All living organisms, whether they be single- or multicellular, actively interact with their surrounding environments and modulate their physiological status to maintain cellular homeostasis. This adaptation process is highly coordinated via diverse signaling pathways, with the involvement of a series of signaling components, including sensors/receptors, kinases, and transcription factors. For microorganisms, the environment is nonfavorable and can be considered a stress. Sensing and responding to a plethora of environmental stresses are key requirements for pathogenic fungi having a saprophytic life cycle, including Cryptococcus neoformans and Aspergillus fumigatus, to colonize successfully and proliferate within the corresponding host system. In either commensal fungi that are able to cause opportunistic infection when the host immune system is compromised (e.g., Candida albicans) or parasitic fungi whose survival depends on the host without saprophytic cycles (e.g., Pneumocystis jirovecii, the dermatophytes, and the microsporidia), the stress-sensing and adaptation process is still required for survival at different host anatomical sites having distinct environmental conditions, during dissemination of organisms, or to counterbalance changes in overall host physiological conditions. These features confer an underlying distinction between pathogenic and nonpathogenic microbes, along with regulation of diverse virulence factors.
Several signaling cascades have been discovered to be involved in the stress response and adaptation of fungi, including the cyclic AMP (cAMP) signaling pathway, Ca2+/calcineurin signaling pathway, protein kinase C (PKC)/Mpk1 (also known as Slt2) mitogen-activated protein kinase (MAPK) pathway, and stress-activated Hog1 MAPK pathway (reviewed in references 15 and 25). Among them, the HOG (high-osmolarity glycerol response) pathway is one of the best-characterized stress response signaling cascades in both fungi and mammals. As the name implies, the HOG signaling pathway was originally identified as playing a pivotal role in maintaining water balance against hyper- and hypoosmotic conditions in the budding yeast Saccharomyces cerevisiae (for earlier reviews, see references 58, 59, 94, and 113).
In pathogenic fungi, however, the HOG pathway plays only a minor role in controlling osmolarity. Instead, it is a major controller governing cellular responses to diverse external stimuli, including osmotic shock, UV irradiation, oxidative and heavy metal stress, and high temperature. Furthermore, the HOG pathway is involved in the growth, morphology, differentiation, and virulence factor regulation of pathogenic fungi, such as capsule and melanin synthesis in C. neoformans and polymorphic transitions in C. albicans (6, 7, 13, 14, 43).
The stress-responsive HOG pathway in fungi consists of two major signaling modules. One is the MAPK module comprising MAPK, MAPK kinase (MAPKK), and MAPKK kinase (MAPKKK), and the other is the two-component system-like phosphorelay system composed of hybrid sensor kinases, histidine-containing phosphotransfer (HPt) protein, and response regulators, which senses and relays environmental signals and subsequently activates the Hog1 MAPK pathway (Fig. 1). Medical mycologists have been interested in investigating the HOG pathway in fungal pathogens mainly because the phosphorelay system is not found in humans. Although it has been suggested that mammals indeed contain some histidine kinases, such as branched-chain α-keto acid dehydrogenase kinase and pyruvate dehydrogenase kinase (19, 129), the multicomponent phosphorelay system observed in fungi has not been discovered in humans. Therefore, perturbation of the HOG pathway by modulating the unique two-component system may provide an unprecedented opportunity to develop novel antifungal therapeutic methods without causing significant toxicity.
General regulatory mechanism of the HOG signaling pathway in fungi. Abbreviations: HK, histidine KD; RR, response regulator receiver domain; S/T kinase, Ser/Thr protein KD; SBD, Ssk2/Ssk22 binding domain; HDS, Hog1 docking site; S/T-Y KD, Ser/Thr and Tyr dual-protein KD.
A report shows that mortality caused by invasive mycotic disease has dramatically increased in the United States, from 1,557 deaths in 1980 to 6,534 deaths in 1997, with Candida, Cryptococcus, and Aspergillus species being mainly responsible for the mycosis-related deaths (77). C. albicans is an ascomycete fungus that can cause both superficial and systemic infections, mainly in immunocompromised patients having AIDS or patients being subjected to anticancer chemotherapy or organ transplants. Particularly, Candida species are the fourth leading cause of nosocomial bloodstream infections in U.S. hospitals (142). C. neoformans is a basidiomycete fungal pathogen that can cause deadly fungal meningitis in both immunocompromised and immunocompetent individuals (60, 71). A. fumigatus is an ascomycete filamentous fungus that is a ubiquitously distributed saprophytic mold. It can generate airborne spores that, when inhaled into the lung, can cause death, especially in immunocompromised patients (57).
This review mainly focuses on the current findings about the role of the two-component system and the HOG signaling pathway in opportunistic human fungal pathogens, including C. albicans, C. neoformans, and A. fumigatus, compared to other yeasts or filamentous fungi in which this pathway has been extensively studied, including S. cerevisiae, Schizosaccharomyces pombe, Neurospora crassa, and Aspergillus nidulans. Therefore, this review aims to provide insights into the commonality and difference in the function and regulatory mechanism of the HOG pathway between pathogenic and nonpathogenic fungi and to help understand the role of its complex signaling network in establishing the virulence of fungal pathogens.
THE GENERAL REGULATORY MECHANISM OF THE HOG PATHWAY
The general regulatory mechanism of the HOG pathway is illustrated in Fig. 1, mainly based on data from S. cerevisiae and S. pombe (for reviews, see references 58, 59, and 94). In fungi, MAPKKKs (Ssk2 and Ssk22 in S. cerevisiae) in the Hog1 MAPK module are activated by the two-component-like phosphorelay system, which is composed of multiple components, including hybrid sensor kinases, the HPt protein, and response regulators. In response to external cues, a hybrid sensor kinase autophosphorylates a His residue in the histidine kinase domain (KD) and subsequently transfers the phosphate to an Asp residue in the response regulator (receiver) domain of the same sensor kinase. Next, the phosphate is transferred to a conserved His residue of an HPt protein, which finally relays the phosphate to an Asp residue of a response regulator. Response regulators either directly activate the expression of downstream target genes as a transcription factor or activate the Hog1 MAPK module (Fig. 1). Since no two-component signaling components have been discovered in mammals, these are considered good candidates as antibacterial or antifungal targets (31, 76).
The activated response regulators (either phosphorylated or dephosphorylated) can interact with the autoinhibitory domain (AID) of Ssk2 or Ssk2-like MAPKKKs, which releases and exposes the Ser/Thr KD for autophosphorylation of the conserved Thr residue in the MAPKKK. Subsequently, a phosphorylated, activated Ssk2-like MAPKKK phosphorylates the conserved Ser and Thr residues of the Pbs2-like MAPKK via binding to the Ssk2/Ssk22-binding domain (SBD). The activated Pbs2-like MAPKK, in turn, dually phosphorylates Thr and Tyr residues in the TGY motif of the Hog1-like MAPK, which is evolutionarily conserved from yeasts to humans. The phosphorylated Hog1 MAPK is translocated into the nucleus and induces the expression of stress-defensive target genes to counteract the incoming environmental stress directly or indirectly via MAPK-activated protein kinases (MAPKAPKs) or transcription factors (Fig. 1). In the case of hyperosmotic stress, genes that encode enzymes that play a key role in glycerol production, such as glycerol-3-phosphate dehydrogenase (Gpd1 and Gpd2) and glycerol-3-phosphatase (Gpp1 and Gpp2), are transcriptionally upregulated to maintain the intracellular water balance. For negative feedback regulation of the HOG pathway upon adaptation to stress, Hog1-like MAPK and/or Pbs2-like MAPKK are dephosphorylated by specific phosphatases and exported out of the nucleus (Fig. 1).
THE TWO-COMPONENT-LIKE PHOSPHORELAY SYSTEM
Hybrid sensor kinases.Sensor kinases have been discovered in prokaryotes and some eukaryotes, including bacteria, slime molds, plants, and fungi, but not in the animal kingdom (17, 76, 92). Indeed, mammals do contain some histidine kinase-like proteins, such as the Nm23 metastasis suppressor, the histone H4 histidine kinase, and the G protein β-subunit kinase, yet sensor-type histidine kinases have not been reported (128). Among the diverse sensor kinases, all of the sensor kinases identified in fungi have a hybrid form consisting of both histidine kinase and response regulator receiver domains in a single polypeptide, in contrast to the majority of bacterial sensor kinases, and therefore are called hybrid sensor kinases (often called hybrid histidine kinases). The fungal hybrid sensor kinases are highly diverse in terms of the number in a single species, functions, and domain structures (Table 1 and Fig. 2). Generally, filamentous fungi harbor more hybrid sensor kinases than yeast types of fungi (Fig. 2). Catlett and coworkers have categorized fungal hybrid sensor kinases into 11 classes (30). However, it becomes more evident that the fungal hybrid sensor kinases appear to be more diverse than this initial estimate (Fig. 2). For example, the C. neoformans Tco2 hybrid sensor kinase contains two histidine KDs and two response regulator receiver domains in a single polypeptide that cannot be classified into any of the 11 classes proposed by Catlett et al. (Fig. 2).
Fungal hybrid sensor kinases. The phylogenetic tree of fungal hybrid sensor kinases was constructed in CLUSTAL W (tree-building method, neighbor joining; mode, best tree; distance, uncorrected [“p”]) with MacVector software (version 7.2.3; Accelrys). The roman numerals on the left are the classes of hybrid sensor kinases categorized by Catlett et al. (30). On the right side are diagrams of known fungal sensor kinases, which were generated by the MacVector software. Each domain was identified by the comparative fungal genomics platform (http://cfgp.snu.ac.kr) and InterPro (http://www.ebi.ac.uk/interpro/). Abbreviations: GAF, GAF domain; HK, histidine KD; RR, response regulator receiver domain; PK, protein KD; TTPH, tetratricopeptide-like helical domain; PAS, PAS/PAC domain; HAMP, HAMP domain; PChr, phytochrome domain; TM, transmembrane domain (predicted by the TMHMM Server v 2.0 [http://www.cbs.dtu.dk/services/TMHMM/]); HEM4, uroporphyrinogen III synthase (HEM4) domain.
