The Aspergillus fumigatus cell wall integrity signaling pathway: drug target, compensatory pathways, and virulence

1The fungal cell wall (CW) is the exoskeleton of fungal cells. Besides its starkness and firmness, it has a very flexible structure, which, therefore, can alter rapidly and efficiently in response to external and internal stimuli. In particular, filamentous fungi adapt their internal pressure and turgor in concert with CW-biosynthesis enzymes, in order to direct hyphal growth, following gradients of nutrients, and chemo attractants (e.g., hormones), or to avoid adverse habitats.

2In recent years, the composition of the fungal CW was extensively studied. It varies among different species but is mostly composed of polymers of sugars, which show a high degree of branching. Although the CW composition varies among different fungal species, there are conserved parts such as a common core composed of branched β-1,3-glucan-chitin. An exception was found in zygomycetes, which contain chitosan instead of chitin. The polysaccharide-based three-dimensional network in Aspergillus fumigatus is completed by the addition of sugar-chains composed of α-1,3-glucan, galactofuran, and mannan, which make the structure rather variable compared to other fungi. Besides the polysaccharide structure, the CW is adorned with a variety of proteins. In A. fumigatus dormant conidia, the cell surface is covered by a rodlet layer, which is composed of regularly arranged RodA hydrophobin proteins. This hydrophobin envelope was found to be essential to immunologically mask the conidia, which are continuously inhaled by humans. Other hydrophobins, which are produced at different developmental stages in A. fumigatus, were associated with drug response and biofilm-growth conditions.

3In the CW biosynthesis, the β-1,3-glucan synthase (Fks1) plays the major role. Fks1 is an integral plasma membrane protein having 16 putative trans-membrane helices. The genome of the majority of fungi normally contains only a single β-1,3-glucan synthase gene. Consequently, the deletion of this gene was supposed to be lethal. However, a recent publication reported the characterisation of an A. fumigatus Δfks1 mutant, which appeared to be viable besides showing a severe growth phenotype.

4The A. fumigatus genome harbors many putative genes responsible for CW modifications. Different β-glucanases and branching enzymes were identified. These genes are more difficult to analyze by classical reverse genetics, because some of them are apparently functionally redundant. An example is given by the β-1,3-glucanosyltransferase (Gel) family, which, in A. fumigatus, is composed of seven different members. Among them, only the deletion of gel4 was reported to be lethal, whereas the Δgel2, Δgel1, and Δgel7 deletion mutant strains were viable.

5For chitin, which is the second most abundant polysaccharide of the CW, a very similar situation was found. Among the eight putative chitin synthase genes identified in A. fumigatus, only the deletion of csmA and csmB led to a significantly altered phenotype, i.e., reduction of the colony radial growth rate and decrease in chitin content, while the mutation of the remaining genes did not result in a significant phenotype, and obvious phenotypical changes of mutants were only reported upon multiple simultaneous gene disruptions.

6The knowledge about genes involved in the biosynthesis of the CW is steadily increasing. In total, more than 30% of the A. fumigatus genome still encodes genes with unknown function. Thus, it is reasonable to assume that in the future further enzymes involved in CW biosynthesis will be discovered, likely having novel enzymatic functions. As an example, it was predicted that A. fumigatus potentially expresses more than 100 different glycosylphosphatidylinositol (GPI) anchored proteins, which are likely to form a bridge between the membrane lipid bilayer and the CW. The majority of these proteins are supposed to play a role in the CW formation, but only very few have been characterized so far.

7From all the different classes of potential antifungal drugs, only three of them have a wide clinical use against invasive and systemic infections caused by A. fumigatus. These compounds belong to the polyenes, azoles, and echinocandins.

8Polyenes and azoles are mainly targeting ergosterol and its biosynthesis, respectively, while echinocandins specifically inhibit β-1,3-glucan formation (Figure 1). However, it is increasingly acknowledged that defense mechanisms against these drugs have common elements. Consistently, a genome-wide screen conducted in Schizosaccharomyces pombe indicated that different ergosterol biosynthesis deficient mutants were also sensitive to β-1,3-glucan synthase inhibitor echinocandins and β-glucanase. A similar observation was made for Candida albicans biofilms, which normally exhibit higher resistance to polyenes depending on ergosterol and β-1,6-glucan synthesis.

