The Aspergillus fumigatus Phosphoproteome Reveals Roles of High-Osmolarity Glycerol Mitogen-Activated Protein Kinases in Promoting Cell Wall Damage and Caspofungin Tolerance

Aspergillus fumigatus is an opportunistic human pathogen causing allergic reactions or systemic infections, such as invasive pulmonary aspergillosis in immunocompromised patients. The mitogen-activated protein kinase (MAPK) signaling pathways are essential for fungal adaptation to the human host. Fungal cell survival, fungicide tolerance, and virulence are highly dependent on the organization, composition, and function of the cell wall. Upon cell wall stress, MAPKs phosphorylate multiple target proteins involved in the remodeling of the cell wall. Here, we investigate the global phosphoproteome of the ΔsakA and ΔmpkC A. fumigatus and high-osmolarity glycerol (HOG) pathway MAPK mutants upon cell wall damage. This showed the involvement of the HOG pathway and identified novel protein kinases and transcription factors, which were confirmed by fungal genetics to be involved in promoting tolerance of cell wall damage. Our results provide understanding of how fungal signal transduction networks modulate the cell wall. This may also lead to the discovery of new fungicide drug targets to impact fungal cell wall function, fungicide tolerance, and virulence.

among the most recent antifungal drugs to be developed (3) and highlight that cellular integrity is a good target for further development of antifungals. Cell wall biosynthesis and the CWI pathway therefore provide excellent targets for further antifungal drug development.
The A. fumigatus CWI pathway has already been characterized, and it is mainly composed of the PkcA-MpkA-Rlm1 module (36,(41)(42)(43)(44)(45). Recently, we demonstrated that the HOG MAPs SakA and MpkC are independently required to mount an adequate response to cell wall damage (29). The ΔsakA and the double ΔsakA ΔmpkC mutants were more sensitive to osmotic and oxidative stresses and to cell wall-damaging agents. Both MpkC:GFP (green fluorescent protein) and SakA:GFP translocated to the nucleus upon osmotic and cell wall stress, with SakA:GFP showing a quicker response. The phosphorylation of MpkA from the CWI pathway in response to osmotic and cell wall stress was dependent on MpkC and SakA, suggesting that both the CWI and HOG pathways collaborate upon exposure to several types of stresses and during cell wall biosynthesis (29,34).
"Omics" high-throughput techniques such as genomics, transcriptomics, metabolomics, and proteomics have revolutionized biological research. Phosphoproteomics is a branch of proteomics important for the characterization of proteins that are modified posttranslationally by the addition of a phosphate group. Phosphorylation is a reversible protein modification that can regulate cell signaling networks through the regulation of protein function, subcellular localization, complex formation, and protein degradation (46). It is estimated that 30% to 65% of all proteins may be phosphorylated (46,47). Phosphoproteomics, in particular, has been utilized to elucidate novel signaling mechanisms (48). Phosphoproteomics provides information about which protein or pathway might be activated, since generally there is a correlation between a change in protein phosphorylation and protein activity. Phosphoproteomics has been used for understanding mitotic exit and cell wall integrity pathways in Saccharomyces cerevisiae (49,50); hyphal morphogenesis in Candida albicans (51); asexual development and multiple stress responses in Beauveria bassiana (52); cellular processes involved in morphology, secretion, protein kinase A regulation, and catabolite repression in A. nidulans (53,54); and the aflatoxin production in Aspergillus flavus (55). Here, we evaluate the global A. fumigatus phosphoproteome exposed to cell wall-damaging agents, identifying 485 proteins putatively involved in the cell wall stress response. Comparative phosphoproteome analyses of the wild-type (WT) and ΔsakA, ΔmpkC, and ΔsakA ΔmpkC mutant strains showed the involvement of the HOG pathway and identified novel protein kinases and transcription factors, which were confirmed by fungal genetics to be involved in promoting CWI signaling and tolerance of cell wall damage.

RESULTS
A. fumigatus phosphoproteome upon cell wall damage. A phosphoproteomics study was used as a first step to understand how protein phosphorylation is involved in the response to the cell wall damage. To achieve this, mycelia were collected from liquid minimal medium before (T0) and after adding 200 mM Congo red (CR) for 30 min (CR30). From the wild-type, ΔsakA, ΔmpkC, and ΔsakA ΔmpkC strains, at T0 and CR30, total protein was extracted, trypsin digested, subjected to TiO 2 phosphopeptide enrichment, and analyzed by liquid chromatography-tandem mass spectrometry (LC-MS/ MS), as illustrated in Fig. 1A.
Principal-component analyses of the wild-type phosphoproteome showed the cluster of untreated samples separated from those exposed to CR (Fig. 1B). A total of 4,499 phosphopeptides were identified with "localization probabilities" higher than 0.75 (75% [see Table S1A in the supplemental material]). A total of 756 phosphopeptides were found to be significantly different between the untreated T0 and the CR-treated CR30 sample (Student t test, P Ͻ 0.05), corresponding to 485 proteins (Fig. 1C). Among these phosphopeptides, three were identified only in T0 samples and 380 were exclusive to CR-treated samples while 373 were observed in both treatments ( Fig. 1E and Table S1B).
