Overview of carbon and nitrogen catabolite metabolism in the virulence of human pathogenic fungi

Summary

It is estimated that fungal infections, caused most commonly by Candida albicans, Aspergillus fumigatus and Cryptococcus neoformans, result in more deaths annually than malaria or tuberculosis. It has long been hypothesized the fungal metabolism plays a critical role in virulence though specific nutrient sources utilized by human pathogenic fungi in vivo has remained enigmatic. However, the metabolic utilisation of preferred carbon and nitrogen sources, encountered in a host niche-dependent manner, is known as carbon catabolite and nitrogen catabolite repression (CCR, NCR), and has been shown to be important for virulence. Several sensory and uptake systems exist, including carbon and nitrogen source-specific sensors and transporters, that allow scavenging of preferred nutrient sources. Subsequent metabolic utilisation is governed by transcription factors, whose functions and essentiality differ between fungal species. Furthermore, additional factors exist that contribute to the implementation of CCR and NCR. The role of the CCR and NCR-related factors in virulence varies greatly between fungal species and a substantial gap in knowledge exists regarding specific pathways. Further elucidation of carbon and nitrogen metabolism mechanisms is therefore required in a fungal species- and animal model-specific manner in order to screen for targets that are potential candidates for anti-fungal drug development.

Introduction

It is estimated that infections caused by the fungal pathogens Candida albicans, Aspergillus fumigatus and Cryptococcus neoformans, kill more people annually than malaria and tuberculosis (Denning and Bromley, 2015; Meyer et al., 2016). These fungi cause a spectrum of diseases, ranging from superficial to systemic and invasive, depending on the underlying disturbance in host defence responses and physiology (Mayer et al., 2013; Abad et al., 2010; Lee et al., 2013).

Pathogenicity, the ability to cause disease, is a multifactorial trait which encompasses a variety of survival and fitness-enhancing factors and pathways that determine the virulence of a pathogen in a given host. Among these, nutrient acquisition and subsequent metabolic processes are a critical virulence determinant as they are essential for promoting fungal fitness, survival, and virulence within the host. Essential nutrients include minerals such as iron and zinc which are required in small amounts, while carbon and nitrogen, the main energy sources for sustaining biosynthetic processes, must be obtained in large quantities from the environment (Ramachandra et al., 2014). Fungi have preference for certain carbon and nitrogen sources that are rapidly metabolised and therefore provide quick energy for growth and niche colonisation (Ruijter and Visser, 1997; Schure et al., 2000). In the presence of these favoured energy sources, the utilisation of alternative, less preferred carbon and nitrogen sources may be repressed, a process known as carbon and nitrogen catabolite repression (CCR, NCR). This review presents an overview of the current literature on CCR and NCR, including sensing and uptake of preferred carbon and nitrogen sources, whilst emphasising the importance of these regulatory mechanisms for the pathogenicity and virulence of C. albicans, A. fumigatus and C. neoformans.

In vivo carbon and nitrogen source availability

In the context of infection, the main carbon sources available to pathogenic fungi are glucose, lactate and acetate, whose availability largely depends on the host niche. In addition, potential nitrogen sources can also serve as carbon sources and are available throughout the human host mainly in the form of proteins. The primary site of infection for A. fumigatus and C. neoformans is the respiratory tract and occurs through inhalation of fungal spores. The airway surface liquid (ASL) of the lung and nasal secretions typically contain less than 0.5 mM glucose in healthy individuals, about 10-fold less than serum (Philips et al., 2003, Baker et al., 2007). In addition, lactate and amino acids have been identified in the ASL of mice and concentrations of these metabolites increase with inflammation (Grahl et al., 2011, Hu et al., 2008). Furthermore, ammonium, a preferred nitrogen source that enters the human body either as dietary free amino acids or is produced locally by bacterial microbiomes of the gastrointestinal (GI) tract (Richardson et al., 2013) and lungs, can also be found in the lung airways. Underlying respiratory diseases, such as cystic fibrosis, significantly increase local ammonium concentrations (Gaston et al., 2002). The lung mucous layer is composed of different glycoproteins such as mucin, which likely present good carbon/nitrogen sources for fungi. Fungal-induced damage to the mucosal barriers allows access to the underlying connective tissue which mainly consists of the structural proteins collagen and elastin (Farnell et al., 2012; Kronstad et al., 2012). A similar situation can be envisaged for C. albicans when switching from a gastrointestinal (GI) and vaginal commensal organism to a pathogenic organism, inducing penetration of the gut and vaginal mucosal membranes and gaining access to the underlying tissue.

Further tissue-induced damage would subsequently result in all three fungi reaching the bloodstream, which typically contains glucose and lactate levels of 4–5.5 mM and 0.5–2mM, respectively, depending on diet, and medical conditions such as diabetes and cystic fibrosis, which increase serum glucose levels (Baker et al., 2006, Ament et al., 1997). Furthermore, interorgan transport ensures the distribution of amino acids throughout the body which are required for protein synthesis and specific metabolic functions (Brosnan, 2003). The concentration of plasma amino acids is around 2.5 mM with glutamine being the most abundant amino acid in the blood (Brosnan, 2003) and in the intracellular amino acid pool of most tissues (Stumvoll et al., 1999). Dissemination via the bloodstream enables organ colonisation; in the case of C. neoformans, access to the cerebrospinal fluid (CSF), crossing of the blood brain barrier (BBB) and invasion of the brain. The organic acids lactate and acetate are produced by bacteria of the human gut or in a tissue-dependent manner, especially during inflammation where infiltrating immune cells primarily produce lactate for ATP generation from glucose (Jiménez-López et al., 2013; Schug et al., 2016) (Borregaard & Herlin, 1982). In the CSF, glucose levels are typically about 60–70% of serum glucose levels (Seehusen et al., 2003). Additionally, the levels of carbon and nitrogen sources in vivo change with corticosteroid treatment, a risk factor for many fungal infections, and upon inoculation with fungal conidia (Beattie et al., 2017). An important area for further research is how medical treatments that pre-dispose individuals to fungal infections alter the nutrient environment of specific organ systems. Thus, in addition to being able to utilize available carbon sources, pathogens must also be able to adapt to a rapidly changing nutrient environment within the host. Therefore, when considering human pathogenic fungi, we must examine metabolism of both glucose and alternative, non-preferred carbon and nitrogen sources. Figures 1 and 2 give an overview of known factors involved in carbon and nitrogen metabolic utilization.

Figure 1.

Glucose sensing, uptake and metabolic pathways in C. albicans, A. fumigatus and C. neoformans (blue boxes: confirmed and putative sensors; green boxes: confirmed and putative transporters; red = metabolic enzymes, solid arrows = confirmed cellular processes, dashed arrows = cellular processes that are not elucidated; yellow = nucleus; purple = mitochondria). Glucose is sensed and taken up into the cell by specific hexose transporters (HXT) and is subsequently phosphorylated by the carbohydrate kinases hexokinase (HXK) or glucokinase (GLK). Further metabolic utilisation occurs via glycolysis and the TCA (tricarboxylic acid) cycle. Glucose uptake and subsequent phosphorylation serve as signals for the carbon catabolite repressor (CCR) Mig1p/CreA/Mig1 to translocate to the nucleus and repress target genes. Several factors, such as the protein kinase SNF, the de-ubiquitinylation complex CreB/CreC or the transcriptional regulator of iron metabolism HapX, have been shown to be involved in CCR by either interacting directly or indirectly, with Mig1p/CreA/Mig1 in a species-dependent manner.