Components involved in the HOG signaling pathway
Hybrid sensor kinases have been first and best characterized in the model budding and fission yeasts S. cerevisiae and S. pombe, respectively. S. cerevisiae contains a single Sln1 hybrid sensor kinase that is localized in the cell surface membrane and senses delicate osmotic changes (73, 99). Deletion of SLN1 results in constitutive dephosphorylation of Ypd1 HPt and the Ssk1 response regulator, which hyperactivates the Ssk2/Ssk22-Pbs2-Hog1 pathway and causes cell lethality by increasing intracellular glycerol levels (73, 99).
S. pombe possesses three hybrid sensor kinases, Mak1, Mak2, and Mak3 (Mcs4-associated kinases, also known as Phk3, Phk1, and Phk2, respectively). All Mak sensor kinases contain PAS/PAC repeats, which have been found in proteins for sensing light, oxygen, and the redox state of cells, but do not have any transmembrane domains, suggesting that they are not transmembrane proteins, in contrast to S. cerevisiae Sln1 (26, 83) (Fig. 2). Among them, Mak2 and Mak3 contain another unusual structural feature at the N terminus, an atypical Ser/Thr KD that is also found in prokaryotes and whose function is not known (26). Notably, activated Mak2 and Mak3 appear to activate the Wis1-Sty1 pathway subsequently in response to oxidative stresses, which is in contrast to the finding that activated Sln1 inhibits the Hog1 MAPK (26, 73, 99). On the other hand, Mak1, a shorter protein than Mak2 and Mak3, mediates the oxidative stress response in a Sty1-independent fashion through the Prr1 response regulator and the Pap1 transcription factor (26). Furthermore, all S. pombe mak1Δ, mak2Δ, and mak3Δ mutants, and even the triple mutant, are viable, unlike the S. cerevisiae sln1Δ mutant (26, 83). However, hybrid sensor kinases that monitor other stress conditions, such as osmosis changes and nitrogen starvation during mating, have not been uncovered in S. pombe. In addition to sensing stress, Mak sensor kinases play another role in the growth and differentiation of S. pombe. Multiple deletion of all three MAK genes causes phenotypes similar to those of the spy1Δ (or mpr1Δ) HPt mutant, exhibiting premature entry into the mitotic phase of the cell cycle and shorter cell length than the wild type, whereas the sty1Δ and wis1Δ mutants are delayed in the initiation of mitosis and exhibit the longer cell shape (8, 78, 139), indicating that each Mak protein plays a redundant role in cell cycle control. Taken together, the evidence shows that the Mak1/2/3 and Spy1 system controls the Wis1-Sty1 pathway in both positive and negative manners.
C. albicans contains three hybrid sensor kinases, Sln1, Nik1 (also known as Cos1), and Chk1 (4, 81, 146). Although C. albicans Sln1 can functionally replace S. cerevisiae Sln1, it is not required for the viability of C. albicans and yet controls growth under hyperosmotic conditions (81, 146). Notably, C. albicans hybrid sensor kinases are involved in hyphal development, which is considered to be a major virulence attribute of the pathogen (146). The sln1Δ mutant is defective in hyphal formation and therefore exhibits reduced virulence in a systemic candidiasis model (146). C. albicans Nik1 is involved in morphological transitions and phenotypic switching but not in osmosensing, although it is structurally homologous to the N. crassa Nik1 osmosensor (4, 127). Disruption of the CHK1 gene causes more severe defects in hyphal development than deletion of SLN1 and NIK1 (146). Accordingly, virulence attenuation is most prominent in the chk1Δ mutant compared to the sln1Δ and nik1Δ mutants (146). Interestingly, double deletion of CHK1 in either the sln1Δ or the nik1Δ mutant partially restores the filamentation and pathogenicity of C. albicans (146). Although the exact regulatory mechanism has not been understood, it appears that there exists some cross talk between these hybrid sensor kinases.
C. neoformans possesses seven proteins that are highly homologous to hybrid sensor kinases, named the Tco1 to Tco7 proteins (two-component-like proteins) (13). Among them, Tco1 and Tco2 play discrete and redundant roles in activating the HOG signaling pathway. Disruption of TCO1 increases resistance to fludioxonil, which is in the phenylpyrrole class of antifungal drugs, and increases melanin biosynthesis, phenotypes which are comparable to a subset of the Hog1-related phenotypes described below (13). However, the tco1Δ mutant is impaired in sexual reproduction due to defective cell-cell fusion during the initial stage of mating, which is the opposite of hog1Δ mutants, which exhibit enhanced mating (13). Therefore, and similar to the situation in S. pombe, a hybrid sensor kinase could modulate the HOG pathway in either a positive or a negative manner in C. neoformans. Disruption of TCO2 also increases fludioxonil resistance, indicating that Tco1 and Tco2 play redundant roles in protecting cells from fludioxonil treatment. Unlike Tco1, however, Tco2 promotes sensitivity to hydrogen peroxide and methylglyoxal but is not involved in mating and production of melanin (13). Tco2 is also involved in osmosensing, but to a lesser extent than Hog1 (13). Interestingly, deletion of TCO1, but not TCO2, attenuates the virulence of C. neoformans (13, 34). Chun and coworkers showed that Tco1 promotes cellular resistance to the hypoxic conditions commonly found in the human host (34). Therefore, it is likely that during host infection Tco1 senses low oxygen levels inside the host and enables the pathogen to adapt to the hypoxic conditions. Notably, phenotypes of the tco1Δ tco2Δ double mutant are similar to only a subset of phenotypes observed in the hog1Δ mutant (13), indicating that other receptors or sensors could exist and modulate the downstream response regulators or directly activate the Hog1 MAPK module by bypassing the two-component system. The functions of other Tco proteins, including Tco3, Tco4, Tco5, Tco6, and Tco7, remain elusive. Although Tco3 is structurally homologous to red light-sensing phytochromes (Fig. 2), its role as a light sensor remains unclear. Recent annotation by the serotype A C. neoformans genome database suggests that the TCO4 open reading frame could be further extended than originally proposed (13), containing two histidine KDs and two response regulator domains, similar to Tco2 (Fig. 2). However, Tco4 does not produce any discernible phenotypes (13). Likewise, neither Tco5 nor Tco7 produces significant stress-related phenotypes. However, the tco7Δ mutant shows some minor hypersensitivity to hydrogen peroxide and methylglyoxal, indicating that Tco7 may play some roles that are redundant with respect to Tco2 (13). The essentiality of Tco6 has been proposed (13) but needs to be experimentally confirmed.
As mentioned previously, filamentous fungi generally have a far greater number of sensor kinases than yeast-type fungi (Table 1 and Fig. 2). N. crassa, A. nidulans, and A. fumigatus have 11, 15, and 14 hybrid sensor kinases, respectively. Catlett and coworkers have reported that some filamentous ascomycetes (Euascomycetes) contain even greater numbers of hybrid sensor kinases: 16 in Gibberella moniliformis, 21 in Cochliobolus heterostrophus, and 20 in Botryotinia fuckeliana (30). However, only a subset has been characterized (Table 1 and Fig. 2). Particularly, sensor kinases studied in filamentous fungi are all involved mainly in morphology and differentiation but not in stress sensing, besides some role in osmosensing. Therefore, stress-sensing hybrid sensor kinases in filamentous fungi remain to be further elucidated. In filamentous fungi, an Sln1-like hybrid sensor kinase, TcsB, has been identified and characterized in A. nidulans and A. fumigatus (42, 49). A. nidulans TcsB is the structural and functional homologue to yeast Sln1 because it has two transmembrane regions at the N terminus (Fig. 2) and its overexpression suppresses the lethality of the S. cerevisiae sln1Δ mutation (49). Unexpectedly, however, the tcsBΔ mutant does not show any detectable phenotypes in stress response and morphology compared to the wild-type strain (49). Similarly, the A. fumigatus tcsBΔ mutant displays almost wild-type phenotypes, except for some minor sensitivity to a detergent, sodium dodecyl sulfate, indicating that another hybrid sensor kinase(s) plays an important role in activating the two-component system of Aspergillus species (42). Indeed, similar to C. albicans and C. neoformans, cytosolic hybrid sensor kinases appear to be more important in controlling the morphogenesis and differentiation of filamentous fungi than transmembrane histidine kinases. A cytosolic hybrid sensor kinase in filamentous fungi was first identified in N. crassa and named Nik-1/Os-1 (3). N. crassa os (osmosensitive) mutants, including the os-1, os-2, os-4, and os-5 mutants, have been isolated by screening of cells displaying sensitivity to osmotic shock (95). N. crassa Nik-1/Os-1 possesses a unique 90-amino-acid tandem motif, called the HAMP domain, in the N terminus. Interestingly, NIK-1 is specifically expressed during the vegetative phase but not in the sexual phase (3). The N. crassa nik-1Δ mutant displays aberrant hyphal development, a phenotype which is more obvious under hyperosmotic conditions (3). A. nidulans contains several cytosolic hybrid sensor kinases. One of them is the TcsA (two-component signaling protein A) sensor kinase, which contains two PAS domains in the N terminus (Fig. 1) and is partially involved in the production of conidia but not in conidiophore formation (138). However, the tcsAΔ mutant does not exhibit any phenotypes during stress responses. The human pathogen A. fumigatus contains an N. crassa Nik-1 homologue, Fos-1/TcsA, although it does not contain 90-amino-acid tandem repeats, as observed in Nik1/Os-1 (3). A. fumigatus lacking Fos-1/TcsA is defective in normal conidiophore development and cell wall assembly (100) and exhibits attenuated virulence in systemic aspergillosis (35).
