|Home | About | Journals | Submit | Contact Us | Français|
The effects of S279F and S279Y point mutations in Candida albicans CYP51 (CaCYP51) on protein activity and on substrate (lanosterol) and azole antifungal binding were investigated. Both S279F and S279Y mutants bound lanosterol with 2-fold increased affinities (Ks, 7.1 and 8.0 μM, respectively) compared to the wild-type CaCYP51 protein (Ks, 13.5 μM). The S279F and S279Y mutants and the wild-type CaCYP51 protein bound fluconazole, voriconazole, and itraconazole tightly, producing typical type II binding spectra. However, the S279F and S279Y mutants had 4- to 5-fold lower affinities for fluconazole, 3.5-fold lower affinities for voriconazole, and 3.5- to 4-fold lower affinities for itraconazole than the wild-type CaCYP51 protein. The S279F and S279Y mutants gave 2.3- and 2.8-fold higher 50% inhibitory concentrations (IC50s) for fluconazole in a CYP51 reconstitution assay than the wild-type protein did. The increased fluconazole resistance conferred by the S279F and S279Y point mutations appeared to be mediated through a combination of a higher affinity for substrate and a lower affinity for fluconazole. In addition, lanosterol displaced fluconazole from the S279F and S279Y mutants but not from the wild-type protein. Molecular modeling of the wild-type protein indicated that the oxygen atom of S507 interacts with the second triazole ring of fluconazole, assisting in orientating fluconazole so that a more favorable binding conformation to heme is achieved. In contrast, in the two S279 mutant proteins, this S507-fluconazole interaction is absent, providing an explanation for the higher Kd values observed.
The emergence of azole-resistant strains of Candida albicans and other Candida species is becoming an increasing problem, especially for immunocompromised patients, due to the prophylactic use of azole drugs in the clinic and to prolonged treatment regimens (9, 36, 41, 45). First developed several decades ago, the azole antifungals retain a prominent position among drugs that can be used for systemic human infections as well as for topical uses and in agriculture. The compounds have a mode of action that involves inhibition of sterol 14-demethylation by binding to the heme of cytochrome P450 51 (CYP51) during ergosterol biosynthesis in fungi, leading to depletion of ergosterol and accumulation of 14-methylated eburicol and also of 14-methylergosta-3,6-diol, the latter through additional activity of sterol C-5-desaturase (18, 28, 50).
Resistance at the level of the drug target (CYP51) has been described, including some evidence of association with overexpression of CYP51, and more extensively, point mutations in the CYP51 protein have been identified as conferring azole resistance (19). Over 140 different amino acid point mutations in the C. albicans CYP51 (CaCYP51) protein have been reported from clinical strains (29), indicating that C. albicans lanosterol 14-α demethylase is highly permissive to structural changes without adversely affecting the catalytic function of this essential protein and drug target (17, 27, 29). Besides the importance of mutations in residues conserved among CYP51 proteins, e.g., R467K and G464S mutations in C. albicans (21, 25, 43, 52), importance in conferring azole resistance has been demonstrated for mutations in residues 428 to 459, constituting a major insertion loop found specifically in fungal CYP51 proteins (2, 7, 31). However, studies on the mechanisms by which these CYP51 mutations affect the biochemical and biophysical properties of the CYP51 protein that confers azole resistance have been limited.
A number of in silico models of the CaCYP51 protein have been published (16, 42, 53), based on published crystal structures for Bacillus megaterium P450BM3 (39) and Mycobacterium tuberculosis CYP51 (37, 38). However, most of the CaCYP51 point mutations observed in azole-resistant clinical C. albicans isolates are not predicted by these models to interact directly or through hydrophobic interactions with the main azole antifungal drugs.
The S279F mutation, located within the second CaCYP51 mutation hot spot, between residues 266 and 287, was first described by Marichal et al. (27). In this study, we examined the effect of the S279F point mutation on the azole sensitivity of CaCYP51, along with that of an analogous S279Y point mutation which results in a functional side chain hydroxyl group like that of serine in addition to the phenyl ring of phenylalanine. Using a combination of azole and lanosterol ligand binding techniques, inhibition of in vitro CaCYP51 activity by fluconazole, and molecular modeling, we discuss the mechanism(s) by which the S279F point mutation confers reduced fluconazole sensitivity on CaCYP51.
