Receptor-mediated signaling in Aspergillus fumigatus

Aspergillus fumigatus is the most pathogenic species among the Aspergilli, and the major fungal agent of human pulmonary infection. To prosper in diverse ecological niches, Aspergilli have evolved numerous mechanisms for adaptive gene regulation, some of which are also crucial for mammalian infection. Among the molecules which govern such responses, integral membrane receptors are thought to be the most amenable to therapeutic modulation. This is due to the localization of these molecular sensors at the periphery of the fungal cell, and to the prevalence of small molecules and licensed drugs which target receptor-mediated signaling in higher eukaryotic cells. In this review we highlight the progress made in characterizing receptor-mediated environmental adaptation in A. fumigatus and its relevance for pathogenicity in mammals. By presenting a first genomic survey of integral membrane proteins in this organism, we highlight an abundance of putative seven transmembrane domain (7TMD) receptors, the majority of which remain uncharacterized. Given the dependency of A. fumigatus upon stress adaptation for colonization and infection of mammalian hosts, and the merits of targeting receptor-mediated signaling as an antifungal strategy, a closer scrutiny of sensory perception and signal transduction in this organism is warranted.


INTRODUCTION
The genus Aspergillus is comprised of environmental filamentous mold fungi which utilize decaying organic matter for metabolic energy and nutrition. Aspergillus fumigatus is the most pathogenic, and is commonly isolated as an agent of human pulmonary infections (Dagenais and Keller, 2009). In healthy individuals, mucociliary clearance and pulmonary immune defences clear the hundreds of conidia inhaled daily (Balloy and Chignard, 2009). However, medical advances in transplantation and anticancer therapies have expanded the immunosuppressed patient population, and the number of individuals infected by opportunistic organisms, such as A. fumigatus, has drastically increased (McNeil et al., 2001;Chamilos et al., 2006). For opportunistic fungal pathogens, the phenomenon of "ready-made" virulence has been postulated, whereby traits which evolved for survival in ecological niches also govern survival in susceptible immunocompromised hosts (Casadevall et al., 2003;Rhodes, 2006).
Beyond residual host immune responses, there are additional obstacles to successful colonization of the mammalian lung, including tolerance of host-facilitated stresses, such as iron starvation (Schrettl et al., 2004(Schrettl et al., , 2007 and alkaline pH (Peñalva and Arst, 2004;Bignell et al., 2005;Peñalva et al., 2008). The requirement for infecting fungi to detect and respond to such extracellular cues is often essential for infectious growth, and in A. fumigatus the fungal receptors through which the extracellular environment is sensed remain largely unknown. This review discusses current knowledge on receptor-mediated signaling in A. fumigatus (Figure 1) and catalogues all of the putative seven transmembrane domain (7TMD) sensors encoded by the A. fumigatus genome (Table A1). Our analysis exposes the vast numbers of uncharacterized A. fumigatus receptor-like proteins.

G PROTEIN COUPLED RECEPTORS (GPCRs) IN A. fumigatus
In silico analyses of fungal genome sequences have identified genes encoding putative GPCR proteins. In the phytopathogenic fungus Magnaporthe grisea, a screen of the predicted proteome using all GPCR sequences at the time available in the GPCR Database (GPCRDB) (Horn et al., 2003) yielded 14 GPCR-like sequences (Kulkarni et al., 2005). A similar exercise applied to A. fumigatus identified 15 putative GPCRs (Lafon et al., 2006).