Recently, a HAMP domain-containing hybrid sensor kinase, NikA, was identified in A. nidulans and was shown to play important roles in sensing fungicide and to be involved in regulating normal asexual sporulation and conidiospore viability, yet it is not implicated in sensing osmolarity and oxidative stresses (51, 137). Finally, A. nidulans has a red light-sensing hybrid sensor kinase called FphA (103). The FphA phytochrome represses sexual development and mycotoxin formation (103). Similar red light-sensing phytochromes, PHY-1 and PHY-2, have been discovered in N. crassa, yet their functions in photosensing remain unclear (47).
Taking the findings together, it seems clear that filamentous ascomycetous fungi harbor a far greater number of sensor histidine kinases than the ascomycetous yeasts such as S. cerevisiae, S. pombe, and C. albicans, regardless of the presence of pathogenicity. The basidiomycetous fungus C. neoformans contains an intermediate number of sensor histidine kinases (a total of seven). Catlett et al. speculated that the large number of sensor histidine kinases found in filamentous fungi may reflect their greater range of environmental niches (30). However, only a few sensor histidine kinases in filamentous fungi have been implicated in sensing a limited spectrum of external signals, compared to the yeast sensor histidine kinases, which were known to sense more diverse environmental signals as described above. Therefore, the hypothesis that the number of sensor histidine kinases is correlated with the diversity of environmental niches that an organism encounters needs to be further addressed in future studies. As a matter of fact, it could be possible that diverse hybrid sensor kinases play a distinct role just as signaling mediators rather than as sensors since the majority of them do not appear to be surface proteins. Moreover, a number of hybrid sensor kinases observed in filamentous fungi and C. neoformans could be evolutionary remnants without any obvious cellular functions.
HPt proteins.HPt protein acts as a key connector in the fungal phosphorelay system by relaying the phosphate group in the Asp residue of the response regulator domain of a hybrid sensor kinase to the Asp residue of a response regulator. The first HPt protein discovered is the SpoB phosphotransferase in the bacterium Bacillus subtilis, which is involved in the regulation of sporulation induction (27). All of the fungi that have been studied have a single HPt protein in their genomes. The structural and functional features of fungal HPt proteins have been extensively studied in S. cerevisiae. Ypd1 is the first HPt protein uncovered in fungi and acts as an essential intermediate phosphorelay protein in the Sln1-Ypd1-Ssk1 two-component system by controlling the osmosensing mechanism of S. cerevisiae (99). Based on X-ray crystallography studies conducted by two research groups, yeast Ypd1 comprises a bundle of four α helices (designated αA to αD) as a core structure and contains a phosphoaccepting His residue on the αC helix (H64) (126, 145). Similar to the Sln1 sensor kinase, disruption of YPD1 results in lethality by overactivation of the HOG pathway. Therefore, disruption of any signaling components downstream of Ypd1 (any of the Ssk1, Ssk2, Pbs2, or Hog1 proteins) or overexpression of the PTP2 phosphatase abolishes the requirement of Ypd1 for cell viability (99).
The essentiality of Ypd1 seems not to be conserved in other fungi. S. pombe has the Mpr1 (multistep phosphorelay 1, also known as Spy1) phosphotransfer protein that is structurally homologous to S. cerevisiae Ypd1 (9, 87). Unlike Ypd1, however, Mpr1/Spy1 is not required for cell viability and can regulate the downstream Mcs4 response regulator in either a positive or a negative manner for stress response and the mitotic cell cycle (9, 87). Interestingly, only oxidative stresses, but no other forms of stress, including osmotic shock, trigger the activation of Mpr1/Spy1 HPt and Sty1 MAPK (87). However, the mpr1Δ mutant itself is as sensitive to oxidative stress as the wild-type strain, indicating that other signaling pathways, including the Sty1-independent Prr1-Pap1 cascade, are involved in the oxidative stress response (87).
In pathogenic fungi, only a limited amount of information about the role of HPt proteins has been accumulated. C. albicans has a single HPt protein, named Ypd1, which is able to functionally replace S. cerevisiae Ypd1 (28). However, the detailed functions of Ypd1 in the growth, differentiation, and virulence of C. albicans remain to be investigated. Similarly a single Ypd1 homologue has been observed in the genome of C. neoformans (13). Disruption of the YPD1 gene is not feasible in C. neoformans, implying that Ypd1 may be essential for the growth of the pathogen (13). Our recent study further confirmed that Ypd1 is required for the viability of C. neoformans. C. neoformans strains in which the native YPD1 promoter is replaced with the copper-regulated CTR4 promoter exhibit growth defects under promoter-repressing conditions (unpublished data).
In filamentous fungi, A. nidulans has a single HPt protein, named YpdA. Similar to C. albicans Ypd1, heterologous expression of A. nidulans YpdA can also suppress the lethality of the S. cerevisiae ypd1Δ mutant (48). The finding that disruption of ypdA is unsuccessful indicates that YpdA may be indispensable for the viability of A. nidulans (48, 137). The ypdA−/ypdA+ heterokaryons show greater sensitivity to increased osmotic pressure, but not to H2O2, than the wild-type strain, indicating that YpdA is involved in the osmosensing signaling of A. nidulans (137). N. crassa has a Ypd1 homologue, named HPT-1 (histidine phosphotransfer protein 1), which also appears to be essential for viability because disruption of the hpt-1 gene is possible only in the os-2Δ (Hog1 homologue) mutant background and not in the wild-type background (16). These data indicate that lethality caused by hpt-1 disruption is mediated through the Os-2 MAPK. No information about the Ypd1 homologue in A. fumigatus is available.
Response regulators.Most fungi contain two classes of response regulators. One class consists of yeast Ssk1-like response regulators containing a response regulator domain at the C terminus, and the other consists of yeast Skn7-like response regulators containing a heat shock factor-type DNA binding domain and a response regulator domain at their N and C termini, respectively.
Based on studies with S. cerevisiae, the two phosphorylated response regulators Ssk1 and Skn7 act differently. Upon exposure to hypoosmotic conditions, both the Ssk1 and Skn7 response regulators are phosphorylated via the Ypd1 phosphotransfer protein. Ssk1 phosphorylated at the Asp554 residue is prevented from interacting with the AID of Ssk2 or Ssk22 MAPKKK. Instead, the phosphorylated and activated Skn7 response regulator is directly required for induction of FPS1, which encodes a MIP family channel protein that functions in glycerol transport for modulating intracellular glycerol levels (131). Under hyperosmotic conditions, the Ssk1 response regulator is dephosphorylated and interacts with Ssk2 MAPKKK, which triggers autophosphorylation of Ssk2 for its activation (73, 96, 99) (Fig. 1). In S. cerevisiae, the Ssk1 and Skn7 response regulators are localized mainly in the cytoplasm and the nucleus, respectively (72). The phosphotransfer protein Ypd1 shuttles between the cytoplasm and the nucleus to relay phosphates from sensor kinases to corresponding response regulators (72).
Skn7 (also known as Pos9) was isolated as a multicopy suppressor of the kre9Δ mutant affected in the assembly of a cell wall β-glucan (24). With a heat shock factor-DNA binding domain at the N terminus and a glutamine-rich transcription activation domain at the C terminus, Skn7 acts as a transcription factor and is predominantly localized in the nucleus (23). Besides the role of osmosensing, Skn7 mediates the oxidative stress response in an Sln1-Ypd1-dependent and -independent manner by inducing a subset of oxidative stress defense genes, such as the TRX2 thioredoxin gene, in collaboration with Yap1 (69, 70, 80). Interestingly, Skn7 interacts with Hsf1 (heat shock factor 1) to induce heat shock genes in response to oxidative stress (70). However, it is unclear how a single HPt protein determines which response regulator is activated by distinguishing incoming signals.
The structural and functional Ssk1 homologue in S. pombe is Mcs4 (mitotic catastrophe suppressor gene 4), which was named for its ability to control the initiation of mitosis (118). However, several differences between Ssk1 and Mcs4 have been noted. S. pombe Mcs4 has both Sty1-dependent and -independent functions. Mcs4 controls the mitotic cell cycle by mechanisms both dependent on and independent of Sty1 (118). In response to multiple environmental stresses, however, Mcs4 is required for Sty1 MAPK activation, which is in contrast to S. cerevisiae Ssk1, which responds to a limited number of stresses (118, 121). Another interesting point is that Mcs4 associates with the Wak1/Wis4 MAPKKK under both unstressed and stressed conditions as an integral component of the Sty1 MAPK pathway (26), which is distinguished from S. cerevisiae Ssk1, which interacts with the Ssk2 MAPKKK only under stressed conditions in the unphosphorylated state (73, 96, 99). S. pombe contains another response regulator, named Prr1, which is homologous to S. cerevisiae Skn7. Prr1 controls resistance to hydrogen peroxide in a Sty1-independent manner (26). However, Prr1 appears to be involved in some downstream actions of the Sty1 MAPK by controlling Sty1-dependent transcription factors such as Atf1 and Pap1 (26). Interestingly, Prr1 represses the sexual development of S. pombe under normal conditions (84).
The two classes of response regulator are widely conserved in pathogenic fungi. C. albicans has two response regulators, Ssk1 and Skn7. Unexpectedly, however, CaSsk1 cannot functionally complement ScSsk1 in osmosensing (29), indicating that the Ssk1 response regulator is functionally divergent between the two species. The major function of Ssk1 is to control the oxidative stress response, heat shock response, and morphological differentiation of C. albicans (29, 32). In addition, cell wall biogenesis is affected by the Ssk1 response regulator (32). Deletion of the SSK1 gene reduces the adherence of C. albicans to human cells by lowering the expression of ALS1 (agglutinin-like sequence 1) (32). The C. albicans Skn7 response regulator is involved not only in yeast-to-hypha transitions but also in responding to certain oxidative stress in an Ssk1- and Hog1-independent manner (123). Accordingly, Skn7 is necessary to confer full virulence on C. albicans, although it is not as important as Ssk1 (123).