The distribution of phosphosites according to amino acid residues followed a pattern typical of eukaryotic organisms, with 99% occurring on serine and threonine and Ͻ1% on tyrosine residues ( Fig. 1F and Table S1B). Approximately half of the phosphoproteins contained a single phosphorylation site, and phosphoproteins containing one to three phosphosites accounted for the vast majority of phosphoproteins detected ( Fig. 1G and Table S1B). FunCat (https://sbi.hki-jena.de/fungifun/fungifun.php) analyses for the 485 proteins showed an enrichment for proteins involved in the regulation of C-compound, translation initiation, osmosensing and response, Ca 2ϩ -mediated signal transduction, MAPK kinase kinase (MAPKKK) cascades, stress responses, encoded protein kinases, actin cytoskeleton, and budding cell polarity and filament formation ( Fig. 2A). This implies that the phosphorylation response to cell wall damage involves stress responses, translation, protein kinases, and different MAPK and calcium-mediated signaling mechanisms. Aspergillus fumigatus Phosphoproteome ® Possible kinase groups targeting each of the 756 phosphorylation sites were predicted by Group-based Prediction System 3.0 (GPS; with medium-stringency cutoff) (Fig. 2B). Based on kinase-substrate relationships, seven serine/threonine kinase families accounted for most of the identified phosphorylation sites (Fig. 2B). The most frequent serine/threonine kinase families (more than 50%) were AGC, TK, and atypical, followed by the CMGC family (15%), which includes MAPKs, such as the HOG (p38) MAPK (Fig. 2B). For example, a frequently observed phosphoserine/threonine kinase was p38 (CMGC/MAPK/p38/Hog) (Fig. 2C). In fact, the analysis of the substrate sequences phosphorylated by p38 MAPKs indicates a high conservation of a serine (S) and a proline (P) in positions ϩ8 and ϩ9 (Ͼ90%) among the putative targets of p38 MAPK (Fig. 2C).
Protein kinases and transcription factors involved in the cell wall damage. We have visually inspected and identified the phosphopeptides that corresponded to the 32 protein kinases and 27 transcription factors (TFs) identified with increased phosphorylation upon CR exposure ( Fig. 2D and E; Table S1C and D). Eleven candidate protein kinase and 5 TF-encoding genes were randomly selected (Table 1), and it was verified whether their corresponding single-gene deletion mutants were more sensitive to cell wall-damaging agents CR, calcofluor white (CFW), and caspofungin and to high-sorbitol-induced osmotic stress. No significant growth differences were observed among these protein kinase null mutants compared to the wild-type strain ( Fig. S1A to D). However, the TF null mutants lacking the gene AFUB_004210, AFUB_009970, AFUB_019790, or AFUB_037220 were more sensitive to CR (Fig. 3A), while mutants lacking AFUB_004210, AFUB_009970, and AFUB_037220 were more sensitive to calcofluor white (CFW), than the wild-type strain (Fig. 3B). All the TF null mutants showed some extent of increased sensitivity to caspofungin (Fig. 3C). The TF null mutants for AFUB_009970 and AFUB_019790 were also more sensitive to osmotic stress (Fig. 3D).
None of the protein kinase null mutants lost the caspofungin paradoxical effect (CPE; the escape of A. fumigatus from caspofungin inhibition at concentrations above the MIC) (3). In contrast, three TF null mutants corresponding to AFUB_009970, AFUB_019790, and AFUB_037220 showed a significantly decreased CPE. Therefore, this TABLE 1 Genes encoding protein kinases and transcription factors that showed phosphorylation when the wild-type strain is exposed to Congo red for 30 min

Protein
Putative function Protein kinases HriA (AFUB_032300) Eukaryotic translation initiation factor 2 alpha kinase activity and role in G 1 cell cycle arrest in response to nitrogen starvation, negative regulation of cytoplasmic translational initiation in response to stress SrrB (AFUB_030660) Response regulator of a two-component phosphorelay system Kin1 (AFUB_010510) Serine/threonine protein kinase involved in regulation of exocytosis SnfA (AFUB_018770) Ortholog(s) has AMP-activated protein kinase activity, ADP ribosylation factor guanyl-nucleotide exchange factor activity CmkA (AFUB_029320) Calcium/calmodulin-dependent protein kinase Prr1 (AFUB_029820) Serine/threonine protein kinase; inhibits pheromone-induced signaling downstream of MAPK, possibly at the level of the Ste12p transcription factor PkhA (AFUB_036500) Phosphoinoside-dependent protein kinase CmkC (AFUB_053520) Calmodulin-dependent protein kinase activity KfsA (AFUB_014350) Checkpoint kinase; required for normal septation in conidiophores, not hyphae; present in hyphae and conidiophores at septa and cortex, not at spindle pole bodies Oca2 (AFUB_059390) Role in negative regulation of transcription by RNA polymerase II, regulation of nitrogen utilization, response to salt stress FpkA (AFUB_059090) Ribosomal S6 kinase (RSK) Transcription factors AFUB_004210 Unknown function AFUB_009270 Unknown function AsgA (AFUB_019790) Zinc cluster protein proposed to be a transcriptional regulator involved in the stress response AreB (AFUB_029020) GATA zinc finger transcription factor, leucine zipper motif, negative regulator of nitrogen catabolism FkhA (AFUB_037220) Forkhead family transcription factor; rate-limiting activator of replication origins; evolutionarily conserved regulator of lifespan; major role in expression of G 2 /M-phase genes phosphoproteomic study has facilitated the identification of novel protein kinases and TFs involved in osmotic, cell wall, and caspofungin stress responses. A. fumigatus HOG MAPKs are important for the phosphoproteomic response to cell wall damage. We previously demonstrated that the MpkC and SakA kinases of the , and sorbitol and grown for 5 days at 37°C. All plates were grown in triplicate, and the average Ϯ SD was plotted. A Student t test using Prism GraphPad (version 6) was applied to confirm the statistical significance difference between treatment and control (P value Ͻ 0.05, ***; Ͻ 0.01, **; Ͻ 0.001, *).

