Analysis of the Protein Kinase A-Regulated Proteome of Cryptococcus neoformans Identifies a Role for the Ubiquitin-Proteasome Pathway in Capsule Formation

Quantitative profiling of the proteome upon modulation of PKA1 expression.Given the virulence defect of a pka1 mutant and Pka1 regulation of secreted virulence factors in C. neoformans, we hypothesized that Pka1 influences the abundance of proteins associated with these functions as well as general processes such as translation and metabolism (19). We therefore evaluated the effect of Pka1 regulation on the proteome by collecting cells of the wild-type (WT) and P_GAL7::PKA1 strains grown under Pka1-repressed (glucose) and Pka1-induced (galactose) conditions at 16 h postinoculation (hpi). To account for any changes associated with the use of different carbon sources (glucose versus galactose) to repress or induce PKA1 expression, the WT strain was grown under the same conditions and comparisons were performed between strains from the same carbon source. We analyzed the samples using quantitative mass spectrometry and identified 3,222 proteins representing 48.1% of the 6,692 predicted proteins from the genome sequence (33). A total of 1,453 and 1,435 proteins were identified and quantified by dimethyl labeling in two or more replicates under Pka1-repressed and Pka1-induced conditions, respectively; 1,176 proteins were found under both conditions. A comparison of Gene Ontology (GO) biological-classification terms for the unique cellular proteins identified under either repressed conditions (277 proteins) or induced conditions (259 proteins) revealed changes in the proteome profiles in the strains responding to modulation of Pka1 activity (Fig. 1). The majority of proteins under both repressed and induced conditions (61% and 63%, respectively) were associated with metabolic, catabolic, and biosynthetic processes or were unknown or unclassified (including hypothetical proteins) (Fig. 1). Overall, the analysis of protein abundance in the PKA1-modulated strain revealed an impact of Pka1 on a diverse array of cellular processes.

• Open in new tab
• Download powerpoint

FIG 1 

Modulation of Pka1 expression changes the proteome profile of C. neoformans. The pie charts indicate unique proteins identified and quantified under Pka1-repressed conditions (glucose-containing medium) and Pka1-induced conditions (galactose-containing medium) and categorized according to their GO term biological classifications (shown on the right).

Identification of Pka1-regulated proteins.We next identified the proteins regulated by Pka1 within the set of 1,176 proteins that were quantified and shared under the repressed and induced conditions. We examined the list of shared proteins and found 302 that passed a nonadjusted Student’s t test with regard to altered abundance in response to Pka1 regulation (see Table S1 in the supplemental material). We also adjusted the significance levels to account for the multiple-testing hypothesis, and this more stringent analysis revealed 40 Pka1-regulated proteins with the highest statistical significance; these proteins are indicated with asterisks in the tables in this article. The broader set of 302 Pka1-regulated proteins identified in our analysis covered a diverse spectrum of GO term biological classifications (14 categories), including metabolic, catabolic, biosynthetic, and cellular processes, transcription and translation, proteolysis and protein folding, oxidation-reduction, response to stress, transport, phosphorylation/dephosphorylation, signaling and signal transduction, hypothetical proteins, and unknown/unclassified proteins. A density plot comparing the proteins identified under Pka1-repressed and Pka1-induced conditions highlighted the influence of modulation of Pka1 activity. Remarkably, this approach revealed that the majority of regulated proteins decreased in abundance upon induction of PKA1 (Fig. 2A). A closer inspection of the GO term biological processes confirmed that the majority of the proteins in all 14 categories showed a decrease in abundance upon PKA1 induction (Fig. 2B). In contrast, induction of PKA1 expression caused an increase in protein abundance for only a small number of proteins in the same categories of metabolic and biosynthetic processes, proteolysis and protein folding, and phosphorylation/dephosphorylation, as well as for hypothetical, unknown, or unclassified proteins (Fig. 2B). The overall patterns were supported by additional analyses performed using GO term classifications and enrichment, as well as KEGG pathway analysis (see Fig. S1 and Tables S1, S2, and S3 in the supplemental material).