C. neoformans also possesses two response regulators, Ssk1 and Skn7 (13). Deletion of the SSK1 gene generates phenotypes highly comparable to those of the hog1Δ mutant, including increased sensitivity to osmotic shock, toxic metabolites like methylglyoxal, UV irradiation, high temperature, and hydrogen peroxide; increased resistance to fludioxonil; enhanced mating capability; and increased production of melanin and capsule (13), strongly indicating that Ssk1 is the major upstream regulator of the Hog1 MAPK in C. neoformans. However, the findings that the hog1Δ mutant shows higher sensitivity to osmotic shock and high temperature than the ssk1Δ mutant and the Hog1 MAPK still can be phosphorylated by osmotic shock in the absence of Ssk1 suggest that another upstream signaling branch may exist (13). In contrast, the skn7Δ mutant exhibits phenotypes both redundant with respect to those of the hog1Δ mutant and distinct from them, including high sensitivity to Na+ salt and certain reactive oxygen species like t-butyl-hydrogen peroxide (but not to hydrogen peroxide), partial resistance to fludioxonil, cell flocculation, and increased melanin synthesis (13, 37, 143). Although some of the phenotypic features of the skn7Δ and hog1Δ mutants overlap, Skn7 is largely independent of the Hog1 MAPK signaling pathway for the following reasons. First, the SKN7 deletion does not change any phosphorylation levels of the Hog1 MAPK (13). Second, Skn7 is not involved in capsule synthesis and sexual differentiation, unlike Ssk1, Ssk2, Pbs2, and Hog1 (12-14). Third, the virulence of the skn7Δ mutant appears to be more severely attenuated than that of the hog1Δ mutant (37, 143).
Filamentous fungi also have two classes of response regulator in their genomes. A. nidulans contains Ssk1 and Skn7 homologues SskA and SrrA, respectively. Some ascomycetous filamentous fungi, including C. heterostrophus, and G. moniliformis, contain additional response regulators such as ChREC1 and GmREC1 (not orthologous to each other), whose functions are not known (30). SskA can functionally replace S. cerevisiae Ssk1, and disruption of sskA generates phenotypes comparable to those of the sakAΔ mutant and abolishes the phosphorylation of SakA in response to either high osmolarity or oxidative stress (48). The two response regulators contribute to the full resistance to osmolarity stress, although SskA plays more dominant roles (137). Instead, SrrA plays a major role in counteracting oxidative damage and cell wall stress than SskA (137). Both SrrA and SskA are essential for maintaining normal asexual formation and conidiospore viability (137). Although A. fumigatus has two putative response regulators homologous to SskA and SrrA, their functions have not been investigated until now. N. crassa contains two response regulators, RRG-1 and RRG-2, which are Ssk1 and Skn7 homologues, respectively. RRG-1 controls osmosensitivity and fungicide resistance in an Os-2-dependent manner (63). Furthermore, the rrg-1Δ mutant was shown to be impaired in producing female reproductive structures and conidial integrity (63), in contrast to the C. neoformans ssk1Δ mutant, which displays enhanced mating phenotypes (13).
In conclusion, diverse and multiple sensor histidine kinases in most yeasts and filamentous fungi, except S. cerevisiae, sense and respond to a variety of environmental stresses and subsequently relay signals to a single HPt protein and a limited number of response regulators. This is a unique situation found in fungi compared to the bacteria and plants that have a far greater number of HPt proteins and response regulators in their genomes than fungi do (30). Therefore, multiple stress signals sensed by distinct sensor histidine kinases could converge into the same HPt-response regulator pathway. However, how each signal coming from each hybrid sensor kinase is distinguished by the same downstream pathway remains elusive. Some sensor histidine kinases, if they have any functions, may employ an unknown signaling pathway even without involvement of HPt and response regulators. On the other hand, the downstream HPt-response regulator pathway could be activated by other receptors or sensors because the response regulator mutants often exhibit stronger phenotypes than multiple mutant forms of sensor kinases.
THE HOG1 MAPK MODULE
Ssk2-like MAPKKK.The regulatory mechanism of Ssk2 has been elegantly studied in S. cerevisiae (96, 132). The Ssk2 MAPKKK is composed of the N-terminal AID and the C-terminal KD (Fig. 1 and 3). Under normal conditions, the AID and KD of Ssk2 interact with each other, which prevents the KD of Pbs2 MAPKK from gaining access to the Ssk2 KD, although the N-terminal regulatory subdomain (RSD-I, amino acids 5 to 54) of Pbs2 constitutively binds to the Ssk2 KD (132). Upon osmotic stress, the Ssk1 response regulator becomes dephosphorylated and then interacts with the N-terminal AID of Ssk2, triggering autophosphorylation of the Thr residue in Ssk2, which is highly conserved throughout the fungal Ssk2-like MAPKKKs (Fig. 1 and 3). Interaction between Ssk1 and the Ssk2 AID enables the C-terminal KD of Ssk2 to bind the Pbs2 KD and ultimately activates the Ser and Thr residues of Pbs2 (Fig. 1). Although the Ssk2 AID spans residues 294 to 413 in S. cerevisiae (96), the region between residues 388 and 413 appears to be highly conserved between fungal Ssk2 MAPKKKs, implying that the 25-amino-acid region may be the core sequence required for binding to and receiving signals from the Ssk1 response regulator (Fig. 3A). In S. cerevisiae, another MAPKKK, named Ssk22, exists to play redundant roles in activating the Pbs2-Hog1 MAPK pathway because the ssk2Δ ssk22Δ double mutant exhibits phenotypes comparable to those of the pbs2Δ and hog1Δ mutants (73, 99).
Comparison of the functional subdomains of Ssk2-like MAPKKKs and Pbs2-like MAPKKs in fungi. The multiple-sequence alignment was generated by using the CLUSTAL W algorithm in the MacVector software. (A) Comparison of the subdomains of fungal Ssk2-like MAPKKKs. The AID (amino acids 294 to 413) and autophosphorylation site (T1460) are indicated based on information from S. cerevisiae Ssk2 (96, 132). (B) Comparison of the subdomains of fungal Pbs2-like MAPKKs. The subdomains of Pbs2-like MAPKKs, including the NES, the Ssk2 binding site, the Sho1 SH3 binding proline-rich motif, the nuclear localization sequence, the Hog1 docking site, and the phosphorylation site, are indicated as shaded regions based on information from S. cerevisiae Pbs2 (73, 132).
Similar to S. cerevisiae, S. pombe contains two MAPKKKs, Wak1 (Wis1-activating kinase 1, also known as Wik1 [Wis1 kinase 1] and Wis4) and Win1. Wak1 shows extensive structural similarity to S. cerevisiae Ssk2 in both the noncatalytic N-terminal region and the catalytic protein KD at the C terminus (118, 121). The wak1Δ mutant shows phenotypes similar to but less severe than those of the wis1Δ and sty1Δ mutants (118, 121). Deletion of the second MAPKKK gene, WIN1, in the wak1Δ mutant genetic background results in more severe phenotypes than each single MAPKKK mutant (115, 119), indicating that Wak1 and Win1 play redundant roles in activating the Wis1-Sty1 pathway, similar to S. cerevisiae Ssk2 and Ssk22. Interestingly, however, Wak1 and Win1 can be distinguished when responding to diverse stresses. In response to osmotic shock and heat shock, Win1, rather than Wak1, is mainly responsible for activation of the Wis-Sty1 pathway (115, 119). In addition, sexual differentiation under nitrogen starvation and viability in stationary phase are controlled by the Wis1-Sty1 pathway in a Win1-dependent and Wak1-independent manner (119). However, for sensing nitrogen starvation and heat shock and sexual conjugation by S. pombe, the Wis1-Sty1 signaling pathway can be activated by unknown signaling components or cascades, other than Wak1 and Win1 MAPKKKs, because the two MAPKKKs are partially dispensable or are not employed to sense the stresses.
In contrast to the model yeasts, most fungal pathogens appear to contain a single Ssk2-like MAPKKK. Two recent studies showed that a single MAPKKK is sufficient to control the Pbs2-Hog1 signaling pathway in C. albicans and C. neoformans (12, 33). In C. albicans, deletion of SSK2, but not STE11, generates phenotypes highly comparable to those of pbs2Δ and hog1Δ mutants in stress response and morphology and completely abolishes phosphorylation and nuclear accumulation of Hog1 in response to external stress (33). Notably, the ssk2Δ mutant exhibits constitutively filamentous morphology, like the hog1Δ mutant, with increased expression of hypha-specific genes such as HWP1 and ECE1, while the ssk1Δ mutant is almost completely defective in hyphal formation (33). This indicates that Ssk2 is negatively regulated by either Ssk1 or other, unknown, signaling components for controlling the morphological differentiation of C. albicans. There is no clear evidence for a direct interaction between Ssk2 and Ssk1 in C. albicans.
A recent study with C. neoformans identified a single Ssk2 MAPKKK via the comparative analysis of meiotic maps between serotype D F1 sibling strains, B-3501 and B-3502, exhibiting differential Hog1 phosphorylation patterns (12). As described below, diverse clinical and environmental isolates of C. neoformans exhibit different patterns of Hog1 phosphorylation, and Hog1 MAPKs in a number of C. neoformans strains are constitutively phosphorylated under normal conditions, which is an unusual phenomenon compared to other fungi (14). Interestingly, C. neoformans strains having higher phosphorylation levels of Hog1 under unstressed conditions show higher stress resistance and virulence. In this process, Ssk2 turns out to be a key interface kinase that determines constitutive phosphorylation levels of Hog1 and stress resistance of C. neoformans (12). The ssk2Δ mutant displays phenotypes highly comparable to those of the ssk1Δ, pbs2Δ, and hog1Δ mutants, including stress resistance, drug resistance, sexual reproduction, and capsule and glucose biosynthesis (12). In addition, deletion of the SSK2 gene completely removes the constitutive phosphorylation of the Hog1 MAPK (12), indicating that Ssk2 is the only MAPKKK for activation of the C. neoformans Hog1 MAPK.
