The sphingolipid pathway regulates Pkc1 through the formation of diacylglycerol in Cryptococcus neoformans.

The sphingolipid biosynthetic pathway generates bioactive molecules crucial to the regulation of mammalian and fungal physiological and pathobiological processes. In previous studies (Luberto, C., Toffaletti, D. L., Wills, E. A., Tucker, S. C., Casadevall, A., Perfect, J. R., Hannun, Y. A., and Del Poeta, M. (2001) Genes Dev. 15, 201-212), we demonstrated that an enzyme of the fungal sphingolipid pathway, Ipc1 (inositol-phosphorylceramide synthase-1), regulates melanin, a pigment required for the pathogenic fungus Cryptococcus neoformans to cause disease. In this study, we investigated the mechanism by which Ipc1 regulates melanin production. Because Ipc1 also catalyzes the production of diacylglycerol (DAG), a physiological activator of the classical and novel isoforms of mammalian protein kinase C (PKC), and because it has been suggested that PKC is required for melanogenesis in mammalian cells, we investigated whether Ipc1 regulates melanin in C. neoformans through the production of DAG and the subsequent activation of Pkc1, the fungal homolog of mammalian PKC. The results show that modulation of Ipc1 regulates the levels of DAG in C. neoformans cells. Next, we demonstrated that C. neoformans Pkc1 is a DAG-activated serine/threonine kinase and that the C1 domain of Pkc1 is necessary for this activation. Finally, through both pharmacological and genetic approaches, we found that inhibition of Pkc1 abolishes melanin formation in C. neoformans. This study identifies a novel signaling pathway in which C. neoformans Ipc1 plays a key role in the activation of Pkc1 through the formation of DAG. Importantly, this pathway is essential for melanin production with implications for the pathogenicity of C. neoformans.

Interestingly, both mammalian and fungal sphingolipid pathways also generate diacylglycerol (DAG). 1 DAG is a glycerolipid that is well known for its function as a second messenger and activator of mammalian classical and novel protein kinase C (PKC) isoforms. Sphingolipid enzymes that catalyze the formation of DAG include sphingomyelin synthase in mammalian systems (7)(8)(9)(10)(11)(12)(13) and Ipc1 (inositol-phosphorylceramide synthase-1) in fungal systems (14). Sphingomyelin synthase and Ipc1 transfer a phosphate head group from phosphatidylcholine or phosphatidylinositol to ceramide or phytoceramide, respectively. Whether the DAG generated by these sphingolipid-metabolizing enzymes regulates cellular processes is, in fact, not known. In mammalian cells, the activation of sphingomyelin synthase correlates with activation and nuclear translocation of NF-B (15), events often regulated by DAG-dependent PKC (16). In fungi, the ability of DAG to activate Pkc1, the fungal homolog of mammalian PKC, is still controversial. Ogita et al. (17) and Simon et al. (18) purified a protein kinase from Saccharomyces cerevisiae that was activated by DAG, whereas other studies using the recombinant Pkc1 protein from S. cerevisiae and Candida albicans suggest that Pkc1 is insensitive to DAG (19 -21).
Cryptococcus neoformans is a pathogenic fungus that infects mainly immunocompromised patients, and it represents the leading cause of fungal meningoencephalitis worldwide (22). In previous studies, we showed that Ipc1 regulates melanin production through the modulation of laccase (23), the enzyme that, in C. neoformans, oxidizes exogenous diphenolic substrates (e.g. catecholamines) to form melanin (24,25). Melanin protects the fungus from the host immune response (26), and melanin-deficient mutants of the fungus are indeed avirulent in animal models of cryptococcosis (27). Interestingly, it has been suggested that melanogenesis by mammalian cells is dependent on the activation of PKC by DAG. Initial reports in mammalian melanocytes showed that DAG increases melanin pigmentation in a PKC-dependent manner (28); and more recently, it was demonstrated that PKC␤, an isoform that is activated by DAG, regulates melanin production in mammalian cells (29). Thus, the generation of DAG by Ipc1 raises the intriguing hypothesis that Ipc1 regulates melanin production in C. neoformans through the formation of DAG and the consequent activation of Pkc1. * This work was supported in part by the Burroughs Wellcome Fund, by Grants AI51924 and AI56168 (to M. D. P.) and Grant HL43707 (to Y. A. H.) from the National Institutes of Health, and by RR17677 Project 2 from the Centers of Biomedical Research Excellence Program of the National Center for Research Resources (to M. D. P.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM  Here, we found that modulation of Ipc1 expression regulates the formation of DAG in C. neoformans. Next, we demonstrated that DAG activates C. neoformans Pkc1 and that deletion of a conserved DAG-binding domain abrogates the activation of Pkc1 by DAG. Finally, inhibiting Pkc1 by both pharmacological and genetic approaches inhibits melanin production, thereby establishing a crucial role for Pkc1 in the activation of this virulence factor required for development of infection by C. neoformans.
