In recent years, lipids have been shown to act as signalling molecules not only in mammalian cells but also in many other eukaryotes. Whereas in mammalian cells lipids regulate cellular functions that play crucial roles in the regulation of pathobiological processes, such as cancer, cardiovascular and neurodegenerative disorders, and inflammation, in the fungus Cryptococcus neoformans lipids play key roles in the regulation of pathogenic traits required for the development of cryptococcosis, an infectious disease particularly frequent in immunocompromised individuals. In this minireview we discuss recent advances in the understanding of lipid metabolism in this important human pathogen, highlighting the potential of fungal lipid enzymatic pathways as promising new drug targets.
Cryptococcal infections frequently manifest themselves as life-threatening meningoencephalitis in vulnerable populations, such as immunocompromised and cancer patients (Perfect & Casadevall, 2002). Well-established techniques to genetically modify this basidiomycetous fungus and the availability of robust animal models have allowed us to examine the contributions of individual fungal genes in host−pathogen interactions. In this respect, Cryptococcus neoformans is a model organism in our understanding of fungal pathogenesis at a molecular and biochemical level.
In the past quarter-century, our thinking of lipids has evolved, from a view of them as structural components of membranes and efficient storers of energy to the recognition that lipids exhibit bioactive properties and play crucial roles in intra- and inter-cellular signalling. Lipids have been implicated in critical cellular events such as cell growth, replication, differentiation, senescence, apoptosis, signal transduction, transcription, and stress responses (Munnik, 2001; Dickson & Lester, 2002; Kolesnick, 2002; Hannun & Obeid, 2002; Rocca & FitzGerald, 2002; Hata & Breyer, 2004; Kanaoka & Boyce, 2004; Wang, 2004; Becker & Hannun, 2005). Indeed, the role lipids play in infectious disease is an emerging field with great potential for pharmacological development (Hanada, 2005; Heung, 2006). This present review surveys our current understanding of how cryptococcal lipid metabolism modulates virulence. Special emphasis is placed on topics concerning lipid signalling pathways and novel targets for the development of antifungal drugs. For the metabolism of fungal sterols, we recommend the review by Odds (2003).
About 10% of the dry weight of C. neoformans is composed of lipids (Kaneko, 1976). Several attempts have been made in various laboratories to identify and quantify the various cryptococcal lipid species, and the reports are often conflicting. Rawat (1984) reported that C. neoformans lacked phosphatidylserine, as well as sphingolipids and cerebrosides. However, this study only extracted 10% of the total lipids from C. neoformans. Later investigations have proved that C. neoformans does indeed contain phosphatidylserine, sphingolipids, and cerebrosides (Itoh, 1975; Kaneko, 1976; Itoh & Kaneko, 1977). Strain-specific differences in lipid profiles have also been noted. There is an inverse relationship between capsule size and total lipid content (Upreti, 1984). Indeed, the polysaccharide capsule of C. neoformans contributes greatly to the total cellular mass and may interfere with some lipid extraction procedures (Itoh, 1975).
Phospholipids represent the largest class of lipids (63% of total lipids) measured in C. neoformans (Itoh & Kaneko, 1977). Approximately 45% of phospholipids are phosphatidylcholine (PC), with phosphatidylethanolamine (PE) and phosphatidylserine (PS) representing an additional 20% and 12%, respectively. Remarkably, the relative levels of these phospholipids do not change during the growth of C. neoformans (Itoh & Kaneko, 1977). The remaining pool of phospholipids consists of phosphatidylinositol (PI), cardiolipin (CL), phosphatidic acid (PA), and unidentified lipids (Itoh & Kaneko, 1977).
Interestingly, an unusual lipid species, pyrophosphatidic acid (pyro-PA), was first identified in C. neoformans nearly 30 years ago (Itoh & Kaneko, 1977). Subsequently, this lipid was identified in some fungi and in plants, but was lacking in mammals (Kaneko, 1976). In vivo labelling of C. neoformans with 32P-orthophophate has shown that the levels of pyro-PA and the closely related PA change dramatically during the growth cycle, suggesting that these compounds may be actively metabolized during growth (Itoh & Kaneko, 1977). However, their functions in the physiopathology of C. neoformans are yet to be elucidated. In plants, the roles of PA and pyro-PA are better understood: the levels of both rise quickly and transiently in response to a number of stresses, including cold, injury, and oxidative stress, nutrient deprivation, and pathogen interactions (reviewed in Wang, 2004). In fact, the increased PA and pyro-PA levels are now understood to signal the initiation of a general stress response by activating protein kinases and stimulating transcription (Munnik, 2001; Zalejski, 2005). In plants, PA is generated by phospholipase D (PLD) hydrolysis of phospholipids; and pyro-PA is generated by phosphorylating PA by a PA-specific kinase (Wissing, 1994). Recently, a cell-associated PLD activity was described in C. neoformans (Ganendren, 2004). The presence of pyro-PA in C. neoformans suggests that PA and pyro-PA may play a role in signalling and stress responses in this organism, and may in turn provide fungal-specific targets for antimycotic drugs. Besides its potential role as a fungal signalling lipid, pyro-PA is important for the mammalian immune response. Upon exposure to pyro-PA, the phospholipase A2 and MAP kinase pathways in macrophages are activated as a defence against fungal cells (Balboa, 1999).
