The roles of zinc and copper sensing in fungal pathogenesis

Highlights • Zinc and copper are essential trace elements required for cell function.• Nutrient Immunity restricts zinc and copper access and mediates toxicity.• Divergent fungi integrate zinc and copper responsive regulons for pathogenesis.


Introduction
The concept of nutritional immunity, in terms of hostdriven iron sequestration, has been appreciated for decades: our bodies maintain extremely low iron cation levels via intracellular sequestration (ferritin, haemoglobin, the hepcidin axis) and expression of extracellular high-affinity iron-binding proteins (transferrin, lactoferrin). Successful pathogens have in turn evolved effective assimilation mechanisms (high-affinity transporters, siderophores and specialised binding proteins) [1]. However, additional layers of nutritional immunity have been revealed in recent years: micronutrient restriction is not limited to iron, with host zinc and manganese sequestration playing key roles in controlling infection [2]. Moreover, in certain host niches, pathogens can be exposed to potentially toxic levels of iron, zinc and copper. Therefore, beyond highaffinity iron uptake systems, microbial pathogens must also possess highly effective homeostatic mechanisms (both assimilation and detoxification) for other trace metals within their hosts [2].
Host nutritional immunity during fungal infection and the pathogen assimilation pathways employed in counterattack have been extensively discussed in a number of recent reviews [3][4][5][6]. In this mini-review we focus on two emerging areas: the adaptive responses of Candida and Cryptococcus species to changes in environmental zinc and copper.

Divergent virulence factor intercalation of the zinc regulon
Zinc is the second most abundant transition metal in the human body but, like iron, its availability to microbial pathogens is strictly limited. Zinc restriction is further compounded during periods of inflammation due to the action of calprotectin. Calprotectin constitutes approximately half the cytoplasmic protein content of neutrophils, is a dominant component of NETs (neutrophil extracellular traps) and elicits antifungal activity via zinc chelation [7]. Calprotectin is a heterodimer of S100A8 and S100A9, members of the S100 family of low molecular weight proteins. Antifungal activity of another S100 family member, psoriasin (S100A7) has recently been reported to elicit antifungal activity via zinc-sequestration [8 ], suggesting that multiple members of this protein family may have similar functions. Furthermore, host cell internalisation and subcellular compartmentalisation further limit accessibility to extracellular and intracellular pathogens, respectively. For example, activated, Histoplasma-infected macrophages shuttle zinc to their Golgi, restricting fungal proliferation and promoting clearance [9].
Interestingly, contemporary pathogenic fungal species appear to simultaneously exhibit remarkably similar and divergent responses to changes in zinc availability. First described in the model yeast Saccharomyces cerevisiae, Zap1 is a zinc finger transcription factor that regulates the expression of zinc transporter-encoding genes and is essential for maintaining zinc homeostasis. Under zinc limitation, Zap1 exhibits positive auto-regulation, rapidly amplifying its own transcription and subsequent protein levels and positively regulating the expression of zinc transporter-encoding genes, including ZRT1, ZRT2 and ZRT3. For a detailed discussion of S. cerevisiae Zap1 function, readers are directed to an excellent recent review by Wilson and Bird [10].
It would appear that Zap1 orthologues function as conserved master regulators of zinc homeostasis in fungi: in the pathogenic fungal species Candida albicans, Candida dubliniensis, Aspergillus fumigatus, and Cryptococcus deuterogattii, Zap1 orthologues (also known as Csr1 in Candida, ZafA in Aspergillus) have been directly demonstrated to regulate the expression of zinc transporter-encoding genes. Given the evolutionary divergence of these species, it is plausible that orthologues of Zap1 regulate zinc homeostasis in many (if not all) fungi [11][12][13][14]. Underscoring the importance of fungal zinc homeostasis during infection, each fungal pathogen tested to date exhibits virulence attenuation or decreased infectivity upon deletion of their respective ZAP1 orthologue in relevant animal models of infection ( Figure 1).
