is dispensable for virulence of C. neoformans.

Discussion
In this study we characterized functions of a novel Hog1regulated kinase, Hrk1, in C. neoformans by molecular and genetic analyses. Hrk1, identified by our prior transcriptome analysis of the HOG pathway, was confirmed to be differentially regulated via the Hog1 MAPK in response to environmental stresses, such as osmotic shock and oxidative stress. Hog1 is required for the high basal levels of Hrk1, probably due to high Hog1-phosphorylation levels under basal conditions. Similar to the HOG mutants, the hrk1D mutants showed increased resistance to fludioxonil by inhibiting over-accumulation of intracellular glycerol and cell swelling. However, Hrk1 appears to have a number of Hog1independent functions. Hrk1 played a minor role in maintaining cellular osmotic balance but in a Hog1-independent manner because deletion of HRK1 exacerbated osmosensitivity of the hog1D mutant. Furthermore, the hrk1D mutant showed distinct oxidative stress response patterns from the hog1D mutant. The hrk1D mutation increases diamide sensitivity of the wild-type and further enhances H 2 O 2 sensitivity of the hog1D mutant. Finally, hrk1D mutation suppresses azole resistance shown by the hog1D mutant without affecting ERG11 expression. Taken together, Hrk1 not only mediates a subset of HOG-related phenotypes, but also appears to be implicated in several HOG-independent phenotypes (see the proposed model for the role of Hrk1 in C. neoformans in Fig. 8).
Hrk1 is structurally close to other p38/Hog1-dependent MAPKAPKs, including Rck1 and Rck2 in S. cerevisiae, and Srk1/Mkp1 and Cmk2/Mkp2 in S. pombe. Although Hrk1 is most homologous to Srk1, the closest C. neoformans homolog to Cmk2, Rck1, and Rck2 is also Hrk1, indicating that Hrk1 may share functions with either of the MAPKAPKs. Furthermore, none of other C. neoformans genes display any significant sequence homology to the Srk1-or Rck2-like MAPKAPKs. Phylogenetic analysis of conserved gene families in the fission and budding yeasts revealed that Rck2 and Srk1 share a common ancestor but Rck1-Rck2 and Srk1-Cmk2 appear to be independently duplicated in S. cerevisiae and S. pombe, respectively [22]. Therefore, Hrk1 is likely to share the common ancestor with Srk1 and Rck2 but is not duplicated in C. neoformans. Although the MAPKAPKs exhibit high similarity to Ca 2+ /calmodulin-dependent kinases, none of them is involved in Ca 2+ -signaling. Srk1 has been identified by microarray analysis to be one of only two protein kinase genes induced by stress in the fission yeast [23]. Srk1 is transcriptionally upregulated in response to a range of stressors, including H 2 O 2 , osmotic and heat shock, and its induction is completely abolished in the sty1D mutant [11]. Rck1 and Rck2 in the budding yeast were initially discovered as suppressors of fission yeast checkpoint mutation (radiation and hydroxyurea sensitivity) [24]. Later Rck2 was also independently identified as a Hog1-interacting protein via yeasttwo hybrid screen and found to be the direct phosphorylation target of Hog1 in response to osmotic shock [14,15]. Similar to SRK1, HRK1 expression is almost abolished in the hog1D mutant. Conversely, HRK1 expression patterns appear to be different from in hog1D, hrk1D, and hrk1D hog1D mutants compared to the wild-type strain (H99) (B). The expression levels of ERG11 relative to ACT1 were measured via phosphorimager (Fuji BAS 2500) (C). The Y-axis indicates ERG11 expression levels relative to ACT1, which is percent ratio of ERG11 expression level vs. ACT1 expression level. The hog1D and hrk1D hog1D, but not hrk1D, mutants showed higher ERG11 expression levels compared to the wild-type strain. doi:10.1371/journal.pone.0018769.g006 SRK1 in response to environmental signals. HRK1 expression is downregulated by osmotic stress, but upregulated by oxidative stress (Fig. 2). Another difference between Hrk1 and Srk1 is its basal expression levels. Basal expression levels of SRK1 appear to be minimal under unstressed conditions [11] whereas basal expression levels of HRK1 are high even under unstressed conditions (Figs. 1B and 2). This discrepancy may result from the fact that C. neoformans Hog1 is constitutively phosphorylated under unstressed conditions whereas S. pombe Sty1 is not.