Triacylglycerols (TAGs) are efficient ways to store energy in eukaryotic cells. TAG synthesis occurs mostly in the endoplasmic reticulum (ER), with some synthesis at lipid particles (LP), yeast organelles specific for TAG storage (reviewed in Sorger & Daum, 2003). Cryptococcus neoformans has an unusually high TAG content; TAG represents nearly one-third of all cryptococcal lipids and nearly 90% of the non-polar lipid fraction (Kaneko, 1976). LP-containing TAG, appearing as ‘oil globules’, accumulate as C. neoformans enters the stationary phase of growth (Itoh, 1975). Interestingly, Schizosaccharomyces pombe mutants defective in TAG synthesis undergo an apoptotic-like cell death as they enter the stationary phase, which underscores the importance of TAG in fungal metabolism (Zhang, 2003). Mycobacterium tuberculosis also accumulates TAG under conditions that induce latency, such as O2 depletion or nitric oxide (NO) exposure (Daniel, 2004).
Although the role of TAG metabolism in the pathogenesis of infectious diseases is not well characterized as yet, there are some indications that the fatty acids from TAG provide an important carbon source for microbes in the nutrient-limiting conditions of the host environment. For instance, a characteristic host environment with nutrient-limiting conditions is the phagolysosome of macrophages. In one investigation, M. tuberculosis isolated from murine lungs were metabolically suppressed, a hallmark of the latent form of the pathogen, compared with M. tuberculosis grown in vitro. Interestingly, respiration in these latent mycobacteria was stimulated by fatty acids, but not by carbohydrates (Bloch & Segal, 1956). The glyoxylate pathway allows microorganisms to utilize alternative carbon sources (such as products of the β-oxidation of fatty acids) to replenish intermediates of the TCA cycle, build carbohydrates, and provide energy. Recent studies have shown that Candida albicans, M. tuberculosis, and C. neoformans increase the expression of genes involved in the glyoxylate pathway, such as isocitrate lyase, when these organisms reside within the phagolysosome of macrophages (McKinney, 2000), (Lorenz & Fink, 2001; Rude, 2002). In both Ca. albicans and M. tuberculosis, disruption of the isocitrate lyase gene (ICL1) attenuated the virulence in murine models of infection (McKinney, 2000; Lorenz & Fink, 2001). Notably, the mutant strains were not as persistent within the organs of their host (Lorenz & Fink, 2002). In contrast, however, disruption of cryptococcal ICL1 did not affect virulence in either a rabbit model of meningitis or a mouse inhalation model, nor did the lack of ICL1 affect cryptococcal growth in macrophages (Rude, 2002). Alternative isocitrate lyase enzymes have been found in both Saccharomyces cerevisiae and M. tuberculosis (Luttik, 2000; Munoz-Elias & McKinney, 2005). However, performing a BLAST search of the C. neoformans var. grubii genomic database (http://cneo.genetics.duke.edu/) with these gene sequences revealed no homologues other than cryptococcal ICL1. This finding, however, does not preclude the importance of the glyoxylate pathway in cryptococcosis, as C. neoformans may possess other pathways to generate glyoxylate independent of isocitrate lyase. Efficient energy storage and metabolism appear to be a common theme among intracellular pathogens, and disruption of these processes may prove useful in controlling and eliminating chronic fungal infections (Lorenz & Fink, 2002; Munoz-Elias & McKinney, 2005).