In Aspergilli, putative GPCRs are classified by homology, and according to a convention established by Lafon et al. (2006) in A. nidulans, into nine groupings. In A. fumigatus, Classes 1 and 2 are comprised, respectively, of two putative pheromone receptors GprA (AFUA_3G14330) and GprB (AFUA_5G07880); Class 3 is comprised of two putative carbon sensors GprC (AFUA_7G04800), GprD (AFUA_2G12640); Class 4 is comprised of three putative nitrogen sensors GprF (AFUA_5G04100), GprG (AFUA_1G11900), and GprJ (AFUA_1G06840); Class 5 of three putative cAMP receptors GprH (AFUA_5G04140), GprI (AFUA_3G00780), and GprL (AFUA_ 3G01750), the latter being unique to A. fumigatus; Class 6 is comprised of a single putative GPCR, GprK (AFUA_4G01350) having a regulator of G-protein signaling (RGS) domain, unique to filamentous fungi; Class 7 includes two putative GPCRs with homology to rat growth hormone-releasing factor receptors (Miller et al., 1999) only one of which is found in A. fumigatus, GprM (AFUA_7G05300); Class 8 is comprised of three putative GPCRs with identity to yeast Izh zinc regulators The pH-response pathway-A shift from acidic to alkaline environmental pH is thought to be detected by the receptor PalH, mediating the phosphorylation, and subsequent ubiquitination of the C-terminally bound arrestin, PalF. This stimulates the proteolytic cleavage of the transcription factor PacC. PacC then undergoes further pH independent cleavage, before translocating to the nucleus. (C) The histidine kinase receptor pathway-The phosphorelay histidine kinases mediate the transduction of specific external or cytosolic stimuli such as cell wall integrity (Fos-1) and osmotic stress (TcsC), stimulating autophosphorylation upon a conserved histidine residue. The activating stimulus for TcsB remains elusive. (D) The cell wall stress pathway-Cell wall receptors detect environmental stress such as cell wall damage/heat shock (MidA) and alternative carbon sources (Wsc1); however, the specific stimuli for receptors Wsc2 and 3 remain elusive. (Dotted lines indicate predictions based on studies in other fungi). (Karpichev et al., 2002;Lyons et al., 2004), two of which are found in A. fumigatus GprO (AFUA_ 3G10570) and GprP (AFUA_6G07160), and Class 9 is comprised of a single putative GPCR, NopA (AFUA_7g01430) having identity to bacterial opsins. The roles of some of these receptors have been identified in other species though in A. fumigatus little is known (Figure 1).
Among the 15 predicted GPCR-like proteins in A. fumigatus, only two, GprC (AFUA_7G04800) and GprD (AFUA_2G12640), have been characterized (Gehrke et al., 2010). GprC and GprD have been noted as having homology to Gpr1p of Saccharomyces cerevisiae which activates the cAMP pathway in response to glucose, as demonstrated by cAMP enzyme immunoassay (Yun et al., 1998;Kraakman et al., 1999). Furthermore, the A. nidulans GprD homologue mediates increase of intracellular cAMP in response to oxygenated polyunsaturated fatty acids (oxylipins), which act as autocrine and paracrine mediators in eukaryotic organisms (Affeldt et al., 2012). Deletion of A. fumigatus GprC and GprD resulted in significant growth impairment under all tested growth conditions and analysis of virulence revealed significant attenuation of virulence for gprD and delayed mortality for gprC in a murine model of aspergillosis (Gehrke et al., 2010). The remainder of the putative A. fumigatus GPCRs remain to be investigated and nothing is known about their molecular linkages to multi-subunit G-proteins. Unlike most Aspergillus spp. where four predicted Gα subunits occur, only three (GpaA, AFUA_1G13140, GpaB, AFUA_1G12930, and GpaC, AFUA_3G12400) have been identified for A. fumigatus (Liebmann et al., 2003), which presumably act via interaction with the Gβ and Gγ subunits (SfaD, AFUA_5G12210 and GpgA, AFUA_1G05210). In the current absence of other identified G protein subunits, or similar proteins, it is thought that the aforementioned five proteins service the entire A. fumigatus GPCR repertoire (Figure 1). Undoubtedly the relevance of A. fumigatus Gβ and Gγ subunits for viability and vegetative growth is significant as sfaD and gpgA gene deletion mutants are extremely impaired for germination and vegetative growth (Shin et al., 2009). Kulkarni et al. (2005) noted, based upon membrane topology, that the number of putative GPCR-like proteins encoded by the M. grisea genome rose to 76 when the criteria were relaxed to include homologs of the Pth11 receptor (DeZwaan et al., 1999) which is required for M. grisea pathogenicity in rice. Applying a more universal approach to A. fumigatus, we used the published genome sequence  to catalogue all A. fumigatus proteins having predicted TMDs (Figure 1). To implement this, we used the TMPRED (Hofmann and Stoffel, 1993) predictive tool to perform an analysis of all 9497 A. fumigatus proteins encoded by the reference genome Af293  http://www.cadre-genomes.org.uk/ Aspergillusfumigatus/Info/Index. In total we identified 6496 proteins having putative TMDs. Among them, 161 proteins were found to encode seven predicted TMDs (Tables 1 and A1). The majority of the predicted 7TMD proteins are of hypothetical function (Table A1).