Similar to C. albicans and C. neoformans, the filamentous fungus A. nidulans contains only a single MAPKKK, named SskB (48). SakA is not phosphorylated in the sskBΔ mutant, which exhibits phenotypes highly comparable to those of the pbsBΔ and sakAΔ mutants (48). N. crassa also possesses a single MAPKKK, called Os-4, which acts upstream of the Os-5 MAPKK and Os-2 MAPK to control osmosensitivity, the heat shock response, and resistance to the antifungal fludioxonil (90). Taken together, the evidence shows that the presence of a single MAPKKK in the HOG pathway appears to be more common in pathogenic fungi, unlike the model yeasts.
Pbs2-like MAPKK.The Pbs2-like MAPKK is the only Ser/Thr protein kinase that is necessary and sufficient for the phosphorylation and activation of the Hog1 MAPK in most fungi. Pbs2 is a major scaffolding protein for the HOG signaling pathway in S. cerevisiae and interacts with multiple signaling components, besides the Ssk2 MAPKKK, containing a long N-terminal nonkinase region with several functional domains (Fig. 3B). In S. cerevisiae, there is an alternative pathway for the activation of Hog1 named the Sho1-dependent pathway (73, 97). Sho1 is a cytoplasmic membrane-localized protein which is less sensitive in osmosensing than the Sln1 hybrid sensor kinase (73, 97, 98). In response to osmotic shock, Sho1 activates the Ste20 PAK-like kinase via Cdc42 GTPase, which subsequently activates the Ste11 MAPKKK with a requirement for Ste50 (98, 104, 109, 133). Ste11 can then activate the Pbs2-Hog1 signaling pathway. Here, Pbs2 can physically interact with Sho1, Ste20, Ste50, and Ssk2 to provide a scaffold in the osmosensing pathway (73, 96, 97, 104, 133). Pbs2 contains a proline-rich domain for binding to the Sho1 SH3 domain, the docking domains for Ssk2/Ssk22 and Hog1, the Ser/Thr KD, the nuclear export signal (NES), and the nuclear localization sequence (Fig. 3B).
Except for the Hog1 docking site and the Ser/Thr KD, however, other subdomains do not appear to be evolutionarily conserved between fungal Pbs2-like MAPKKs, indicating that the regulatory mechanism of Pbs2 may be divergent in fungi (Fig. 3B). For example, the proline-rich motif of S. cerevisiae Pbs2 is not widely conserved in other fungi. In agreement with this finding, most of the fungal pathogens studied thus far either do not contain the Sho1-like osmosensor or contain Sho1, which indeed plays Hog1-independent roles. In C. neoformans, none of the Sho1-like membrane protein has been discovered in its genome (15). The findings that deletion of SSK2 completely abolishes Hog1 phosphorylation and phenotypes of the ssk2Δ mutant are almost identical to those of the hog1Δ mutant in both C. neoformans and C. albicans further suggest that the Ssk1-Ssk2-dependent pathway is the major signaling branch for activation of Hog1 in the human fungal pathogens studied to date (12, 33).
S. pombe contains a Pbs2 homolog named Wis1, the deletion of which generates phenotypes almost identical to those of the sty1Δ mutant (118, 139), suggesting that Wis1 operates in the same pathway with the Sty1 MAPK. The Wis1 MAPKK appears to be activated by bifurcated upstream signaling branches, depending on the type of stress. One is the Wak1 and Win1 MAPKKKs associated with the Mak1/2/3-Mpr1-Mcs4 phosphorelay system, and the other is an unknown signaling cascade(s) that acts directly on the Wis1 MAPKK, as witnessed by the finding that sty1Δ and wis1Δ mutants show hypersensitivity to high temperature and osmotic shock whereas the mcs4Δ, win1Δ, and wak1Δ mutants exhibit less severe defects (118). The latter cascade includes the Pyp1 and Pyp2 tyrosine phosphatases, the inhibition of which strongly activates the Wis1-Sty1 pathway (87).
The functions of Pbs2 are highly conserved in most pathogenic and nonpathogenic fungi. C. albicans contains a structural and functional Pbs2 homologue named Pbs2, which physically interacts with and activates Hog1 by phosphorylation (10, 33). The pbs2Δ mutant exhibits phenotypes highly comparable to those of the hog1Δ mutant, and the additional deletion of HOG1 in the pbs2Δ mutant does not confer any additive effects. In addition, Hog1 phosphorylation is completely absent in the pbs2Δ mutant, further indicating that Hog1 is the only target MAPK for Pbs2 in C. albicans (10). Notably, the N-terminal domain of Pbs2 containing the putative Sho1-binding proline-rich domain and NES is not essential for the functionality of Pbs2 (10), suggesting that the Sho1 signaling pathway is not required for the Hog1 signaling cascade in C. albicans.
C. neoformans also contains a Pbs2-like MAPKK (14). Deletion of the PBS2 gene totally abolishes Hog1 phosphorylation, and therefore the phenotypes of the pbs2Δ mutant are highly comparable to those of the hog1Δ mutant (14). However, certain phenotypes, such as sensitivity to hydrogen peroxide, are stronger in the pbs2Δ mutant than in the hog1Δ mutant, indicating that another Pbs2 target protein, other than Hog1, could exist in C. neoformans (14). In agreement with this finding, the virulence defect is more accentuated in the pbs2Δ mutant than in the hog1Δ mutant in a systemic cryptococcosis model (14).
Similar to C. albicans and C. neoformans Pbs2, A. nidulans PbsB also does not harbor both a proline-rich SH3 binding domain and an SBD at its N terminus (48). Despite the structural difference between yeast Pbs2 and PbsB, PbsB can complement the yeast pbs2Δ mutant in an Ssk2/Ssk22-dependent manner (48), indicating that PbsB might have a novel Ssk2-binding site. The pbsBΔ mutant shows phenotypes similar to those of the sskBΔ and sakAΔ mutants and does not phosphorylate SakA in response to osmotic and oxidative stresses (48), indicating that PbsB is the bona fide MAPKK for SakA/HogA. Interestingly, the A. nidulans hogAΔ (sakAΔ) mutant shows less severe sensitivity to osmotic shock than the sskAΔ, sskBΔ, and pbsBΔ mutants, indicating that Pbs2B appears to have target MAPKs other than HogA (48), which is similar to the case in C. neoformans. A. nidulans indeed contains another Hog1 orthologue, named MpkC (48). Although the mpkCΔ mutant does not exhibit any discernible phenotypes, overexpression of mpkC suppresses the osmosensitivity of the hogAΔ mutant, implying that MpkC may be another target for PbsB (48). N. crassa has a Pbs2-like MAPKK, named Os-5, which is necessary and sufficient for phosphorylation of the Os-2 MAPK (90).
Hog1-like MAPK.The HOG1 gene was first identified in S. cerevisiae by random screening of osmoregulation-defective mutants (22). In this screening, the PBS2 gene (originally named HOG4) was also isolated. Hog1 is a Ser/Thr and Tyr dual-protein kinase with the evolutionarily conserved TGY motif. Subsequently, the p38 MAPK was identified as a mammalian counterpart of yeast Hog1. Expression of p38 is strongly induced in monocytes and macrophages in response to the endotoxin lipopolysaccharide (53). Heterologous expression of mammalian p38 MAPK rescues normal osmoresistance of the S. cerevisiae hog1Δ mutant (53), indicating that the p38/Hog1 MAPK is functionally conserved from yeasts to humans. Phosphorylation of both Thr and Tyr residues is required for full activation of Hog1. Therefore, dephosphorylation of either the Thr or the Tyr residue by protein tyrosine phosphatase (PTP) or protein phosphatase type 2C (PP2C) leads to inactivation of Hog1 (113). Upon dual phosphorylation, yeast Hog1 is translocated into the nucleus to induce the expression of target genes, depending on the activity of the small G protein Gsp1 and an importin β homologue, Nmd5 (46, 108). Upon either removal of or adaptation to stress, Hog1 is exported out of the nucleus. Notably, Hog1 kinase activity is required for nuclear export, but not for nuclear import (46, 108). However, whether expression of Hog1 target genes is induced by a phosphorylated or dephosphorylated form of Hog1 in the nucleus remains elusive. Indeed, there is a report showing that induction of Hog1 target genes coincides with Hog1 dephosphorylation during the transition from high osmolarity to lower osmolarity (136).
Although the Hog1 MAPK is required for long-term adaptation to high osmolarity through transcriptional activation of the stress defense genes, it is directly involved in an immediate relief of osmotic stress in a spatially and temporally distinct manner (102). In response to hyperosmotic stress, most DNA binding proteins, except histones and elongating RNA polymerase II, are rapidly dissociated from chromatin due to the increased ion concentrations in the nucleus. To immediately counteract this situation, Hog1 is directly recruited to and interacts with the Nha1 Na+/H+ antiporter and the Tok1 potassium channel to achieve an intracellular ion concentration sufficient for resuming the association of appropriate transcription factors for induction of stress defense genes. After the immediate early response to osmotic shock, Hog1 is involved in the transcriptional induction of a number of stress defense genes, including the Ena1 Na+ extrusion pump, for long-term adaptation (102). Although the phenomenon is reported in S. cerevisiae, whether the same event occurs in other fungi has not been examined. Furthermore, whether Hog1 is involved in a similar type of early immediate response to other stresses remains to be further investigated.
During osmostress, the cell cycle of S. cerevisiae is arrested at different checkpoints, G1 and G2, in a Hog1-dependent manner (5, 18, 36, 45). For cell cycle arrest at G1 during osmostress, Hog1 is directly targeted to and stabilizes the cyclin-dependent kinase inhibitor Sic1 (45, 149). For cell cycle arrest at G2 during osmostress, Hog1 phosphorylates the Hsl1 checkpoint kinase, which stabilizes the Swe1 kinase and downregulates Clb-associated Cdc28 activity (5, 36). Therefore, activated Hog1 prevents premature entry into mitosis before cells adapt to hyperosmotic conditions.