DAG Kinase Assay-A total of 10 8 cells were collected from cultures of the C. neoformans wild-type H99 and GAL7::IPC1 strains. Neutral lipids were extracted from the cell pellets as described by Bligh and Dyer (31) and dried down. Three-fourths of each sample were used for lipid measurement, whereas one-fourth was used for inorganic phosphate measurement. Briefly, the lipids were incubated at room temperature for 45 min in the presence of 7.5% ␤-octyl glucoside and 25 mM dioleoylphosphatidylglycerol micelles, 2 mM ATP (mixed with [␥-32 P]ATP), and a membrane preparation of Escherichia coli overexpressing DAG kinase (6 g) in a final volume of 100 l. After the lipids were extracted by the Bligh and Dyer method, the reaction products were separated by TLC in chloroform/methanol/acetone/acetic acid/H 2 O (50:20:15:10:5), and the radioactivity associated with the phosphatidic acid and phosphorylated ceramides was measured by scintillation counting. Ceramide and DAG levels were quantitated using external standards and were normalized to phosphate as described previously (32,33).
Isolation of the C. neoformans PKC1 Gene-The C. neoformans PKC1 gene was first identified by blasting the S. cerevisiae homolog in the C. neoformans Genome Project Database of var. neoformans serotype D strain JEC21. 2 Three sequences (502511B03.x1, 502488F12.x1, and 502342F06.x1) showed very high homology to the S. cerevisiae PKC1 gene and corresponded to the 3Ј-region of the gene containing the serine/threonine kinase domain, which is highly conserved among all fungal and human homologs. These sequences were used to design the following primers to amplify a 1589-bp fragment of the PKC1 gene from cDNA of C. neoformans var. grubii serotype A strain H99: PKC-antisense-5 (5Ј-GAT GAA TTC ATA GAA TCC CTC ACA GAT TCA CTC-3Ј) and PKC-antisense-3 (5Ј-GTA CTC GAG CTA GTG TCA GTA GCA TTA CCG ATC-3Ј), which contain EcoRI and XhoI sites (boldface and underlined), respectively. The DNA fragment was then used to screen both a genomic DNA library of C. neoformans var. grubii serotype A strain H99 and a cDNA library of C. neoformans var. neoformans serotype D strain B3501 to isolate, clone, and sequence the PKC1 gene (GenBank TM /EBI Data Bank accession numbers AY373758 and AY373759).
Expression of Recombinant Pkc1, App1, and ⌬C1-Pkc1 Proteins-To produce the recombinant Pkc1 protein, the cDNA of PKC1 from serotype D was amplified using primers XO-SeroD (5Ј-CTA TCT CGA GGG CAT CCA ATG ATC CCA AGG CCA AG-3Ј) and SeroD-H3 (5Ј-GAT TAA GCT TCT AGG TCT GTG CTG CAG CCC AAG GAG-3Ј9), which contain XhoI and HindIII sites (boldface and underlined), respectively. The PCR fragment was digested with XhoI and HindIII and cloned into the pBAD/HisB vector (Invitrogen), placing it in-frame downstream of the ATG start codon and the His and Xpress sequence tags. The resulting plasmid was digested with NcoI and HindIII, yielding a fragment containing the ATG start codon, the His and Xpress tags, and PKC1 cDNA sequence. The fragment was blunted and subcloned into HindIIIdigested and blunted vector pYES2 (Invitrogen) downstream of the GAL1 promoter, yielding the pYES2/GAL1::Xpress::PKC1 plasmid. To produce the recombinant App1 (antiphagocytic protein-1) protein, the APP1 cDNA was subcloned from the pBAD/His-App1 vector (34) into pYES2 to generate the pYES2/GAL1::Xpress::APP1 plasmid.
The pYES2/GAL1::Xpress::PKC1, pYES2/GAL1::Xpress::APP1, and pYES2/GAL1::Xpress::⌬C1-PKC1 plasmids were transformed into S. cerevisiae strain JK9-3d␣. Transformants were selected on YNB ura Ϫ agar plates containing 2% glucose, purified, randomly chosen, and then grown in YNB ura Ϫ broth containing 2% glucose for 24 h in a 30°C shaker. The cultures were pelleted and washed twice with sterile distilled deionized water. Cells were then inoculated into YNB ura Ϫ broth containing 2% galactose and incubated for 16 h in a 30°C shaker. Cell pellets were collected and stored at Ϫ20°C. To lyse the cells, the cell pellets were resuspended in 200 l of modified radioimmune precipitation assay buffer (150 mM NaCl, 1% Nonidet P-40, 0.5% deoxycholic acid, 1 mM phenylmethylsulfonyl fluoride, 20 g/ml chymostatin, 20 g/ml leupeptin, 20 g/ml antipain, 20 g/ml pepstatin A, and 50 mM Tris, pH 7.5). Acid-washed glass beads (Sigma) were added in a volume equal to three-fourths of the cell suspension, and the cells were homogenized in a mini-Beadbeater (Biospec Products, Inc., Bartlesville, OK) three times for 45 s at 4°C. Samples were centrifuged at 2500 ϫ g for 10 min at 4°C, and the supernatant was transferred to a sterile tube. The total protein concentration of the cell lysates was measured using the method of Bradford (35).