Figure 2.

Amino acid and ammonium sensing, uptake and metabolism in C. albicans, A. fumigatus and C. neoformans (blue boxes: confirmed and putative sensors; green boxes: confirmed and putative transporters; solid arrows = confirmed cellular processes, dashed arrows = cellular processes that are not elucidated; yellow = nucleus; purple = mitochondria). Proteins are degraded into peptides or proteins by secreted proteases (SAP = secreted aspartyl proteases). Peptides are internalised via oligopeptide (OPT) or peptide (PTR) transporters. Amino acids and ammonium are sensed and taken up by respective transporters and signals are relayed via several predicted pathways including cAMP and protein kinase A (PKA). Amino acids serve as precursors for TCA (tricarboxylic acid) cycle intermediates to generate ATP. Transcription factors (Gat1p, Gln3p/AreA, NmrA/Gat1) are activated and translocate to the nucleus where they regulate the catabolic and anabolic utilisation of different nitrogen sources.

Carbon catabolite metabolism in human pathogenic fungi

Glucose sensing and uptake

Given the importance of glucose in fungal metabolism, fungi have evolved sensitive systems for the sensing and uptake of glucose. In the well-studied model organism, S. cerevisiae, there are two major types of glucose sensing mechanisms: the G-Protein Coupled Receptor (GPCR) system, with the receptor Gpr1p and G-protein Gpa2p, and the hexose transporter gene family members Snf3p and Rgt2p [reviewed in (Peeters and Thevelein, 2014)]. Here we use this well-studied system as a basis for discussion but note that unlike the human pathogenic fungi discussed herein, S. cerevisiae is a Crabtree positive organism, i.e., it produces ethanol in aerobic conditions and high external glucose concentrations rather than accumulating biomass via the tricarboxylic acid (TCA) cycle. Thus, additional opportunities exist to discover glucose sensing and uptake mechanisms in the most commonly encountered human pathogenic fungi. In C. albicans, Gpr1p/Gpa2p regulate hyphal morphogenesis through activation of the cAMP-PKA (protein kinase A) pathway (Miwa et al., 2004, Maidan et al., 2005a). Whether this receptor is directly sensing glucose is unclear, however evidence suggests that this receptor also responds to methionine to stimulate hyphal formation in response to carbon source (Maidan et al., 2005b). The role of the Gpr1p sensing system in virulence was assessed in two studies, where Maidan and colleagues (2005a) reported reduced hyphal formation, tissue invasion and virulence of a gpr1−/− strain when injected intravenously in an inbred mouse model of systemic candidiasis, whereas Miwa et al. (2004) observed wild type virulence of intravenously-injected hetero- and homozygous mutants of GPR1 and GPA2 using an outbred mouse strain in a systemic model of candidiasis. Additionally, each study used a different cell density (optical density) inoculum, thus this signalling pathway may be important for virulence in a host-specific or dose-dependent context. In C. neoformans, Gpa2 functions upstream of cAMP-PKA signalling, and is required for cAMP signalling in response to glucose, however, the upstream sensor associated with this G protein remains unknown (Xue et al., 2006) and deletion of GPA2 did not affect virulence in an inhalation model of murine infection using both immune competent and immune deficient (severe combined immunodeficiency, SCID) mice (Li et al., 2007).

The genome of A. fumigatus encodes at least 15 GPCRs, of which only GprC, GprD and GprK, have been characterised (Grice et al., 2013, Jung et al., 2016). GprC and GprD show amino-acid sequence similarity to ScGpr1p, and deletion of either of these receptors results in a severe, temperature-dependent growth defect independent of carbon source (Gehrke et al., 2010, Han et al., 2004). Although no carbon-dependent growth defect of GprD was observed, metabolomic studies demonstrated that loss of the homologue receptor in A. nidulans altered glucose and central carbon metabolism (de Souza et al., 2013). Deletion of gprC resulted in delayed mortality whereas deletion of gprD attenuated virulence, when administered intranasally, in a neutropenic mouse model of infection, although the mechanism behind this reduction in virulence is unknown (Gehrke et al., 2010). GprK plays a role in carbon sensing, with the null mutant displaying increased germination under carbon starvation conditions and decreased growth on pentose sugars. Although this mutant showed significantly less invasion into A549 cells in vitro, there was no significant difference in survival using a wax moth larvae model, thus the role of this receptor in pathogenesis remains unclear (Jung et al., 2016).

The second major glucose sensing mechanism in S. cerevisiae uses high and low affinity hexose transporters (HXT), Snf3p and Rgt2p, respectively. These two proteins do not actively transport glucose but rather act as sensors for glucose concentration and transduce signals that regulate the expression of HXT genes appropriate for the glucose environment (Forsberg and Ljungdahl, 2001, Ozcan et al., 1998). In C. albicans, there are 20 predicted hexose transporters (HGT1-20), one of which encodes an Snf3p/Rgt2p homologue, Hgt4p, which senses glucose and subsequently regulates the expression of 6 HGT genes and is partially responsible for filamentation in response to this stimulus (Brown et al., 2006, Fan et al., 2002). Loss of HGT4 results in an increased median survival of mice when injected intravenously into an immunocompetent systemic model of candidiasis, supporting the hypothesis that sugar sensing is an important aspect of C. albicans virulence (Brown et al., 2006). Furthermore, HGT6 and HGT12 were highly expressed in a mouse model of oral candidiasis (OPC) when compared to the in vitro control condition (Fanning et al., 2012), HGT7 was significantly expressed in a mouse model of intra-abdominal candidiasis (Cheng et al., 2013) and HGT2, 4, 7, 12 and 17 were induced during an infection time course experiment in mice with systemic candidiasis (Wenjie et al., 2015). These results indicate that sugar scavenging, sensing and/or transport are important during early and late infection at various sites of the mammalian host. The exact role played by these HGT genes in C. albicans virulence remains to be determined.

The most closely related proteins to Snf3p/Rgt2p/Hgt4p in C. neoformans are Hxs1 and Hxs2. Transcriptome analysis of C. neoformans strains isolated from the cerebrospinal fluid (CSF) of 2 AIDS patients, showed high induction of Hxs1 when compared to the in vitro control condition (Chen et al., 2014). Hxs1 was shown to be a high affinity glucose transporter that did not complement the glucose sensing defect of an S. cerevisiae SNF3/RGT2 mutant, nor did loss of HXS1 affect expression of hexose transporters, suggesting that it may not act as a glucose sensor (Liu et al., 2013). Deletion of HXS2 was not possible but a role in glucose sensing is questioned given the low expression of this gene in high and low glucose (Liu et al., 2013). Sugar sensing and uptake systems and their relationship with virulence have not been characterized in A. fumigatus, however 17 putative hexose transporter genes have been identified in A. nidulans (Wei et al., 2004). Among these 17 hexose transporters, 15 predicted homologues are found in A. fumigatus. In summary, significant opportunities exist to better define the roles of sugar sensing and uptake in human fungal virulence.