The S. pombe Hog1 homologue Sty1 (also known as Spc1 and Phh1) responds to more diverse environmental stimuli, including osmosis change, heat shock, oxidative challenges, carbon starvation, and UV irradiation, than S. cerevisiae Hog1 and is also involved in growth control and sexual differentiation (64, 78, 120). In contrast to S. cerevisiae Hog1, Sty1 promotes the onset of mitosis upon osmostress (120). The S. pombe sty1Δ mutant is delayed in the onset of mitosis (G2-mitosis transition), exhibits a longer cell shape, and is defective in G1 arrest during mating upon nitrogen starvation (120). The delayed G2 arrest observed in the sty1Δ mutant is even more aggravated under hyperosmotic conditions (120).
Among the signaling components of the HOG pathway of fungal pathogens, the Hog1 MAPK is the best characterized. The Hog1 MAPK of fungal pathogens was first characterized in C. albicans. C. albicans Hog1 was identified by its ability to functionally complement the osmosensitive phenotype of the S. cerevisiae hog1Δ mutant (116). The homozygous hog1Δ null mutant exhibits hypersensitivity to osmotic shock due to a lack of the ability to accumulate intracellular glycerol in response to high external osmosis (116). Furthermore, the C. albicans hog1Δ mutant is impaired in the oxidative stress response by displaying hypersensitivity to hydrogen peroxide and menadione (a superoxide anion generator) and in adaptation to UV light (7). C. albicans Hog1 appears to contribute to regulation of the oxidative stress response in a manner independent of the Cap1 bZIP transcription factor, an S. cerevisiae Yap1 homologue, because CAP1 mutation in the hog1Δ mutant background provides additional sensitivity to oxidative stress (7). Besides regulating stress responses, Hog1 plays an important role in controlling diverse morphological differentiation of C. albicans in several ways. First, Hog1 is required for maintaining normal cell morphology and a normal cell cycle under hyperosmotic conditions. The hog1Δ mutant is not able to complete the last stage of cytokinesis (cell wall separation) and exhibits defective budding patterns under osmostress conditions (6). Second, Hog1 represses the filamentous growth of C. albicans (6). The hog1Δ mutant shows hyperfilamentous phenotypes with concomitant upregulation of hypha-specific genes (6, 44). The hyperfilamentous morphology of the C. albicans hog1Δ mutant is thus similar to the elongated cell morphology of the S. pombe sty1Δ mutant. Third, chlamydospore formation is abolished in the hog1Δ mutant (7, 43). Chlamydospores are thick-walled spores formed in response to microaerophilic, low-temperature, and no-light environments under certain nutritional conditions such as cornmeal agar. As a result of some or all of these phenotypes, the virulence of the homozygous hog1Δ null mutant is significantly attenuated compared to that of the wild-type strain in a systemic candidiasis model (6).
Hog1 is involved in even more diverse cellular processes in C. neoformans. Deletion of HOG results in many phenotypic outcomes, including hypersensitivity to osmotic shock, UV irradiation, oxidative stress, high temperature, and the toxic metabolite methylglyoxal; resistance to fludioxonil; increased production of melanin and capsule; and enhanced pheromone production during mating (14). Most interestingly, however, Hog1 was found to be differentially regulated in diverse clinical and environmental isolates (14, 66). In some C. neoformans strains, including serotype D JEC21, which was a platform strain for serotype D genome sequencing, Hog1 is rarely phosphorylated under normal conditions and becomes phosphorylated upon a stress response, similar to Hog1 homologues in other fungi (14). In contrast, a major portion of C. neoformans strains, such as H99 and B-3501, which are platform strains for serotype A and serotype D genome sequencing, respectively, under normal conditions contain constitutively phosphorylated Hog1, which becomes dephosphorylated upon stresses such as osmotic shock (14). Remarkably, dephosphorylation of Hog1 upon stress requires Hog1 kinase activity because the kinase-dead Hog1 mutant is incapable of dephosphorylating and activating Hog1 (14). This unusual, constitutive Hog1 phosphorylation pattern may confer additional phenotypes on the serotype A H99 strain over the serotype D JEC21 strain, such as increased stress resistance, elevated mating by overproduction of pheromone, and enhanced melanin and capsule biosynthesis (14). This kind of Hog1 phosphorylation pattern has never been observed in organisms other than C. neoformans. The differential Hog1 phosphorylation results not from structural differences in Hog1 MAPKs but from differential upstream signaling components, including Ssk1 MAPKKK (12, 14). When SSK2 genes are exchanged between the B-3501 and JEC21 strains, Hog1 in strain JEC21 becomes constitutively phosphorylated, which results in the increased stress resistance and virulence of strain JEC21 (12).
It remains elusive what selective advantages C. neoformans has with a constitutively phosphorylated Hog1 during evolution and development of virulence attributes. One possible explanation is that C. neoformans Hog1 is preferentially recruited into the nucleus by constitutive phosphorylation under normal conditions and cross talks with other signaling pathways, as can be seen in pleiotropic effects of HOG1 disruption. As mentioned before, if the dephosphorylated form of Hog1 is responsible for induction of target genes, it could be advantageous for cells to have already phosphorylated and nucleus-concentrated Hog1 to confer an immediate response to stress.
Hog1-like MAPKs are also highly conserved in filamentous fungi. A. nidulans has SakA (stress-activated kinase A, also known as HogA), which is structurally and functionally homologous to yeast Hog1 (54, 65). Similar to Hog1 MAPKs in other fungi, except C. neoformans Hog1, SakA is activated by phosphorylation in response to osmotic and oxidative shock (65). Unlike Hog1 in pathogenic and nonpathogenic yeasts, however, A. nidulans SakA/HogA is transcriptionally regulated in response to salt shock, along with other genes in the HOG pathway, including pbsA, ptpA, and msnA (54). In agreement with this finding, the sakAΔ mutant exhibits growth retardation in a salt- and heat-dependent fashion, with abnormal morphologies such as defective septum formation (54). Notably, and in this case similar to the C. neoformans hog1Δ mutant, the sakAΔ mutant also shows enhanced, premature sexual development in an SteA-dependent fashion, although the phosphorylation and expression levels of SakA are unaltered during sexual differentiation (65). Although the sakAΔ mutant does not show any morphological defects during asexual development, SakA is required for maintaining conidiospore viability against heat shock and oxidative stress (65). N. crassa has a Hog1 homologue, called Os-2, which promotes resistance to osmotic and heat shock in an Os-4 MAPKKK- and Os-5 MAPKK-dependent manner (90, 151). Os-2 also confers sensitivity to the phenylpyrrole class of fungicides, such as fludioxonil and fenpiclonil (151). The phenylpyrrole fungicides were shown to trigger the HOG pathway to overaccumulate intracellular glycerols, affecting normal conidial and hyphal growth and eventually causing cells to swell and burst (151).
Similar phenomena have also been observed in other fungi, including C. albicans and C. neoformans, but not in all fungi, such as S. cerevisiae, for the phenylpyrrole chemicals (61, 66, 91). This could be explained by the fact that S. cerevisiae contains only a single hybrid sensor kinase that cannot recognize the fungicides. It appears that the class III sensor kinases containing HAMP domains (Fig. 2) are mainly responsible for sensing the phenylpyrrole drugs (13, 91, 147). Indeed, heterologous expression of the C. neoformans hybrid sensor kinase Tco1 renders S. cerevisiae sensitive to fludioxonil (13). However, other mechanisms for sensing and responding to this drug should exist, for the following reasons. First, the C. neoformans Tco2 hybrid sensor kinase without any HAMP domains also promotes sensitivity to fludioxonil (13). Second, a C. neoformans strain containing both Tco1 and Tco2, such as the JEC21 strain, is completely resistant to fludioxonil, like S. cerevisiae (66). Indeed, constitutive phosphorylation levels of Hog1 were shown to be tightly correlated with fludioxonil sensitivity in C. neoformans (12). For example, C. neoformans strains having higher constitutive Hog1 phosphorylation levels, like the H99 strain, exhibit sensitivity to fludioxonil, whereas strains having lower or no Hog1 phosphorylation, like the JEC21 strain, are resistant to fludioxonil (12, 66). Taken together, the evidence shows that the detailed mechanism and domains for fludioxonil sensing are not clear yet.
DOWNSTREAM REGULATORS OF THE HOG PATHWAY
Downstream of the HOG pathway, two types of signaling component relay stress signals and ultimately induce the expression of stress defense genes to counteract and adapt to external stimuli. One is another protein kinase, named MAPKAPK, and the other transcription factors. Notably, Hog1 itself is actively involved in the expression of target genes by associating with some transcription factors and is recruited to their target promoter sites.
MAPKAPK.In mammals, the p38 MAPK activates either transcription factors or MAPKAPK to induce the expression of target genes. MAPKAPK2 is the p38-regulated kinase which subsequently phosphorylates both CREB and ATF1, which are targeted to protein kinase A (PKA) (130). In S. cerevisiae, Rck1 and Rck2 were identified as Hog1-regulated MAPKAPKs. Originally, the S. cerevisiae RCK1 and RCK2 genes were isolated as suppressors of S. pombe checkpoint mutants (38) and found to inhibit meiosis because both rck1Δ and rck2Δ mutants exhibit premature entry into the meiotic process (105). Rck2 is the direct phosphorylation target of Hog1 following osmotic shock, and overexpression of RCK2 can rescue the osmosensitivity of hog1Δ and pbs2 mutants (21, 134). Interestingly, however, disruption of either RCK1 or RCK2 does not result in hyperosmosensitivity but causes hypersensitivity to oxidative stresses (20), indicating that the two MAPKAPKs play a key role in activating a subset of Hog1-responsive genes. Similarly, S. pombe contains two Rck1 and Rck2 homologues, Srk1 (Sty1-regulated kinase 1, also known as Mkp1) and Mkp2, respectively. Both Srk1 and Mkp2 physically interact with Sty1, although Srk1-Sty1 binding is stronger than Mkp2-Sty1 binding (11, 125). In agreement with S. cerevisiae Rck1, Srk1 inhibits meiotic processes of S. pombe, but Mkp2 does not play any role in meiosis (11).