In mammalian systems, the role that diacylglycerol (DAG) plays as a second messenger in the activation of protein kinase C (PKC) has been understood for more than a quarter of a century: phospholipase C (PLC) liberates the polar phosphoryl head group from a phospholipid, generating the hydrophobic DAG, which in turn binds to and activates the serine/threonine kinase PKC (reviewed in Becker & Hannun, 2005). In yeast, the role of DAG signalling in PKC activation is less clear. Some investigators have reported that Pkc1 from S. cerevisiae is activated in the presence of DAG, similarly to the classical isoforms of mammalian PCK (Ogita, 1990), while other investigators have claimed that DAG and other phospholipids play no role in PKC activation in yeast (Watanabe, 1994) (reviewed in Schmitz & Heinisch, 2003). Pkc1 has also been identified and characterized in Ca. albicans and found not to be activated by DAG (Paravicini, 1996). In C. neoformans, one protein kinase C gene was identified by its homology to S. cerevisiae Pkc1 and cloned as PKC1 (Heung, 2004). Its amino acid sequence is about 60% similar to Pkc1 from S. cerevisiae and Ca. albicans, and it shares a 31% similarity with the DAG-dependent mammalian PKC-ɛ. Interestingly, cryptococcal Pkc1 has a C1 domain, characterized by two tandem repeating patterns of histidine and cysteine residues (Hurley, 1997), which show greater similarity in key consensus residues with the C1 domain of DAG-dependent mammalian isoforms (Heung, 2004). In mammalian PKC, the C1 domain is responsible for DAG binding to PKC. The Pkc1 from S. cerevisiae and Ca. albicans, as well as the DAG-independent mammalian PKC isoforms lack many of these key residues (Heung, 2004). A purified recombinant cryptococcal Pkc1 protein showed increased kinase activity in the presence of several subspecies of DAG, as well as with PS. This increased kinase activity was abolished when the cDNA encoding the C1 domain of Pkc1 was deleted (Heung, 2005). Interestingly, the fungal sphingolipid phytoceramide was found to inhibit kinase activity independent of the C1 domain. Therefore, it appears that, in C. neoformans, DAG can activate PKC, which in turn functions to maintain cell-wall integrity. Compromising this pathway in C. neoformans, by deleting the region encoding the C1 domain of Pkc1, prevents association of the virulence factor laccase with the cell wall and leads to cell-wall defects such as in the formation of capsule microfibrils (Heung, 2005). The discovery of the DAG-Pkc1 pathway in C. neoformans provides new insight into how this pathogen regulates key virulence factors, such as melanization and capsule production, and opens new areas in which to develop antifungal agents that can disrupt multiple systems in cryptococcal pathobiology.
Diacylglycerol can also control the transcription of cryptococcal genes. The expression of the antiphagocytic protein App1 was found to be upregulated under conditions associated with increased intracellular DAG levels (Luberto, 2003). A luciferase reporter gene fused to the promoter region of APP1 showed increased luciferase activity in the presence of DAG (Mare, 2005). Two putative consensus sequences in the promoter region of APP1 were identified by BLAST-searching the promoter region in ‘WWW Promoter Scan’ (bimas.dcrt.nih.gov/molbio/proscan/index.html). The putative AP-2 sequence was found to act as a repressor of APP1 transcription, because deletion of this cis-acting site element increased the transcription of APP1. On the other hand, the putative AFT site was shown to enhance transcription, and mutations at this site abolished the DAG-dependant increase of APP1 transcription. A putative ATF2 gene was also identified and deleted. Loss of the putative Atf2 transcription factor resulted in the insensitivity of APP1 transcription to DAG levels (Mare, 2005). These studies provide solid genetic evidence that DAG does act as a bioactive lipid messenger in C. neoformans, although further work will be needed in order to elucidate fully the mechanisms by which DAG controls transcription.
Phospholipase activity has been shown to be involved in microbial pathogenicity in a number of studies. Pseudomonas aeruginosa secretes an enzyme, PLC-H, which is associated with vascular permeability and organ damage, and purified PLC-H is sufficient to cause death when injected into mice (Berk, 1987). The protozoan parasite Toxoplasma gondii may use its PLA2 to migrate between host cell compartments (Saffer & Schwartzman, 1991) (please refer to Fig. 1 for phospholipase nomenclature.) Among fungal pathogens, the phospholipase B of Ca. albicans is best characterized. Supernatants from phospholipase B-producing strains of Ca. albicans can damage epithelial cells in vitro significantly more than those from isogenic phospholipase-deficient strains, and loss of phospholipase B activity is associated with decreased invasiveness of the fungal strain (reviewed in Ghannoum, 2000). Disruption of phospholipase B genes in Ca. albicans produces strains with attenuated virulence in mouse models of candidiasis, with no observable defects in growth or germination in vitro (Leidich, 1998).
Sites of hydrolysis for different phospholipases. PLA1 and PLA2 are distinguished based on their ability to hydrolyze the ester bonds at the sn-1 or sn-2 positions, respectively. Phospholipase B (PLB) can hydrolyze at one, the other, or both positions. Lysophopholipase (LPL) hydrolyzes the remaining fatty acid from the lysosphospholipid, while lysophospholipid transacylase (LPTA) can esterify a free fatty acid to lysophospholipids, forming phospholipids with two acyl chains.