• Open in new tab
• Download powerpoint

FIG 2 

Modulation of Pka1 expression influences the abundance of 302 proteins. (A) Density scatterplot of proteins whose abundance changed upon repression or induction of PKA1 expression. Proteins were identified in two or more replicates under Pka1-repressed (glucose) conditions and Pka1-induced (galactose) conditions. Normalized fold change values for the P_GAL7::PKA1 strain versus the WT strain are presented. Statistical analysis performed using a Student’s t test identified 302 proteins whose results were significantly different (P < 0.05) under the two sets of conditions (green labels). A subset of 40 of these proteins showed a significant difference following correction for multiple-hypothesis testing using the Benjamini-Hochberg method (FDR = 0.05) (red labels). (B) Numbers of Pka1-regulated proteins in different GO term categories that showed changes in abundance upon induction of PKA1 expression. The proteins were classified based on GO terms for biological processes.

Several clusters of interactions are predicted by network mapping of Pka1-regulated proteins.Given the diverse impacts of Pka1 on the C. neoformans proteome, including components of several metabolic pathways and cellular processes, we next evaluated the predicted interaction network of Pka1-regulated proteins using the STRING (Search Tool for Retrieval of Interacting Genes/Proteins) database. We mapped the identified Pka1-regulated proteins and found that 1,509 protein-protein interactions were predicted (Fig. 3). Several clusters of protein-protein interactions were prominent in the interaction network map, with the greatest number occurring among proteins associated with the ribosome and translation (Table 1; see also Table S3 in the supplemental material). A second interaction cluster included proteins associated with the ubiquitin-proteasome pathway (UPP). Although this category was noted in the KEGG analysis, the network mapping highlighted the prominence of these proteins and revealed their association in the network with the proteins for translation (Table 2). Additional clusters in the network were more diverse and included interactions between proteins associated with metabolic and biosynthetic processes, phosphorylation/dephosphorylation, oxidation-reduction, and other functions. Interactions were also predicted for proteins in categories associated with stress, signaling, secretion, and virulence (Table 3). This group also included proteins with functions in protein sorting and folding, e.g., an ortholog of the chaperone GrpE, a prefoldin alpha subunit, and a peptidyl-prolyl cis-trans-isomerase. With regard to virulence functions, our analysis revealed PKA regulation of the Cig1 protein (heme acquisition) and an acid phosphatase, both of which were previously observed in the PKA-regulated secretome, as well as chaperones, catalase, urease, and a superoxide dismutase (29, 34–37). Assays of the latter two enzymes and immunoblot analysis performed with anti-Hsp70 antibody supported the observed differences in the proteome data (see Fig. S2A, B, and C in the supplemental material). Similarly, our previous finding (that Pka1 induction reduced secreted urease activity) was consistent with the lower protein level for the enzyme observed in the current proteome analysis (Table 3) (28). Protein-protein interactions were not predicted for ~85 proteins in the data set. In general, the protein-protein interaction network for Pka1-regulated proteins emphasized the significant impact of Pka1 on the regulation of ribosomal proteins and translation, the UPP, and metabolism. Furthermore, the diverse classification of hypothetical proteins provided an opportunity for novel protein discovery in the context of PKA regulation of the proteome.

• Open in new tab
• Download powerpoint

FIG 3 

STRING analysis reveals an impact of Pka1 regulation on a diverse set of proteins as well as a prominent influence on translation and on the ubiquitin proteasome pathway. STRING was used to visualize predicted protein-protein interactions for the identified 300 Pka1-regulated proteins (http://string-db.org) using the corresponding proteins from C. neoformans strain JEC21 in the database (79). Several clusters were identified from the network mapping, and examples include (i) proteins associated with ribosomes and translation, including small-subunit ribosomal protein S13e (CNAG_01153) and eukaryotic translation initiation factor 3 subunit F (CNAG_06563); (ii) proteins associated with the proteasome, including 26S proteasome regulatory subunit N2 (CNAG_06175) and nascent polypeptide-associated complex subunit α (CNAG_04985); and (iii) diverse proteins associated with metabolism, phosphorylation, and virulence, including urease (CNAG_05540) and acetyl-coenzyme A (CoA) C-acetyltransferase (CNAG_02918). Nodes directly linked to the input are indicated with colors other than white, and nodes of a higher iteration/depth are indicated with white. Edges or predicted functional links consist of up to eight lines, with a different color representing each type of evidence (e.g., neighborhood, gene fusion, cooccurrence, coexpression, experiments, databases, text mining, and homology).