Immunoprecipitation of Pkc1, App1, and ⌬C1-Pkc1-To prepare antibody-bead complexes, 4 l of anti-Xpress antibody (Invitrogen) were combined with 200 l of protein G-Sepharose 4FF (50% slurry; Amersham Biosciences) in a 1-ml final volume of modified radioimmune precipitation assay buffer and incubated for 1 h at 4°C on a rocking platform. The antibody-bead complexes were washed twice with modified radioimmune precipitation assay buffer and collected by centrifugation at 10,000 ϫ g for 1 min at 4°C. To preclear the cell lysate, an aliquot containing 150 g of total protein was placed in a 1-ml final volume of modified radioimmune precipitation assay buffer with 200 l of protein G-Sepharose (50% slurry), rocked at 4°C for 1 h, and then centrifuged at 10,000 ϫ g for 1 min at 4°C. The precleared supernatant was added to the antibody-bead complexes, and immunoprecipitation was carried out for 16 h at 4°C on a rocking platform.
In Vitro Kinase Assay-Recombinant protein conjugated to protein G-Sepharose 4FF was washed twice with kinase buffer (50 mM Tris, pH 7.4, 10 mM sodium orthovanadate, 10 mM sodium fluoride, 5 mM MgCl 2 , 1 mM phenylmethylsulfonyl fluoride, 0.01% leupeptin, 0.01% aprotinin, and 0.01% trypsin/chymotrypsin inhibitor). A 20-l aliquot was pelleted and resuspended in 15 l of kinase buffer. The reaction was started by the addition of 60 l of reaction mixture (20 mM Tris, pH 7.4, 10 mM MgCl 2 , 0.2 mg/ml histone type III-SS (Sigma H-4524), 1ϫ Triton X-100/ lipid mixed micelles, 50 M unlabeled ATP, and 5 Ci/reaction [␥-32 P]ATP). Triton X-100/lipid mixed micelles were prepared as described previously (36) with lipids at the following final concentrations: 8 mol % phosphatidylserine, 2.5 mol % DAG, 2.5 mol % C 6 -ceramide and C 6 -phytoceramide. All lipids were from Avanti Polar Lipids (Alabaster, AL), except for the short-chain ceramides, which were synthesized by the Medical University of South Carolina Lipidomics Core Facility. The reaction was carried out at 30°C for 10 min and terminated by the addition of 4ϫ Laemmli buffer. Samples were boiled for 5 min, loaded 2 Available at www-Sequence.Stanford.edu/group/C.neoformans. onto a 12% SDS-polyacrylamide gel, and separated by SDS-PAGE. Gels were fixed in a 50% methanol and 10% glacial acetic acid solution, rehydrated, and dried on gel blot paper. Radioactively labeled phosphorylated histone was visualized by autoradiography and quantitated using the PhosphorImager system and by ImageQuant analysis (Amersham Biosciences). Specific activity is defined as picomoles of histone phosphorylated per min/mg of lysate, although it may vary according to the level of expression of the recombinant protein. The moles of phosphorylated histone were calculated using a standard curve of the reaction mixture containing both unlabeled ATP and [␥-32 P]ATP, which had been spotted onto gel blot paper and scanned using a PhosphorImager concurrently with the dried gels. The amount of total lysate corresponding to the amount of recombinant immunoprecipitated Pkc1 used in the assay was determined by subjecting an aliquot of the immunoprecipitate with a standard curve of the total lysate containing the recombinant protein to Western blotting. The reaction time for the kinase assay and the concentration of immunoprecipitated protein used were chosen after performing a time course and protein titration to ensure that the measurements were within the linear range of kinase activity.
Laccase Activity-Laccase activity was assayed as described previously with minor modifications (37,38). Cells were inoculated from fresh cultures in YPD medium into YP broth with 2% glucose or raffinose and 2% galactose and incubated at 30°C for 16 h in a shaking incubator. Cells were pelleted, washed once with liquid YNB medium without sugar, and resuspended in 10 ml of the same medium. The serine/threonine kinase inhibitors staurosporine, bisindolylmaleimide I, and calphostin C (all from Calbiochem) and the tyrosine kinase inhibitors genistein (BIOMOL Research Labs Inc., Plymouth Meeting, PA) and herbimycin A (Calbiochem) suspended in Me 2 SO were added to the cell suspensions at the indicated concentrations, whereas Me 2 SO alone was added to untreated cells. Cells were incubated for an additional 5 h at 30°C, pelleted, washed once with sterile distilled deionized water, and resuspended in 50 mM sodium phosphate at pH 5.0. Then, 1 ϫ 10 7 cells/ml were incubated in a 30°C water bath for 30 min with 1 mM 2,2Ј-azinobis(3-ethylbenzothiazoline-6-sulfonic acid). The cells were pelleted in a microcentrifuge at 14,000 rpm for 30 s, and the supernatants were analyzed at 420 nm. The A 420 nm of a sample containing no cells was used as the blank. The specific activity of laccase is defined as nanomoles of 2,2Ј-azinobis(3-ethylbenzothiazoline-6-sulfonic acid) oxidized (⑀ 420 nm ϭ 36 mM Ϫ1 cm Ϫ1 ) (39) per min/10 7 cells. To ensure that the addition of the inhibitors did not affect cell viability, colonyforming units were measured after each treatment by plating cells on YPD plates and comparing with untreated samples.