9Polyenes belong to the oldest known antifungal compounds. Amphothericin B (AmB), first isolated from Streptomyces nodosus, is the most commonly used polyene against systemic infections. AmB possesses a name-giving amphoteric character, which allows the binding of the lipophilic compound to the cytoplasmic membrane. The mode of action of polyenes is not fully understood. They theoretically bind to all kinds of sterols, with a significant higher affinity to ergosterol (; Figure 1A). The fungicidal activity of AmB is potentially based on the formation of channels in the cell membrane, which lead to increased membrane permeability and leakage of small molecules, inhibition of aerobic and anaerobic respiration, and accumulation of reactive oxygen species (ROS) (; Figure 1A). It was suggested that eight AmB molecules are assembled to a ring-like structure, showing both a hydrophobic interaction to ergosterol with its conjugated double-bond system and forming a hydrophilic pore with its inward directed hydroxyl groups. Thus, the sensitivity toward polyenes is dependent on the composition of membranes, with regard to sterol content, and chemical structure of phospholipid fatty acyl chains.

10Surprisingly, despite the long-term use of polyenes, resistance to AmB and others polyene drugs remains rare. Nevertheless, there are resistant species of Candida (C. glabrata, C. tropicalis, C. lusitaniae), which are characterized by an intrinsically low ergosterol content. The resistance to AmB in Aspergillus terreus was attributed to a higher level of catalase expression that counteracts the AmB-induced formation of ROS. Additionally, resistant environmental Aspergillus flavus strains were isolated, which were characterized by an alteration of CW composition and higher levels of α-1,3-glucan.

11The group of azoles comprise a plethora of synthetic compounds, characterized by the presence of either a 2-N-containing imidazole ring or a 3-N-containing triazole ring with complex side-chains of at least one halogenated phenyl group. In general, azoles show a broad spectrum of antifungal activity, a good toleration by humans and a relative low cytotoxicity, making them the major agents for treatment, and prevention of Aspergillus infections. There are many licensed azole antifungal drugs, but only four, namely fluconazole, itraconazole, posaconazole, and voriconazole, are mainly used for treatment of fungal infections. They all act by inhibiting the fungal cytochrome P450 14α-sterol demethylase, which leads to a decrease in ergosterol in the fungal cytoplasmic membrane and thus to inhibition of growth (Figure 1B). In A. fumigatus, two genes have been assigned to this enzymatic function: cyp51A and cyp51B (erg11). The deduced proteins catalyze the oxidative removal of the 14α-methyl group of lanosterol or eburicol, respectively, (Figure 1B). However, single deletions of genes involved in either demethylation or desaturation of ergosterol intermediates such as cyp51A, cyp51B, and erg3A, erg3B and erg3C, respectively, showed differences in relative sterol composition, but further tests indicated that neither of these genes is individually essential for A. fumigatus survival and virulence.

12The main mechanism of resistance of A. fumigatus against azole antifungal drugs is based on a mutation of the cyp51A gene locus with a conversion of leucine at position 98 into histidine (L98H) in combination with the appearance of a 34-bp tandem repeat in cyp51A promoter region (TR34). In addition to TR34/L98H, also other hot spots for point mutations in the cyp51A gene have been reported, e.g., the exchange of glycine at pos. 54, with different impact on azole resistance in clinical isolates. Recently, a novel azole resistance mechanism was reported, which was caused by a mutation in the CCAAT-binding transcription factor (TF) complex subunit HapE. HapE interacts with the Hap-complex, which is, besides other functions, also important for adaptation to iron starvation and iron excess stress.

13The third and most recent class of clinically used anti-mycotic drugs are echinocandins. These molecules are composed of a cyclic hexapeptide core linked to a lipid side chain. Echinocandins act as non-competitive inhibitors of the β-1,3-glucan synthase (Figure 1C;; ). Caspofungin was the first clinically applied echinocandin (CANCIDAS®, caspofungin acetate), which specifically targets the fungal CW assembly. Inhibition of β-1,3-glucan synthesis results in inhibition of growth, increased osmotic sensitivity, and can even lead to lysis of cells. Besides their specificity, the use of echinocandins in clinical therapy is influenced by two drawbacks: the emergence of resistant strains, and the occurrence of the so-called paradoxical effect, which makes the drug less effective when used at high concentrations.