Characterization of Pka1-regulated novel proteins.Our quantitative proteomic analysis revealed that 64 (21%) of the 302 candidates for Pka1-regulated proteins were conserved hypothetical proteins. To better understand the potential functions of these proteins, we used bioinformatic analyses to predict subcellular location, secretion, protein function, and possible enzymatic classification and to identify homologues (see Table S4 in the supplemental material). The subcellular location prediction tools TargetP and TMHMM identified 14 proteins associated with the mitochondrion, along with one protein containing a transmembrane helix. We noted that mitochondrial functions also appeared prominently among the metabolic and biosynthetic proteins. None of the hypothetical proteins contained a predicted glycosylphosphatidylinositol (GPI) anchor, and only one contained an N-terminal signal peptide for conventional secretion. ProtFun identified proteins with cellular roles related to central intermediary metabolism, energy metabolism, amino acid biosynthesis, biosynthesis of cofactors, translation, regulation and transcription, regulatory functions, transport and binding, purines and pyrimidines, and the cell envelope. Forty candidate enzymes were identified, and eight were classifiable as lyases, three as ligases, and three as isomerases. Additional GO term categories were associated with growth factors, hormones, transcription and transcriptional regulation, immune response, voltage-gated ion channels, and structural proteins. Lastly, HMMER identified protein homologues, including NAD-, RNA-, and carbohydrate-binding proteins, enzymes (methyltransferase, glutathione-S-transferase, phosphatases, and a protein tyrosine kinase), a heat shock protein, translation/ribosomal proteins, and proteins associated with transcription and transcriptional repression. These results revealed a diverse classification of hypothetical proteins, along with the potential for novel protein discovery in the context of PKA regulation of the proteome. Interestingly, the two Hsp12 homologues that we identified are known to influence growth on amphotericin B and to be regulated by the cAMP and high-osmolarity glycerol (HOG) signaling pathways in C. neoformans (16, 38). Some of the identified proteins may also contribute to virulence and proliferation in vertebrate hosts. For example, this may be the case for CNAG_03007, which is a homologue of the CipC protein, because the corresponding transcript was previously found to be highly abundant in cryptococcal cells obtained from the central nervous system of infected rabbits (39).

Connections between inhibition of the proteasome, PKA, and capsule production.As mentioned above, we found that induction of Pka1 expression decreased the abundance of multiple components of the UPP in C. neoformans, perhaps as part of a PKA-regulated remodeling of proteostasis. Given that one of the most dramatic consequences of elevated PKA activity is the enlargement of the polysaccharide capsule, we investigated the impact of a proteasome inhibitor, bortezomib, on capsule production in the WT strain, the pka1Δ and pkr1Δ mutants, and the P_GAL7::PKA1 strain. This analysis revealed a reduction in the diameter of cell-associated capsule for the WT strain, the pkr1Δ mutant, and the induced P_GAL7::PKA1 strain in the presence of the inhibitor (Fig. 4A and B). The cell sizes of the pkr1Δ mutant and the induced P_GAL7::PKA1 strain were also decreased relative to the sizes of the WT cells upon treatment. No clear differences in capsule or cell size were observed in the absence of PKA1 expression using the pka1Δ mutant strain and the P_GAL7::PKA1 strain grown under the repressed condition. These findings are consistent with the previously observed changes in cell size upon modulation of PKA1 expression (28). We also grew the WT strain, the pkr1Δ mutant, and the P_GAL7::PKA1 strain in the presence of a range of bortezomib concentrations to identify the levels that were inhibitory for capsule formation (Fig. 4C). We found that capsule thickness was first noticeably impacted at 10 µM bortezomib for the WT strain grown under both glucose and galactose capsule-inducing conditions. Conversely, capsule thickness and cell size remained unchanged for the pkr1Δ mutant and the induced P_GAL7::PKA1 strain with a 10 µM concentration of the inhibitor. A reduction in capsule and cell size was noted for these strains once the concentration of inhibitor reached 15 to 20 µM, and a concentration of 25 µM completely eliminated capsule production. Overall, these results are compatible with the conclusion that the regulatory connection between the cAMP/PKA signaling pathway and capsule formation includes a role for the UPP.