Construction of Antisense PKC1-The DNA fragment used to isolate the PKC1 gene from C. neoformans was digested with EcoRI and XhoI and cloned into the EcoRI and XhoI sites of the pJMM170-3 vector (kindly provided by Dr. Gary Cox, Duke University Medical Center, Durham, NC) (40). The resulting plasmid, named pURA5/GAL7:: PKC1-AS (where AS is antisense), contains (i) the GAL7 promoter from the C. neoformans serotype D strain, which regulates transcription of the 3Ј-5Ј PKC1 mRNA under galactose-inducing conditions, and (ii) the URA5 gene as the selectable marker. To enable transformation and selection of the plasmid in C. neoformans, the URA5 gene in pURA5/GAL7::PKC1-AS was replaced with the nourseothricin acetyltransferase gene (NAT1) (41), which confers resistance to the antibiotic nourseothricin (Werner BioAgents, Jena-Cospeda, Germany). First, the NAT1 gene under the control of the C. neoformans actin promoter was amplified from the pNAT1 vector (a gift from Dr. John Perfect, Duke University Medical Center) using primers XB-NAT-F (5Ј-CTA ATC TAG AGC GAG GAT GTG AGC TGG AGA GCG G-3Ј) and XB-NAT-R (5Ј-CGC GTC TAG AGA AGA GAT GTA GAA ACT AGC TTC C-3Ј), which contain XbaI sites (boldface and underlined); subcloned into the pCR2.1-TOPO vector (Invitrogen); excised by digestion with XbaI; and blunted. Plasmid pURA5/GAL7::PKC1-AS was digested with SacI and XbaI, blunted, dephosphorylated, and ligated with the NAT1 fragment. This pNAT1/GAL7::PKC1-AS construct was then transformed into both C. neoformans wild-type and GAL7::IPC1 strains by biolistic transformation (42) to produce the C. neoformans IPC1/PKC1-AS and GAL7::IPC1/PKC1-AS strains, respectively. Transformants were grown in YPD medium containing 100 mg/liter nourseothricin. Three stable nourseothricin-resistant transformants of each strain (IPC1/PKC1-AS, transformants 16, 18, and 19; and GAL7::IPC1/PKC1-AS, transformants 5, 27, and 31) were chosen randomly and grown in YP broth containing 2% galactose, glucose, or raffinose for 24 h at 30°C. Total RNA was extracted from each culture and used for analysis of PKC1 mRNA levels by reverse transcription-PCR using primers PKC-5 (5Ј-GAG GGA GAC GAG GAA AAA ATC-3Ј) and PKC-antisense-3 (5Ј-GTA CTC GAG CTA GTG TCA GTA GCA TTA CCG ATC-3Ј). Laccase activity and melanin production by the transformants were also analyzed as described above.
Statistical Analysis-Statistical analysis was performed using Student's t test.

Ipc1 Regulates Melanin Production by C. neoformans-In
previous studies, we generated a C. neoformans GAL7::IPC1 strain in which the IPC1 gene is regulated by the galactoseinducible GAL7 promoter (23). This strain allows up-regulation of IPC1 by growing cells in galactose medium (inducing conditions) and down-regulation of IPC1 by growing cells in glucose or raffinose medium (repressing conditions). Using the GAL7::IPC1 strain, we showed that Ipc1 regulates the activity of laccase, the melanin-producing enzyme of C. neoformans, as measured by an in vitro assay (23). Thus, we wondered whether Ipc1 modulation would affect the production of melanin pigment in vivo (growing cells). The C. neoformans wild-type (H99) FIG. 1. Melanin production by the C. neoformans wild-type and GAL7::IPC1 strains on L-DOPA agar plates when Ipc1 is up-regulated (galactose) and down-regulated (raffinose). A, the GAL7::IPC1 strain produced more melanin than the wild-type (WT) strain in the presence of galactose and less melanin in the presence of raffinose. No significant difference was observed in melanin production by the wild-type strain grown in galactose or raffinose. B, Ipc1 activity increased when the GAL7::IPC1 strain was grown in galactose and decreased when the GAL7::IPC1 strain was grown in raffinose compared with the wild-type strain. Data are representative of at least three independent experiments. and GAL7::IPC1 strains were grown on agar containing L-DOPA, which is a substrate for laccase, with either galactose or raffinose. The production of melanin was visualized by the formation of the characteristic brown pigment. As shown in Fig. 1A, the GAL7::IPC1 strain produced more melanin compared with the wild-type strain on L-DOPA/galactose agar, whereas it produced less melanin compared with the wild-type strain on L-DOPA/raffinose agar. No significant difference in melanin production was detected with the wild-type strain when grown in galactose or raffinose. We confirmed that, under these experimental conditions, Ipc1 activity in the GAL7::IPC1 strain was up-and down-regulated, respectively, compared with the wild-type strain (Fig. 1B).