Basal expression levels of HRK1 are mainly governed by the Ssk1 response regulator and Hog1 MAPK. Interestingly, though, HRK1 can be strongly induced in response to oxidative and osmotic stress in the ssk1D mutant background, strongly indicating that unknown signaling component(s), other than Ssk1, can activate Hog1 and subsequently induce expression of Hrk1. In response to osmotic stress (1 M NaCl), HRK1 expression is slowly recovered to the wild-type levels in the ssk1 mutant although it is not known whether the same signaling component(s) is involved in both the oxidative and osmotic stress response. These findings correspond to our previous observation that constitutive phosphorylation of Hog1 almost disappears in the ssk1D mutant but its phosphorylation can be induced by 1 M NaCl [6]. This signaling component(s) should exist upstream of the Ssk2 MAPKKK and Pbs2 MAPKK since Hog1 phosphorylation is not induced in the ssk2D or pbs2D mutant [5,8]. The identity of the Ssk1-independent signaling component(s) upstream of the Ssk2-Pbs2-Hog1 MAPK module will be elucidated in future studies.
In addition to different expression patterns, functions of Hrk1 appear to be distinct from those of other Hrk1 orthologs, such as Srk1, Cmk2, Rck1 or Rck2. Although Srk1 is directly phosphorylated by Sty1 and translocated into the nucleus in S. pombe, the srk1D mutant does not exhibit any discernable stress-related Figure 7. The role of Hrk1 in capsule and melanin production and virulence of C. neoformans. (A) Each C. neoformans strain were spotted and cultured on DME medium for capsule production at 37uC for 2 days. Capsule was visualized by staining with India ink and observed by microscopy (Bar, 10 mm). Quantitative measurements of the relative capsule volume (right panel). The packed volume of the cells was measured by calculating the ratio of the length of packed cell volume phase/length of total volume phase. Statistical differences in relative capsule size between strains were determined by Bonferroni's multiple comparison test. Error bars indicate the standard deviations. *, P,0.001. (B) Each strain was spotted and grown on Niger seed medium (glucose 0.1, 0.5, and 1%) at 30uC or 37uC for 5 days. (C) For virulence assay, groups of six-to eight-week old female A/J mice were infected with 5610 4 cells in 0.05 ml PBS by intranasal inhalation. Percent survival (%) was monitored daily until all mice were sacrificed. The hrk1D mutant is as virulent as the WT strain. doi:10.1371/journal.pone.0018769.g007 phenotypes. Instead, Srk1 was found to repress meiosis and sporulation in the fission yeast whereas Cmk2/Mkp2 is not involved in the mating and meiotic processes [11,20]. Cmk2 is involved in oxidative stress response (against H 2 O 2 and sodium arsenite) in S. pombe [10]. Similarly, although Rck2 is phosphorylated by Hog1 in response to osmotic stress in S. cerevisiae and overexpression of RCK2 restores salt resistance of the hog1D mutant, deletion of either RCK2 or RCK1, or both does not cause any obvious osmosensitive phenotypes and does not influence activity of glycerol-3-phosphate dehydrogenase and catalase [14,15]. Furthermore, the rck1D, rck2D, or rck1D rck2D mutants do not show any other abnormalities in growth, morphology, conjugation, and actin cytoskeleton regulation [24]. Similar to Srk1, however, both Rck1 and Rck2 repress meiosis and sporulation [12]. Rck1 and Rck2 are also involved in oxidative stress response [16], similar to S. pombe Cmk2. The rck2D mutant is more sensitive to tBOOH and cadmium than WT whereas the rck1D mutant is only slightly more sensitive to tBOOH [16]. In contrast, Hrk1 plays a substantial role in response and adaptation to diverse environmental stresses and antifungal drugs. Hrk1 is required for intracellular glycerol synthesis via the HOG pathway and thereby controls sensitivity to the antifungal drug fludioxonil (Fig. 4). Hrk1 is required for resistance to both osmotic stress and H 2 O 2 in the absence of the Hog1 MAPK. In addition, Hrk1 mediates resistance to diamide in a Hog1-independent manner ( Fig. 4 and 5). In conclusion, Hrk1 plays more significant roles in environmental stress response than Srk1 and Cmk2 in S. pombe or Rck1 and Rck2 in S. cerevisiae.