Only a limited amount of information about the role of MAPKAPKs in the HOG pathway in fungal pathogens is available. Both C. albicans and C. neoformans have Rck2 homologues in their genomes. Interestingly, a recent microarray analysis revealed that induction of the C. albicans Rck2 homologue is significantly reduced in the hog1Δ mutant upon osmotic stress, indicating that Rck2 is transcriptionally regulated by Hog1 (44). Our recent transcriptome analysis of the HOG pathway in C. neoformans also showed that the Rck2 homologue is transcriptionally regulated, dependent on Hog1 and Ssk1 (unpublished data). Similarly, S. pombe Srk1 was also shown to be transcriptionally regulated by Sty1 (125). Taken together, the evidence shows that Rck1/2-like MAPKAPKs play a key role, at least partially, in relaying signals from the Hog1 MAPK.
Transcription factors.As downstream regulators of the HOG pathway, transcription factors have been more intensively characterized than MAPKAPKs in the model yeasts. Regardless of the presence of a single hybrid sensor kinase, S. cerevisiae contains multiple Hog1-dependent transcription factors, including Msn2/4, Sko1, Hot1, Sko1, Smp1, and Sgd1 (Table 1). Msn2 and Msn4 are two redundant Zn2+ finger transcription factors that mediate a general stress response through binding to the stress response element (75). Although the nuclear localization of Msn2/4 is negatively regulated by the cAMP/PKA signaling pathway (55, 124), expression of Msn2/4-dependent osmoresponsive genes requires Hog1 (110). The expression of MSN2 itself is regulated by Sko1, which is another Hog1-dependent ATF/CREB transcription factor (also known as Acr1) (101). Upon osmotic stress, Hog1 converts Sko1, which is a transcriptional repressor, into a transcription activator in association with the Tup1 repressor (101). Interestingly, the Hog1-Sko1-Tup1 transcription activator complex recruits SAGA histone acetylase and SWI/SNF nucleosome-remodeling complexes to activate osmoresponsive genes, thereby regulating a global change in gene expression (101). However, the Msn2/4 and Sko1 transcription factors are not sufficient to express all of the Hog1-dependent osmoresponsive genes. The major portion of the Msn2/4-independent osmoresponsive genes was found to be regulated by the Hot1 and Msn1 transcription factors (111). The Hot1 transcription factor recruits Hog1 to express a subset of Hog1-responsive genes that are involved in glycerol metabolism (2). Another Hog1-dependent transcription factor is Sgd1, which contains a leucine zipper region and is essential for cell viability (1). Overexpression of SGD1 suppresses the osmosensitivity observed in the hog1Δ and pbs2Δ mutants by restoring the expression of the glycerol phosphate dehydrogenase gene GPD1 (1).
In S. pombe, three transcription factors, Atf1, Pap1, and Prr1, are mainly regulated by the Sty1 MAPK. Atf1 is a bZIP transcription factor that is highly homologous to mammalian ATF2 (141). Atf1 forms a heterodimer with another bZIP transcription factor, Pcr1, and is a nucleus-localized protein, anchoring Sty1 in the nucleus (50, 67). Atf1 mediates most of the stress-induced phenotypes regulated by the Wis1-Sty1 signaling pathway, except for oxidative stress (41, 82, 141). To mediate the oxidative stress response, Sty1 mainly utilizes Pap1, which is closely related to the mammalian c-Jun and S. cerevisiae Yap1 transcription factors (82, 135). The nuclear localization of Pap1 in response to oxidative stress is indeed dependent on Sty1 (135).
Downstream transcription factors of the HOG pathway are not well characterized in fungal pathogens relative to upstream signaling components. A search for homologues of known Hog1-dependent transcription factors of the model yeasts reveals that only the Sgd1 transcription factor is widely conserved in human fungal pathogens (Table 1). However, the function of Sgd1 in fungal pathogens is completely unknown. Although C. albicans has a Yap1/Pap1 homolog, Cap1, it is not dependent upon Hog1 for its activation (7, 44). Interestingly, C. albicans contains Msn2- and Msn4-like transcription factors Mnl1 and CaMsn4, respectively, and yet they are not involved in diverse stress responses, unlike S. cerevisiae Msn2/4 (89). However, it has been recently reported that Mnl1 is required for long-term adaptation of C. albicans to weak acid stress (106). The only notable conservation between C. albicans and S. cerevisiae is the function of the Sko1 transcription factor. Sko1 is transcriptionally regulated by Hog1 upon osmotic stress, although it has a distinct role in the cell wall damage response (44, 107).
Both A. fumigatus and C. neoformans lack obvious Msn1, Msn2/4, Sko1, and Hot1 homologues in their genomes. Instead, the Atf1 transcription factor is conserved in the two fungal species, but not in C. albicans. Recently, the function of Atf1 has been characterized in Aspergillus oryzae (114). The atfBΔ mutant exhibits higher stress sensitivity than the wild-type strain (114), indicating that the function of the Atf1 transcription factor may be conserved in filamentous fungi. The Atf1 transcription factor in C. neoformans has also been characterized. Unlike S. pombe Atf1, C. neoformans Atf1 is required for the oxidative stress response by modulating the expression of thioredoxins (79). However, whether the oxidative stress response controlled by the HOG pathway is mediated through Atf1 is unclear. Our recent transcriptome analysis indicates that Atf1 is transcriptionally regulated by Hog1 in C. neoformans (unpublished data).
FEEDBACK REGULATION OF THE HOG PATHWAY
Upon either adaptation to or removal of external stress, cells resume normal cellular conditions by inactivating the HOG pathway. Two negative feedback regulatory systems have been reported. One is the protein phosphatase-mediated system, and the other is the ubiquitin-proteasome system. Protein phosphatases dephosphorylate the phosphorylated signaling components, including Pbs2 and Hog1, to shut down the HOG pathway. The ubiquitin-proteasome system degrades the signaling components in the two-component system (117). The negative feedback regulatory mechanisms of the HOG pathway have been well characterized in the model budding and fission yeasts. In contrast, the roles of phosphatases and ubiquitin-proteasome in the HOG pathway of fungal pathogens are largely unknown. Particularly, it should be noted that an external stimulus often causes inactivation of protein phosphatases, which results in activation of the HOG pathway by bypassing of the two-component system, indicating that protein phosphatases are involved in both positive and negative regulation of the HOG pathway.
Protein phosphatases.There are three classes of protein phosphatases based on the target phosphorylated amino acid residues of protein kinases. The first class includes PTPs that are capable of dephosphorylating only phosphotyrosine. S. cerevisiae has three PTPs, Ptp1 to Ptp3, while S. pombe contains two PTPs, Pyp1 and Pyp2. The second class includes PP2Cs that dephosphorylate phosphothreonine and phosphoserine. S. cerevisiae has five enzymatically active PP2Cs, Ptc1 to Ptc5, while S. pombe contains three PP2Cs, Ptc1 to Ptc3. The final class includes dual-specificity phosphatases (DSPs), which are able to dephosphorylate both phosphotyrosine and phosphothreonine. S. cerevisiae has two DSPs, Msg5 and Sdp1. Among these, only PTPs and PP2Cs, but not DSPs, are involved in the regulation of the HOG pathway.
Among the PTPs, Ptp2 and Ptp3 were found to be important in negatively regulating Hog1 in S. cerevisiae and Ptp2 plays a major role (62, 144, 150). Overexpression of either PTP2 or PTP3 suppresses the lethality of cells caused by hyperactivation of Hog1 (144, 150). In S. pombe, Pyp1 plays a major role not only in inactivating Sty1 MAPK as feedback regulation but also in activating Sty1 in response to heat treatment (88).
Among the PP2Cs, Ptc1 inactivates the HOG pathway by dephosphorylating the phosphothreonine residue of Hog1 (140). Although Ptc2 and Ptc3 inhibit Hog1 similarly, they appear to act differently from Ptc1 in two respects (148). First, Ptc1 is required for maintaining basal Hog1 activity and dephosphorylating Hog1 during adaptation to stress, whereas Ptc2 and Ptc3 limit the maximum levels of Hog1 activity (140, 148). Second, Ptc1 only indirectly interacts with Hog1 by using an adaptor protein, Nbp2, while Ptc2 and Ptc3 do not interact with Nbp2 (74).
In contrast to the well-studied protein phosphatases in the model yeasts, none of phosphatases have been characterized in detail in the human fungal pathogens. C. albicans contains Ptp2/3 and Ptc1/2 homologues, while both C. neoformans and A. fumigatus contain only Ptp2 (no Ptp3) and Ptc1/2 homologues (Table 1). In conclusion, the widely conserved Ptp2, Ptc1, and Ptc3 homologues may play a role in regulating the HOG pathway of fungal pathogens, but their functions remain to be elucidated.
The ubiquitin-proteasome system.The ubiquitin-proteasome system in negative feedback regulation of the two-component phosphorelay system has been investigated only in S. cerevisiae. Upon adaptation to stress, the unphosphorylated and activated Ssk1 response regulator is ubiquitinated by the Ubc7 E2 protein and an unknown ubiquitin ligase (E3) on the membrane of the endoplasmic reticulum. Ubiquitinated Ssk1 is then degraded by the 26S proteasome, which prevents Ssk1 from activating Ssk2 MAPKKK (117). However, the Ubc7 E2 binding domain of the Ssk1 response regulator has not been determined.
The ubiquitin-proteasome system has not been analyzed in fungal pathogens. Genome database searches have revealed that the Ubc7 E2 protein is highly conserved in fungal pathogens since C. neoformans and C. albicans contain Ubc7 homologues (CNAG_02238 and orf19.7329, respectively). Ubc7 appears to be also highly conserved in filamentous fungi, including Afu5g04060 in A. fumigatus, AN8258 in A. nidulans, and NCU03623 in N. crassa. Therefore, the function of the ubiquitin-proteasome system may be conserved in fungal pathogens.