Ipc1 Regulates DAG and Phytoceramide Levels in C. neoformans-We next investigated the mechanisms by which Ipc1 may regulate melanin formation in C. neoformans. Because mammalian melanogenesis is suggested to be dependent on a DAG-PKC pathway and because Ipc1 catalyzes the production of DAG, we wondered whether Ipc1 may control the activation of fungal Pkc1 through the modulation of DAG. First, we determined whether Ipc1 regulates the level of DAG in C. neoformans. The wild-type and GAL7::IPC1 strains were grown under inducing and repressing conditions, and total DAG levels in each sample were measured by in vitro DAG kinase assay. As shown in Fig. 2A, DAG levels in the GAL7::IPC1 strain increased significantly by ϳ24% when IPC1 was up-regulated and decreased by ϳ29% when IPC1 was down-regulated. There was no significant change in the level of DAG in the wild-type strain grown in galactose or glucose. The level of phytoceramide (fungal ceramide), which is a substrate of the Ipc1 reaction, was also measured by DAG kinase assay. As shown in Fig. 2B, phytoceramide increased by 3-fold when IPC1 was down-regu-  c.), and C. albicans (C.a.) reveals a conserved pattern of histidine and cysteine residues (boldface). The C1A domain of human PKC␦ also contains key residues that maintain the DAG-binding groove (red) and the surrounding hydrophobic wall (blue). The majority of these residues are conserved in C. neoformans Pkc1 and absent in the Pkc1 proteins from S. cerevisiae and C. albicans. C, shown is the expression construct of Pkc1 fused to the Xpress tag in the pYES2 vector, which was transformed into the recipient S. cerevisiae JK9-3d␣ strain. The pYES2 vector utilizes the galactose-inducible GAL1 promoter. D, the Pkc1 protein was expressed and immunoprecipitated from total cell lysates using anti-Xpress antibody conjugated to protein G-Sepharose beads as described under "Experimental Procedures." Western blot analysis with anti-Xpress antibody detected a single band at 120 kDa.

FIG. 2. Regulation of DAG and phytoceramide levels in C. neoformans by Ipc1.
A, down-regulation of IPC1 decreased DAG levels, and up-regulation of IPC1 increased DAG levels in the GAL7::IPC1 strain compared with the wild-type (WT) strain. *, p Ͻ 0.02. B, down-regulation of IPC1 caused a buildup of cellular phytoceramide. *, p Ͻ 0.02. Data are shown as picomoles of DAG or phytoceramide/nmol of P i . lated in the GAL7::IPC1 strain (glucose), but there was no notable decrease in phytoceramide levels when IPC1 was upregulated (galactose).
Isolation and Expression of the PKC1 Gene-Because the above data demonstrated that Ipc1 regulates the levels of DAG and phytoceramide (an activator and inhibitor of mammalian PKC, respectively) (43, 44), we next wondered whether these two lipids regulate the activity of C. neoformans Pkc1. To measure Pkc1 activity, it was necessary to produce and isolate the recombinant Pkc1 protein.
First, the C. neoformans PKC1 gene was identified and isolated from genomic DNA and cDNA libraries of serotype A strain H99 and serotype D strain B3501. Strands of the open reading frame and 5Ј-and 3Ј-untranslated regions were sequenced, revealing that the PKC1 gene is 3784 base pairs in length with 10 intronic sequences, yielding a predicted protein of 1086 amino acids (GenBank TM /EBI Data Bank accession numbers AY373758 and AY373759). Compared with mammalian PKCs, C. neoformans Pkc1 was most similar to the DAGdependent novel (e.g. PKC␦) and classical (e.g. PKC␤II) isoforms at 31% and least similar to the DAG-independent atypical isoforms (e.g. PKC) at 24%. The Pkc1 homologs from S. cerevisiae and C. albicans had similarities of 63 and 57%, respectively. By performing a BLAST search of the Conserved Domain Database, 3 it was determined that C. neoformans Pkc1 contains a conserved serine/threonine kinase domain (active site) as well as three regulatory motifs: HR1 (Rho-binding domain in S. cerevisiae Pkc1 (45) and mammalian PKC-related kinases (46)), C2 (calcium-binding domain in mammalian PKCs), and C1 (DAG-binding domain in mammalian PKCs) (Fig. 3A).
The C1 domain of mammalian and fungal PKCs consists of two tandem DAG-binding domains designated C1A and C1B (47), each of which represents a zinc finger motif due to a conserved pattern of histidine and cysteine residues (shown in boldface in Fig. 3B) (48,49). It has been suggested that only one of these zinc finger motifs (C1A or C1B) is required for the activation of PKC by DAG (47). These C1A and C1B motifs of mammalian and fungal PKCs can be classified as "typical" or "atypical" to refer to C1 domains that bind or do not bind DAG, respectively. Typical C1 domains, found in the DAG-dependent classical and novel PKC isoforms, contain three consensus residues (Pro 11 , Gly 23 , and Gln 27 ) that maintain the structure of the DAG-binding pocket as well as five consensus residues at positions 8, 13, 20, 22, and 24, which compose a hydrophobic wall around the binding site, as illustrated in Fig. 3B for the C1A motif (47). Atypical C1 domains, found in the DAG-independent atypical mammalian PKCs, lack two or more of these key residues (47). Based on these criteria, the C1A domain of C. neoformans Pkc1 is classified as a typical domain because it has conservation of the three DAG-binding residues as well as four of the five hydrophobic residues (Fig. 3B). In contrast, the C1A and C1B domains of the S. cerevisiae and C. albicans Pkc1 proteins lack the majority of these important residues (Fig. 3B) and are therefore classified as atypical (47,50  vector (Fig. 3C). The resulting pYES2/GAL1::Xpress::PKC1 construct was transformed into the recipient S. cerevisiae JK9-3D␣ strain. Transformants were screened by Western blot analysis using anti-Xpress antibody. As depicted in Fig. 3D, under inducing conditions (galactose), a single band was detected at ϳ120 kDa, the predicted size of Pkc1 including the Xpress epitope tag, whereas no expression was detected under repressing conditions (glucose). The Pkc1 protein was immunoprecipitated from the total cell lysate using anti-Xpress antibody conjugated to protein G-Sepharose beads (Fig. 3D). The isolated recombinant C. neoformans Pkc1 protein was then used for in vitro kinase assays.