Carbon catabolite repression (CCR)

In many fungi, carbon source utilization is regulated via a finely-tuned system that allows preferential utilization of glucose and preferred sugars before other available carbon sources. This system is called Carbon Catabolite Repression (CCR) and is regulated in part by the C₂H₂ zinc-finger transcription factor Mig1p/CreA/Mig1, which represses the expression of genes required for catabolism of less preferable carbon sources, as well as gluconeogenic genes, and nutrient acquisition genes in response to carbon starvation (Ruijter and Visser, 1997; Zaragoza et al., 2000; Caza et al., 2016). CCR in S. cerevisiae, regulated by the transcriptional repressor Mig1p and the corepressors Tup1p and Ssn6p, and CreA in A. nidulans and Aspergillus niger, has been studied exhaustively and is reviewed in (Kim et al., 2013, Gancedo, 1998, Klein et al., 1998).

Transcriptional studies of Mig1p and Tup1p in C. albicans, indicate that each transcription factor regulates unique sets of genes, in addition to shared genes between both repressors (Murad et al., 2001). Unique Mig1p targets and those shared by Mig1p and Tup1p were largely annotated as carbohydrate uptake and catabolism factors, implicating these repressors in CCR. Furthermore, microarray studies of Mig1p identified this regulator to have predominantly repressing functions, with only 0.8% of measured genes down-regulated upon loss of Mig1p, versus ~12.5% of measured genes whose mRNA levels increased in response to Mig1p loss (Murad et al., 2001).

In C. neoformans, Mig1 regulates the expression of amino acid, heme and carboxylic acid biosynthesis genes, TCA cycle and electron transport genes. The loss of Mig1 results in increased sensitivity to oxidative stress, respiratory inhibitors and antifungal drugs. Interestingly, a large number of genes are regulated by Mig1 under iron limitation, suggesting a role for this protein in iron homeostasis or metabolic adaptation to low iron (Caza et al., 2016). Indeed, a relationship of Mig1 with HapX, a bZIP transcription factor required for adaptation to iron limiting and iron excess conditions and virulence in several fungal pathogens (Jung et al., 2010, Schrettl et al., 2010, Gsaller et al., 2014, Chen et al., 2011, Lopez-Berges et al., 2012), was identified in C. neoformans (Caza et al., 2016). Although current studies on Tup1 provide little evidence as to whether Tup1 is involved in CCR (Lee et al., 2007, Lee et al., 2005, Lee et al., 2009), a role for Tup1 in positive regulation of iron homeostasis genes was identified, including CIG1, a gene encoding the putative heme binding protein (Lee et al., 2009), that was also identified to be positively regulated by Mig1 (Caza et al., 2016). Although interactions have yet to be identified between Mig1 and Tup1 in C. neoformans, a potential mechanistic link between these regulatory proteins in low iron conditions is an intriguing possibility.

In filamentous fungi, the Mig1p homologue, CreA, has been well studied in A. nidulans where the presence of glucose results in CreA translocation to the nucleus and subsequent repression of genes encoding enzymes required for the utilization of alternative carbon sources (Kulmburg et al., 1993, Tamayo et al., 2008, Garcia et al., 2004; Brown et al., 2013; Ries et al., 2016). Unlike CCR in yeasts, the Tup1p homologe, RcoA does not play a role in CCR in A. nidulans at the level of gene expression, however, loss of RcoA results in changes in nucleosome positioning at some CreA regulated promoters (Hicks et al., 2001, Garcia et al., 2008). Notably, loss of CreA in A. nidulans results in a leaky lethal phenotype (Dowzer and Kelly, 1991), resulting in an extremely sick strain. Full lethality of this gene deletion has been reported in several other filamentous fungi including Fusarium oxysporum (Jonkers and Rep, 2009), Penicillium chrysogenum (Cepeda-García et al., 2014) and Colletotrichum gloeosporoides (Bi et al., 2015), whereas in Neurospora crassa (Sun and Glass, 2011), Trichoderma reesei, and A. fumigatus (Beattie et al., 2017), generation of a creA-genetic null mutant is viable, underscoring divergence of the role of this transcription factor across fungal species. In A. fumigatus, the loss of creA results in a carbon and nitrogen-independent growth defect on solid media, with carbon-dependent growth defects in liquid culture. Similar to C. neoformans, loss of creA also results in increased sensitivity to mitochondrial inhibitors and the antifungal drug, voriconazole. It is unclear if the increased susceptibility to these inhibitors is a direct effect, or a result of increased drug permeability due to altered cell wall composition in the creA null mutant. Interestingly, in contrast to C. albicans, where the predominant changes in gene expression of MIG1 indicated a repressor only function for Mig1p, RNA-sequencing of the A. fumigatus creA genetic null mutant revealed nearly equal numbers of genes whose transcript levels decreased and increased in glucose medium compared to the wild-type parent (Beattie et al., 2017). While work to characterize the direct targets of CreA is ongoing, these results suggest that CreA may play a role in transcriptional activation in addition to repression in A. fumigatus as previously observed in Trichoderma reesei (Portnoy et al., 2011) and A. nidulans (Mogensen et al., 2006).

The variable essentiality of Mig1/CreA homologues across fungal species suggests functional divergence of this transcription factor. Indeed, primary amino acid sequence alignments of Mig1/CreA between yeasts and filamentous fungi show that only the C₂H₂-zinc finger domain is conserved (Figure 3). Thus, understanding the role of these transcription factors and their interacting partners in each species is a promising approach to better understand and define the metabolic similarities and differences of between fungal species and how these differences impact metabolism and growth in vivo.

Figure 3.

ClustalW alignment of CreA/Mig1p/Mig1 amino acid sequences from A. fumigatus (AfCreA), A. nidulans (AnCreA), S. cerevisiae (ScMig1p), C. albicans (CaMig1p) and C. neoformans (CnMig1p). Percent identity is shown with shading, where darker shading indicates higher percent identity. Zinc-finger domains are outlined in black, based on SMART domain prediction of ScMig1p.

Additional factors involved in CCR

In S. cerevisiae, glucose repression is dependent on the hexokinase, Hxk2p, which is required for phosphorylation of Mig1p by Snf1p (Herrero et al., 1998, Ahuatzi et al., 2007). In A. fumigatus, the hexokinases, hxkA and to a larger extent glkA, appear to also play a role in CCR, as loss of these proteins increase isocitrate lyase (ICL, glyoxylate cycle) activity in the presence of glucose and ethanol (Fleck and Brock, 2010). The expression of C. albicans HXK2 is regulated in response to glucose, however, null mutants of this gene have yet to be characterized and a role in CCR remains undefined. In C. neoformans, the deletion of both hexokinases, HXK1 and HXK2, results in the inability to utilize glucose, however their role in CCR has not yet been investigated. The loss of both hexokinases together, but not single deletions, results in reduced virulence when inoculated intranasally in an immunocompetent mouse model (Price et al., 2011).