14Until today, only little is known about naturally occurring resistance mechanisms against echinocandins. The most common mechanism is over-expression of the fks1 gene in clinical isolates of A. fumigatus. Additionally, mutation of the Fks1 protein by substitution of a serine in position 678 by a proline (S678P) makes A. fumigatus resistant to caspofungin. Apart from these changes, the adaptation of A. fumigatus to caspofungin is accompanied by a change in the CW sugar composition. As reported for other fungi, the exposition of A. fumigatus to caspofungin led as well to a decrease in β-glucan content and to an increase of chitin. In particular, these mechanisms seem to be related to the paradoxical effect exerted by this drug. Potential mechanisms that have been suggested to induce chitin synthesis include manipulation of signaling pathways, such as the up-regulation of protein kinase C-encoding gene pkcA, a key component of the cell wall integrity (CWI) pathway, and elements acting in the calcineurin pathway.

15Several of the major findings about signaling pathways were firstly reported in the model fungus Saccharomyces cerevisiae. The advances in genome sequencing allowed the identification of similar or identical signaling pathways in different fungal species, revealing the presence of highly conserved proteins that can be potentially used as antifungal drug targets. In particular, general signaling cascades such as the mitogen activated protein kinases (MAPKs), calcineurin, cAMP, and target of rapamycin (Tor) pathways are highly conserved in the fungal kingdom.

16Among the different signaling pathways, the MAPK CWI signaling plays a major role in CW maintenance. The CWI signal pathway in A. fumigatus is composed of a highly conserved module formed by three MAPKs, namely Bck1 (MAPK kinase kinase), Mkk2 (MAPK kinase), and MpkA (MAPK), which sequentially phosphorylate each other (Figure 2;; ). Upon phosphorylation, MpkA moves into the nucleus, where it likely activates transcriptional regulators. The deletion of one of the three kinases of the MpkA module led to the lack of MpkA phosphorylation, which is the bottleneck in the activation of the entire pathway. The three mutants appear phenotypically identical (Figure 2).

17The lack of phosphorylation of MpkA yields mutants with a typically impaired CW, showing compact colonies and reduced filamentation. Additionally, mutants of all three MAPKs are equally sensitive to all known CW-disturbing compounds, which highlight their incompetence to recover from CW stress.

18RNA-seq analysis performed on the ΔmpkA mutant of A. fumigatus showed that the deletion of this important kinase globally changed the expression of genes putatively involved in primary metabolism. Many genes involved in sugar and amino acid metabolism were differentially regulated, whereas many well-known CW biosynthesis genes were not affected. This unexpected finding suggested that the CWI pathway regulates CW biosynthesis by directly affecting the synthesis of sugars, which are the essential bricks for the CW formation. As a consequence, the role of the CWI pathway seems to be not only the mere structuring of the CW, but it is likely involved in fine-tuning the balance between stress responses and energy consumption in cellular processes required for growth and development. The latter statement is strengthened by the involvement of the CWI pathway in stress responses that are not directly connected to the CW biosynthesis. As an example, MpkA was activated during iron depletion. When A. fumigatus grows under iron starvation conditions, the fungus produces secondary metabolites named siderophores, which are derived from ornithine metabolism. Siderophores are secreted to bind external iron from the environment. The production of such metabolites is rather energy-consuming. The lack of mpkA in A. fumigatus increased the global siderophore production during iron depletion by positively regulating the metabolic flux toward the ornithine metabolic pathway. Living cells use ornithine and its precursor arginine for polyamine production. The functions of these small aliphatic molecules like putrescine and spermidine are still a matter of debate, but it was shown that their concentration accompanies certain developmental transitions or exposure to stress conditions. In fact, the synthesis of putrescine and spermidine has been proven to be essential for life. As a result, the increased production of siderophores occurring during iron starvation conditions negatively influenced the global polyamine production and storage. These results suggested that during iron starvation, a functional MpkA is required to act as a central regulator to balance essential polyamine formation and de novo siderophore production. Additionally, the latter mechanism seems to be independent from the well-known HapX-SreA iron regulation system in A. fumigatus, suggesting that the role of this kinase is epistatic to HapX-SreA in the stress response process.