• Open in new tab
• Download powerpoint

FIG 4 

Inhibitors of the proteasome or of translation or glycosylation reduce capsule formation. Capsule diameter was examined with differential interference contrast (DIC) microscopy and India ink staining. (A) Capsule formation in the WT strain and the P_GAL7::PKA1 strain grown in low-iron medium (LIM) {either with glucose [LIM (D)] or with galactose [LIM (G)]} in the absence or presence of bortezomib (BTZ; 50 µM). (B) Capsule formation in the pka1Δ and pkr1Δ mutants in LIM (D) in the absence or presence of bortezomib (50 µM). (C) Capsule formation was examined in the WT strain, the pkr1Δ mutant, and the induced P_GAL7::PKA1 strain grown in LIM (D) or LIM (G) in the absence or presence of a range of bortezomib concentrations: 0, 1, 5, 10, 15, 20, and 25 µM. Cells were incubated at 30°C for 48 h, mixed with India ink, and examined by DIC microscopy to visualize the capsule. (D) The WT strain was grown in LIM with or without 50 µM bortezomib (New England Biolabs) and cycloheximide (CHX; 1 µg/ml) for the times indicated. (E) The WT strain was grown in defined LIM in the absence or presence of tunicamycin (TM; 0.1 µg/ml) after incubation for 24 h at 30°C. The bar indicates 10 µm (×100 magnification).

It is interesting that proteasome inhibition reduced capsule formation whereas PKA activation causes an enlarged capsule. These results suggest that a balance in the levels of expression of the UPP must be achieved by regulated PKA activity, perhaps in conjunction with regulation of translation, to properly orchestrate capsule production. To explore this possibility, we tested the translation inhibitor cycloheximide for an influence on capsule formation and found that although cell growth and capsule formation were inhibited, the cells eventually recovered their ability to make capsule (Fig. 4D). Interestingly, the observed recovery of capsule formation occurred even in the presence of a concentration of bortezomib that otherwise prevented capsule formation. Although additional biochemical studies are needed, these observations support the conclusion that a balance between protein synthesis and degradation has an impact on capsule formation.

The proteasome inhibitor bortezomib influences the growth of C. neoformans strains with altered PKA expression or activity.We next determined the influence of bortezomib on the growth of the WT and the P_GAL7::PKA1 strains, and on the growth of the pka1Δ and pkr1Δ mutants, in liquid media in the presence or absence of exogenous cAMP. Initially, we observed that the regulated P_GAL7::PKA1 strain showed slower growth than the WT strain in galactose medium without bortezomib but that it eventually reached a similar culture density by 72 h (Fig. 5A). This result suggests that the activation of PKA1 expression negatively influences growth, possibly because of the reduction in the abundance of translation functions observed in the proteome analysis (Table 1). The addition of bortezomib negatively impacted the growth of both the P_GAL7::PKA1 and WT strains, although substantial growth still occurred. In contrast, addition of both bortezomib and cAMP completely blocked the growth of both strains. These results indicated that activation of PKA by addition of cAMP enhanced susceptibility to bortezomib even in the regulated strain upon galactose-induced expression of PKA1. Similar results were observed when the two strains were grown with glucose as the carbon source (Fig. 5B). The two strains had the same growth profile in the absence of bortezomib (with or without cAMP), but addition of the drug again extended the lag phase and allowed eventual growth at the level seen with the untreated cultures. The inclusion of cAMP with bortezomib again eliminated the growth of both strains. Addition of cAMP alone did not influence the growth of any of the strains (data not shown).