PalH: A PUTATIVE 7TMD pH SENSOR
During colonization of the mammalian lung A. fumigatus is exposed to a range of microenvironments, of likely differing pHs, not only within the pulmonary niche but also following phagocytosis by macrophages or ingestion by neutrophils and exposure to their vacuole contents (Levitz et al., 1999;Newman, 1999;Reeves et al., 2002;Ibrahim-Granet et al., 2003). Versatility of metabolism and physiology is required to survive such extremes, including appropriate pH-responsive gene expression for nutrient acquisition and survival (Bignell et al., 2005). In the model ascomycete and occasional pathogen A. nidulans, the PacC transcription factor governs gene expression in response to extracellular pH (Tilburn et al., 1995;Diez et al., 2002) and is vital for mammalian pathogenicity (Peñalva and Arst, 2004;Bignell et al., 2005;Peñalva et al., 2008). Under alkaline conditions, a signaling cascade involving seven proteins is involved in activation of PacC. A putative pH sensor, PalH, has 7TMDs and a cytoplasmic C-terminus (Negrete-Urtasun et al., 1997, 1999, which interacts with a cognate arrestin encoded by palF (Herranz et al., 2005;Hervas-Aguilar et al., 2010). Unlike canonical GPCR receptors, PalH is not thought to act via interaction with G-protein subunits (Kroeze et al., 2003). When an alkaline response is triggered, PalF is phosphorylated and subsequently ubiquitinated in a PalH-dependent manner (Herranz et al., 2005), leading to PalB-mediated, signal dependent, proteolytic cleavage of the pH-responsive transcription factor PacC (Penas et al., 2007;Rodriguez-Galan et al., 2009). Subsequent translocation of the truncated PacC protein, from cytoplasm to nucleus, permits alkaline adaptation via differential expression of genes required to enable growth under alkaline extracellular conditions (Tilburn et al., 1995;Mingot et al., 1999Mingot et al., , 2001Espeso and Arst, 2000;. In A. fumigatus the amino acid residues crucial for PalH and PalF interaction are conserved, and in split-ubiquitin analyses the proteins enter into close proximity (Bertuzzi and Bignell, 2011;Bignell, 2012). We have also recently demonstrated the requirement for A. fumigatus PalH for murine infection (Bertuzzi et al., in preparation).

HISTIDINE KINASE SENSORS IN A. fumigatus
Histidine kinases (HK) are phosphorelay protein sensors which transduce extracellular signals. HKs are common in the fungal kingdom, and apparently absent in humans (West and Stock, 2001). Amongst archaea, bacteria and fungi, two classes of HK (two-component and hybrid) are found. The former class of two-component receptor systems predominate in bacteria and archaea, whereby autophosphorylation of the HK protein precedes transfer of the phosphoryl group to a conserved aspartate residue in a second protein, termed the response regulator (RR) (Li et al., 2012). HK activities have been associated with both the osmo-and peroxide-regulatory pathways in multiple fungi, and have been most extensively characterized in S. cerevisiae (Santos and Shiozaki, 2001). However, RR proteins are not abundantly encoded by fungal genomes; Skn7 and Ssk1 are two examples of such proteins, which in S. cerevisiae and C. albicans, account for the entire RR cohort of these species (Kaserer et al., 2009;Oide et al., 2010). The fungal phosphotransfer relay can involve three proteins, as exemplified by the S. cerevisiae HOG1 MAPK phosphorelay, where an HK (Sln1), a histidine phosphointermediate (Ypd1) and an RR (Ssk1) collectively mediate a multistep phosphotransfer (Kaserer et al., 2009). Fungal HKs most commonly fall into the hybrid class of regulators which utilize a single polypeptide. This protein possesses both a Histidine kinase A (HiskA) and a receiver domain (REC) containing a conserved aspartate residue (Li et al., 2012). Other domains, such as the ATP-binding HATPase_c domain (Dago et al., 2012) are also found; however, as these proteins are largely uncharacterized for A. fumigatus, the functional relevance of domain organization is unknown. The composition, and/or relative positioning, of additional domains provides the basis for sub-classification of HKs (Catlett et al., 2003), presented for 12 A. fumigatus HKs in Table A2. Amongst these, only three have been studied: the two-component system proteins A, B, and C (TcsA/Fos-1 AFUA_5G10240, TcsB AFUA_2G00660 and TcsC AFUA_2G03560).