Despite this apparent conservation of function as a regulator of zinc homeostasis and virulence, there may be differences in the Zap1-regulons of different contemporary fungal species. Cryptococcus neoformans and C. deuterogattii are closely related species that express three major virulence factors for infectivity: capsule formation, melanin production and growth capacity at 37 8C. Interestingly, while C. deuterogattii ZAP1 is dispensable for capsule and melanin production [14], C. neoformans cells lacking the ZAP1 orthologue (ZAP104) are deficient in capsule, melanin, and a/a cell fusion, suggesting rewiring [15,16 ]. While we cannot rule out that C. deuterogattii Zap1 and zinc homeostasis may influence melanin and capsule formation under conditions not examined by Schneider et al. [14], Liu et al. and Jung et al. independently demonstrate ZAP104 involvement in virulence factor expression (melanin and capsule synthesis) in C. neoformans [14,15,16 ]. These three studies together suggest that Zap1-mediated zinc homeostasis in C. neoformans, but not C. deuterogattii, may impinge on virulence factor (capsule/melanin) expression, and raise questions about the reasons for phenotypic differences. An examination of Zap1 and Zap104 protein sequences immediately reveals structural differences: where the Zap1 N-terminus encodes a single zincbinding domain, Zap104 encodes two [14]. Whether these structural differences lead to differing sensitivities to zinc availability or differences in promoter binding remain to be explored.
In the Candida genus, Zap1 function has been investigated in three relatively closely related species [17,18]: C. albicans, C. dubliniensis and C. tropicalis. Despite diverging from basidiomycetes approximately 500 mya [19], C. albicans, like C. neoformans, appears to regulate major virulence attributes (hyphal morphogenesis, adhesion and biofilm maturation) via Zap1 (note that Zap1 in Candida is also known by the common name Csr1). Kim et al. [20] first reported that ZAP1 deletion in C. albicans precluded hyphal morphogenesis under a range of environmental conditions. Moreover, overexpression of ZRT1 or ZRT2 (encoding the two predicted plasma membrane zinc transporters of C. albicans) promoted filamentous growth of the zap1D mutant on certain media, suggesting that the morphogenic defect may be due to perturbed Micronutrient sensing and fungal pathogenesis Ballou and Wilson 129  Phylogenetic and functional relationships of Zap1 orthologues in fungi. The predicted amino acid sequences of Zap1 orthologues were downloaded from FungiDB, Aspergillus Genome Database, Saccharomyces Genome Database and Candida Genome Database and aligned using Phylogeny.fr. Note that in all reported cases, Zap1 regulates the expression of zinc transporters and is required for virulence or infectivity in animal models of infection, but that confirmed pathogenicity factor expression is species-specific. cellular zinc homeostasis rather than direct Zap1-regulation of hypha-formation.
A recent study of C. dubliniensis found that, in this close relative of C. albicans, Zap1 also positively regulates ZRT1, ZRT2 and PRA1 [12 ], reinforcing the concept that Zap1 is the universal regulator of zinc assimilation in the fungal kingdom ( Figure 1). However, unlike C. albicans, the C. dubliniensis zap1D mutant was not defective for hypha formation. On the other hand, Zap1 does regulate filamentation in C. tropicalis [21].
Using a powerful combination of large-scale mutant library screening, transcriptional profiling, chromatin immunoprecipitation, regulatory network analysis and infection modelling, the group of Aaron Mitchell found that Zap1 regulates extracellular matrix production by C. albicans biofilms and governs the expression of adhesin molecules. The zap1D mutant efficiently formed biofilms; however, these structures produced highly elevated levels of b-glucan, indicating that Zap1 negatively regulates extracellular matrix production -a key attribute in biofilm maturation [11]. Zap1 occupies ZRT1, ZRT2 and PRA1 promoters and positively regulates their transcription, in agreement with Zap1 acting as the master regulator of zinc homeostasis in C. albicans. However, overexpression of these zinc assimilation factors in the zap1D mutant had no effect on b-glucan production, suggesting that cellular zinc homeostasis per se may not directly regulate extracellular matrix synthesis, and that Zap1 possesses additional regulatory function [11]. Indeed, Zap1 was also found to be required for the regulation of genes involved in adhesion -another key element of biofilm formation [22]. Despite this regulatory role, the zap1D mutant itself exhibited wild type levels of adhesion. Therefore, to test whether Zap1 functions redundantly in adherence regulation, these authors tested the effect of ZAP1 overexpression in strains lacking other transcription factors important for adhesion. This approach demonstrated that ZAP1 overexpression restored the adhesive capacity of multiple adhesion-defective transcription factor mutants. This was most likely due to Zap1-dependent transcriptional induction of CSTAR (cell surface targets of adherence regulators) genes, as ZAP1 overexpression was found to increase the expression of these target genes in mutants lacking three direct regulators of adhesion (zcf28D, try2D and try3D). Therefore, in addition to zinc assimilation, C. albicans appears to have intercalated key virulence attributes -adhesion and biofilm maturation -into its Zap1 regulon. It should be noted that, in these studies, zinc limited media was not used, and Zap1 may regulate different subsets of genes, depending on the environmental zinc status.