Aspergillus fumigatus Phosphoproteome
® HOG pathway play a role in the response to osmotic, oxidative, and cell wall stresses, in addition to influencing CWI MpkA signaling (29). To assess the involvement of MpkC and SakA in the modulation of phosphorylation cascades activated by cell wall damage, the phosphoproteome of the ΔsakA, ΔmpkC, and double ΔsakA ΔmpkC mutants was examined in the absence (T0) or presence of CR for 30 min (CR30).
When the CR30-treated ΔsakA mutant was compared to CR30-treated WT, a total of 4,835 phosphopeptides were identified (location probability Ͼ75% [Table S2A]); 409 phosphopeptides corresponding to 268 proteins show significant differences between mutant and WT (Student t test with P Ͻ 0.05) (Fig. 4A). A prevalence of phosphorylation events was observed (67 phosphopeptides with total phosphorylation exclusive for the mutant strain and 284 with increased phosphorylation), with only 33 phosphopeptides with total dephosphorylation (exclusive to WT) and 25 with decrease of phosphorylation in the mutant strain ( Fig. 4A and Table S2B). For the same comparison without CR treatment (T0), a total of 3,741 phosphopeptides were selected (location probability Ͼ75%) with only 27 phosphopeptides significantly different between ⌬sakA and WT strains (P Ͻ 0.05) (Table S3A and B).
For the comparison between ⌬mpkC and WT strains after 30-min CR incubation, 4,638 phosphopeptides were identified (location probability Ͼ75% [Table S4A]); 157 phosphopeptides corresponding to 118 proteins were observed with significant differences between mutant and WT (Student t test with P Ͻ 0.05) ( Fig. 4A and Table S4B). The phosphorylation profile of the ΔmpkC strain was the opposite of ΔsakA strain, with a major reduction of phosphorylation (37 phosphopeptides with total dephosphorylation, 94 with decreased phosphorylation), and 6 phosphopeptides with increased phosphorylation and 20 with total phosphorylation (Table S4B). For the comparison between the controls (T0), a total of 3,675 phosphopeptides were identified (location probability Ͼ75%), with 157 phosphopeptides significantly differing between ⌬mpkC and WT strains (P Ͻ 0.05) (Table S5A and B).
In the ⌬mpkC ⌬sakA response to CR for 30 min, 4,855 phosphopeptides were identified (localization probability Ͼ75%), and 142 phosphopeptides corresponding to 111 proteins have significant phosphorylation change between mutant and WT ( Fig. 4C and Table S6A and B). The double mutant also showed a profile of increased phosphorylation, as observed for the WT and ⌬sakA strains in the CR 30-min response, with 26 phosphopeptides with total dephosphorylation, 12 with decreased phosphorylation, 70 with increased phosphorylation, and 34 with total phosphorylation (Fig. 4C). For the comparison without CR (T0), 3,741 phosphopeptides were observed, with only 54 significantly different between mutant and WT (Table S7A and B).
To determine Gene Ontology (GO) categories overrepresented in the mutant phosphoproteomes, we carried out a functional categorization by FungiFun analysis (Fig. 4D to F). For all mutants, at least 65% of the phosphopeptides have known function and less than 35% are hypothetical proteins (Fig. 4D). For ⌬sakA mutants, most of the proteins with increased phosphorylation upon CR exposure are related to protein binding, epidermal growth factor (EGF) receptor signaling pathway, DNA conformation modification, and DNA or RNA binding. For the group of proteins with decreased phosphorylation upon CR exposure, most of them are protein kinases and proteins related to modification by phosphorylation (phosphorylation, dephosphorylation, or autophosphorylation) (Fig. 4E). For the ⌬mpkC mutant, proteins with increased phosphorylation are mainly related to RNA topology, RNA binding, and rRNA processing, while proteins with decreased phosphorylation are related to cytoskeleton, such as actin cytoskeleton; budding, cell polarity, and filament formation; and spindle pole body/centrosome and microtubule cycle (Fig. 4F). The functional analysis of the double mutant phosphoproteome in CR response by FungiFun did not show enrichment of any pathway (data not shown). However, by using GO enrichment, we were able to identify phosphorylation of proteins related to metabolism, signal transduction, transcription/ translation, cell cycle, cytoskeleton, and protein degradation. Enrichment for protein kinases/phosphatases and proteins related to intracellular transport, for example, has been identified in the double mutant with both increased and decreased phosphorylation (Fig. 4G).
A Venn diagram was generated for correlating the phosphorylation profile between each strain treated with CR (Fig. 4H). Although each strain had a unique set of phosphoproteins (245 for WT, 110 for ΔsakA mutant, 21 for ΔmpkC mutant, and 32 for All of them showed more than 60% characterized proteins. (E to G) Functional categorization of approximately 75% of proteins with known function differentially phosphorylated in ⌬sakA, ⌬mpkC, and ⌬sakA ⌬mpkC strains during CR stress. (H) Venn diagram of total differentially phosphorylated proteins (n ϭ 621) representing the number of common phosphoproteins between strains in each intersection (each strain has a set of differentially phosphorylated proteins after CR treatment compared to its untreated control). Thus, the WT strain has a total set of 431 phosphoproteins from which a unique set of 245 is not shared with any other condition, 69 are shared only with the ⌬sakA mutant, 13 are shared only with the ⌬sakA ⌬mpkC mutant, and 38 are shared only with the ⌬mpkC mutant, while 16 phosphoproteins are shared simultaneously by the four strains.