The AMP-activated protein kinase Snf1p plays an important role in CCR in S. cerevisiae through regulating the cellular localization of Mig1p (De Vit et al., 1997, Ostling and Ronne, 1998). Snf1p belongs to a highly conserved family of serine/threonine protein kinases, with an essential role in mediating the response to glucose abundance/depletion thus regulating cellular energy homeostasis (Mizuno et al., 2015). Regulation of Mig1p by Snf1p has not yet been characterized in C. albicans, but a homozygous mutant of Snf1p is lethal (Petter et al., 1997). Characterization of a heterozygous SNF1 mutant revealed a growth defect, however carbon source utilization was not altered, and the mutant maintained wild-type virulence in an immunocompetent systemic model, despite a hyper-filamentous phenotype in vitro (Petter et al., 1997). Similar to S. cerevisiae, re-localization of A. nidulans CreA to the cytoplasm following transfer to glucose starvation conditions, as well as expression of cellulase and hemicellulase-encoding genes is dependent on SnfA (Brown et al., 2013). However, the role of SnfA in CCR has yet to be fully defined in filamentous human pathogens. In C. neoformans, generation of a SNF1-null mutant is viable and experimental evidence implicates this kinase in the regulation of CCR (Hu et al., 2008a). Interestingly, Snf1 appears to have divergent roles across serotypes within the species (Hu et al., 2008a) but in all strains, Snf1 is essential for virulence in both pulmonary and systemic immunocompetent models of cryptococcosis (Hu et al., 2008a, Yang et al., 2010).

Studies of CCR in A. nidulans also identified three additional factors which regulate glucose repression in this organism: CreB, CreC and CreD (Hynes and Kelly, 1977, Kelly and Hynes, 1977). CreB, a deubiquitination enzyme (Lockington and Kelly, 2001), and CreC, a WD-40 repeat protein (Todd et al., 2000), form a complex that is required for proper glucose repression (Lockington and Kelly, 2002). Mutation of either of these proteins yields strains that have growth defects in both repressing and de-repressing conditions (Hynes and Kelly, 1977, Lockington and Kelly, 2001, Ries et al., 2016). Although CreB has deubiquitination activity, the role of ubiquitin in the regulation of CreA is unclear. Ries and colleagues (Ries et al., 2016) identified an ubiquitin smear on a CreA immunoprecipitation Western blot, however, mass spectrometry to identify post-translational modifications on CreA found several phosphorylated peptides, but no ubiquitinylated peptides in repressing or de-repressing conditions (Alam et al., 2016). CreD, on the other hand, is an arrestin and PY motif containing protein that appears to regulate CCR through opposing the action of the CreB/CreC complex. Yeast two-hybrid studies of CreD identified a physical interaction between CreD and the ubiquitin ligase, HulA, suggesting that this protein may be involved in ubiquitination of targets, potentially sharing targets with the CreB/CreC complex (Boase and Kelly, 2004). Homologues of CreB, CreC and CreD are all present in A. fumigatus but a role of these proteins in virulence has not yet been determined. Homologues of these three proteins are also present in C. albicans and include the ubiquitin hydrolase Upb13p (CreB orthologue) with a predicted role in biofilm formation (Nett et al., 2009; Bonhomme et al., 2011); the alpha subunit of the COPI (coat protein) vesicle coatomer complex (CreC orthologue) which functions in vesicle and retrograde vesicle-mediated transport between the ER (endoplasmic reticulum) and the Golgi (Bonhomme et al., 2011) and lastly, the membrane protein Rod1p (CreD orthologue) which has a role in drug tolerance and is under the transcriptional control of Rgt1p, which is also involved in the regulation glucose transporter-encoding genes (Sexton et al., 2007). A recent study described the involvement of A. oryzae CreD in glucose-induced endocytosis of a maltose transporter, whose expression is regulated by CCR, through working as an adaptor protein for HulA, required for the ubiquitination of target plasma membrane transporters and subsequent endocytosis (Tanaka et al., 2017). It is therefore an intriguing possibility that the role of CreB, CreC and CreD may be one in CCR-related vesicle transport, responding to a change in available nutrients, rather than directly targeting CreA. No homologues of these three proteins have been identified in C. neoformans.

Role of CCR in virulence

Progress towards understanding the utilization of potential carbon sources during an infection has been made using transcriptional studies of C. albicans, A. fumigatus and C. neoformans exposed to immune cells or in infection models (Fradin et al., 2005, Lorenz et al., 2004, Barelle et al., 2006, Chen et al., 2014). Similarities between the response of these fungi to phagocytosis include the induction of metabolic pathways required during carbon starvation, including gluconeogenesis, fatty acid metabolism and the glyoxylate shunt, many of which are targets of the CCR system in these fungi. However, phagocytosis represents only one host niche, and during infection, occupation of several niches is likely required for full pathogenesis. This is demonstrated by reporter fusion studies, where Barelle and colleagues (2006) reported the upregulation of the glyoxylate cycle genes in phagocytosed cells, however, the population of cells in the kidney during systemic infection were largely expressing glycolytic genes and some fungal cells even expressed gluconeogenic and glyoxylate cycle genes. The aforementioned studies indicate that fungal cells experience a variety of micro-environmental conditions in vivo, and the ability to adapt to and thrive under several carbon conditions is critical for virulence (Barelle et al., 2006). Furthermore, characterisation of gene expression in fungal strains isolated directly from human subjects, as has recently be done for 2 C. neoformans isolates extracted from the CSF of two AIDS patients (Chen et al., 2014), can also provide novel insights into infection-related processes. Gene expression of C. neoformans in human CSF was more similar to the in vitro YPD control condition than in ex vivo CSF, an observation that was predicted to be due to varying nutrient availability, pH and epigenetic changes that occur specifically in the human host (Chen et al., 2014). Such studies are valuable as they characterise the genetic response of fungal pathogens that are metabolically active during human host infection, and do not rely on animal models or in vitro conditions that simulate mammalian environments, therefore providing insight into in vivo infection processes in the human host.

These studies highlight the potential for differential metabolic requirements across fungal pathogens, likely dependent on the host niche occupied by each species and the models used to test each mutant. Further support for niche specific metabolic requirements comes from Hu and colleagues (Hu et al., 2008a), who reported differential expression of several core carbon metabolism genes in C. neoformans, including the high affinity glucose transporter HXT1 and phosphoenolpyruvate carboxykinase PCK1, which were highly expressed early in pulmonary infection, but not in CSF infection, suggesting low glucose availability in the lungs early in infection. The same study identified the upregulation of carbon starvation and alternative carbon utilization pathways, such as beta-oxidation, the glyoxylate shunt and the utilization and production of acetyl-coA (Hu et al., 2008a, Steen et al., 2003). Furthermore, deletion of the high-affinity glucose transporter HXS1, required for growth in low glucose, results in attenuated virulence of C. neoformans in a pulmonary model of cryptococcosis (Liu et al., 2013).