Regardless of the stress-related and virulence factor-related phenotypes, Hrk1 was found to be dispensable for virulence of C. neoformans, suggesting that single antifungal therapy with any Hrk1 inhibitors will not be effective for treatment of cryptococcosis. Nevertheless, Hrk1 could be a good target for combination antifungal therapy with azole drugs. Inhibition of Hrk1 increases fluconazole-sensitivity of C. neoformans. Notably, mutation of HRK1 completely suppresses azole-resistance of the hog1D mutant. Indeed the hrk1D hog1D double mutant displayed increased susceptibility to both fluconazole and ketoconazole. Since the hrk1D mutant exhibits nearly wild-type levels of amphotericin B resistance, Hrk1 does not appear to be directly involved in ergosterol biosynthesis. Supporting this, ERG11 expression was not altered by deletion of HRK1. These results indicate that simultaneous inhibition of the HOG pathway and Hrk1 along with combination of either polyene or azole drugs could be a good combination antifungal therapy for treatment of cryptococcosis. In future studies, it needs to be further investigated how Hrk1 promotes azole-resistance in C. neoformans.
In conclusion, Hrk1 plays both redundant and distinct roles with the HOG pathway in environmental stress and antifungal drug response and virulence factor production in C. neoformans. Hrk1 is a protein kinase downstream of the Hog1 MAPK and yet has the potential to be modulated by other signaling pathways, which will be elucidated in future studies.

Ethics Statement
All animal experiments were done at the University of Minnesota in strict accordance with good animal practice as defined by the National Institutes of Health Office of Animal Welfare (OLAW), and the Association for the Assessment and Accreditation of Laboratory Animal Care (AAALAC). All experiments were reviewed and approved by the University of Minnesota Institutional Animal Care and Use Committee (IACUC) under protocol number 0712A22250. However, Hrk1 appears to play several Hog1-independent roles. Independent of the HOG pathway modulating ergosterol biosynthesis, Hrk1 promotes susceptibility to polyene and azole drugs. Furthermore, Hrk1 is partly involved in promoting resistance to oxidative damaging agents, such as diamide, and melanin biosynthesis, which is in stark contrast to Hog1. In conclusion, Hrk1 plays both Hog1-dependent and -independent roles in C. neoformans. doi:10.1371/journal.pone.0018769.g008 Strains and growth conditions C. neoformans strains used in this study are listed in Table 1 and were cultured in YPD (yeast extract-peptone-dextrose) medium. Agar-based DME (Dulbecco's modified Eagle's) medium (Invitrogen, Carlsbad, CA) for capsule production, Niger seed or L-DOPA medium for melanin production, and V8 mating medium were prepared as previously described [25]. For stress and antifungal drug sensitivity, cells were grown as follows. Each strain was incubated overnight at 30uC in YPD medium, washed, serially diluted (1 to 10 4 dilutions) in dH 2 O, and spotted (4 ml) onto solid YPD medium. To test osmosensitivity, cells were spotted on solid YP (no glucose) or YPD medium containing 1 M or 1.5 M of NaCl or KCl. To test sensitivity to antifungal drugs, cells were spotted on solid YPD medium containing the indicated concentration of fludioxonil (PESTANALH; Sigma), fluconazole (Sigma), ketoconazole (Sigma), or amphotericin B (Sigma). To examine oxidative stress, the cells were spotted on YPD containing the indicated concentration of H 2 O 2 (Junsei), diamide, tert-butyl hydroperoxide (tBOOH), or menadione. To test sensitivity to UV, cells spotted on solid YPD medium were exposed to UV irradiation for 0.2 (480 J/m 2 ) or 0.3 (720 J/m 2 ) min using a UV Stratalinker TM (Model 2400, Stratagene). To test temperature sensitivity, plates were incubated at 30, 37, and 40uC. Each plate was incubated for 2-5 days, and photographed.