Despite the significance of the HK Fos-1 for detection of extracellular stresses, this hybrid HK has been previously predicted as possessing no TMDs, implying a cytosolic presence (Pott et al., 2000). However, our TMPRED analyses predicted TMDs for all of the three HKs, with Fos-1 possessing a single TMD ( Table A2). Deletion of the fos-1 gene leads to a ∼66% reduction in conidiation after 48 h in liquid YG medium, as well as heightened resistance to the cell wall-degrading enzyme mix novozym 234, suggesting the role of fos-1 in cell wall assembly (Pott et al., 2000).
fos-1 mutants were found to have normal morphology, germination, osmotic and oxidative stress tolerance, and antifungal susceptibilities. Subsequent transcriptional analyses found a significant increase in fos-1 expression, relative to in vitro growth, during the first 72 h of infection in a murine model of pulmonary aspergillosis (Zhang et al., 2005), and reduced virulence of A. fumigatus in a systemic murine model of infection (Clemons et al., 2002).
In a study addressing the role of oxidative stress in A. fumigatus pathogenicity, Du et al. (2006) characterized the A. fumigatus TcsB protein, a putative homolog of Sln1 in S. cerevisiae. In A. nidulans, TMPRED analysis predicted 2 TMDs for TcsB at the N-terminus (Furukawa et al., 2002), though in A. fumigatus, our prediction extends this to 4TMDs (Table A1). Unlike S. cerevisiae where deletion of sln1 is lethal (Maeda et al., 1994), an A. fumigatus tcsB mutant is viable, demonstrates normal morphology, and is as tolerant as the wild type to increased temperatures, various cell wall damaging agents, and poor nitrogen/carbon sources. The only phenotype discernable for the mutant was a minor sensitivity to SDS (Du et al., 2006). This data suggests a nonessential role for TcsB, or redundancy of function with other, as yet uncharacterized protein(s).
It is believed that group III HK mediate resistance to high osmotic pressure via the high osmolarity glycerol pathway (HOG). For this reason, the characterization of the sole A. fumigatus group III hybrid HK TcsC, classified as such on the basis of putative HAMP (HK, adenyl cyclase, methylaccepting chemotaxis protein, phosphatase) domains, was investigated (McCormick et al., 2012). The significance of the HAMP domains, based upon studies of other sensor proteins and signaling is postulated as providing the means to couple input and output since HAMP domains of integral membrane hybrid HKs are found in close proximity to the membrane-spanning segment (Parkinson, 2010). It is speculated that in response to extracellular signals, such as altered osmolarity, a conformational rearrangement is triggered which prompts activation of an output domain (Parkinson, 2010). In A. fumigatus, deletion of the tcsC gene resulted in an extended white colony rim and a reduced number of extending hyphae. However, unlike the A. nidulans homologue nikA (Hagiwara et al., 2009), no detrimental effects on sporulation and conidial growth were observed. In the presence of nitrate as a nitrogen source a significant reduction in radial growth was detectable, and furthermore, compared to the control strain, growth of tcsC at 2% O 2 abolished sporulation and prompted a dome-shaped morphology indicative of oxygen starvation. A strong inhibition of growth resulted from exposure to hyperosmotic stress (1.2 M sorbitol, 1 M KCl, or 1 M NaCl) but sensitivity to calcofluor white, amphotericin B, posaconazole and caspofungin, extremes of pH/temperature, or oxidative stress were reportedly normal.
In a comparative analysis of wild type and tcsC virulence, no discernable differences in pathogenicity analysis in a murine model of invasive aspergillosis were detected (McCormick et al., 2012).

CELL WALL RECEPTORS
The fungal cell wall is essential for viability and an important target of antifungal drugs (Latgé et al., 2005;Latgé, 2007;Walker et al., 2010). In fungi a conserved MAPK signaling module is responsible for maintaining cellular integrity, shape and resilience to environmental stresses (Lee and Levin, 1992;Levin, 2005;Lesage and Bussey, 2006). In S. cerevisiae, such MAPK signaling (Chen and Thorner, 2007) is initiated through stress detection at five integral membrane receptors Wsc1-3, Mid2, and Mtl1 (Lodder et al., 1999). This promotes guanine nucleotide exchange factor (GEFs-Rom1 and Rom2)-mediated nucleotide exchange upon the GTPase Rho1, facilitating the regulation of numerous downstream effectors (Zu et al., 2001). In a quest to find equivalent cell wall sensors in A. fumigatus, Dichtl et al. (2012) performed BLAST analyses to reveal three previously uncharacterized open reading frames with domain structures similar to those of Wsc1-3 (Af Wsc1, AFUA_4G13670, Af Wsc2, AFUA_3G07050, and Af Wsc3, AFUA_5G09020 respectively). Bioinformatic analyses predicted the presence of characteristic cell wall integrity (CWI) sensor N-terminal WSC domains with downstream, though truncated, ser/thr rich regions, and (with the exception of Wsc2) transmembrane domains. In common with the S. cerevisiae sensors a short cytosolic C-terminus was also predicted for two of the sensors (Dichtl et al., 2012).