Although these clues to in vivo carbon source utilization have identified expression of targets of carbon catabolite repression, Mig1p/Mig1 is dispensable for virulence (as measured by murine mortality) in an immunocompetent systemic model of C. albicans and an immunocompetent pulmonary model of C. neoformans, suggesting that at least in the models tested, the global de-repression of carbon catabolite genes is not detrimental to growth within the host or the expression of required virulence factors (Zaragoza et al., 2000, Caza et al., 2016). Interestingly, although MIG1 deletion had no effect on virulence in a murine inhalation model of cryptococcosis, fungal burden in the blood of MIG1 deletion inoculated animals was significantly higher than wild type inoculated animals at the time of death (Caza et al., 2016). This suggests that although mitochondrial metabolism in this mutant is perturbed, the metabolic state provides a growth or survival advantage in this niche (Caza et al., 2016).

In contrast to the pathogenic yeasts, A. fumigatus CreA plays a role in virulence in a host-specific manner. The absence of CreA results in a significant reduction in virulence of A. fumigatus in a triamcinolone model of IPA highlighted by qualitative differences in immune effector cell function, however, in a neutropenic model of IPA highlighted by quantitative differences in leukocyte populations, the creA-null mutant causes 100% mortality with a one-day increase in median survival compared to wild type. Notably, the defect in virulence in the triamcinolone model is not due to a defect in initiation of infection, as both wild type and the creA null mutant germinate and grow to the same extent over the first 48 hours of infection. Thus, it is fitness in the established infection site microenvironment where the creA null mutant fails (Beattie et al., 2017). In support of this, established biofilms of the creA null mutant are unable to proliferate in low oxygen conditions, a feature of fungal lesions that has been characterized in the triamcinolone model of IPA (Grahl et al., 2011, Beattie et al., 2017). Thus the role of this transcription factor in virulence depends on the context of the host immune status and microenvironmental conditions, and is not required for in vivo growth per se, but for adaptation to the established infection site microenvironment. CreA was thus termed a disease progression factor (DPF) for genes dispensable for infection initiation but critical for infection persistence and progression. It is hypothesized that other key regulators of fungal and host metabolism are likely DPFs given the profound changes that occur in nutrient availability during the course of microbial infections.

In addition, however, the implication of several CCR regulated pathways in vivo and the apparent lack of a role for yeast Mig1p during infection using standard genetic null mutants raises an interesting question as to whether a genetic null mutant is the best way to assess the role of a transcriptional repressor in vivo. If de-repression of alternative carbon source utilization in vivo is required for growth, perhaps a constitutively repressing Mig1p/CreA mutant may yield additional insights to the role of CCR in vivo. Studies in S. cerevisiae have identified a phosphorylation site on serine 311, that when mutated, results in constitutive repression of target genes (Ahuatzi et al., 2007). While it is unclear whether regulatory mechanism of Mig1p homologues in pathogenic fungi are conserved, creation of these types of mutants may give more insight to the role of CCR and the importance of metabolism of alternative carbon sources in various infection models.

The importance of understanding carbon source utilization in vivo is exemplified by studies that show the carbon source and nutrient environment influences virulence determinants and efficacy of drug treatments. Ene and colleagues showed that growth on the physiologically relevant carbon source, lactate, changes fungal cell wall architecture and structure, which affects the physical properties and results in increased resistance to osmotic stress, cell wall perturbing agents and antifungal drugs, a phenomenon that occurs across Candida species (Ene et al., 2012). Further, cells grown on lactate or amino acids were more virulent in an intravenous model of candidiasis (Ene et al., 2012), and lactate-grown cells were phagocytosed less efficiently than glucose-grown cells, and increased IL-10 and IL-17 immune response from PBMCs (Ene et al., 2013). Similar observations in A. fumigatus have revealed altered cell wall composition based on growth medium, where growth on yeast extract resulted in reduced beta- and alpha-glucan levels and increased chitin when compared to cells grown in RPMI. These differences in cell wall composition translate to differences to an increased minimal effective concentration (MEC) of caspofungin in RPMI when compared to yeast extract, and increased efficacy of combinatorial treatment with caspofungin and nikkomycin Z in RPMI versus yeast extract (Clavaud et al., 2012). Furthermore, for C. albicans, glucose, as well as galactose, fructose and sucrose, stimulate hyphal growth, a morphological feature that is considered to be a major virulence factor of this organism (Hudson et al., 2004, Maidan et al., 2005b). These changes in virulence attributes and drug resistance are important to understand, especially from the perspective of treatment, given that lab tests for drug susceptibility are performed in RPMI, a glucose-rich media, which given the above results, may not be an accurate reflection of the susceptibilities of strains within a patient. Therefore, understanding the metabolic state of cells within a host not only opens up the possibility of novel metabolic drug targets, but may also help in developing drug susceptibility testing that is more reflective of the in vivo status of the fungus. Table 1 lists all aforementioned CCR-related factors and their importance in virulence.

Table 1.

Factors involved in glucose sensing and metabolism and their role in virulence, when compared to the respective wild-type (WT) strain, in the human pathogenic fungi Aspergillus fumigatus, Cryptococcus neoformans and Candida albicans (NC = not characterized).

S.c protein	Role/function	A. fumigatus		C. neoformans		C. albicans
		Homolog	Virulence	Homolog	Virulence	Homolog	Virulence
Glucose sensing and uptake
Gpr1p	GPCR	GprC	Delayed (Gehrke et al., 2010)	Gpr1 (NC)	NC	Gpr1p	Context dependent (Miwa et al., 2004; Maidan et al., 2005b)
Gpr1p	GPCR	GprD	Attenuated (Gehrke et al., 2010)	-	-	-	-
-	GPCR	GprK	WT (wax moth larvae model; Jung et al., 2016)	-	-	-	-
Gpa2p	G protein	NC	NC	Gpa2	WT (Li et al., 2007)	Gpa2p	WT (Miwa et al., 2004)
Snf3p	Glucose sensor	NC	NC	Hsx1	Delayed (Liu et al., 2013)	Hgt4p	Delayed (Brown et al., 2006)
Rgt2p	Glucose sensor	NC	NC	Hsx2 (essential)	NC
Carbon catabolite repression
Mig1p	Transcriptional repressor of CCR	CreA	Attenuated (Beattie et al., 2017)	Mig1	WT (Caza et al., 2016)	Mig1p	WT (Zaragoza et al., 2000)
Tup1p	Co-repressor of CCR	Rco1	NC	Tup1	Delayed (Lee et al., 2009)	Tup1p	Avirulent (Murad et al., 2001)
Additional factors involved in CCR
Snf1p	Protein kinase	SnfA	NC	SnfA	Avirulent/delayed (Hu et al., 2008a; Yang et al., 2010)	Snf1p (Essential)	NC/Heterozygous mutant - WT (Petter et al., 1997)
Hxk2p	Hexokinase	HxkA/GlkA	NC	Hxk1/Hxk2	Avirulent -double mutant only (Price et al., 2011)	Hxk2p	NC
-	De-ubiquitination enzyme	CreB	NC	-	-	Ubp13p	NC
-	WD-40 repeat protein	CreC	NC	-	-	Cop1p (C3_01720C_A)	NC
-	Arrestin	CreD	NC	-	-	Rod1p	NC