Quantitative reverse transcriptase-PCR (qRT-PCR) analysis
The qRT-PCR for quantitatively measuring relative expression levels of HRK1 was performed with primers listed in Supplementary Table S1. cDNAs were synthesized using the SuperScript II reverse transcriptase system with three independent sets of total RNA samples that were prepared for the previous DNA microarray analysis of the HOG pathway in C. neoformans [9]. Relative gene expression of HRK1 was calculated by the threshold cycle 2 2DDCT method [26] from three biological replicates and three technical replicates. Transcript levels of the ACT1 gene were used for normalization.

Disruption of the HRK (Hog1-regulated kinase) gene
To characterize the function of Hrk1, the HRK1 gene was deleted in the genetic background of C. neoformans serotype A strain H99 (MATa) and KN99a (MATa) with a disruption cassette generated by overlap PCR and by biolistic transformation as previously described [5,27]. Primers for amplification of the 59 and 39 flanking regions of the HRK1 gene and Nat r selectable markers are described in supplemental Table S1. Stable transformants were selected on YPD medium containing nourseothricin. The hrk1D mutant strain was confirmed by both diagnostic PCR and by Southern blot analysis using a gene-specific probe prepared by primers listed in supplemental Table S1 (Fig. S2). To construct the hrk1D hog1D double mutant, the hog1D mutant (YSB64) was biolistically transformed with the hrk1D disruption cassettes containing a Neo r selectable marker and confirmed by diagnostic PCR and Southern blot analysis (Fig. S3).
To authenticate hrk1 mutant phenotypes, the hrk1D+HRK1 complemented strains were constructed as follows. First, the 4.2-kb fragment containing the complete ORF of the HRK1 gene was amplified by PCR with primers B32 and B33 (Table S1) and cloned into the pCR2.1-TOPO vector (Invitrogen), generating pCR-HRK1. After confirming the DNA sequence, the HRK1 insert was subcloned into plasmid pJAF12 (NEO r ), generating plasmid pNEOHRK1. For the targeted re-integration of the wild-type HRK1 allele into its native locus, pNEOHRK1 was linearized by BbsI digestion and biolistically transformed into the hrk1 strain YSB270 (Table 1).

Southern, Northern, and Western blot analysis
For Southern blot analysis, genomic DNAs were isolated from each strain using CTAB (Cetyl trimethylammonium bromide, Sigma) extraction buffer as described previously [28]. Isolated genomic DNA was digested with the indicated restriction enzymes (i.e. XbaI and XhoI for checking deletion of the HRK1 gene and BamHI for checking the hrk1Dhog1D double mutant). The digested genomic DNAs were separated by 1% agarose gel-electrophoresis, denatured in 0.4 N NaOH, transferred to the nylon membrane (GE) in 0.4 N NaOH and 1 M NaCl, and fixed by 1200 J/m 2 UV exposure. The membrane was hybridized, washed, and developed as described before [28]. For Northern blot analysis, each strain was grown to mid-logarithmic phase in YPD medium at 30uC. Total RNAs were isolated using Trizol (Invitrogen) or Ribo-EX (Geneall) as described before [28]. Northern blot analysis was performed with 10 mg of total RNA from each strain. Electrophoresis, membrane transfer, hybridization, and washing as described before [28]. Gene specific probes were prepared by PCR amplification with gene specific primers listed in Table S1. Western blot analysis for monitoring Hog1 phosphorylation was performed as previously described [5]. To detect phosphorylated C. neoformans Hog1, rabbit p38-MAPK specific antibody (Cell Signaling) and anti-rabbit IgG horseradish peroxidase-conjugated antibody were used as primary and secondary antibodies, respectively. The blot was developed using the ECL Western Blotting Detection System (Amersham Bioscience). Subsequently, the blot was stripped and used for detection of Hog1 with a rabbit polyclonal anti-Hog1 antibody (Santa Cruz Biotechnology) as a loading control.

Assays for capsule and melanin production and mating
Qualitative visualization of capsule and melanin production was performed as described previously [25,29]. Mating assay was performed as described previously [25,29].