To discern subcellular localization, ectopically integrating vectors were applied to generate four putative CWI sensor-GFP fusions, Wsc1-3, and MidA. From these, localization of all C-terminally tagged sensors was observed at the fungal surface. Additionally a strong presence was observed in vacuoles, though this was dismissed as a by-product of over expression or misfolding. Phenotypic analyses of single and double mutants identified a significant impairment of radial growth in the case of a wsc1 wsc3 double mutant. These findings were further exacerbated in a triple mutant wsc1 wsc3 midA. Furthermore, in all mutants lacking wsc1, provision of glycerol as carbon source lead to a significant reduction in radial growth on minimal media (Dichtl et al., 2012). Previously, mutants lacking members of the CWI MAPK pathway have demonstrated a clear sensitivity to echinocandins and azole antifungals (Fujioka et al., 2007;Dirr et al., 2010). Extending this analysis to the A. fumigatus mutant phenotypes revealed a single relevant phenotype, namely the heightened sensitivity of the wsc1 mutant to the echinocandin, caspofungin (Dichtl et al., 2012).
To study stress-induced activation of intracellular signaling, effects on growth and MpkA phosphorylation were analyzed in the presence of the cell wall perturbing agent, calcoflour white, or following heat shock (48 • C). None of the Wsc mutants were found to be sensitive to cell wall perturbation or heat shock, however, midA was highly sensitive to all of these stresses. In agreement with phenotypic data, calcofluor white-induced MpkA phosphorylation was significantly reduced in the midA mutant compared with wild type, while phosphorylation of MpkA was not diminished in mutants lacking wsc1 or wsc1 and wsc3 (Dichtl et al., 2012). In S. cerevisiae, the Wsc1 cell wall sensor mediates signaling of alkaline stress via the CWI MAPK module (Serrano et al., 2006); no evidence for such a role in A. fumigatus was obtained. Thus, while Mkk2 null mutants are sensitive to alkalinization of the medium (Dirr et al., 2010), the identity of the activating cell wall sensor remains unknown.
Taken together these findings suggest that MidA is the sole cell wall perturbation sensor, while Wsc1 is required for glycerol carbon source assimilation. Furthermore, a compensatory role between Wsc1 and Wsc3 with regards to efficient growth and conidiation has been demonstrated. Despite these observations, a role for Wsc2 has yet to be identified, while the putative CWI pH sensor remains elusive.

RECEPTOR-MEDIATED SIGNALING DURING A. fumigatus INFECTION: RELEVANCE FOR THERAPEUTIC STRATEGY
Drugs which target GPCR function account for >50% of currently licensed drugs (Davies et al., 2008). It therefore follows that fungal GPCRs are likely to be similarly responsive to chemical perturbations. This fact, coupled with the absolute requirement for some GPCRs in fungal growth make a compelling case for these proteins as antifungal drug targets. Although the pharmaceutical market is dominated by GPCR-active molecules, the discovery of most of these agents was made on the basis of functional activity in high throughput screens, only later were the targets and modes of action clarified (Filmore, 2004). In the post-genomic era, with confidence in the pharmaceutical relevance of such proteins, drug discovery can become targetdriven. An expanding repertoire of technologies to probe 7TMD protein activities provides the basis upon which to confront functional studies of the uncharacterized receptors in A. fumigatus and to screen for inhibitory molecules. It has recently been suggested that considering seven transmembrane receptors as disordered proteins able to allosterically respond to a number of binding partners, is useful in understanding the plasticity of function exhibited by such proteins (Kenakin and Miller, 2010). Conformational changes which occur in response to extracellular ligands and/or stimuli expose cytosolic signaling domains and present three distinct arenas open to perturbation: extracellular sensing/ligand binding, cytosolic surfaces, and intramembrane domains. In order to prioritise the most promising candidates for drug development, a crucial experiment will be to assess the requirement of such receptors, via regulatable promoters, for sustained viability of established fungal mass in murine models of infection (Gossen and Bujard, 1992).

ACKNOWLEDGMENTS
We gratefully acknowledge the assistance of Arshad Khan in automating the genome-wide TMPRED analyses.