Similar to Ca. albicans, C. neoformans produces an extracellular phospholipase B (Plb1), encoded by the PLB1 gene. The Plb1 enzyme has three activities: (i) phospholipase B (PLB) cleaves the ester bond between a fatty acid in either the sn-1 or sn-2 position and the glycerol backbone of phospholipids, generating free fatty acids and a lysophospholipid; (ii) lysophospholipase (LPL) activity from Plb1 can liberate the remaining fatty acid from lysophospholipids; and iii) lysophospholipid transacylase (LPTA) activity can esterify fatty acids back onto the glycerol backbone, generating phospholipids and lysophospholipids (Fig. 1). The cryptococcal Plb1 amino acid sequence is 36% homologous to Plb1 and Plb2 from Ca. albicans, and has an additional glycosylphosphatidylinositol (GPI)-anchoring motif at the C-terminus (Djordjevic, 2005).
Since its identification, cryptococcal phospholipase activity has been positively correlated with virulence in murine models of infection (Chen, 1997; Cox, 2001). In 2000, a single glycoprotein with PLB, LPL, and LPTA enzymatic activities was purified from the secretions of C. neoformans var. grubii (Chen, 2000). Although these activities were maintained at mammalian physiological temperatures (37°C), they were only present under acidic conditions (pH 4.0–5.0). All phospholipids tested, with the notable exception of PA, were found to be substrates, although dipalmitoyl-(C16:0)PC and dioleoyl-(C18:1)PC were the preferred substrates. Cryptococcal phospholipase activity was first observed as an extracellular activity, but fractionation experiments revealed that most (c. 85%) of the total PLB activity of C. neoformans is cell-wall-associated, presumably via GPI-linkage, while only a small portion is secreted into the supernatant (Santangelo, 1999; Djordjevic, 2005). The mechanism by which C. neoformans regulates the secretion of Plb1 is unknown at this time. Growth at 30°C in an acidic environment appears to be the optimal condition for secretion of PLB activity. Interestingly, PLB secretion is enhanced under tissue culturing conditions (Dulbecco's modified eagle medium with 5% CO2), although there were strain-specific differences observed in secretion (Wright, 2002). Antibodies to Plb1 have been detected in serum from patients with cryptococcal infection, providing evidence for its expression in vivo, and underscoring the clinical importance of this enzyme (Santangelo, 2005).
Disruption of the plb1 locus results in a complete abolition of PLB activity of C. neoformans, and significant attenuation of LPL and LPTA activities (Cox, 2001). Further investigations have revealed the existence of another secreted enzyme in C. neoformans var. grubii, namely LYSO1p, which also has LPL and LPTA activities (Coe, 2003). It is noteworthy that disruption of the lyso1 locus significantly attenuates LPL and LPTA activities, and, in spite of the presence of the intact Plb1 locus, lower amounts of secreted Plb1 were found in the lyso1 deletion mutants. This suggests that Plb1 and Lyso1, instead of having independent functions, may act interdependently in order to maintain full phospholipase activities in C. neoformans. A third cryptococcal phospholipase protein (Lpl1) with only LPL and LPTA activities has been identified in both C. neoformans var. grubii and var. gattii (Wright, 2004). Of clinical importance is the recognition that this enzyme has significant LPL and LPTA activities at physiological pH (7.0). N-terminal amino acid sequencing of Lpl1 revealed this protein to be different from that of Plb1 and the gene product of Lyso1.
Animal studies utilizing deletion mutants of PLB1 provide compelling evidence that phospholipase activity is a virulence factor in cryptococcal infection. A/Jcr mice intranasally challenged with a plb1 mutant survived significantly longer than mice challenged with wild-type or reconstituted strains (Cox, 2001). Intratracheal challenges of CBA/J mice showed significantly lower fungal growth in the lungs of Δplb1-infected mice, with lower pulmonary inflammation (Noverr, 2003a). Likewise, the plb1 strain does not survive in the cerebrospinal fluid as well as wild-type or reconstituted strains in a rabbit meningitis model (Cox, 2001).
The presence of a functional Plb1 gene appears to enhance cryptococcal growth and survival inside phagocytic cells. Cox (2001) described a delay in fungal budding upon phagocytosis of the plb1 mutant compared with the wild-type strain. Likewise, the plb1 mutant showed a significant growth defect inside the amoeba, Acanthamoeba castellanii, compared with strains with a functional PLB1 gene (Steenbergen, 2001). This suggests that C. neoformans may have originally evolved extracellular phospholipases as a defence against environmental predators, such as the amoeba, which consequently confer cross-resistance to antifungal properties of macrophages.