In summary, it would appear that beyond zinc homeostasis, the Zap1 transcription factor plays different roles in different fungal species and, at least in C. neoformans and C. albicans, these are directly associated with the expression of 'classical' pathogenicity factors. So why might fungal pathogens place control of key virulence attributes within their Zap1 regulon? One likely possibility is the relative zinc levels found in the natural reservoirs of these yeasts, compared to the infected host. Both arboreal environments and the mammalian digestive tract are relatively high zinc environments, whilst nutritional immunity creates extreme zinc depletion within infected tissues. Hardwiring virulence factor expression into their zinc starvation-responses may contribute to the high pathogenic potential of these species, similar to the situation for iron [23].

Dynamic regulation of response to nutritional copper immunity
In contrast to zinc nutritional immunity, which functions primarily via micronutrient sequestration by the host, copper nutritional immunity appears to represent a highly dynamic system [3,23]. During C. albicans bloodstream infection, serum copper levels rise early in infection [24]. Upon kidney colonisation, an early spike in copper levels is followed by rapid copper sequestration. For C. neoformans, while the lungs are a high copper environment, alveolar macrophages induce host CTR1 but repress the ATP7A transporter for phagosomal copper compartmentalization, resulting in copper starvation upon engulfment [25]. Consistent with this, C. neoformans requires the copper transporter CTR4 for survival within alveolar macrophages but it is dispensable for growth in the lungs [26 ,27]. Although cerebral spinal fluid (CSF) copper levels are estimated at 100 mM under certain conditions, the CSF is generally copper poor and induces copper transporter expression [27,28]. Likewise, upon crossing the blood-brain-barrier, the low copper availability of the brain necessitates CTR1 or CTR4 for C. neoformans virulence [27]. Responding to this dynamic environment requires a coordinated transcriptional response on the part of the fungus.
Structurally, fungal copper responsive transcription factors fall into three classes typified by S. cerevisiae Ace1 and Mac1 and S. pombe Cuf1 (Figure 2). All maintain an N-terminal copper responsive (R/K)GRP motif and conserved N-terminus [29 ]. In ScAce1 and ScMac1 residues 1-40 encode a zinc-binding domain. ScAce1 residues 40-110 encode CXC and CX 2 C motifs necessary to coordinate a tetra-copper cluster for binding MRE promoter motifs. Mac1 is characterised instead by dual cysteine rich C-terminal motifs, REP-I (CXCX 4 CXCX 2 CX 2 H) and REP-II (CXCX 4 CXCX 2 CX 2 H), which sense copper and stabilise DNA binding, respectively [30,31]. Specificity is achieved through division of labour for high and low copper responses: ScMac1 regulates the expression of the CTR1 and CTR2 copper importers in response to copper limitation, while copper toxicity triggers Mac1 degradation and assembly of the ScAce1 tetra-copper cluster, activating DNA-binding and induction of CUP1 and CRS5 metallothioneins and SOD1, which in turn further negatively regulates Mac1 [32][33][34]. Likewise, the C. glabrata Ace1 orthologue Amt1 is required for high copper and metallothionein expression and C. glabrata encodes a putative Mac1 homolog with overall structural similarity to ScMac1 [35,36] (Figure 2).
SpCuf1 is representative of the third class, with a single, structurally distinct REP motif: CXCX 3 CXCX 2 CX 2 H. SpCuf1 is required for growth on limited copper through binding CuSE motifs in the CTR4 promoter; in the presence of high copper, Cu-REP interaction induces a conformational change masking the N-terminal nuclear localisation signal and inhibiting function [37]. S. pombe lacks metallothineins, but detoxifies copper through a dual function SOD chaperone, PCCS [38].
C. albicans responds to both low and high copper levels via CaMac1, which induces either Cu-SOD1 in high copper or Mn-SOD3 in low copper, along with the CTR1 copper importer; both CaMAC1 and CTR1 are required for hyphal growth, and SOD1 is required for virulence [24,[39][40][41].