19Besides siderophore production, the CWI pathway seems to be an important regulator of secondary metabolism in A. fumigatus. Transcriptome data revealed that more than 50% of the identified secondary metabolite gene clusters were affected by the lack of mpkA. This also included the production of important virulence determinants such as gliotoxin and melanins. Melanins are a class of brown-pigmented molecules, which are often associated with the CW. Their main role is to protect the organisms from exogenous stresses, contributing to the first line of defense against external hazards. A. fumigatus produces two kinds of melanins: the dihydroxynaphthalene (DHN) melanin, which is responsible for the characteristic gray-greenish color of the conidia, and pyomelanin, which is produced during tyrosine catabolism. The expression of the DHN-melanin biosynthesis gene cluster was found to be decreased in the ΔmpkA mutant strain. However, these mutants were still able to produce DHN-melanin, suggesting that the role of the CWI pathway in the regulation of the cluster is significant but not decisive. Concerning pyomelanin, a higher production of this compound was found in the ΔmpkA mutant strain. These data confirmed that pyomelanin exerts a protective role, potentially compensating CW stress effects in a dysfunctional CWI background.

20As summarized here, the A. fumigatus CWI pathway is required for many physiological states, and the CW maintenance is not its only function. It is still a matter of debate whether the CWI signaling is actively involved in the regulation of such a variety of processes or whether the observed phenotypes of corresponding deletion mutant strains are a consequence of the strong physiological impairment derived by the lack of a functional phosphorylation cascade. The meaning of the CWI signaling in A. fumigatus for virulence is unclear. The lack of mpkA did not affect virulence in a cyclophosphamide-cortisone-acetate murine infection model. By contrast, the Δmkk2 mutant strain showing the same phenotypes as the ΔmpkA mutant in vitro, displayed reduced virulence in a murine systemic infection model. These results suggest that in spite of the strongly reduced growth, the presence of neutrophil granulocytes is needed to clear the infection caused by these mutants.

21Comparative analysis of signaling cascades between different fungi highlighted that the central core of signaling, i.e., kinases, are conserved, while upstream (receptors), and downstream (TFs) elements are more diverse. In particular, receptor families are conserved, but vary in number. This suggests that activation of signaling is species-specific.

22One of the most studied class of receptors acting upstream to the CWI signaling are the so-called WSC receptors (CWI and stress response components), which are characterized by the presence of highly repetitive domains containing serine and threonine residues. The WSC receptors also contain a putative carbohydrate-binding domain, which suggests that they could function as a bridge between the CW itself and the cytoplasmic membrane. Studies on the S. cerevisiae Wsc1 receptor suggested that these receptors work as a nanospring, able to detect/sense pulses derived from mechanical stress. BLAST analysis revealed that the genome of A. fumigatus contains four putative genes coding for WSC receptors (wsc1, wsc2, wsc3, and mid2; ). All these proteins were found in the cytoplasmic membrane. However, their activity in A. fumigatus appears to be partially redundant, and a clearer CW phenotype was observed when multiple genes were simultaneously deleted. Additionally, the function of wsc2 could not be related to the CWI signaling.

23Other receptors putatively acting upstream of MAPK signaling pathways are the G protein coupled receptors (GPCRs). The number of these receptors also varies between different species. Deletion of the gprC and gprD genes in A. fumigatus produced mutant strains impaired in CW-related stress response.

24In S. cerevisiae, receptors activating the MAPK signaling cascade are connected to the MAPK module via different signaling elements. BLAST analysis revealed that those elements are also conserved in the A. fumigatus genome. Two of them, Rho1 and PkcA, are supposed to be essentials; thus, their action should not only be related to the CWI signaling (Figure 2; ). Rho1 belongs to the Ras homolog family, which is a group of important signaling proteins. In A. fumigatus, Rho1 is localized in the hyphal tip, and it is misplaced by addition of farnesol, which suggests that this Ras protein interacts with putative membrane receptors to transduce external stimuli and to activate signaling cascades. A more evident phenotype was reported for the Rho guanyl-nucleotide exchange factor Rom2, which clearly localizes to the hyphal tips and septa, and its reduced expression strongly effects CW shape and stability.