The mechanism by which phospholipase and/or transacylase activities of Plb1 contribute to the virulence of C. neoformans is unknown at this time. There are no detectable changes in other cryptococcal virulence factors, including capsule production, laccase activity, urease activity, growth at 37°C, and auxotrophy (Cox, 2001). A secreted phospholipase could potentially help C. neoformans to cross lipid barriers, such as lung surfactant or endothelial cells. Indeed, human surfactant is rich in dipalmitoyl phosphatidylcholine, the preferred substrate of cryptococcal Plb1 (Fisher, 2005). However, there is no measurable PLB activity under neutral pH conditions, as would be found in the lungs or vasculature of the host. The acid pH optimum of PLB suggests that the enzyme might be active within the acidic environment of the macrophage phagolysosome, in which it may aid in the lysing of that compartment. Indeed, disruption of the phagolysosome compartment is a characteristic of C. neoformans-infected macrophages (Cox, 2001). However, Noverr (2003a) found no significant changes in the viability of macrophages co-cultured with a Δplb1 strain compared with strains carrying a functional PLB1 gene. Further investigations are required in order better to understand the function of Plb1 in the interaction of C. neoformans with macrophages.
Another potential role of phospholipase activity could relate to disruption of host signalling via lipid second messengers. The pseudomonal PLC-H has been found to attenuate host neutrophil respiratory oxidative burst by modulating a host-PKC-dependant pathway (Terada, 1999). Lysophospholipids, particularly lysophosphatidylcholine, have been shown to stimulate mammalian PKC at lower (<20 μM) concentrations, and to inhibit it at higher (>30 μM) concentrations (Oishi, 1988). The release of free fatty acids by PLB and LPL activities could also alter eicosanoid signalling in the host (see below).
The presence of an extracellular phospholipase enzyme in a number of clinically important fungi, including C. neoformans, Ca. albicans, Aspergillus fumigatus and A. flavus, and the recognition of the role that phospholipases have on the pathogenesis of infection suggest that inhibitors to these activities may be potent antifungal agents. Indeed, Hanel et al. (1995) have found that synthetic compounds blocking Ca. albicans phospholipase activity decreased the tissue invasiveness of the fungus, and, in combination with fluconazole, protected mice from a lethal inoculum. A few potent cryptococcal phospholipase inhibitors have been identified and tested. Alexidine dihydrochloride inhibited secreted PLB activity, and, to a lesser extent, secreted and cell-associated LPL and LPTA activities. Another compound, 1,12 bis-(tributylphosphonium)-dodecane, is a more selective inhibitor of phospholipase B activity, with no significant inhibition of LPL or LPTA activities. Both compounds also inhibited the growth of C. neoformans and Ca. albicans at low micromolar concentrations (Ganendren, 2004). In contrast, dioctadecyldimethylammonium bromide, a compound that selectively inhibited LPL and LPTA activity but spared PLB activity, had no effect on cryptococcal or candidal growth (Ganendren, 2004). Taken together, these studies suggest that PLB may play a critical role in the normal physiology of C. neoformans and other pathogenic fungi, and, together with the work of Hanel (1995), reinforce Plb1 as a critical target for pharmacological development of new antifungals.
Sphingolipids are vital components of both mammalian and fungal membranes. Besides contributing to the stability of membranes through biophysical interactions with phospholipids and sterols, it has become increasingly clear that the metabolites of the sphingolipid pathways serve crucial roles as second messengers in a vast number of biological processes, including stress responses, cell-cycle regulation, host immune responses, and apoptosis (Dickson & Lester, 2002; Hannun & Obeid, 2002).
Although all eukaryotic organisms produce sphingolipids, differences have been noted between mammals and yeast with respect to their biosynthetic pathways (Mandala, 1998). In mammals, the neutral lipid ceramide serves as a common precursor for the synthesis of complex sphingolipids by the addition of polar head groups to ceramide. A phosphorylcholine moiety is transferred from phosphatidylcholine (PC) to ceramide by the mammalian-specific enzyme sphingomyelin synthase, generating a molecule of sphingomyelin and DAG (Fig. 2). The mammalian enzymes sphingomyelin synthase and sphingomyelinase (responsible for hydrolyzing the phosphorylcholine group from sphingomyelin to generate free ceramide) are being extensively studied in cancer models because of their ability to regulate two important second messengers, ceramide and DAG (Kolesnick, 2002).