Pkc1 Is Activated by DAG and Inhibited by Phytoceramide-To determine whether the activity of Pkc1 may be regulated by DAG and/or phytoceramide, kinase assays were performed to test the ability of recombinant C. neoformans Pkc1 to phosphorylate histone substrate in the presence or absence of DAG or phytoceramide delivered in Triton X-100/lipid mixed micelles. In the absence of lipids, the specific activity of Pkc1 was 31.5 pmol/min/mg. In the presence of the DAG subspecies dipalmitoylglycerol and dimyristoylglycerol, Pkc1 activity was increased by 1.7-and 2-fold, respectively (Fig. 4A). A third DAG subspecies, palmitoyloleoylglycerol, caused a 1.5-fold increase in Pkc1 activity (data not shown). Interestingly, Pkc1 activity also increased in the presence of phosphatidylserine (Fig. 4A), a known lipid cofactor of mammalian PKCs.
To ensure that the kinase activity detected was specific to Pkc1, a different Xpress-tagged protein from C. neoformans (App1) was immunoprecipitated from cell lysates of S. cerevisiae strain JK9-3d␣ expressing the pYES2/GAL1::Xpress::APP1 vector and used in the same kinase assay. As shown in Fig. 4B, App1 did not cause any notable phosphorylation of histone in the presence or absence of DAG, as expected. The specific activity of App1 in the absence of lipid was estimated at 7.1 pmol/min/mg.
Next, the effect of ceramides, which inhibit certain mammalian PKC isoforms (43,44), on Pkc1 activity was evaluated. As shown in Fig. 4C, the short-chain ceramide analogs C 6 -phytoceramide and C 6 -ceramide decreased Pkc1 activity by ϳ50% compared with the untreated sample.
Deletion of the C1 Domain Abolishes the Activation of Pkc1 by DAG-To investigate a possible mechanism by which DAG activates Pkc1, we focused on the putative DAG-binding domain (C1), which was deleted from the pYES2/GAL1::Xpress::PKC1 expression vector (Fig. 5A). The resulting pYES2/GAL1::Xpress:: ⌬C1-PKC1 plasmid (Fig. 5A) was transformed and expressed in S. cerevisiae JK9-3D␣ cells. The production of an ϳ104-kDa protein (⌬C1-Pkc1) was confirmed by Western blot analysis of total cell lysate with anti-Xpress antibody (Fig. 5B). Next, the ⌬C1-Pkc1 protein was immunoprecipitated and tested for kinase activity in the presence or absence of DAG and phytoceramide. The specific activity of ⌬C1-Pkc1 in the absence of lipids was similar to that of wild-type Pkc1 at 25.9 pmol/min/mg. However, as shown in Fig. 5C, deletion of the C1 domain abolished the activation of Pkc1 by DAG. Interestingly, deletion of the C1 domain did not affect the repression of Pkc1 by phytoceramide.
Inhibition of Pkc1 Inhibits Laccase Activity and Melanin Production-Because the above data suggested that Ipc1 may regulate Pkc1 through the production of DAG and because DAG-PKC signaling may control melanogenesis in mammalian models (28,29), we investigated whether Pkc1 plays a role in the regulation of melanin formation by C. neoformans. First, a pharmacological approach was used in which laccase activity was measured in C. neoformans wild-type and GAL7::IPC1 cells grown in galactose or glucose upon treatment with mammalian serine/threonine kinase inhibitors such as staurosporine, bisindolylmaleimide I, and calphostin C.
Staurosporine is a nonspecific inhibitor of PKC that targets the ATP-binding site (51). As shown in Fig. 6A, treatment of C. neoformans with increasing concentrations of staurosporine (1, 5, and 10 M) caused a dose-dependent decrease in laccase activity. At the highest concentration of staurosporine (10 M), there was an ϳ65% decrease in laccase activity in the wild-type strain in both galactose and glucose compared with vehicle controls. The GAL7::IPC1 strain grown in galactose showed a more modest, but significant 33% decrease in laccase activity. The GAL7::IPC1 strain grown in glucose exhibited the most inhibition at 78% (Fig. 6A).
Bisindolylmaleimide I is specifically directed against the ATP-binding site of PKC (52). Like staurosporine, bisindolylmaleimide I induced a dose-dependent decrease in laccase activity (Fig. 6B). However, bisindolylmaleimide I was more potent than staurosporine. Wild-type cells treated with 10 M bisindolylmaleimide I showed a 72% decrease in laccase activity. GAL7::IPC1 cells in galactose showed a 48% decrease, and GAL7::IPC1 cells in glucose had a decrease of 84% (Fig. 6B).
Calphostin C is a specific inhibitor of PKC that blocks the binding of DAG to PKC (53). As illustrated in Fig. 6C, treatment with calphostin C also caused a dose-dependent decrease in laccase activity and proved to be more potent than staurosporine and bisindolylmaleimide I in the inhibition of laccase activity. Indeed, the laccase activity of the wild-type strain decreased by 83% upon treatment with 10 M calphostin C, whereas the laccase activity of the GAL7::IPC1 strain decreased by Ͼ85% (Fig. 6C).