Aspergillus fumigatus Phosphoproteome ® ΔsakA ΔmpkC mutant), among the three mutant strains the ΔsakA strain showed the highest similarity to the wild-type phosphoproteome during CR exposure, with 69 shared phosphoproteins. The low number of common phosphoproteins among all conditions (16 phosphoproteins) in addition to the low number of common phosphoproteins between the ΔmpkC and the double ΔsakA ΔmpkC mutants (only 2 phosphoproteins) may indicate their independent roles in distinct pathways involved in the CR response.
Interaction networks of identified A. fumigatus phosphoproteins. Pairwise analyses identified the differentially phosphorylated proteins after exposure to CR in the different strains. This showed that the phosphoproteomes of wild-type and ΔsakA strains were functionally enriched for serine/threonine protein kinase domains and RNA recognition motif domains, respectively (see Table S1E for less representative functional enrichments). However, the phosphoproteomes for the ΔmpkC and the double mutant ΔsakA ΔmpkC strains did not have any functional enrichment. The individual sets of differentially phosphorylated proteins in wild-type, ΔsakA, ΔmpkC, and the double mutant ΔsakA ΔmpkC strains were combined, yielding 595 proteins which were used to generate a single functional protein association network. The network was composed of 336 nodes and 1862 interactions between nodes (Fig. 5), revealing the enrichment for protein kinases, RNA recognition motif domains, and MAPK signaling pathways (Table S1E). The change of the color pattern in the nodes of Fig. 5A to D highlights the overall network rewiring in terms of differentially phosphorylated proteins for each strain (WT and ΔsakA, ΔmpkC, and ΔsakA ΔmpkC mutants) in the presence of CR compared to their respective controls (untreated).
Protein kinases and phosphatases play essential roles in eukaryotic signal transduction pathways. The network analysis showed enrichment for protein kinase and phosphatase domains; hence, a subnetwork highlighting only differentially phosphorylated kinases and phosphatases was extracted ( Fig. 6 and Table S1E). The phosphorylation profiles among strains after exposure to CR changed dramatically. In the wild-type strain (Fig. 6A), most of the kinases and phosphatases in the network showed increased phosphorylation compared to the untreated control, including all of the highly connected kinases except for the mitogen-activated protein kinase MpkB (AFUA_6G12820). There was an overall decrease of differentially phosphorylated kinases and phosphatases in ΔmpkC, ΔsakA, and ΔsakA ΔmpkC strains in comparison to the wild-type CR response ( Fig. 6B to D). In the ΔmpkC strain, MpkB was the only kinase with increased phosphorylation (Fig. 6B). In contrast to the other strains, the ΔsakA mutant presented a higher abundance of kinases and phosphatases with decreased phosphorylation after CR exposure (Fig. 6C). The double ΔsakA ΔmpkC mutant presented a specific set of differentially phosphorylated proteins, which were not modulated in the same manner in any of the other networks (Fig. 6D). This included the increased phosphorylation of the serine/threonine protein kinase SchA (AFUA_1G06400) and protein phosphatase 2C PtcA (AFUA_1G15800), plus the decreased phosphorylation of the MAP kinase kinase kinase BckA (AFUA_3G11080).
Phosphosite validation. Two approaches were adopted to determine the effects of the identified protein phosphorylation events on A. fumigatus CWI: (i) the solved protein structures of available homologues deposited in the Protein Data Bank (PDB) were assessed and (ii) serine/tyrosine residues were mutated to alanine/phenylalanine, or to phosphomimetic aspartate and functionality was assessed. These phosphomutations were introduced by using a CRISPR-Cas9 system. Only a single three-dimensional (3D) structure was available for RckA Rck2p , since no solved structures for the phosphorylated regions of the other chosen proteins were deposited in the PDB. We concentrated our attention on a few proteins, aiming to validate the results obtained by the phosphoproteomics analysis (Fig. 7). Therefore, RckA, MpkB, and SrrB were selected for functional characterization to validate the phosphoproteomics observations. RckA Rck2p (AFUA_2G03490) is the homologue of the S. cerevisiae Rck2p calcium/ calmodulin-dependent protein kinase involved in response to oxidative and osmotic stresses. A single phosphorylation site was identified in RckA (477-MDFRSPGAI-485), which was phosphorylated only when the wild type was exposed to CR ( Fig. 7A and B). A BLASTP search performed for RCK1 homologues with solved structures deposited in the PDB identified a calcium/calmodulin-dependent serine/threonine kinase from The whole set of differentially phosphorylated proteins from the wild-type (WT), ΔsakA, ΔmpkC, and ΔsakA ΔmpkC strains during CR stress was combined for the generation of a general protein association network. Each node represents a protein that is differentially phosphorylated in at least one strain under CR stress. Node colors are divided into gray nodes, for proteins which were not differentially phosphorylated, and heatmap-colored nodes (blue-white-red), for proteins which were differentially phosphorylated in each strain in relation to its control (the same strain without any stress condition). Node shapes represent molecular functions and are divided into triangles (phosphatases), squares (kinases), and circles (other functions). Each edge represents a functional protein association retrieved from the STRING server (medium confidence threshold of 0.4 for the interaction score), and node sizes represent the degree of each node (number of edges connected to the node). It is important to note that not all the differentially phosphorylated proteins are present in the network, as many proteins did not present any functional associations within the whole set. Changes in the color patterns of nodes from panels A to D represent changes in the differential phosphorylation of the proteins comprising the network in WT, ΔsakA, ΔmpkC, and ΔsakA ΔmpkC strains, respectively, during CR stress.