Nitrogen catabolite metabolism in human pathogenic fungi

Ammonium and glutamine sensing and uptake

The preferred nitrogen sources for most microorganisms are ammonium and glutamine as they are easily assimilated and readily used for protein synthesis (Wong et al., 2008). Ammonium and glutamine metabolism are intrinsically linked, with glutamine being synthesised from ammonium and glutamate by the enzyme glutamine synthetase (Margelis et al., 2001). Fungi have evolved several sophisticated nitrogen sensory and uptake systems which allow the cell to scavenge the preferred nitrogen sources, even when present in limiting concentrations. The genome of C. albicans encodes 2 ammonium permeases MEP1 and MEP2, that allow growth under low ammonium concentrations, with Mep2p shown to be crucial for pseudohyphal growth induction through activating downstream mitogen activated protein (MAP) kinase and cAMP-dependent pathways (Biswas and Morschhäuser, 2005; Dabas et al., 2009). Homologues of Mep1p and Mep2p are encoded in the genome of A. fumigatus (MeaA, Afu1g10930) but have not been characterised to date in this fungus. Similar to C. albicans, low concentrations of extracellular ammonium induce pseudohyphal growth in C. neoformans, a process that relies on the two ammonium permeases Amt1 and Amt2 (Lee et al., 2012). Neither Amt1 nor Amt2 were required for C. neoformans pathogenicity when immunocompetent mice were infected by nasal inhalation (Rutherford et al., 2008).

In C. albicans, the amino acid sensor Csy1p was shown to sense and induce the transport of glutamine (Brega et al., 2004). Csy1p plays an important role in the induction of pseudohyphal growth in serum-based media (Brega et al., 2004) and together with the general amino acid permease (AAP) and major nitrogen source sensor Gap2p (Kraidlova et al., 2011) activates the PKA (protein kinase A) pathway which is important for virulence in pathogenic fungi (Donaton et al., 2003; Kraidlova et al., 2016). A direct role of the homozygous CSY1 deletion strain in virulence has not been published (Brega et al., 2004). The genome of A. fumigatus encodes several AAPs, including the orthologue of Gap2p (Afu7g04290), but none have been characterized so far and their role in virulence remains unknown (Sugui et al., 2008; Morton et al., 2011; Gehrke et al., 2010; Oosthuizen et al., 2011). Similar to A. fumigatus, the genome of C. neoformans encodes 10 AAPs whose expression is temperature- and nitrogen source-dependent, with Aap4 and Aap5 (both transporting glutamine) being important for growth during high temperature, oxidative stress and in vivo virulence in an immunocompetent murine model of pulmonary cryptococcosis (Fernandes et al., 2015; Cruz Martho et al., 2016).

The main nitrogen nutrient sensor, and regulator of downstream signalling pathways is the target of rapamycin (TOR) pathway (extensively reviewed in Rohde et al., 2008). The TOR pathway consists of one or two (depending on the fungal species) Tor protein kinases which are activated by amino acid-, including glutamine, derived signals and regulate a wide variety of cellular processes (Crespo et al., 2002; De Virgilio et al., 2006; Su et al., 2012; Staschke et al., 2010). The genomes of C. albicans, A. fumigatus and C. neoformans encode one TOR kinase which is essential for viability. In C. albicans, Tor1p controls NCR, the expression of Gcn4p, morphogenesis, cell adhesion, and condition-specific biofilm formation, whereas in A. fumigatus, amino acid transport and metabolism are under the regulatory control of TorA (Bastidas et al., 2009; Baldin et al., 2015; Cruz et al., 1999; Lee et al., 2012). Deletion of the C. albicans gene SIT4, homologue of the S. cerevisiae PP2A-like protein phosphatase Sit4p which mediates NCR repression by TORC1, significantly affected yeast and hyphal growth and resulted in a strain with attenuated virulence in an immunocompetent mouse model of systemic candidiasis (Lee et al., 2004). In C. albicans, Tor1p-mediated signalling is controlled by the small GTPase Rhb1p and the GTPase-activating protein Tsc2p, both of which are thought to control the expression of the ammonium permease Mep2p (Tsao et al., 2009). Rhb1p homozygous deletion mutants are viable although a role in infection of this strain has not been investigated yet. Similar to C. albicans, the A. fumigatus PP2A-like protein phosphatase SitA, was shown to be important for the cell wall integrity (CWI) pathway, adhesion and the deletion of the corresponding gene resulted in an avirulent strain when administered by nasal instillation in a neutropenic mouse model (Bom et al., 2015). Similarly, deletion of the GTPase RhbA, homologue of C. albicans Rhb1p, resulted in a strain with attenuated virulence when in a neutropenic mouse of pulmonary aspergillosis (Panepinto et al., 2003). In C. neoformans, Sin1, the orthologue of the S. cerevisiae TORC2 complex component Avo1p, and Ypk1, which is involved in the TOR-dependent phosphorylation of ribosomal proteins in S. cerevisiae, were shown to be important for virulence in an immunocompetent murine model of infection when administered via tail vein injection (Lee et al., 2012). Additional TOR-mediated, nitrogen metabolism-related signalling processes have not been studied in C. neoformans.

Nitrogen catabolite repression (NCR)

The preferential utilisation of easily assimilated nitrogen sources, such as ammonium and glutamine, is known as NCR. In A. nidulans, NCR is mediated by the co-repressor NmrA, with the deletion of the corresponding gene resulting in the partial de-repression of genes encoding enzymes required for the utilisation of alternative nitrogen sources such as asparagine and urea in the presence of favoured nitrogen sources (Andrianopoulos et al., 1998). The genome of A. fumigatus encodes a homologue of A. nidulans NmrA (Afu5g02920) although it has not been characterised and a role in virulence has not been studied to date. Homologues of NmrA do not appear to present in the genomes of C. albicans and C. neoformans.

Upon the depletion of the preferred nitrogen sources ammonium and glutamine, the fungus switches from nitrogen anabolism to catabolism by inducing the expression of catabolic genes required for the scavenging and utilisation of energetically less favoured nitrogen sources (Tudzynski, 2014). The GATA-type transcription factor Gat1p/AreA/Gat1 induces the expression of catabolic genes in C. albicans, A. fumigatus and C. neoformans. The functions as well as the role in virulence of Gat1p/AreA/Gat1 diverge in all three fungal species. Deletion of GAT1 in C. albicans still allowed the use of a number of nitrogen sources except for isoleucine, tyrosine and tryptophan (Limjindaporn et al., 2003). The genome of C. albicans encodes a second GATA-type transcription factor Gln3p, which shares overlapping functions with Gat1p, regulating the expression of genes encoding enzymes involved in nitrogen metabolism and ammonium or amino acid uptake (Liao et al., 2008). Both transcription factors have been shown to be crucial for virulence in an immunocompetent murine model of systemic disseminated candidiasis (Limjindaporn et al., 2003; Liao et al., 2008), which may be due to the inability of the homozygous deletion mutants to induce filamentation in a nitrogen source-dependent manner (Liao et al., 2008). Deletion of areA in A. nidulans and A. fumigatus abolishes growth on a wide range of nitrogen sources except for ammonium and glutamine (Langdon et al., 1995; Hensel et al., 1998). AreA was shown to contribute to, but not be essential for A. fumigatus virulence in a neutropenic murine model of pulmonary aspergillosis, with the deletion mutants presenting a delayed growth phenotype during infection (Hensel et al., 1998). In C. neoformans, deletion of GAT1 resulted in severely reduced growth on most amino acids (except arginine and proline), urea and ammonium (Kmetzsch et al., 2011). Knockout of GAT1 in C. neoformans was shown to be important for the production of capsular glucuronoxylomannan, but did not affect virulence in an immunocompetent murine model of pulmonary cryptococcosis (Kmetzsch et al., 2011; McClelland et al., 2005).