The CaCYP51 gene (UniProtKB accession number P10613) was isolated by PCR from the CYP51 cDNA in YEp51 (44), and like the case for the first CaCYP51 yeast expression construct to probe the reaction mechanism of this enzyme, it contained an altered codon for Ser263 (CTG to TCA) to ensure correct expression in Escherichia coli. The CaCYP51 gene was cloned into and expressed using the pCWori+ expression plasmid (21). The S279 mutations were generated by Genecust (Dudelange, Luxembourg) from the supplied wild-type pCWori+:CaCYP51 plasmid, altering the codon at amino acid position 279 from TCC to TTT for the S279F mutation and to TAC for the S279Y mutation. Gene integrity was confirmed by DNA sequencing.
The wild-type, S279F, and S279Y pCWori+:CaCYP51 constructs were transformed into competent E. coli DH5α cells and expressed as previously described (49). Recombinant CaCYP51 proteins were isolated according to the method of Arase et al. (3), except that 2% (wt/vol) sodium cholate and no Tween 20 were used in the sonication buffer. The solubilized CaCYP51 proteins were purified by affinity chromatography using Ni2+-nitrilotriacetic acid (Ni2+-NTA) agarose as previously described (6, 49). Ni2+-NTA agarose-purified CaCYP51 proteins were used for all subsequent spectral determinations. Protein purity was assessed by SDS-polyacrylamide gel electrophoresis.
Cytochrome P450 concentrations were determined by reduced carbon monoxide difference spectroscopy according to the method of Estabrook et al. (10), with carbon monoxide being passed through the cytochrome P450 solution prior to the addition of sodium dithionite to the sample cuvette. An extinction coefficient of 91 mM−1 cm−1 (33) was used to calculate cytochrome P450 concentrations from the absorbance difference between 446 and 490 nm. Absolute spectra were determined from 700 nm to 300 nm (light path, 4.5 mm), using 5 μM CaCYP51 in 0.1 M Tris-HCl (pH 8.1) and 25% (wt/vol) glycerol for the oxidized protein, the 10 mM sodium dithionite reduced protein, and the reduced carbon monoxide-P450 complex as previously described (6). The absolute reduced carbon monoxide spectra of the three CaCYP51 proteins were measured at 45-s intervals to determine the formation rate of the reduced CO-P450 adduct. All spectral determinations were made using a Hitachi U-3310 UV-visible (UV-Vis) spectrophotometer (San Jose, CA).
Aqueous lanosterol (0.1% [wt/vol]) was prepared as previously described (49). Lanosterol was progressively titrated against 5 μM CaCYP51 in a quartz semimicrocuvette (light path, 4.5 mm), with equivalent amounts of 5% (vol/vol) Tween 80 added to the reference cuvette, also containing 5 μM CaCYP51. The absorbance difference spectrum from 500 nm to 350 nm was determined after each incremental addition of lanosterol (up to 140 μM). The lanosterol saturation curve was constructed from the ΔA385–419 derived from the difference spectra, including corrections for changes in sample volume. The substrate binding constant (Ks) was determined by nonlinear regression (Levenberg-Marquardt algorithm) using the Hill equation. Spin state change was calculated using an extinction coefficient of 118 mM−1 cm−1, derived from the low- to high-spin-state transition of Mycobacterium smegmatis CYP164A2 (48).
Reconstitution assays were performed as previously described (49), with each assay mixture containing 15 μg lanosterol, 100 μg dilaurylphosphatidylcholine (DLPC), 2.5 nmol CaCYP51, 10 nmol truncated Δ33 yeast cytochrome P450 reductase (CPR) (47), 3 U yeast glucose-6-phosphate dehydrogenase, 2 μmol disodium glucose-6-phosphate, 100 μmol potassium phosphate, and 2 μmol β-NADPHNa4 (final pH, ~7.2). Sterol metabolites were recovered by extraction with n-hexane followed by derivatization with N,O-bis(trimethylsilyl)trifluoroacetamide and tetramethylsilane (BSTFA/TMS) prior to analysis by gas chromatography-mass spectrometry (GC-MS) (46). Fifty percent inhibitory concentration (IC50) determinations for fluconazole were obtained by varying the concentration of azole in 5 μl of dimethyl sulfoxide (DMSO) added to the assay mixture prior to incubation at 37°C and addition of β-NADPHNa4.