Aspergillus fumigatus Phosphoproteome ® Cryptosporidium parvum. The crystal structure of full-length CpCDPK3 (cgd5_820; PDB accession no. 3LIJ) protein in complex with Ca 2ϩ and AMP-PNP (adenylylimidodiphosphate), at 1.9-Å resolution, was used as a template to build three homology models for RckA with respect to the wild-type S481 residue (corresponding to S309 in  (CR 10 min). The subset of differentially phosphorylated kinases and phosphatases from the wild-type (WT), ΔsakA, ΔmpkC, and ΔsakA ΔmpkC strains during CR stress was extracted from the general protein association network. Each node represents a protein that was differentially phosphorylated in at least one strain under CR stress. Node colors are divided into gray nodes (proteins that were not differentially phosphorylated) and heatmap-colored nodes (blue-white-red; proteins which were differentially phosphorylated in each strain in relation to its untreated control). Node shapes represent molecular functions and are divided into squares (kinases) and triangles (phosphatases). Each edge represents a functional protein association retrieved from the STRING server (medium confidence threshold of 0.4 for the interaction score). Node size is representative of the number of edges connected to the node. Note that not all the differentially phosphorylated proteins are present in the network, as for many proteins no functional associations were available. Changes in the color patterns of nodes from panels A to D represent changes in the differential phosphorylation of the proteins comprising the network in WT, ΔsakA, ΔmpkC, and ΔsakA ΔmpkC strains, respectively, during CR stress. C. parvum) and the two phosphomutations on the S481 residue (S481A and S481D). The pairwise alignments used for homology modeling were refined by structural inspection to avoid gaps (by deletion or insertion) inside secondary structure elements. Models were analyzed to reveal satisfactory atomic contacts, chemical environment, and stereochemistry as well as high-resolution structures (above 2.0 Å), and they indicate that residue S481 is inserted at the middle of the short loop of the calcium-binding EF-hand motif (Fig. 8A).
The three RckA models with different S481 residues were used in long-trajectory molecular dynamics simulations, to assess differences in structural stability and conformation. RckA models containing S481 and S481A were more stable, while the highest main chain conformational change was observed in the S481D model, whose aspartic residue moves from a buried to a solvent-accessible conformation, thus providing an additional calcium-binding residue. This protein shows different patterns of calcium binding and protein activity from the other two (Fig. 8A).
The ΔrckA, RckA S481D , and RckA S481A strains have growth rates and conidiation comparable to the wild-type strain (Fig. 8B). The ΔrckA and RckA S481A strains were more sensitive to CR-and CFW-induced cell wall damage, while the RckA S481D strain was more resistant to CFW (Fig. 8C and D). Surprisingly, all three RckA mutant strains were more sensitive to caspofungin and completely or partially lost the CPE (Fig. 8E). There were no differences in osmotic stress tolerance among the wild-type, ΔrckA, and RckA S481A strains. However, the RckA S481D strain was more resistant to high sorbitol concentrations than the wild-type or other mutant strains (Fig. 8F). MpkA phosphory- lation in the ΔrckA mutant upon exposure to CR was similar to the wild-type strain (Fig. 8G). The RckA S481A mutant had increased MpkA phosphorylation levels that decreased after exposure to CR for 10 and 30 min but increased about twice at 60 min of exposure (Fig. 8G). The RckA S481D mutant has higher levels of MpkA phosphorylation than the wild-type strain upon exposure to CR (Fig. 8G). Taken together, the data suggest that rckA mutants impact MpkA phosphorylation upon cell wall damage, strongly suggesting that RckA is important for the CWI pathway.
MpkB (AFUA_6G12820) is the homologue of S. cerevisiae Fus3p mitogen-activated serine/threonine protein kinase, involved in mating. A single phosphorylation site was identified in MpkB (179-FMTEYVATR-187) that was phosphorylated only when ΔmpkB was exposed to CR ( Fig. 7C and D). The ΔmpkB strain has reduced growth and conidiation (28), while the MpkB Y481F mutant has similar growth rates and conidiation as the wild-type strain (Fig. 9A). The ΔmpkB and MpkB Y183F strains were more sensitive to CR, CFW, caspofungin (also exhibiting reduction in the CPE), and osmotic stress (Fig. 9B). MpkA phosphorylation in the ΔmpkB mutant was 10-fold less than for the wild type after 10-min exposure to CR (Fig. 9D), while in the MpkB Y183F mutant MpkA phosphorylation was 4-fold higher in untreated cells, but 2-fold lower after 10-min exposure to CR, than in the wild-type strain (Fig. 9D). After many attempts, we were not able to construct an MpkB Y183D mutant.