Comparison of sphingolipid biosynthesis in mammals and fungus. The early steps in sphingolipid synthetic pathway are shared in common between mammals and fungi, from the condensation of serine and palmitoyl-CoA to the formation of sphinganine. In yeast, Sur1 catalyzes the hydroxylation of sphinganine to form phytosphingosine, the first step in the production of fungal-specific sphingolipids. Gene names are written in italics and capitalized. Inhibitors are italic and underlined. Abbreviations: DAG, diacylglycerol; MIPC, mannosylinositolphosphorylceramide; M(IP)2C, mannosyldiinositolphosphorylceramide; PC, phosphatidylcholine; PI, phosphatidylinositol.
In contrast, fungi use the more hydroxylated sphingolipid, phytoceramide, to serve as the hydrophobic core for the synthesis of more complex sphingolipids. The fungal-specific enzyme inositol-phosphorylceramide synthase (IPC synthase or Ipc1) transfers a phosphorylinositol group from phosphatidylinositol (PI) to phytoceramide, forming inositol-phosphorylceramide (IPC) and DAG (Dickson & Lester, 2002). (Fig. 2)
The IPC1 gene from S. cerevisiae (also called AUR1) has homologues in many other fungi, including the pathogenic fungi Ca. albicans, Ca. glabrata, and C. neoformans (Heidler & Radding, 2000). Like many other enzymes in the fungal sphingolipid pathway, Ipc1 is an ideal drug target because it is unique to fungal metabolism, and thus cross-reactivity of Ipc1-inhibitory compounds with host cells is eliminated. Indeed, IPC1 was first recognized as an essential gene encoding resistance to aureobasidin A (AbA), an antifungal cyclic depsipeptide isolated from Aureobasidium pullulans that was found to inhibit growth of a variety of fungi including Ca. albicans, C. neoformans, and S. cerevisiae (Takesako, 1991). Other drugs inhibiting Ipc1, such as khafrefungin and rustmicin, also have antifungal properties against Ca. albicans and C. neoformans (Heidler & Radding, 2000; Nakamura, 2003). Some of these drugs, such as AbA and khafrefungin, do not inhibit mammalian sphingolipid biosynthesis in vitro (Del Poeta, unpublished data; Nakamura et al., 2003). Besides being important targets in their own right, inhibitors of fungal sphingolipid synthesis can help to increase the sensitivity of fungi to other established antifungal agents. Treatment with fumonisin B1, an inhibitor of phytoceramide synthase, caused Ca. albicans to become hypersensitive to a variety of antifungal agents by disrupting the proper localization and functionality of the drug efflux pump Cdr1p (Mukhopadhyay, 2004), suggesting that compounds targeting the sphingolipid-metabolizing enzymes may increase the antimicrobial effect of other drugs by augmenting their intracellular concentrations.
In C. neoformans, the sphingolipid pathway has been shown to play an important role in the pathobiology of cryptococcal infections, including the expression of virulence factors and the ability of C. neoformans to withstand environmental stresses. Our laboratory has genetically engineered a strain of C. neoformans whereby expression of the IPC1 gene is under the control of a galactose-inducible promoter. Suppressing Ipc1 by growth on glucose causes a decrease in laccase activity, with a corresponding decrease in the rate that and extent to which this strain produces melanin (Luberto, 2001). Further investigations have shown that Ipc1 can modulate the levels of DAG and phytoceramide, which in turn can activate and inhibit, respectively, Pkc1 in C. neoformans (Heung, 2004). In addition, when Ipc1 expression is suppressed with glucose, there is a significant growth defect in vitro under acidic conditions (Luberto, 2001). Likewise, suppression of Ipc1 results in a defect in intracellular growth in J774.16 murine macrophage-like cells. In contrast, when this strain is grown on galactose, Ipc1 expression is increased compared with wild-type, and this strain replicates more quickly in the phagolysosome than its parental strain (Luberto, 2001). Wild-type C. neoformans increases transcription of the IPC1 gene when inside the phagolysosome, providing further evidence for its role in the intracellular growth and pathogenicity of C. neoformans (Fan, 2005). Indeed, suppression of Ipc1 activity significantly attenuated virulence in both a murine intranasal model of infection and a rabbit model of cryptococcal meningitis (Luberto, 2001).