25Concerning PkcA (protein kinase C), there are no studies describing its function in A. fumigatus, but the repression of this gene in the closely related fungus Aspergillus nidulans affected the production of penicillin, which supports the involvement of the CWI pathway in the regulation of secondary metabolism. However, although these results suggested that the role of Rho1 and PkcA is conserved in fungi, the link between membrane receptors and the activation of the CWI pathway was shown in model yeasts, but not experimentally confirmed in Aspergillus species until now.

26As mentioned before, the MAPK module is supposed to modulate the activity of transcriptional factors, associated with CWI signaling. Transcriptional regulators that act downstream of the CWI signaling pathway in S. cerevisiae were identified by BLAST analysis also in the A. fumigatus genome, but were mostly not analyzed in detail yet. Deletion of two genes coding for putative zinc-finger TFs, named dvrA and ace2, resulted in phenotypes affecting the CW. In particular, mutants lacking dvrA were more resistant to nikkomycin Z, a well-known chitinase inhibitor, and more virulent compared to the wild-type strain. A similar phenotype was observed for the Δace2 strain, which displayed abnormal pigmentation, as well as increased virulence. Nonetheless, besides these phenotypes, it remains to be shown whether these TFs act downstream of the CWI signaling pathway in a MAPK-dependent manner.

27Aspergillus fumigatus is challenged by a multitude of external stimuli, each of them needing an appropriate response. Although several different signal transduction pathways exist to initiate the required transcriptional changes, they would not be able to respond to environmental signals in a balanced way if they only acted separately in a linear manner. Therefore, an interaction between the pathways is likely. For S. cerevisiae, cross talk between the CWI pathway and other signal transduction pathways under different conditions have been described, but in A. fumigatus this interesting and complex field of research is just emerging.

28In S. cerevisiae, the connection between the high osmolarity glycerol (HOG) and the CWI signaling pathway has been elucidated in detail. Under CW stress, the pathways can display co-regulatory roles depending on the stress-inducing agent. It was shown that Hog1 (the yeast SakA ortholog) was activated in the absence of Slt2 (the yeast MpkA ortholog) under conditions causing CW stress, indicating an inhibitory effect of the CWI signaling on the HOG pathway.

29Candida albicans also shows the previously mentioned paradoxical effect in response to high dose exposure to caspofungin. When analyzing the response to this antifungal drug, the involvement of several signal transduction cascades was discovered, including CWI, HOG, and calcium/calmodulin-dependent calcineurin signaling, suggesting an interaction of these three pathways. For A. fumigatus the participation of calcium-mediated signaling in the paradoxical growth in response to caspofungin has been described as well. Calcium signaling involves the Ca2+-binding protein calmodulin and the serine/threonine protein phosphatase calcineurin. Deletion of the calcineurin subunit A-encoding gene cnaA (also named calA) in A. fumigatus resulted in the loss of the paradoxical growth phenotype under caspofungin stress, which was attributed to the transcriptional regulation of chitin synthase-encoding genes by this signal transduction pathway. This finding shows that the caspofungin response in A. fumigatus, apart from CWI and HOG signaling, also involves the calcium signal transduction cascade suggesting an interaction between all three pathways in this filamentous fungus, as it was described for C. albicans. This conclusion was supported by the finding that in A. nidulans the constitutive over-expression of protein kinase C (PkcA) can in part restore the wild-type phenotype of a ΔcnaA deletion mutant. In addition, CnaA affects CWI signaling by regulating MpkA phosphorylation. Moreover, PkcA has an influence on the transcription of calcium-related processes as well as on the maintenance of normal intracellular calcium levels. For the basidiomycete Cryptococcus neoformans, an interaction of the CWI pathway and calcium signaling has also been described. This suggests that the cross talk between both pathways could be common to different fungal species.

30Another survey shows that not only inhibitors of the calcineurin signal transduction cascade but also rapamycin, the inhibitor of the Tor signal transduction pathway, dramatically increased the effect of caspofungin on A. fumigatus. As another component involved in the response to caspofungin, TOR might as well interact with CWI, HOG, or calcium signaling to coordinate the response to this drug. Altogether, this model raises the question whether in A. fumigatus the cross talk between signaling pathways is more pronounced than it is thought today.