SrrB (AFUA_2G15010) is a putative response regulator of a two-component phosphorelay system. Two phosphorylation sites (1462-ARSQTIGSQGE-1472) were identified as phosphorylated in both wild-type and ΔsakA strains exposed to CR ( Fig. 7E and F). The SrrB S1466 -1469A mutant had comparable growth rates and conidiation as the wild-type strain (Fig. 9A). The SrrB S1466 -1469A strain was more sensitive to CR, CFW, and caspofungin (with a reduction in the CPE) and had growth comparable to the wild-type strain under high-osmotic-stress conditions (Fig. 9C). MpkA phosphorylation in the SrrB S1466 -1469A mutant upon exposure to CR for 10 min was 3-fold less than in the wild-type strain (Fig. 9D).
Taken together, these results suggest that the RckA S481A mutation mimics the loss-of-function mutation, while the RckA S481D mutation allows the mutant strain to be more efficient in coping with cell wall and osmotic stress conditions. However, for all the other phosphomutants there was an increase in the sensitivity to cell walldamaging agents and a reduction in the MpkA phosphorylation upon CR stress, suggesting these phosphosites could be important for the MpkA modulation and CWI pathway regulation.

DISCUSSION
The filamentous fungal cell wall is essential for survival and serves as a primary means by which fungi interact with their environment. However, relatively limited information is available regarding the major regulatory events involved in cell wall biosynthesis and repair. The A. fumigatus module PkcA-MpkA-RlmA is essential for the CWI (36,(41)(42)(43)(44)(45). To learn more about the main regulatory posttranslational events that occur during CWI activation, we have investigated the global phosphoproteomics of A. fumigatus during cell wall damage caused by CR. A total of 756 high-confidence phosphorylation sites were identified in the A. fumigatus wild-type strain. A total of 380 were exclusive to CR-treated samples, and 3 were observed only in the control, while 373 were observed in the presence and absence of CR. These phosphopeptides correspond to 485 proteins. According to FunCat annotation enrichment and the interaction network analyses of this novel phosphoproteome, the proteins phosphorylated in response to CR were involved in a variety of important biological processes, including signal transduction (osmosensing and response, Ca 2ϩ -mediated signal transduction, and MAPKKK cascade), stress response, protein kinases, actin cytoskeleton, and budding cell polarity and filament formation. These results emphasize the importance of osmotic stress, calcium metabolism, and MAPKs in the A. fumigatus CWI pathway, as previously demonstrated (29,34,(78)(79)(80)(81)(82).
We also investigated the role played by the A. fumigatus HOG MAPKs MpkC and SakA. These two MAPKs are important for the MpkA phosphorylation during cell wall damage (29). Recently, SakA was demonstrated to physically associate with MpkC upon osmotic and CR-induced cell wall stresses (34). We were not able to identify a phosphorylation pattern that explains how MpkC modulates SakA function. Under cell wall stress, SakA also physically interacts with the CWI MAPK, MpkA, and the PtcB MAPK phosphatase, suggesting a close connection between the HOG and CWI pathways. Cross-talk between the HOG and CWI pathways has been observed in several fungal species, including S. cerevisiae, Candida albicans, Fusarium graminearum, and Pyrenophora graminea (56)(57)(58)(59)(60)(61)(62)(63)(64). Now, we show that the absence of these two HOG pathwayassociated MAPKs, MpkC and SakA, abolishes the remodeling of the phosphoproteome seen in the wild-type fungus in response to cell wall damage, and we also reveal novel proteins potentially involved in cell wall stress response. Our global phosphoproteome network analysis showed an enrichment for protein kinases, RNA recognition motif domains, and the MAPK signaling pathway. In contrast to the wild-type strain, there was an overall decrease of differentially phosphorylated kinases and phosphatases, among them RckA Rck2p , MpkB Fus3p , and SrrB, in ΔmpkC, ΔsakA, and ΔsakA ΔmpkC strains. We introduced phosphomutations in these proteins and assessed their tolerance of cell wall-damaging agents and osmotic stress conditions. Although it is very difficult to assign a single role for a specific protein kinase or phosphatase in the CWI pathway, the validation of our results by introducing corresponding phosphomutations in two protein kinases (RckA and MpkB) and a putative response regulator of a two-component phosphorelay system (SrrB) strongly indicated the importance of the osmotic stress and the cross-talk among HOG and MpkB Fus3p pathways in the CWI pathway. This is reflected in the fact that most of the A. fumigatus putative homologues of the components of the three MAPK pathways were differentially phosphorylated or dephosphorylated in the presence or absence of cell wall damage (Fig. 10).
We discovered several protein kinases and TFs involved and validated their importance in the A. fumigatus cell wall damage response. None of the 11 kinase null mutants were found to be more sensitive or resistant to several cell wall or osmotic stresses. Hence, further experimental validation will help to determine which kinases are important to the osmotic stress and cell wall damage response pathways. In contrast, all five TF null mutants showed cell wall stress-related phenotypes. This included the AreB (AFUB_029020) GATA zinc finger transcription factor, which negatively regulates nitrogen catabolism (65). The A. nidulans areB null mutant has significant phenotypic effects on the utilization of specific carbon sources, confirming its role in the regulation of carbon metabolism (66). Considering the importance of carbon source assimilation in the remodeling of the cell wall could explain why in A. fumigatus AreB is important for the cell wall damage response. Another TF was AsgA (AFUB_019790), the homologue of S. cerevisiae Asg1p, a zinc cluster protein proposed to be a transcriptional regulator involved in the stress response. The S. cerevisiae Δasg1 mutant has a respiratory defect and is sensitive to CFW (67). It is possible that A. fumigatus AsgA is similarly involved in some aspects of the cell wall damage response. In S. cerevisiae, the forkhead family transcription factor Fkh2 is involved in the activation of early-firing replication (68) and Fkh2 mutations suppress the checkpoint-defective phenotype of both HOG and cell wall integrity MAPK mutations (63). Therefore, the A. fumigatus homologue, FkhA, could be a putative CWI checkpoint for cell cycle progression in A. fumigatus.