NCR is subject to negative regulation and in S. cerevisiae, this is mediated by the two GATA transcription factors Dal80p and Gzf3p. Both regulators are predicted to antagonise the functions of Gln3p and Gat1p by competing for the same binding site in NCR-regulated genes (Coffman et al., 1997; Svetlov and Cooper, 1998; for an extensive review please refer to Magasanik and Kaiser, 2002). The genome of C. albicans encodes only one homologue of S. cerevisiae Gzf3p and Dal80p, termed GZF3, whose deletion resulted in altered colony morphology at 30°C and 37°C (Homann et al., 2009). A role for this transcription factor in virulence has not been investigated yet. In A. nidulans, the homologue of Dal80p is the GATA zinc finger transcription factor AreB, which was predicted to antagonise the activation of AreA under nitrogen-limiting or – starvation conditions (Wong et al., 2009). The mechanistic nature underlying AreB-mediated regulation of AreA is currently unknown. In addition, in the presence of ammonium and glutamine, NmrA interacts with AreA, preventing the induction of genes required for the utilisation of non-preferred nitrogen sources; whereas in the absence of ammonium and glutamine, NmrA dissociates from AreA, allowing nitrogen catabolite gene activation (Andrianopoulos et al., 1998; Han et al., 2016). Transcription of nmrA and areA is inversely regulated, with both genes responding to available nitrogen sources in an opposite manner, whereas overexpression of nmrA prevented AreA function irrespective of the nitrogen source (Wong et al., 2007). Furthermore, nmrA transcription was shown to be dependent on MeaB, a bZIP DNA-binding protein, with transcriptional expression patterns of both nmrA and meaB mirroring each other (Wong et al., 2007). The expression of meaB was repressed under nitrogen starvation conditions and induced in the presence of preferred and alternative carbon sources (Wong et al., 2007). In A. fumigatus, the homologue of MeaB (Afu3g10930) has not been characterised. In C. neoformans, additional regulatory proteins involved in NCR have not been described and protein BLAST analysis of the aforementioned S. cerevisiae, C. albicans and A. nidulans factors against the C. neoformans genome database did not generate any significant hits.

Similar to CCR discussed above, the aforementioned studies highlight substantial differences in the regulation of NCR and the role it plays in establishing and maintaining an infection in all three human fungal pathogens. In Aspergillus spp., there appears to be a clear division between the utilisation of preferred and non-preferred nitrogen sources, with both processes being regulated by two independent factors which also control the function of each other (Adrianopolous et al., 1998). In C. albicans, two independent transcription factors also regulate nitrogen metabolism, although a clear division between NCR and NC induction, as seen in Aspergillus spp., is not observed. Both transcription factors rather have overlapping and specific functions regulating genes encoding enzymes required for the utilisation of a wide range of nitrogen sources, thereby creating a complex regulatory network underlying nitrogen metabolism in this fungus (Limjindaporn et al., 2003; Liao et al., 2008). In C. neoformans, only one transcription factor involved in NCR has been characterised that is required for growth on both preferred and non-preferred (with the exception of proline and arginine) nitrogen sources (Kmetzsch et al., 2011). It remains to be determined how Gat1 regulates the utilisation of different nitrogen sources and whether a second NCR transcription factor exists that could be involved in these processes.

Additional factors involved in ammonium and glutamine acquisition

In C. neoformans, the enzyme urease, which converts urea into ammonium, is an important virulence factor, that is thought to contribute, together with additional factors, to the transmigration of the blood-brain-barrier (BBB) (Shi et al., 2010). Indeed, disruption of the C. neoformans urease-encoding gene resulted in attenuated virulence in both an immunocompetent murine inhalation and intravenous infection model of cryptococcosis (Cox et al., 2000). The exact mechanism by which urease production enables dissemination into the brain is not known but ammonia production could increase adherence or toxicity to capillary endothelial cells (Olszewski et al., 2004; Shi et al., 2010). The genomes of both C. albicans and A. fumigatus also encode ureases but a role for these enzymes in infection has not been determined.

A key feature and major virulence factor of pathogenic fungi is their ability to secrete different classes of proteases which degrade proteins, therefore releasing peptides and/or amino acids which can subsequently be taken up and metabolised (Monod et al., 2002) when extracellular nitrogen sources are scarce. In C. albicans, ten aspartic proteinases (Sap1p-Sap10p) are secreted in an infection stage-dependent manner and have been shown to mediate various virulence attributes such as adhesion, biofilm formation and host tissue penetration (Naglik et al., 2003; 2004; 2008). Furthermore, they present various immunomodulatory properties, thus allowing evasion of the host immune responses (Gropp et al., 2009; Bras et al., 2012; Wu et al., 2013). Single deletions of SAP1, SAP2 and SAP3 resulted in attenuated virulence when injected intravenously into immunocompetent mice (Hube et al., 1997) whereas the simultaneous deletion of SAP4, SAP5 and SAP6, led to reduced growth in protein-rich medium and attenuated virulence in immunocompetent mouse and guinea pig models of systemic candidiasis (Sanglard et al., 1997). Sap8p is essential for regulating CWI and biofilm formation through the proteolytic cleavage of Msb2p, a sensor of the Cek1p-MAPK pathway, a factor that is important for virulence (Puri et al., 2012).

The genome of A. fumigatus is predicted to encode at least 99 secreted proteases (Machida et al., 2005; Nierman et al., 2005) of which different mixtures are secreted in a nitrogen source-dependent manner (Farnell et al., 2012). Protease secretion is under the control of NCR mediated by the transcriptional regulator, PrtT, which was shown to not be essential for virulence when neutropenic and cortisone acetate-treated mice models of pulmonary aspergillosis (Bergmann et al., 2009; Sharon et al., 2009). Similar to C. albicans, proteases modulate host immune responses (Behnsen et al., 2010) and are thought to contribute to A. fumigatus virulence by facilitating host tissue degradation and invasion (Gautam et al., 2006; Balenga et al., 2015; Kogan et al., 2004). Deletion of the alkaline serine protease alp1/aspf13 led to significantly reduced virulence in nasally infected neutropenic mice (Kolatukuddy et al., 1993), although a more recent study showed that disruption of alp1 had no significant effect on virulence, when the strain was administered intranasally, in a cortisone acetate mouse model (Behnsen et al., 2010). This discrepancy may be due to the use of different mouse models or fungal strains. A high degree of redundancy is likely to exist amongst these secreted proteases as no other secreted protease has been shown to play an important role in virulence in A. fumigatus.