• Open in new tab
• Download powerpoint

FIG 5 

The proteasome inhibitor bortezomib impairs the growth of strains with altered PKA expression or activity. Growth assays were performed for the WT and P_GAL7::PKA1 strains (A and B) and the pka1Δ and pkr1Δ mutants (C and D) in liquid minimal medium with galactose (MM + G) (A and C) or glucose (MM + D) (B and D) at 37°C. Where indicated, bortezomib was added at 50 µM and cAMP was added at 5 mM. Each experiment was performed a minimum of three times in triplicate. The averages and standard errors of data from an experiment representative of three replicates are shown for each time point.

The influence of bortezomib was examined further by comparing the growth of a pka1Δ mutant lacking the main catalytic subunit to that of a pkr1Δ mutant lacking the regulatory (cAMP-binding) subunit of PKA (Fig. 5C and D). The levels of growth of these strains were similar on galactose (Fig. 5C) and glucose (Fig. 5D) without additions, but treatment with bortezomib alone or in combination with cAMP almost completely eliminated the growth of the pkr1Δ strain on either carbon source. In contrast, the pka1Δ strain in galactose or glucose media showed delayed growth on bortezomib, and the inclusion of cAMP caused a further delay. We note that the Pkr1 regulatory subunit and a second catalytic subunit, Pka2, are still present in the pka1Δ mutant. The responsiveness of the mutant may therefore indicate a contribution of Pka2 to the observed response to cAMP and bortezomib. Overall, these experiments demonstrated that bortezomib impairs the growth of C. neoformans and that activation of PKA by addition of exogenous cAMP or loss of the regulatory subunit of PKA increases susceptibility. In general, the regulated P_GAL7::PKA1 strain may not be fully activated for PKA in galactose or fully repressed by glucose because addition of cAMP caused further growth inhibition by bortezomib. Additionally, this strain showed poorer growth with bortezomib in galactose medium versus glucose medium, suggesting that activation of PKA1 expression increases susceptibility.

ER stress caused by tunicamycin impacts the growth of C. neoformans strains with altered PKA expression or activity.We next hypothesized that the observed influence of PKA on the abundance of proteins for translation and the UPP might be part of a larger impact of PKA activity on proteostasis and protein trafficking. We tested this theory by examining susceptibility to tunicamycin, an antibiotic that provokes ER stress by inhibiting the first step in the lipid-linked oligosaccharide pathway (40). Specifically, we tested the WT and the P_GAL7::PKA1 strains, as well as the pka1Δ and pkr1Δ mutants, in liquid media in the presence and absence of exogenous cAMP. Initially, we found that the regulated P_GAL7::PKA1 strain again showed slower growth than the WT strain on galactose medium (Fig. 6A) as shown in Fig. 5A. The influence of growth on galactose appeared more pronounced in this experiment, perhaps due to the lower growth temperature (30°C) that was employed. The lower temperature was used because we noted that tunicamycin had a more pronounced influence on susceptibility at 30°C whereas the impact of bortezomib was greater at 37°C. Tunicamycin impaired the growth of both the WT and P_GAL7::PKA1 strains on galactose, and addition of cAMP did not have a notable influence, although the growth of the strains was already minimal (Fig. 6A).

• Open in new tab
• Download powerpoint

FIG 6 

The glycosylation inhibitor tunicamycin impairs the growth of strains with altered PKA expression or activity. Growth assays were performed for the WT and P_GAL7::PKA1 strains (A and B) and the pka1Δ and pkr1Δ mutants (C and D) in liquid minimal medium with galactose (MM + G) (A and C) or glucose (MM + D) (B and D) at 30°C. Where indicated, tunicamycin was added at 0.5 µg/ml and cAMP was added at 1 mM. Each experiment was performed a minimum of three times in triplicate. The averages and standard errors of data from an experiment representative of three replicates are shown for each time point.

