By exploiting the biosynthetic pathways of raft lipid constituents, in this study we demonstrate that fluctuations in either sphingolipid or ergosterol levels result in increased drug sensitivity and morphological defects in Candida albicans cells. We show that any change in either ergosterol composition by conditionally disrupting ERG1 or in sphingolipid composition by homozygously disrupting its biosynthetic gene IPT1 leads to improper surface localization of a major ABC (ATP-binding cassette) drug efflux protein, Cdr1p. Results suggest that sterol/sphingolipid-rich membrane microdomains play an important role in positioning and functional maintenance of the integral efflux protein. The impaired ability of erg1/ipt1 mutant cells to efflux drugs mediated through Cdr1p appears to be the main cause of increased drug sensitivity of Candida cells.
The existence of discrete membrane microdomains, known as lipid rafts within a lipid bilayer, predominantly composed of sphingolipid and sterol, are well documented in several systems [1–5]. Interestingly, the acquisition of the MDR (multidrug resistance) phenotype is also accompanied by the up-regulation of lipids and protein synthesis that constitute lipid rafts [6–9]. A recent study by Martin and Konopka  has demonstrated the presence of highly polarized ergosterol-rich microdomains, which are specifically present during morphogenesis of Candida albicans. They suggest that lipid rafts contribute to hyphal morphogenesis and probably to virulence of this pathogen. We had earlier observed that the interactions between membrane lipid raft constituents are important determinants of drug susceptibilities of C. albicans cells . We now show a close interaction between PM (plasma membrane) ergosterol and sphingolipids in C. albicans cells, wherein the depletion of either of the two resulted in impaired functionality of a major drug efflux pump Cdr1p . This new evidence suggests that the major ABC (ATP-binding cassette) drug efflux pump protein of C. albicans probably resides in these microdomains and any perturbation in the raft lipid constituents affects its positioning and functionality.
Disruption of ERG1 and IPT1 genes
To understand the interactions between ergosterol and sphingolipids, we disrupted two genes of C. albicans; ERG1 (GenBank® accession no. U69674), which encodes for squalene epoxidase (EC 188.8.131.52) and converts squalene into 2,3-oxidosqualene [12–14] (Figure 1A), and IPT1 (GenBank® accession no. AY884203), which is responsible for the formation of M(IP)2C (mannosyl di-inositol diphosphoceramide) from MIPC (mannosyl inositol phosphoceramide) [15,16] (Figure 1C). URA-blaster approach was employed to disrupt both the genes [17,18]. While homozygous disruption of IPT1 was easily confirmed, both alleles of ERG1 could not be disrupted. It turned out that the ERG1 of Candida like that of Saccharomyces cerevisiae is an essential gene . We, therefore, achieved conditional disruption of the second allele of ERG1 by using regulatable MET3 promoter [20,21]. The conditional mutant strain ΔE4.2.7 (MET3p-ERG1/erg1Δ::hisG), when grown with methionine/cysteine (M/C) in the media, suppressed the formation of ergosterol.
Characterization of the mutants by HPLC and MS analysis
Conditional erg1/ERG1 and homozygous ipt1/ipt1 mutants show accumulation of squalene and MIPC respectively
HPLC and MS analyses were used to confirm the compositional consequence of IPT1 and ERG1 disruptions. Detailed HPLC analyses revealed that ergosterol levels in the heterozygous strain (ERG1/erg1) were lower when compared with the WT (wild-type) CAF2-1. The conditional disruptant strain (ΔE4.2.7), after growth in the presence of M/C, showed no detectable characteristic ergosterol peak, but instead showed a high peak of squalene (Figure 1B). Thus the ΔE4.2.7 strain, which conditionally lacks functional squalene epoxidase (Erg1p), is unable to synthesize ergosterol and instead accumulates squalene (Figure 1B). In the case of the homozygous IPT1 disruptant (ipt1/ipt1), MS revealed an absence of M(IP)2C peak and instead showed the accumulation of MIPC (Figure 1D).
erg1 and ipt1 mutant cells show increased sensitivity to drugs
In view of the fact that alterations in lipid composition affect drug susceptibilities [11,22,23], we examined whether the depletion of ergosterol and concomitant accumulation of squalene in the erg1 mutant and M(IP)2C loss in the ipt1 mutant affected drug sensitivities of disrupted cells. Interestingly, erg1 conditional mutant under repressing conditions displayed increased sensitivity to drugs like fluconazole, ketoconazole, cycloheximide and terbinafine, compared with WT (Figure 2A). Loss of M(IP)2C in ipt1 disruptant, as well, resulted in increased sensitivity to similar drugs (Figure 1B). The ipt1 disruptants were additionally found to be sensitive to drugs like 4-nitroquinoline oxide, O-phenanthroline and sulphomethuron methyl (results not shown). Of note the erg1 conditional strain became sensitive to polyenes like nystatin and amphotericin B (Figure 2A). The sensitivity of erg1 conditional strain to polyenes in the absence of ergosterol is interesting since polyenes, owing to their lethality, are known to specifically interact with membrane ergosterol. The reverse results support increasing evidence that suggests that membrane lipids other than ergosterol could also influence polyene sensitivity [24–27].