Proteases secreted by C. neoformans also have immunomodulatory properties (Chen et al., 1996) and some have been found to play a role in modulating cell wall integrity (CWI) pathway components that are important for virulence (Bien et al., 2009). The metalloproteinase Mpr1 was shown to be required for establishing an infection in the central nervous system (CNS) by promoting migration of C. neoformans across the brain endothelium (Vu et al., 2014). Deletion of MPR1 resulted in attenuated virulence due to reduced brain fungal burden in an immunocompetent mouse model of pulmonary and systemic cryptococcosis (Vu et al., 2014). Similarly, C. neoformans serine proteases are thought to increase the permeability of the blood brain barrier (BBB) therefore facilitating CNS invasion (Xu et al., 2014).

Role of NCR in virulence

The utilisation of preferred nitrogen sources affects various virulence determinants in human pathogenic fungi. In C. albicans, limiting ammonium levels induce filamentous growth, a prominent virulence factor (Biswas and Morschhäuser, 2005). Furthermore, several amino acids, including glutamine, promote alkalinization of the extracellular environment, triggering hyphal morphogenesis and ensuring fungal survival in environments with a range of different pHs which is predicted to be important during infection establishment (Vylkova et al., 2011). In C. neoformans, growth in the presence of ammonium results in reduced capsule size but increases cellular melanin content (Kronstad et al., 2011) that has been determined to be an important virulence factor (Jacobson, 2000). In A. fumigatus however, a relationship between the utilisation of preferred and non-preferred nitrogen sources and the elaboration of virulence factors has not been extensively studied.

In general, carbon metabolism has received more attention than nitrogen metabolism, although several studies investigating gene expression in pathogenic fungi during early pulmonary infection showed induction of genes required for amino acid biosynthesis and transport, thereby suggesting nitrogen starvation is an important stress response faced by fungi during infection establishment (McDonagh et al., 2008; Hu et al., 2008). Further characterisation of nitrogen metabolism in C. albicans, A. fumigatus and C. neoformans during later infection time points and in different host niches is therefore required in order to further dissect the role of nitrogen acquisition during in vivo host infection and its relevance for fungal virulence determinants. NCR is intrinsically embedded in a complex regulatory network that finely tunes nitrogen metabolism, adapting the fungal cells to different types and concentrations of nitrogen sources, influencing virulence factor and potentially disease outcome. Table 2 summarises all aforemention factors that are important for NCR as well as their role in virulence in all three human pathogenic fungi.

Table 2.

Factors involved in ammonium and amino acid sensing and metabolism and their role in virulence, when compared to the respective wild-type (WT) strain, in the human pathogenic fungi Aspergillus fumigatus, Cryptococcus neoformans and Candida albicans (NC = not characterized).

		Homolog	Virulence	Homolog	Virulence	Homolog	Virulence
Nitrogen sensing and uptake
Mep3p	Ammonium permease	MeaA	NC	Amt1	WT (Rutherford et al., 2008)	Mep1p	NC
Mep2p	Ammonium permease	Afu1g10930	NC	Amt2	WT (Rutherford et al., 2008)	Mep2p	NC
Ssy1p	Amino acid sensor	-	-	-	-	Csy1p	NC
Gap1p	Nitrogen sensor and amino acid permease	Afu7g04290	NC	Aap4	Avirulent (Cruz Martho et al., 2016)	Gap2p	NC
				Aap5
Tor1p, Tor2p	Protein kinase, major nitrogen sensor	TorA	NC	Tor1	NC	Tor1p	NC
Sit4p	Protein phosphatase	SitA	Avirulent (Bom et al., 2015)	NC	NC	Sit4p	Attenuated (Lee et al., 2004)
Rhb1p	GTPase	RhbA	Attenuated (Panepinto et al., 2003)	NC	NC	Rhb1p	NC
Nitrogen catabolite repression (NCR)
-	Co-repressor	NmrA	NC	-	-	-	-
Gat1p	GATA-type transcription factor	AreA	WT (Hensel et al., 1998)	Gat1	WT (McClelland et al., 2005)	Gat1p	Avirulent (Limjindaporn et al., 2003)
Gln3p	GATA-type transcription factor	-	-	-	-	Gln3p	Attenuated (Liao et al., 2008)
Gzf3p	GATA-type transcription factor	AreB	NC	-	-	Gzf3p	NC
-	Transcription factor	MeaB	NC	-	-	-	-
Additional factors involved in NCR
Dur1p	Urease	Afu1g04560 (UreB)	NC	Ure1	Attenuated (Cox et al., 2000)	Dur1p	NC
Dur2p		Afu2g12900 (UreD)				Dur2p
Yps3p	Aspartic proteinase	OpsB	NC	-	-	Sap1p	Attenuated (Hube et al., 1997)
Yps3p	Aspartic proteinase	OpsB	NC	-	-	Sap2p	Attenuated (Hube et al., 1997)
Yps3p	Aspartic proteinase	OpsB	NC	-	-	Sap3p	Attenuated (Hube et al., 1997)
Yps3p, Bar1p	Aspartic proteinases	OpsB	NC	-	-	Sap4p, Sap5p, Sap6p	Attenuated/Triple mutant only (Hube et al., 1997)
-	Metalloproteinase	Mep/Aspf5	NC	Mpr1	Attenuated (Vu et al., 2014)	-	-
Urc2p	Transcription factor	PrtT	WT (Bergmann et al., 2009; Sharon et al., 2009)	NC	NC	Zcf19p	NC

Future perspectives

The aforementioned studies highlight the importance of carbon and nitrogen metabolism mechanisms in fungal pathogenicity and virulence as they are required for infection establishment and progression and modulating production and activity of virulence factors such as fungal cell wall structure/content and melanin, or in the case of yeast, phenotypic dimorphism. The current understanding on when and where specific pathways are required during an infection and their mechanistic nature, presents a major knowledge gap and detailed characterisation of these pathways in in vivo-relevant conditions is expected to lead to the identification of unique targets or strategies for anti-fungal drug development. In this respect, fungal species-specific niches and mechanisms need to be studied, preferably in multiple clinically relevant animal models, as the severity and degree of an infection varies with the underlying immune system disturbance. In contrast to A. fumigatus, C. albicans is often isolated from blood cultures, supporting crucial differences in infection exerted by different fungal pathogens. The role of CCR and NCR in this process remains unclear as insufficient data exists, but systemic dissemination is likely to rely on additional factors such as fungal morphology and the progression and/or the development of the immune status of the patient. A. fumigatus rapidly germinates in the lung airways into hyphae that may be more difficult to disseminate via the blood stream when compared to yeast cells. Systemic dissemination of A. fumigatus is a rare event, that is not seen in mice models and typically happens very late in affected human patients. Nevertheless, comparative studies with so called non-pathogens may be revealing as shown for differences in metabolic flexibility between S. cerevisiae and C. albicans due to the presence/absence of ubiquitination sites on enzymes required for the utilisation of alternative carbon sources (Sandai et al., 2012).