CROSS TALK WITH OTHER SIGNALING PATHWAYS
To confer an appropriate cellular response to a given stimulus and maintain signaling specificity, the corresponding signal transduction pathway should be properly insulated by preventing cross talk with other, unrelated, signaling pathways. The HOG pathway plays a key role in preventing cross talk with other signaling pathways. In S. cerevisiae, the hog1Δ mutation leads to activation of the pheromone response Fus3/Kss1 MAPK pathway through the Sho1 branch of the HOG pathway upon exposure to osmotic stress (93). Furthermore, the hog1Δ mutation results in activation of the filamentation/invasion Kss1 MAPK pathway, causing the formation of long cellular projections (39). The C. neoformans HOG pathway plays a similar role in preventing cross talk with other signaling pathways. Interestingly, however, the C. neoformans HOG pathway appears to interact with and prevent abnormal activation of other signaling pathways under normal conditions. The pheromone response MAPK is hyperactivated with increased production of pheromone by mutation of the HOG1 gene, even without any osmotic stress (14). Furthermore, the defective melanin production observed in cells having mutations in genes of the cAMP signaling pathway, such as GPA1 (Gα subunit), CAC1 (adenylyl cyclase), and PKA1 (PKA), can be completely restored to normal by the hog1Δ mutation, indicating that the cAMP signaling pathway is repressed by the HOG pathway (14). These phenotypes are found only in some C. neoformans strains having higher levels of constitutive phosphorylation of Hog1, such as serotype A strain H99 (14), indicating that constitutively phosphorylated Hog1 is important for repression of the pheromone response MAPK and cAMP signaling pathways.
Cross talk of the HOG pathway with other signaling pathways is also evident in C. albicans. The Cst20-Ste11-Hst7-Cek1 MAPK pathway, which is homologous to the S. cerevisiae Ste20-Ste11-Ste7-Fus3/Kss1 MAPK pathway and is involved in filamentous growth and cell wall biogenesis of C. albicans (10, 43, 112), is hyperactivated by mutation of the HOG1 gene (10, 43). The findings that involvement of the Cek1 MAPK pathway in cell wall biogenesis is partly mediated by Hog1 and chlamydospore defects observed in the hog1Δ mutant can be suppressed by inhibition of the Cek1 MAPK pathway further imply that the two signaling pathways are interconnected (43). Interestingly, however, hyperfilamentous phenotypes observed in the hog1Δ mutant are independent of the Cek1 MAPK and Efg1 signaling pathways, which are two major signaling pathways that control the filamentation of C. albicans (43), indicating that the HOG pathway interacts with another signaling pathway(s) to repress the filamentation of C. albicans.
Cross talk between the HOG pathway and the Ca2+/calcineurin signaling pathway has also been reported. In response to calcium treatment, S. cerevisiae Hog1 is transiently phosphorylated, but its phosphorylation levels are dramatically increased and sustained by inhibition of the Ca2+/calcineurin-Crz1 pathway (e.g., the calcineurin cnb1Δ mutant) (122). However, this sustained Hog1 phosphorylation is not observed in the ssk1Δ cnb1Δ double mutant, indicating that the calcineurin pathway negatively regulates the HOG pathway via inhibition of the Sln1-Ypd1-Ssk1 two-component system (122). In C. neoformans, it has been reported that the antifungal activity of fludioxonil via hyperactivation of the HOG pathway is synergistically enhanced by inhibition of the calcineurin pathway, including the calcineurin mutation (cna1Δ and cnb1d mutants) or FK506 (a calcineurin inhibitor) treatment (66). However, Hog1 phosphorylation patterns in the cna1Δ mutant are equivalent to those in the wild-type strain (66), suggesting that the calcineurin pathway either acts independently of the HOG pathway or interacts with the downstream regulators of the HOG pathway.
Cross talk of the HOG pathway with the PKC/Mpk1 cell wall integrity pathway has been suggested. In S. cerevisiae, the Hog1 MAPK and PKC/Mpk1 MAPK pathways are oppositely regulated upon osmostress (58, 59). In response to hypotonic shock, the Mpk1 MAPK is activated by phosphorylation in a Pkc1-dependent manner, whereas the Hog1 MAPK is activated by hypertonic shock (40). However, several lines of evidence suggest that the two MAPK pathways are highly coordinated to maintain cell wall integrity in response to osmotic changes. First, the dephosphorylation of the Hog1 MAPK in response to hypoosmotic shock is dependent on the PKC/Mpk1 pathway (40). Second, the Mpk1 MAPK is transcriptionally induced upon exposure to high osmolarity in a Hog1-dependent manner (52). The Mpk1 MAPK is also involved in maintaining the cell wall integrity of C. neoformans and C. albicans (a Mkc1 MAPK) (68, 85, 86). In C. albicans, Mkc1 is phosphorylated upon exposure to oxidative stress (hydrogen peroxide) in a Hog1-dependent manner (10). However, the correlation between the Mpk1 and Hog1 MAPK pathways in C. neoformans has not been addressed. Taken together, the evidence shows that the HOG pathway plays a key role in interconnecting diverse signaling pathways in the complex fungal signaling network.
CONCLUSIONS
The HOG signaling pathway, in connection with the two-component phosphorelay system, is evolutionarily conserved in fungal pathogens and yet exhibits distinct divergence from the model budding and fission yeasts. Mycologists often adopt the HOG pathway of S. cerevisiae as a major model system to explain and compare the HOG pathway of a fungal pathogen. However, the Sty1 MAPK pathway in S. pombe appears to be a more appropriate model for fungal pathogens than S. cerevisiae, for the following reasons. First, S. pombe harbors multiple hybrid sensor kinases with diverse subdomains, except histidine kinase and receiver domains, like pathogenic fungi. Second, S. pombe not only responds to multiple external stimuli, including osmotic shock, oxidative stress, heat shock, and UV irradiation, but also controls the differentiation and morphology of cells. Similar pleiotropic roles of the HOG pathway have also been observed in fungal pathogens. Third, the Wak1-Wis1-Sty1 pathway is mainly controlled by the two-component system in S. pombe, as in pathogenic fungi, whereas the HOG pathway in S. cerevisiae is controlled by multiple upstream branches, including the Sho1 osmosensing pathway.
Regardless of the similarities between fungal pathogens and fission yeast, pathogenic fungi also have developed distinct and divergent features in the function and regulation of the HOG pathway. Some pathogenic fungi, such as A. fumigatus and C. neoformans, contain far more hybrid sensor kinases than fission yeast, which may reflect diversity in their environmental and biological niches because the pathogens survive and proliferate in both natural environments and human hosts. In addition, the involvement of the HOG pathway in the production of virulence factors and in the in vivo virulence of fungal pathogens cannot be adequately explained in nonpathogenic model yeasts. Therefore, it is obvious that the function and regulatory mechanism of the HOG pathway in each fungal pathogen ought to be investigated in the corresponding species.
Thus far, the function and regulatory mechanism of the HOG pathway in fungal pathogens have been best characterized in C. albicans and C. neoformans, as summarized in Fig. 4. To further elucidate the HOG pathway in fungal pathogens, the following topics need to be the focus of future studies. First, regardless of the presence of diverse and multiple hybrid sensor kinases in pathogenic fungi, the comprehensive and integrative understanding of the function of all hybrid sensor kinases is far from complete. Therefore, more extensive mutant construction and phenotypic analysis of hybrid sensor kinases should be performed. The lack of phenotypes observed in some hybrid sensor kinase mutants, such as the C. neoformans tco3Δ, tco4Δ, tco5Δ, and tco7Δ mutants, may result from a limited number of phenotypic assays available in the corresponding system. The recent report of functions for 97% of the S. cerevisiae genes suggests that phenotypes can be elucidated (56). Therefore, more diverse phenotypic assays need to be designed and performed to look for the proper role, if there is any, of each hybrid sensor kinase. Furthermore, structural and functional domain analyses are required to identify the detailed roles of various subdomains, including HAMP, GAF, PAS/PAC, and PK (Fig. 2), found in hybrid sensor kinases. Since the histidine kinase and receiver domains are commonly found in all hybrid sensor kinases, the signaling or ligand-sensing region must lie in other subdomains of hybrid sensor kinases. Second, the function of the HPt protein needs to be elucidated by the construction of either null mutants or conditional mutants of fungal pathogens. Third, domain analysis is also required for identifying the regulatory mechanism of the key interface protein, Ssk2-like MAPKKK, between the two-component system and the Hog1 MAPK module because the known subdomains of S. cerevisiae Ssk2 are not widely conserved in fungal pathogens. This will explain why the Ssk1 response regulator often differentially regulates the Ssk2 MAPKKK in fungal pathogens. Fourth, downstream regulators of the Hog1 MAPK, including MAPKAPKs and transcription factors, need to be further investigated. This is particularly important because downstream signaling components of the Hog1 MAPK appear to be highly divergent in fungal pathogens. Therefore, novel Hog1-dependent regulators could be essential for the role of Hog1 in controlling the virulence of fungal pathogens, providing an excellent opportunity to develop a novel antifungal therapy. Finally, the negative feedback regulatory systems of the HOG pathway, including the PTP- and PP2C-mediated system and the ubiquitin-proteasome system, should be characterized in fungal pathogens.
Proposed mechanism of the HOG pathway in fungal pathogens. The upper panel describes the C. albicans HOG pathway, and the lower panel illustrates the C. neoformans HOG pathway. In each panel, the left side shows unstressed conditions while the right side illustrates stressed conditions. temp., temperature; MG, methylglyoxal.
ACKNOWLEDGMENTS
I thank Alex Idnurm and Fritz Mühlschlegel for critically reading the manuscript.
This work was supported by the Korea Research Foundation grant funded by the Korean Government (MOEHRD, Basic Research Promotion Fund) (KRF-2007-331-C00223) and in part by the Korea Science and Engineering Foundation (KOSEF) grant funded by the Korea government (MEST) (R11-2008-062-02001-0).
FOOTNOTES
↵▿ Published ahead of print on 24 October 2008.
- American Society for Microbiology
REFERENCES