Drug-resistance profiles were determined by using spot assay
Cdr1p function is impaired in erg1 and ipt1 disrupted cells
Resistance to antifungals could be visualized as a gradually evolving process, wherein different mechanisms may appear during the course of chemotherapy [22,28]. The main mechanisms of antifungal resistance include alterations of the target enzyme of ergosterol biosynthetic pathway lanosterol 14α-demethylase or ERG11 (point mutations), which lead to reduced affinity to fluconazole [29,30]. Reduced intracellular accumulation of drugs (due to rapid efflux) is another prominent mechanism of drug resistance in Candida cells. Most commonly, genes encoding drug efflux pumps belonging to ABC and MFS (major facilitator superfamily) are overexpressed in azole-resistant Candida isolates [31,32]. Accordingly, it has been well documented by several groups that clinical azole-resistant isolates of C. albicans display transcriptional activation of genes encoding ABC (Cdr1p and Cdr2p) or MFS (CaMdr1p) efflux pump proteins [31–33]. Invariably, resistant Candida cells, which show enhanced expression of efflux pumps encoding genes, also show simultaneous increase in the efflux of drugs thus implying a causal relationship between efflux pump encoding gene expression levels and intracellular concentration of the drug [11,33,34]. Interestingly, both erg1 conditional mutant and ipt1 mutant showed impaired energy-dependent efflux of R6G (rhodamine 6G), a well-known substrate of Cdr1p (Figures 3A and 3B). It would seem that the reduced drug-efflux activity probably contributes to hyper-susceptibility of these mutant cells.
R6G efflux in mutants and their morphogenetic studies on solid and liquid media
Considering the fact that R6G is a substrate of Cdr1p, whose efflux is reduced in erg1 and ipt1 disrupted cells, we examined how changes in Candida lipids could affect Cdr1p functioning. We had earlier observed that as compared with MFS transporter CaMdr1p, ABC transporter Cdr1p is relatively more sensitive to lipid fluctuations . It is thus possible that the depletion of ergosterol and accumulation of squalene, in conditional erg1 mutant strain, and accumulation of M(IP)C2 in ipt1 mutants might particularly affect Cdr1p protein partitioning within PM and in turn affect its functioning. This contention was confirmed by Western-blot analysis, wherein Cdr1p appeared to be poorly expressed in the PM fractions of erg1 conditional mutant (ΔE4.2.7) and ipt1 disrupted cells (Figure 2C).
Whether the small amount of Cdr1p was due to limited expression of protein or to the poor surface localization of Cdr1p was further resolved by confocal analysis. For this, we expressed Cdr1p as a GFP (green fluorescent protein)-tagged protein by integrating CDR1-GFP cassette to CDR1 locus using an SAT1 marker, in both WT and ERG1/erg1 conditional cells and ipt1 disrupted cells . Confocal microscopic images confirmed that GFP-tagged Cdr1p was poorly localized on the surface of conditional erg1 and ipt1 mutant strains. The typical rimmed fluorescent appearance of Cdr1p–GFP, which was seen in WT, was absent from both conditional erg1 and ipt1 mutant cells. These mutant cells rather showed fluorescence which appeared to be intracellularly trapped (Figure 2D). These results suggest that the fluctuations of lipid raft constituents result in improper surface localization of Cdr1p leading to its impaired functionality.
ERG1 and IPT1 disruptants show defect in hyphae formation
Furthermore, both hypersensitive erg1 and ipt1 mutants were unable to form hyphae in different media (Figures 3C and 3D). In this context, our results are in agreement with the recent report where the presence of polarized membrane domains rich in ergosterol and sphingolipid has been linked to morphogenesis and virulence in C. albicans .
In conclusion, both ergosterol and sphingolipids are two major constituents of membrane microdomains (membrane rafts) and blocking synthesis of either of the two results in increased drug susceptibilities and defects in hyphae formation in Candida cells. Considering our results and recent reports, it appears that the major ABC drug efflux pump protein, Cdr1p, of C. albicans may be preferentially associated within raft microdomains, which remains to be demonstrated experimentally.
Seventh Yeast Lipid Conference: Independent Meeting held at Swansea Clinical School, Swansea, Wales, U.K., 12–14 May 2005. Organized and Edited by D. Kelly, S. Kelly and D. Lamb (Swansea, U.K.).
We thank J. Morschhauser and N. Akhtar Gaur for the gift of GFP construct tagged with SAT1 marker and R. Serrano for PM-ATPase antibodies. Our thanks to Dr R.A. Vishwakarma (National Institute of Immunology, New Delhi, India) for providing excellent MS facilities. We also thank A. Mukhopadhyay (National Institute of Immunology) for providing us with confocal facilities. We are indebted to Ranbaxy Laboratories (New Delhi, India) for providing unlimited amounts of fluconazole. This work was partially supported by grants (to R. Prasad) from Department of Biotechnology, India [BT/PR3825/MED/14/488 (a)/2003 and BT/PR4862/BRB/10/360/2004], and the European Commission, Brussels (QLK-CT-2001-02377). A part of this work was presented at the 7th Yeast Lipid Conference, Swansea, Wales, U.K. R. Pasrija and T.P. thank the Council of Scientific and Industrial Research (CSIR), India, for providing research fellowships.
These authors contributed equally to this work.