In addition to producing phytoceramide, which serves as a backbone for the inositol-containing sphingolipids, C. neoformans also produces ceramide, which can also be glycosylated to form glucosylceramide (GlcCer). In mammals and plants, the role of GlcCer is well studied; however, the apparent lack of GlcCer in the model fungus S. cerevisiae has dampened investigations of its role within the fungal kingdom (Leipelt, 2001). It is known that GlcCer in fungi mediates immune responses of plants and insects. The antimicrobial compounds defensins produced by insects and plants can inhibit the growth of fungi by binding GlcCer and disrupting membrane integrity. The mechanisms of action of human defensins are not fully elucidated, but it appears that human defensins may form pores by non-specifically binding to anionic phospholipids in the microbe plasma membrane (Chen, 2005). Interestingly, plant- and insect-derived defensins do not bind to human-derived GlcCer, highlighting important structural differences between mammalian and fungal GlcCer (Thevissen, 2004). The enzyme glucosylceramide synthase (GCS) was disrupted in Ca. albicans; the null mutant showed no growth defect, suggesting that, unlike IPC1, GCS1 is not an essential gene (Leipelt, 2000). What effect the loss of GCS/GlcCer has on the virulence of Ca. albicans has not been reported at this time. It is noteworthy that sera from patients infected with C. neoformans have antibodies that bind to cryptococcal GlcCer. In addition, sera from patients with histoplasmosis, aspergillosis, and paracoccidioidomycosis also reacted with cryptococcal GlcCer (Rodrigues, 2000). These antibodies inhibit cryptococcal growth in vitro, presumably by interfering with yeast budding (Rodrigues, 2000). Genetic disruption of the GCS1 gene produces a cryptococcal strain that is markedly less pathogenic than its parental strain in a murine inhalation model of infection. Phenotypic screening of this mutant shows increased sensitivity to alkaline stresses (in contrast with suppression of Ipc1 expression, which shows increased sensitivity to acidic stresses) (Rittershaus, 2005). The mechanisms underlying the role of GlcCer in infection are currently being investigated.
A new and exciting field in the study of C. neoformans was triggered by the discovery that this fungal pathogen has the ability to synthesize eicosanoids (Noverr, 2001), which are potent lipid-soluble hormones used by mammals to modulate and coordinate immune responses. While this field of study is still in its infancy, it could potentially change our fundamental understanding of the cryptococcal pathogen−host interaction.
Eicosanoids, which include prostaglandins (PG) and leukotrienes (LT), are a class of oxygenated lipids derived from C20 polyunsaturated fatty acid precursors such as arachidonic acid. In mammalian systems, eicosanoid production begins with the release of arachidonic acid from the sn-2 position of a phospholipid by a cytosolic phospholipase A2 (cPLA2). Prostaglandin synthesis is initiated by the activity of cyclooxygenase on arachidonic acid to produce a ring structure, forming prostaglandin H (PGH) (Fig. 3). PGH can then be further modified to produce specific classes of prostaglandins, such as PGE2 and PGF2α. Leukotrienes are synthesized by mammalian 5-lipoxygenase, which adds a hydroperoxy group to arachidonic acid, producing 5-hydroperoxyeicosatetraenoic acid (5-HpETE), the common precursor to all leukotrienes. Cysteinyl leukotrienes (CysLT), which include LTC4, LTD4, and LTE4, are characterized by the presence of a cysteinyl residue on the leukotriene (reviewed in Noverr, 2003b).
Biosynthesis of prostaglandins (PG) series-2 and leukotriene (LT) series-4 eicosanoids from arachadonic acid. The depicted pathways are based on mammalian systems; there is little evidence describing fungal production of eicosanoids. Cysteinyl leukotienes (CysLT) include LTC4, LTD4, and LTE4.
The immunomodulatory properties of both prostaglandins and leukotrienes have been studied intensively in mammalian systems. The biological activity of prostaglandins are mediated through the family of transmembrane, G-protein-coupled prostanoid receptors, such as the EP2 receptor (Hata & Breyer, 2004). As an agonist for the EP2 receptor, PGE2 is proinflammatory by activating the innate immune system and stimulating reactive-oxygen species (ROS) production in response to lipopolysaccharide (LPS). However, PGE2 appears to have anti-inflammatory properties as well. It has been shown to deactivate macrophages, probably through the EP4 receptor, and to interfere with macrophage antigen presentation and cytokine release (reviewed in Hata & Breyer, 2004). PGE2 inhibits the production of interleukin-2 (IL-2) and interferon-gamma (IFN-γ) by T cells, thus indirectly promoting a Th2 cytokine response (Rocca & FitzGerald, 2002). The CysLT are recognized as playing critical roles in the development of asthma and allergic reactions. Specifically, LTE4 promotes the recruitment of eosinophils to the lungs and also favours a Th2 response (Kanaoka & Boyce, 2004).
Considering the role eicosanoids play in intercellular communications in multicellular organisms, the discovery of eicosanoids in unicellular eukaryotic microorganisms may seem surprising. Nevertheless, Noverr (2001, 2003a) have found that both C. neoformans and Ca. albicans secrete prostaglandins, such as PGE2 and CysLT, and that supplementing the media with arachidonic acid increased eicosanoid production. The purified fungal prostaglandins had biological activity at low nanomolar concentrations in mammalian systems in vitro, including suppression of lymphocyte proliferation, suppression of TNF-α, and enhancement of IL-10 production.