Binding of fluconazole, voriconazole, and itraconazole to the CaCYP51 proteins was performed as previously described (20, 49), using split cuvettes with a 4.5-mm light path. Stock solutions (0.2 mg ml−1) of fluconazole, voriconazole, and itraconazole were prepared in DMSO and progressively titrated against 5 μM wild-type, S279F, and S279Y CaCYP51 proteins in 0.1 M Tris-HCl (pH 8.1) and 25% (wt/vol) glycerol. The difference spectra from 500 nm to 350 nm were determined after each incremental addition of azole, and binding saturation curves were constructed with the ΔApeak−trough against azole concentration. The dissociation constants of the enzyme-azole complex (Kd) were determined by nonlinear regression (Levenberg-Marquardt algorithm) using a rearrangement of the Morrison equation for tight ligand binding (26, 30).
Fluconazole binding studies were also performed in the presence of 80 μM lanosterol to establish whether substrate-saturated CaCYP51 (5 μM) altered the observed Kd for fluconazole. The lanosterol was mixed with the CaCYP51 protein 5 min prior to commencing fluconazole binding. Where ligand binding was no longer tight, the Michaelis-Menten equation was used to fit the data.
In addition, fluconazole displacement by lanosterol was studied by incubating 5 μM CaCYP51 with fluconazole (0, 1, 2, 3, 4, 6, 8, 10, and 12 μM) for 5 min prior to the addition of 80 μM lanosterol and subsequent monitoring of the type I binding difference spectrum (500 to 350 nm) over 10 min. A fluconazole displacement plot of ΔA385–419 against fluconazole concentration was constructed.
Curve fitting of numerical data was performed using the computer program ProFit 6.1.12 (QuantumSoft, Zurich, Switzerland). Amino acid sequence alignments were performed using the computer program ClustalX, version 1.8 (http://www.clustal.org/).
All chemicals were obtained from Sigma Chemical Company (Poole, United Kingdom). Growth media, sodium ampicillin, isopropyl-β-d-thiogalactopyranoside (IPTG), and 5-aminolevulenic acid were obtained from Foremedium Ltd. (Hunstanton, United Kingdom). Ni2+-NTA agarose affinity chromatography matrix was obtained from Qiagen (Crawley, United Kingdom).
Structural modeling of the wild-type, S279F, and S279Y CaCYP51 sequences was performed using an automated homology modeling pipeline built with the Biskit structural bioinformatics platform (12), which scans the entire Protein Data Bank (PDB) library (http://www.pdb.org/pdb/home/home.do) for candidate homologies. The pipeline workflow incorporates the NCBI tool platform (51), including the BLAST program (1), for similarity searching of sequence databases. T-COFFEE (32) was used for alignment of the test sequence with the template. Homology models were generated over 10 iterations of the MODELLER program (11). All models were visualized using the molecular graphics program Chimera (35). CYP homologues are selected by the pipeline as those with the highest identity to the test sequence. Therefore, each of the CaCYP51 sequences may be modeled on a different group of homologues (Table 1), depending on the identity shared with available CYP structures, resulting in a model which reflects the secondary and tertiary structural differences incurred by amino acid substitutions. This modeling strategy has already been used successfully to explain the emergence of azole resistance mediated by Mycosphaerella graminicola CYP51 (31). The S279 substitutions led to omission of the human microsomal P450 1A2 homologue in favor of P450 3A4. The other pipeline-selected homologues used in the modeling of the three CaCYP51 proteins were closely related P450s from cyanobacteria, Streptomyces coelicolor, M. tuberculosis, and Trypanosoma brucei.
Fluconazole was docked by superimposition within the active site cysteine pocket, with the unprotonated N atom coordinated by the heme iron atom (31) according to the heme position and fluconazole binding in PDB structure 1EA1 (CYP51 from M. tuberculosis in complex with fluconazole ), as this structure provided the highest homology for any of the determined structures cocrystallized with fluconazole. Residues within given distance ranges of the docked azole were identified using the surface zone function in the Chimera molecular graphics program (35).