To determine whether these decreases in laccase activity were specific to inhibition of a serine/threonine kinase, the C. neoformans wild-type and GAL7::IPC1 strains grown in glucose or galactose were treated with tyrosine kinase inhibitors such as genistein and herbimycin A at the same concentrations (1, 5, and 10 M). As shown in Fig. 6D, genistein treatment did not affect laccase activity. Similar results were obtained with herbimycin A (data not shown).
To verify that the above inhibitors were nontoxic to C. neoformans, treated cells were tested for cell viability. There was no significant change in the number of colony-forming units of C. neoformans cells upon treatment with staurosporine, bisindolylmaleimide I, calphostin C, genistein, or herbimycin A, indicating that these inhibitors are nontoxic to the cells at the concentrations used (data not shown).
Next, to confirm the direct involvement of Pkc1 in the regulation of melanin production by C. neoformans, Pkc1 was specifically down-regulated by expression of an antisense PKC1 transcript. A PCR fragment corresponding to the 3Ј-region of the PKC1 gene containing the serine/threonine kinase domain (see "Experimental Procedures") was used to create the pNAT1/GAL7::PKC1-AS construct (Fig. 7A). This construct was transformed into the C. neoformans wild-type (H99) and GAL7::IPC1 strains. Nourseothricin-resistant transformants of each strain were isolated and named IPC1/PKC1-AS and GAL7::IPC1/PKC1-AS strains, respectively. Next, three inde- pendent transformants for the IPC1/PKC1-AS strain (transformants 16, 18, and 19) and the GAL7::IPC1/PKC1-AS strain (transformants 5, 27, and 31) were randomly selected and grown in galactose for expression of antisense PKC1 and in glucose or raffinose as a control. As illustrated in Fig. 7 (B and C), expression of antisense PKC1 significantly decreased laccase activity by Ͼ40% in both IPC1/PKC1-AS and GAL7::IPC1/ PKC1-AS compared with the corresponding parental strains. No decrease in laccase activity was observed when the transformants were grown in glucose or raffinose medium (data not shown). As determined by reverse transcription-PCR, the level of PKC1 mRNA was indeed reduced when transformants were grown in galactose compared with the corresponding parental strains, whereas PKC1 mRNA levels did not change when transformants were grown in glucose or raffinose (data not shown).
To verify that down-regulation of Pkc1 inhibited melanin production by C. neoformans cells, the IPC1/PKC1-AS (transformants 16, 18, and 19) and GAL7::IPC1/PKC1-AS (transformants 5, 27, and 31) strains were grown alongside their parental strains (wild-type and GAL7::IPC1) on L-DOPA/galactose agar plates, and melanin pigment formation was monitored. As shown in Fig. 7 (D and E), melanin production by both IPC1/ PKC1-AS and GAL7::IPC1/PKC1-AS strains was significantly reduced compared with the corresponding parental strains. As a control, empty antisense vector was transformed into the wild-type strain. The resulting transformants did not exhibit any significant change in melanin production compared with the wild-type strain under the experimental conditions used (data not shown). DISCUSSION In this study, the signaling pathway by which the sphingolipid enzyme Ipc1 regulates melanin production by C. neoformans was investigated. The results suggest that the sphingolipid pathway controls the activation of Pkc1 via Ipc1 through the regulation of DAG and phytoceramide. Additionally, evidence is provided that the C1 domain of Pkc1 is required for activation by DAG. Finally, the results demonstrate that Pkc1 is essential for melanin formation in C. neoformans. Taken together, these results provide evidence for a novel signaling pathway (Ipc1-DAG/phytoceramide-Pkc1) that plays a crucial role in melanin production and in the pathogenicity of C. neoformans. Relatively little is known about the regulatory pathways leading to the formation of melanin in C. neoformans. Recent studies show that the G␣ protein-adenylyl cyclase-cAMP-protein kinase A signaling cascade may regulate melanin formation (38,54,55). However, studies in mammalian cells suggest that melanin production is regulated by a DAG-PKC mechanism (28,29,56). Here, we extend our previous observations that C. neoformans Ipc1 controls the formation of melanin (23) and provide evidence for a key role for the DAG-Pkc1 pathway. We found that modulation of Ipc1 causes significant changes in the level of DAG in C. neoformans. These changes were not dramatic, but a possible explanation may involve the modulation of subspecies of DAG by Ipc1. An interesting study by Schneiter et al. (57) showed that the distribution of phosphatidylinositol (substrate for Ipc1) within fungal cells is heterogeneous. In particular, using mass spectrometry, they showed that different phosphatidylinositol subspecies, defined by the composition of fatty acids in position 1 or 2 of the glycerol backbone, are localized in different subcellular compartments (i.e. Golgi versus plasma membrane). Because the Ipc1 enzyme is localized in the Golgi (58), it may be exposed to a distinct population of phosphatidylinositol subspecies. As a consequence, Ipc1 may regulate the production of only one or a few DAG subspecies, which may represent a small subset of the total DAG found in the cell. Another possibility could be the selectivity of Ipc1 for phosphatidylinositol subspecies, as observed for other enzymatic reactions in yeast and mammalian cells, such as the acylation of glycosylphosphatidylinositols in C. neoformans (59) and the hydrolytic activity of phospholipase C (60). Importantly, the hypothesis of species selectivity is strengthened by the observation that DAG subspecies differentially activated Pkc1. In particular, dimyristoylglycerol was the most potent activator, followed closely by dipalmitoylglycerol (Fig. 4A). Palmitoyloleoylglycerol also activated Pkc1, but less significantly compared with the other two DAG subspecies (data not shown).