Intracellular glycerol measurement
Each strain was grown to mid-logarithmic phase. Cells were subcultured into YPD liquid medium with or without fludioxonil (final 10 mg/ml) for 48 hrs at 30uC with shaking. After incubation, cell morphologies were observed by microscope and photographed. To measure the intracellular glycerol content, cells were collected by centrifugation, washed with distilled water and then extracted by boiling at 100uC for 10 minutes. After centrifugation, the glycerol concentration in the supernatant was measured using a Glycerol UV-method kit according to the manuals (Roche). The same amount of each sample was lyophilized to measure cellular dry weight for normalization.

Virulence assay
For animal infections, overnight cultures of C. neoformans strains H99, hrk1D (YSB270), hrk1D+HRK1 (YSB568) were grown in YPD broth at 30uC. The resulting cultured yeast cells were pelleted and washed 3 times in sterile PBS and concentrations were adjusted to 1610 6 cells/ml for each strain based on hemocytometer count. Virulence studies were performed using the murine inhalation model. Groups of six-to eight-week old female A/J mice (Jackson Labs, Bar Harbor, MA) were anesthetized by intraperitoneal pentobarbital injection. Mice were infected with 5610 4 cells in 0.05 ml PBS by intranasal inhalation. Initial concentrations were confirmed by plating serial dilutions on YPD agar and enumerating colony-forming units (CFUs). Mice were monitored daily and those showing signs of severe morbidity (weight loss, abnormal gait, extension of the cerebral portion of the cranium) were sacrificed by CO 2 inhalation. Homogenized lung tissue was plated onto YPD agar or YPD media containing nourseothricin (200 mg/ml) or neomycin (100 mg/ml) from representative animals from each infection to verify the genotype of the recovered cryptococcal strains. All animal experiments were reviewed and approved by the University of Minnesota Institutional Animal Care and Use Committee (IACUC) under protocol number 0712A22250. Survival data from the mouse experiment were analyzed by Analyse-It for Excel. gene in YSB270 (hrk1D a) strain. Primers for the first-round and second-round PCR are indicated as bent arrows. Through recombination between 59 and 39 flanking region of HOG gene, the intact HOG1 gene is displaced with neomycin-resistant gene (NEO). (B) The correct genotype of the hrk1D hog1D double mutants was confirmed by Southern blot analysis using genomic DNA digested with the restriction enzyme BamHI. (PPT) Figure S4 Western blot analysis of Hog1 phosphorylation in WT and hrk1D mutant. The WT (strain H99) and hrk1D (YSB270) mutant strains were grown to the mid-logarithmic phase and exposed to 1 M NaCl in YPD medium for the indicated amount of time and total protein extracts were prepared. The dual phosphorylation status of Hog1 (T171 and Y173) was examined by using anti-dually phosphorylated p38 antibody (Hog1-P). Subsequently the same blots were stripped and reprobed with polyclonal anti-Hog1 antibody as a loading control (Hog1). (PPT) Figure S5 Hrk1 does not play a role in resistance to high temperature, UV irradiation, and cell membrane/wall integrity destabilizers. Each C. neoformans strain indicated below was grown overnight (about 16 hrs) at 30uC in liquid YPD medium, 10-fold serially diluted (1-10 4 dilutions), and spotted (4 ml of dilution) on YPD agar containing the indicated concentrations of SDS, DTT, Congo red, and H 2 O 2 . To test genotoxic DNA damaging stress, cells were spotted on solid YPD medium and exposed to 400 J/m 2 of UV by using a UV crosslinker (UVP). To test temperature sensitivity, plates were incubated at 38uC for 3 days. The wild-type H99, ssk1D (YSB261), hog1D (YSB64), skn7D (YSB349), hrk1D (YSB270), hrk1D+HRK1 (YSB883), and hrk1D hog1D (YSB988)] strains were used. (PPT) Figure S6 Hrk1 is not required for sexual differentiation in serotype A C. neoformans strain. Each MATa and MATa strains were co-incubated on V8 medium (pH 5.0) for up to 15 days at room temperature in the dark: WT a6WT a (H99 and KN99a), hog1D a6hog1D a (YSB64 and YSB81), hrk1D a6hrk1D a (YSB270 and YSB874). The images were photographed after 7 and 15 days. (PPT)