31In the last years, several transcriptomics studies of A. fumigatus signal transduction mutants were published. The generated data suggest potential interactions of the CWI signaling cascade with other signal transduction pathways. For example, in a microarray analysis of the ΔmpkA strain, two calcium/calmodulin dependent kinases were found to be differentially regulated in the mutant compared to the wild type under stress conditions induced by glucanex, which lyses the CW. Furthermore, a microarray hybridisation approach comparing the transcriptional profile of the calcineurin mutant ΔcalA with the wild type, found the two MAP kinase kinase-encoding genes mkk2, involved in CWI signaling, and pbs2, involved in the HOG pathway, to be significantly down-regulated in the mutant strain. Data indicate once more a potential co-regulation of central signaling pathways in A. fumigatus. The increasing number of transcriptome analyses will lead to the identification of further interactions between signaling cascades, and most likely reveal the complexity of signal transduction in A. fumigatus.

32It is common praxis in infection biology to test the virulence of mutant strains in mouse infection models, in order to define whether genes have a potential function as virulence determinant. Among the different A. fumigatus mutant strains having a defect in the CW, more than 20 were affected in virulence (Table 1). Surprisingly, many mutant strains with defects in CW biosynthesis enzymes were still virulent in the applied infection models. As an example, both the α-1,3-glucan synthase mutant strains, Δags1, and Δags2, were still pathogenic, although their α-1,3-glucan content was reduced by 50% compared to the wild-type strain. Moreover, the Δags3 mutant was reported to be even hyper-virulent. To our knowledge, only the deletion of the genes encoding for a β-1,3-glucanosyltransferase (gel2) and a α-1,2-mannosyltransferase (mnt1) resulted in a decrease of virulence.

33As reported above, structuring of the CW depends on different signaling pathways. The observation that these pathways are connected to each other is increasingly acknowledged. Deletion of a single gene often affected more than one specific signaling pathway. In fact, many of the reported mutants with impaired CW, which displayed decreased virulence, were obtained by deleting genes putatively involved in signaling, such as ΔcalA, ΔrasA, and B, and ΔpkaC mutants. Recently, it was reported that the deletion of ptcB, a putative protein phosphatase 2C, positively affected the phosphorylation status of both MpkA and SakA, resulting in decreased virulence. Similarly, blocking of CWI signaling led to reduced virulence. However, blocking of CWI signaling also resulted in various physiological alterations, which are apparently not directly connected to the CW (e.g., alteration of secondary metabolism; ).

34A major question concerns the selection of suitable and effective targets for future antifungal drugs. The three clinically important classes of antifungal drugs used against A. fumigatus target either enzymatic steps involved in cell membrane/CW biosynthesis or ergosterol as part of the cytoplasmic membrane. These drugs show some limitations either because they cannot really clear A. fumigatus infection, or because resistant strains were isolated and are therefore of major clinical concern. The increase of life-threatening mycoses accompanied with the lack of effective drugs, has fostered the search for new, broad-spectrum fungicidal agents. Drug efficacy could also be increased by the reformulation of existing antifungals as well as the search for synergistically acting compounds that can be used for treatment and prophylaxis. As an example, the effectiveness of caspofungin in prophylaxis is still a matter of debate, and there are studies indicating that caspofungin did not decrease the mortality rate of patients with diagnosed invasive aspergillosis. In future, because of their low toxicity, echinocandins might well be used in a combinatorial antifungal therapy with other synergistically acting drugs. In line, it was already reported that the combination of caspofungin with azoles or AmB increased caspofungin activity in vitro. Additionally, the combination of these drugs improved the efficacy of treatments of patients with severe fungal infections, in which the first line therapy failed.

35The fungal CW still remains a very powerful target for antifungal drugs. However, recent studies suggest that the signaling pathways responsible for CW formation have not been completely elucidated. It remains to be shown which signaling pathways act as compensatory pathways that decrease the effectiveness of drug treatments. The identification of such pathways could lead to the discovery of new targets and new modes of action that can be exploited to potentiate the efficiency of known drugs, and to improve prophylaxis against invasive mycoses.

36The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.