Interestingly, the sensitivity of C. neoformans Pkc1 to DAG may be due to the structure of its C1 domain, which appears to be that of a typical DAG-binding domain. In contrast, the C1 domains of S. cerevisiae and C. albicans are classified as atypical or DAG-insensitive. Studies using recombinant Pkc1 proteins from S. cerevisiae and C. albicans indicate that these serine/threonine protein kinases are not activated by DAG in vitro (19 -21), although there is some indirect evidence that Pkc1 from S. cerevisiae is activated by DAG (17,18,61,62). We must note that the conflicting results concerning DAG activation of the fungal Pkc1 proteins could be attributed to the different experimental conditions used to assess kinase activity. Unlike the other studies, this study used a Triton X-100/ lipid mixed micelle system for the delivery of DAG and other lipids. Additionally, because only certain DAG subspecies may be able to activate Pkc1, as discussed above, perhaps the choice of DAG subspecies accounts for the different outcomes, as already suggested by Marini et al. (62). On the other hand, in C. neoformans, the DAG-Pkc1 signaling mechanism may be specifically used by this pathogenic fungus to produce melanin, which is not produced by S. cerevisiae or C. albicans. Thus, the fundamental differences in both C1 domain structure and cellular function among the fungal Pkc1 proteins indicate that Pkc1 from C. neoformans is truly a distinct class of fungal protein kinase C from that of S. cerevisiae and C. albicans.
The regulation of Pkc1 by Ipc1 may be mediated not only by the production of DAG, but also by the modulation of phytoceramide. Down-regulation of Ipc1 in vivo caused a significant buildup of phytoceramide in C. neoformans cells (Fig.  2B), and both phytoceramide and ceramide inhibited Pkc1 activity in vitro (Fig. 4C). Studies in mammalian cells suggest that ceramide may inhibit PKC directly by binding at the C1 domain and thereby competing with DAG (43) or indirectly through the activation of a phosphatase that dephosphorylates PKC (44). Because phytoceramide alone can inhibit C. neoformans Pkc1 in the kinase assay, a direct mechanism of action is likely, at least in this particular fungal system. However, phytoceramide was still able to inhibit Pkc1 even when the C1 domain was removed (Fig. 5C), although this inhibition was attenuated, suggesting that phytoceramide binds at a different regulatory domain or is able to regulate Pkc1 through more than one domain. These observations support a model in which Ipc1 functions as a molecular switch for the activation of Pkc1 in C. neoformans by regulating at the same time and in opposite directions the levels of DAG (activator) and phytoceramide (inhibitor). Such possibilities clearly warrant further investigation.
The observation that up-regulation of Ipc1 did not decrease phytoceramide below the level found in wild-type cells under the same conditions (Fig. 2B) is intriguing. This result could be explained by a compensatory up-regulation of the sphingolipid enzymes upstream of Ipc1 to maintain a constant level of phytoceramide. Indeed, the mRNA level of the LCB2 (longchain base 2) gene, which encodes the catalytic subunit of Spt1 (serine palmitoyltransferase-1), the first enzyme in sphingolipid synthesis, was increased when Ipc1 was up-regulated in the GAL7::IPC1 strain. 4 The activation of the Spt1 enzyme upon up-regulation of Ipc1 may account for an increase in the 4 M. Del Poeta, unpublished data.
FIG. 8. Proposed model for the regulation of melanin production by the sphingolipid pathway in C. neoformans. De novo sphingolipid production in both fungi and mammals is initiated by the enzyme Spt1, which consists of two subunits, Lcb1 and Lcb2. The fungus-specific enzyme Ipc1 regulates the levels of phytoceramide and DAG, which inhibit and activate Pkc1, respectively. In turn, Pkc1 activates laccase and regulates melanin production and pathogenicity of C. neoformans. PI, phosphatidylinositol; IPC, inositol phosphorylceramide; PC, phosphatidylcholine; SMS, sphingomyelin synthase.
de novo synthesis of phytoceramide, thereby maintaining a constant level of intracellular phytoceramide.
In conclusion, our results establish a novel role for the sphingolipid enzyme Ipc1 in the regulation of fungal Pkc1. Ipc1 has previously been recognized as a promising target for antifungals because it is a fungus-specific enzyme (14). Now, because the newly defined Ipc1-Pkc1 pathway appears to be crucial for melanin formation in C. neoformans, targeting Ipc1 and its downstream effectors for the development of new antifungal drugs is even more attractive (Fig. 8). Importantly, with the demonstration that Ipc1 activates Pkc1 through the formation of DAG, this study provides the first biochemical evidence that a link between sphingolipid-derived DAG and protein kinase C activity exists and could represent a paradigm that is applicable to mammalian systems.