Interestingly, cryptococcal Plb1 plays a role in eicosanoid production. The Δplb1 mutant produces less prostaglandins than its parental strain. However, supplementing the medium with free arachidonic acid, but not arachidonic acid esterified to PC, completely restores the defect in prostaglandin production. These results suggest that Plb1 is required for prostaglandin synthesis in the absence of exogenous free arachidonic acid (Noverr, 2003a). In mammalian systems, cPLA2 releases arachidonic acid from phospholipids, providing substrates for prostaglandin and leukotriene synthesis. PLB has inherent PLA2 activity, and could potentially liberate arachidonic acid from host cells as well, resulting in deregulation of the host cytokine profile. When mice were infected with Δplb1, there were lower levels of PGE2, PGF2α and CysLT in the lungs compared with mice infected with the parental strain. There was also a corresponding decrease in leukocytes, eosinophils and macrophages infiltrating the lungs (Noverr, 2003a). It is tempting to speculate that Plb1 might aid in the pathogenicity of C. neoformans once it has been internalized into the acidic environment of the host phagolysosomes. Plb1, which works optimally at acid pH, could release arachidonic acid from host membranes, leading to the synthesis of eicosanoids by either C. neoformans or the host, that inhibit the antifungal activities of macrophages and promote a non-productive Th2 inflammatory response that helps to ensure fungal survival (Cox, 2001; Noverr, 2003a).
At this time, it is not known whether the changes in pulmonary eicosanoid levels observed in Δplb1-infected mice are host- or fungal-derived. Separating the relative roles that Plb1 and eicosanoid production play in virulence will be difficult without a better understanding of the fungal pathways of eicosanoid synthesis. The ability of C. neoformans and Ca. albicans to produce prostaglandins and leukotrienes implies the presence of cyclooxygenase and 5-lipoxygenase activities. Indeed, cyclooxygenase inhibitors block production of prostaglandins in C. neoformans and other fungi (Noverr, 2003a, b). However, identification of fungal cyclooxygenase enzymes has eluded us until recently: three cyclooxygenase-like enzymes with homology to mammalian cyclooxygenase PGH synthase were isolated and characterized in A. nidulans and A. fumigatus (Tsitsigiannis, 2005). A better understanding of fungal eicosanoid production will lead to the identification of useful targets for pharmacological development. Interestingly, cyclooxygenase inhibitors had a toxic effect on both C. neoformans and Ca. albicans (Noverr, 2001). This implies an essential role of cyclooxygenases in fungal biology that may potentially be exploited.
The contribution of fungal lipid metabolism to the development of cryptococcosis is still not fully understood. Recent studies clearly suggest that C. neoformans can actively metabolize lipids to promote its pathogenicity in mammalian hosts. Extracellular phospholipases may directly contribute to virulence by causing host tissue damage. Eicosanoid production could potentially subvert the host immune response and lead to a chronic fungal infection. Sphingolipid metabolism may favour intra- and extra-cellular growth of fungal cells in the host. Lipid metabolites can also indirectly aid in pathogenicity by modulating the expression of fungal virulence factors, such as melanin and App1. The current antifungal repertoire is limiting and fungal resistance is increasing (Perfect & Casadevall, 2002). In addition, relapse of cryptococcosis is common, even among the immunocompetent, forcing patients to receive antifungal prophylaxis. Therefore, there is a clinical need to develop potent fungicidal drugs that have no cross-reactivity with humans. The studies reviewed here have identified several fungal-specific targets for the development of novel antimicrobial agents that can aid in the treatment of this yeast and potentially of other pathogenic fungi.
Maurizio Del Poeta is supported by the Burroughs Wellcome Fund, the National Institute of Health (NIH) grant AI56168, and the Centers of Biomedical Research Excellence (COBRE) Program of the National Center for Research Resources grant RR17677 (Project 2). Maurizio Del Poeta is a Burroughs Wellcome New Investigator in Pathogenesis of Infectious Diseases.
(2000) Purification and characterization of secretory phospholipase B, lysophospholipase and lysophospholipase/transacylase from a virulent strain of the pathogenic fungus Cryptococcus neoformans. Biochem J 347: 431–439.
(2004) Induction of a novel class of diacylglycerol acyltransferases and triacylglycerol accumulation in Mycobacterium tuberculosis as it goes into a dormancy-like state in culture. J Bacteriol 186: 5017–5030.
(2005) APP1 transcription is regulated by inositol-phosphorylceramide synthase 1-diacylglycerol pathway and is controlled by ATF2 transcription factor in Cryptococcus neoformans. J Biol Chem 280: 36055–36064.