CaACE1/CUP2 is additionally required for growth on high copper and regulates CUP1 and CRD2 metallothioneins [42,43]. C. dubliniensis does encode MAC1 and AMT1 homologues that presumably act in a similar fashion, however their targets remain uncharacterised (Figure 2).
Cryptococcus Cuf1 is required for growth on both low and high copper and is structurally distinct from SpCuf1 and ScMac1/Ace1: in addition to the Cuf1 REP motif, a second cysteine-rich repeat (CCX3CX4CXCX3CCXCCXC) is conserved in both C. neoformans and C. deneoformans, estimated to have diverged 18-24 mya [44,45]. For the outbreak strain C. deuterogattii, Cuf1 sequences in multiple isolates (R265, LA55, RAM5) have undergone truncations that specifically excise the copper responsive REP motifs (Figure 2), although these motifs are maintained in C. decagattii (IND107), C. gattii (EJB2, Ru294, WM276) and C. bacillisporus (CA1873, Ca1280). The relationship between copper and Cuf1 activity has not yet been reported in these organisms.
For both C. neoformans and C. deneoformans, growth in low copper requires CTR1/CTR2 (CNAG_07701) and CTR4 Micronutrient sensing and fungal pathogenesis Ballou and Wilson 131  Phylogenetic and structural relationships of copper responsive transcription factors in fungi. The predicted amino acid sequences of the indicated orthologs were downloaded from FungiDB, Broad, NCBI, Saccharomyces Genome Database, Pombase, and Candida Genome Database and aligned using Phylogeny.fr. Structural schematics were constructed based on sequence analysis and functional analysis is based on the reported literature, discussed in the text. [25,26 ,46,47]. While CTR1 is constitutively expressed, CTR4 is specifically induced by low copper in a CUF1dependent manner. During growth on high copper, CUF1 mediates CMT1 and CMT2 metallothionein expression [46]. Similarly, Lin et al. report that C. deneoformans Mac1/ Cuf1 is required for growth on both high and low copper [48]. This dual activity represents a rewiring in comparison to Cuf1 activity in S. pombe, where high copper leads to the sequestration of Cuf1 in the cytoplasm [37]. In addition to growth, copper levels directly influence Cryptococcus virulence through capsule and melanin via Cuf1independent and Cuf1-dependent mechanisms. In both species, Cuf1 is required for melanin production in a low copper environment, likely because melanin synthesis requires the LAC1 copper-dependent oxidase and a cuf1D mutant is unable to acquire sufficient copper from the environment [46,[48][49][50]. Lac1 copper loading is mediated by the CCC2 P-type copper transporting ATPase, ATX1 copper chaperone, and the ClC-A chloride channel. However, in copper-replete conditions, Cncuf1D melanin defects persist in the absence of growth defects [16 ,47], and the cuf1D null is more readily phagocytosed [51]. Although a complete transcriptional analysis of Cuf1 and copper homeostasis during virulence factor induction has not been performed, CTR1 and CTR4, but not CUF1, are induced by the temperature shift sufficient to induce capsule during growth in DMEM [51,52]. In C. deneoformans, excess copper induces filamentation and is dependent on Ccc2 activity [48]. Interestingly, a C. albicans gpa2D/D mutant is deficient for CTR1 and FRE7 copper importer expression and inappropriately expresses CRD2, suggesting a role for cAMP/PKA in copper homeostasis and hyphal growth. In C. neoformans, cAMP/PKA regulates melanin and capsule via integration with the pHresponsive transcription factor Rim101 and ESCRT [53,54]. Cnrim101 and ESCRT mutants vps25 and rim20 are sensitive to low copper, although capsule and immune evasion defects in these mutants are likely due to changes in the cell wall rather than defects in expression [51,55]. CnCCC2 and ATX1 are also required for growth on low iron, likely due to the interaction between copper and iron uptake [49].
In summary, Candida and Cryptococcus species experience significant shifts in zinc and copper availability upon transitions between commensal or environmental and infective stages. It would appear that certain successful fungal pathogens have hardwired virulence factor expression into their metal ion sensing machinery and may use the metal ion environment of the host as a key signal. A major challenge for the future will be to understand how these signals are integrated, and whether we can therapeutically target these pathways to treat fungal infections.