Our global phosphoproteome analysis of the fungal response to cell wall damage has provided the identification of several novel proteins that are important for cell wall stress tolerance. These results will provide opportunities to understand how the signal Aspergillus fumigatus Phosphoproteome ® transduction networks that modulate the integrity of the cell wall function. In addition, our results highlight potential new targets for interference in fungal signaling events and cell wall remodeling to promote the efficacy of existing antifungals while impeding virulence.
Fungal growth, protein extraction, digestion, and phosphoenrichment. A. fumigatus conidia (1 ϫ 10 7 ) from the wild-type (WT) and mutant strains (⌬sakA, ⌬mpkC, and ⌬sakA ⌬mpkC strains [ Table 2]) (72) were inoculated in YG medium {0.5% yeast extract, 1% dextrose; 0.1% trace elements [22.0 g/liter ZnSO 4 , 11 g/liter boric acid, 5 g/liter MnCl 2 , 5 g/liter FeSO 4 , 1.6 g/liter CoCl 2 , 1.6 g/liter CuSO 4 , 1.1 g/liter (NH4) 2 MoO 4 ]} and incubated for 16 h under 200-rpm rotation and 37°C. After that, 300 g/ml of Congo red (CR) was added, followed by incubation for 10 min (or time zero, as control [T0]). Mycelia were then filtered using a vacuum system and frozen by liquid nitrogen. Frozen mycelia were macerated, and protein was extracted by the addition of TNE buffer (50 mM Tris-HCl [pH 7.5], 140 mM NaCl, 5 mM EDTA, phosphatase inhibitor cocktail [PhosphoStop; Roche], EDTA-free protease inhibitor cocktail [Roche], and 0.1 mM phenylmethylsulfonyl fluoride [PMSF]) and incubation for 15 min under agitation following centrifugation at 10,000 ϫ g for 20 min. The supernatant was collected, and total protein was quantified by the Bradford assay. Five hundred micrograms of protein was purified by 3-kDa-molecular-weight- The gradient elution started at 5% mobile phase B, was increased to 35% at 60 min and 80% at 65 min, and was held at 80% for 5 min before a 25-min reequilibration at 5%. The eluting peptides were interrogated with an Orbitrap Fusion mass spectrometer running a data-dependent LC-MS/MS method with the Top Speed decision selection. FTMS1 (Fourier Transform Mass Spectrometry) spectra were collected using the following parameters: scan range 350 to 1,800 m/z, resolving power 120k, automatic gain control (AGC) target 4E5, and maximum injection time of 50 ms. ITMS2 (Ion Trap Tandem Mass Spectrometry) spectra were collected using the following parameters: rapid scan rate, collisioninduced dissociation (CID) NCE (normalized collision energy) 35, 1.6 m/z isolation window, AGC target 1E4, and maximum injection time of 50 ms. MS2 (Mass Spectrometry) precursors were selected for a 3-s cycle. Precursors with an assigned monoisotopic m/z and a charge state of 2 to 7 were interrogated. Precursors were filtered using a 60-s dynamic exclusion window.
Phosphopeptide identification and protein analysis. Raw mass spectrometric data were analyzed in the MaxQuant environment (version 1.6.0.16). The MS/MS spectra were matched against the Aspergillus fumigatus UniProt FASTA database. Enzyme specificity was set to trypsin, and the search included cysteine carbamidomethylation as a fixed modification and N-acetylation of protein, oxidation of methionine, and/or phosphorylation of Ser, Thr, and Tyr residues (STY) as variable modifications. Up to two missed cleavages were allowed for protease digestion, and peptides had to be fully tryptic.
The analyses of phosphopeptide, phosphorylation sites, and number of phosphoresidues were performed using filter tools in the software Microsoft Office Excel. Student's t test (two-tailed, 95% [P Ͻ 0.05]) was performed for identification of differentially phosphorylated peptides between CR and T0 wild-type samples, CR mutant and CR WT samples, or T0 mutant and T0 wild-type samples. The total phosphopeptides identified were used for kinase site prediction by the software GPS 3.0, according to the default parameters, including medium threshold. For analysis of conservation between kinase family  Table S8A and B in the supplemental material.
The single guide RNAs (sgRNAs) were linked into the CRISPR-Cas9 vector pFC330 according to the tails of two primers. Previously, PCR fragments were amplified using the Pfu X7 polymerase with the primers from Table S8A and using pFC902 as a template. The "biobrick" fragments were further fused to the digested pFC330 by uracil-specific excision reagent (USER) fusion and USER cloning (74). Plasmids were purified with the GenElute plasmid miniprep kit (Sigma-Aldrich) and confirmed by restriction analysis. The selection of the transformants was performed using MM as selective medium, because the CRISPR-Cas9 vector (pFC330) contains pyrG as an auxotrophic marker.