The WT and P_GAL7::PKA1 strains grew to similar levels on glucose, and addition of tunicamycin (with or without cAMP) had only a slight negative impact on the growth of the regulated P_GAL7::PKA1 strain but severely inhibited the WT strain (Fig. 6B). That is, a reduction in the expression of PKA1 under conditions of growth of the regulated strain in glucose eliminated the susceptibility to tunicamycin, suggesting that a low level of Pka1, perhaps resulting in low PKA activity, abrogated the ER stress provoked by the drug. In contrast, the growth of the WT strain was markedly impaired by addition of tunicamycin with or without cAMP. An influence of exogenous cAMP was not observed in this experiment, compared to similar treatment with bortezomib, and this may be because the growth of the WT strain was already quite impaired by treatment with tunicamycin. Importantly, a comparison of the levels of growth of the P_GAL7::PKA1 strain on galactose (Fig. 6A) versus glucose (Fig. 6B) revealed a much greater inhibitory influence of tunicamycin under the inducing condition with galactose. We also observed that tunicamycin treatment reduced capsule elaboration, as also seen with bortezomib and cycloheximide (Fig. 4E).

We also tested the growth of the pka1Δ and pkr1Δ mutants with or without tunicamycin and exogenous cAMP (Fig. 6C and D). Most notably, the pkr1Δ mutant was quite susceptible to tunicamycin, with similar levels of impairment seen with and without cAMP on galactose or glucose medium. The pka1Δ strain showed slightly impaired growth in the presence of tunicamycin, and addition of cAMP actually improved the growth of the strain on galactose but not on glucose (compare Fig. 6C to D). This result could again indicate a potential role for Pka2 and/or an influence of Pkr1. It was noteworthy that added cAMP had little influence on the susceptibility of the pka1Δ mutant compared with its influence on bortezomib susceptibility for this strain, while the pkr1Δ mutant was highly susceptible, as seen with bortezomib. In general, a reduction in PKA activity, in either the pka1Δ mutant or the P_GAL7::PKA1 strain grown on glucose, protected the strains against ER stress provoked by tunicamycin. These results are consistent with the observed high susceptibility of the pkr1Δ mutant to tunicamycin. Taken together, our results revealed that connections between ER stress caused by tunicamycin and the cAMP/PKA pathway exist through the influence of the Pkr1 regulatory subunit on PKA activity or as a consequence of an additional function of Pkr1 and through the expression or activity of the Pka1 protein. Finally, we tested whether there was synergy between tunicamycin and bortezomib by using checkerboard assays to examine drug susceptibility (41). We did not observe synergy for these drugs or for bortezomib and fluconazole, an antifungal drug known to provoke ER stress (42 and data not shown).

Phenotypes of deletion mutants encoding PKA-regulated proteins.We compared our list of Pka1-regulated proteins with the C. neoformans gene deletion set to gain a more in-depth appreciation of the impact of PKA1 modulation on the proteome (43, 44). In total, we linked our Pka1-regulated proteins with 59 deletion mutants that had previously been assessed for the production of virulence factors and growth at 37°C and that had been recently tested by chemical genetic analysis for sensitivity or resistance to small molecules (see Table S5 in the supplemental material) (43, 44). Of these 59 genes, the mutants encoded by 3 displayed a defect in melanin production, and the corresponding genes encoded a ubiquitin-like protein (CNAG_02827), a class E vacuolar protein-sorting machinery protein, HSE1 (CNAG_05882), and an uncharacterized protein (CNAG_01644) (43). These results are consistent with the known regulation of melanin production by PKA (19). Twenty-nine deletion mutants showed growth phenotypes, including sensitivity to H₂O₂, FeCl₃, sirolimus, and amphotericin B, upon challenge by small molecules. Several deletion strains displayed sensitivity to cell wall stressors, including SDS, Congo red, and caffeine, suggesting a role in cell wall integrity and potential capsule attachment. Mutants for several genes showed resistance to H₂O₂ or azole drugs such as fluconazole, suggesting roles for the proteins in protection against oxidative stress or ergosterol biosynthesis. These observations are interesting given the impact of redox activity on disulfide bond formation in the ER. Interestingly, three mutants showed an impact of ER stress triggered by tunicamycin. These mutants had defects in a translation-associated argonaute protein (CNAG_04609) or a glycosyl hydrolase (CNAG_05077) that enhanced susceptibility and in a phosphoketolase (CNAG_02230) that led to resistance. These observations further support the idea of a connection between PKA and ER function as well as of other potential links between PKA regulation and specific proteins that warrant further study.