Protoplastization and genetic transformation. Protoplastization of A. fumigatus was achieved using protocols previously described (73,75). For transformation, 5 to 10 l of the 100 M stock donor oligonucleotide solution (Table S2)  . The protoplasts were plated on osmotic and selective TM medium and incubated at 37°C until transformants appeared on the transformation plates (approximately 1 week). Transformants were restreaked four times in MM, genomic DNA was extracted, and the PCR fragments for each gene were sent for DNA sequencing to confirm the presence of the respective mutation (the list of primers used for sequencing is in Table S8A).
Protein-protein network analysis. For network analysis, phosphoproteomic data were filtered for differences in phosphorylation level of Ͼ2-fold or ϽϪ2-fold and P value of Ͼ0.05. Next, data were filtered and analyzed through ad hoc R scripts (total number of differentially expressed phosphoproteins ϭ 621) for the generation of Venn diagrams (VennDiagram package obtained from the following CRAN: https://cran.r-project.org/web/packages/VennDiagram/index.html) representing the intersections of the differentially phosphorylated proteins between the previously described studied strains. The list of differentially phosphorylated proteins and functional enrichment was obtained using the STRING server (https://string-db.org/), medium confidence threshold of 0.4 for interaction scores, using Aspergillus fumigatus as a model organism, and the generated functional protein association network presented a total of 336 nodes and 1,862 interactions as 26 phosphoproteins could not be found on the STRING server and the other 259 phosphoproteins did not present any association scores within the given set. The association network was then reprocessed through ad hoc R scripts for acquiring only "AFUA_" IDs for each of its nodes and was finally plotted/analyzed in Cytoscape software (https:// cytoscape.org/). For network visualization, the fold change of phosphorylation levels was transformed to log 2 and loaded in Cytoscape for color mapping as a heatmap scale (blue-white-red), and the node degree was mapped to node size. Node shapes were also mapped to protein functional domains representing three classes: kinases, phosphatases, and other protein domains.
Sequence alignment and homology modeling procedures. The search for RCK1 homologues with solved structures deposited in the Protein Data Bank (PDB; http://www.rcsb.org) was performed using the web server BLASTp (Basic Local Alignment Search Tool, for proteins) algorithm (https://blast.ncbi .nlm.nih.gov/Blast.cgi?PAGEϭProteins). A protein with the PDB ID 3LIJ (crystal structure of full-length CpCDPK3 [cgd5_820] in complex with Ca 2ϩ and AMP-PNP, at 1.9-Å resolution-a calcium/calmodulindependent serine/threonine kinase from Cryptosporidium parvum) was detected and here used as a template to build three homology models for RCK1, with respect to the wild-type residue (S481) and two mutations on residue 481 (S481A and S481D), using Modeller v. 9.2 (76). The pairwise alignments used as input for Modeller were previously performed using the web server Clustal Omega (v. 1.2.4; https:// www.ebi.ac.uk/Tools/msa/clustalo/) and subsequently refined by structural inspection of the template, thus avoiding gaps (by deletion or insertion) inside secondary structure elements. Models were analyzed/ validated regarding atomic contacts, chemical environment, and stereochemistry, using the web server SAVES v.5.0, at http://servicesn.mbi.ucla.edu/SAVES/.
MD procedures. Molecular dynamics (MD) calculations were carried out using the Biovia Discovery Studio software (77), founded on the CHARMm molecular mechanics force field engine. We used the standard dynamics cascade. Molecular dynamics was done with no restraints using as input the homology models built for RCK1 (Afu2g03490), a calcium/calmodulin-dependent protein kinase from A. fumigatus, with respect to two mutations on residue S481 (S481A and S481D). At the first step, energy minimization was carried out using steepest descent with 100 maximum steps and a root mean square (RMS) gradient of 1.0. At the second step of the MD protocol, energy minimization was performed setting the conjugated gradient as algorithm, with 500 maximum steps and an RMS gradient of 0.1. For the heating step, we used a simulation time of 50 ps, with 2-ps time steps and initial and target temperatures of 50 and 300 K, respectively. The equilibration phase was carried out with 100 ps and target temperature of 300 K. On the production step, a simulation time of 1,000,000 ps (1,000 ns) was considered, saving results at 1,000-ps intervals (snapshots). We used an implicit solvent dielectric constant of 80 (simulation in water), with a nonbond higher cutoff distance of 12.0 and nonbond lower cutoff distance of 10.0, with a leapfrog Verlet dynamics integrator.
Phenotypic characterization of A. fumigatus null mutant strains. A. fumigatus conidia (10 5 ) of wild-type (CEA17) and null mutant strains (Table 2) were inoculated on solid minimal medium (MM; 1% glucose, 6 g/liter NaNO 3 , 0.52 g/liter KCl, 1.52 g/liter KH 2 PO 4 , 0.52 g/liter MgSO4·7H 2 O, 1 ml/liter of trace elements, pH 6.5, 2% agar) with different concentration of CR, CFW, sorbitol, and caspofungin and grown for 5 days at 37°C. The radial growth of each strain in the presence of the molecules was normalized by the growth of each strain in MM without any molecule. All plates were grown in triplicate, and the average Ϯ SD was plotted. A Student t test using Prism GraphPad (version 6) was applied to confirm the statistical significance difference between treatment and control (P value Ͻ 0.05).

SUPPLEMENTAL MATERIAL
Supplemental material is available online only.