The volume of the heme cavity in the wild-type and variant protein models was determined using Pocket-Finder, written by Alasdair Laurie and Richard Jackson, University of Leeds, United Kingdom (http://www.modelling.leeds.ac.uk/pocketfinder/), which is a pocket detection algorithm based on Ligsite (13). The S279Y substitution leads to a slight constriction of the heme cavity.
Ser-279 is not conserved in all Candida species CYP51 proteins (Fig. 1A) or other fungal CYP51 proteins (Fig. 1B). However, Ser-279 is adjacent to a highly conserved leucine residue (Leu-280) among fungal CYP51 proteins. Ser-279 substitution in other fungal CYP51 proteins is predominantly with another aliphatic amino acid (Asn, Gln, Thr, Ala, or Glu), not an aromatic amino acid (Phe, Tyr, Trp, or His), suggesting that under normal circumstances, steric considerations do not favor an amino acid with a bulky side chain group at this position without an external selection pressure, such as prolonged therapeutic exposure to azole antifungal agents.
Following heterologous expression in E. coli, the proteins were extracted using cholate with sonication (3) and yielded 270 (±37), 195 (±22), and 323 (±24) nmol per liter of culture for the wild-type, S279F, and S279Y CaCYP51 proteins, respectively, as determined by carbon monoxide difference spectroscopy (10). The wild-type CaCYP51 expression level of 270 nmol per liter of culture was comparable to those obtained previously (5, 23, 49). Purification by chromatography on Ni2+-NTA agarose resulted in 54%, 57%, and 30% recoveries for native wild-type, S279F, and S279Y CaCYP51 proteins, respectively. SDS-polyacrylamide gel electrophoresis confirmed the purity of the Ni2+-NTA agarose-purified CaCYP51 proteins to be >90% as assessed by staining intensity, with an apparent molecular mass of 61 kDa, close to the predicted value of 61,221 Da for the wild-type protein including a 4-His C-terminal extension.
The absolute spectra of the resting oxidized forms of the wild-type, S279F, and S279Y CaCYP51 proteins (Fig. 2A) were similar. Wild-type CaCYP51 had α, β, Soret (γ), and δ spectral bands at 565, 536, 417, and 356 nm, respectively, compared to 569, 535 to 537, 418, and 354 nm for the two S279 mutant CaCYP51 proteins. These spectral characteristics were typical for a low-spin ferric cytochrome P450 enzyme (6, 14). Addition of 6 μM fluconazole to the oxidized CaCYP51 proteins (Fig. 2B) caused a red shift of 3 to 4 nm in the Soret peak, to 421 nm.
One-electron reduction with 10 mM sodium dithionite (Fig. 2C) did not alter the wavelength of the heme Soret peak of the three CaCYP51 proteins (417 to 418 nm). Reduced carbon monoxide spectra (Fig. 2D) confirmed that the wild-type, S279F, and S279Y CaCYP51 proteins all bound carbon monoxide to produce the red-shifted heme Soret peak at 446 nm characteristic of P450 enzymes, indicating that the CaCYP51 proteins were expressed in the native form. The formation of the reduced CO-P450 adducts for 5 μM S279F and S279Y CaCYP51 proteins was 2.5- and 1.7-fold slower than that with the wild-type CaCYP51 protein.
Progressive titration of the three CaCYP51 proteins with lanosterol gave type I difference spectra with a peak at 385 nm and a trough at 419 nm (Fig. 3A, B, and C). Type I binding spectra occur when the substrate or another molecule displaces the water molecule coordinated as the sixth ligand to the low-spin hexa-coordinated heme prosthetic group, causing the heme group to adopt a high-spin penta-coordinated conformation (14). Lanosterol binding to the S279F and S279Y CaCYP51 proteins produced 3.3- and 2.9-fold more intense type I binding spectra (Fig. 3D) than those obtained with the wild-type CaCYP51 protein, indicative of 3.3- and 2.9-fold larger changes in spin state. In addition, the S279F and S279Y CaCYP51 proteins had 2-fold higher affinities for lanosterol (Ks values of 7.1 and 8.0 μM, respectively) than the wild-type CaCYP51 protein (Ks value, 13.5 μM).
The presence of 80 μM lanosterol caused a 12% change in the spin state of the oxidized form of the wild-type CaCYP51 protein, from a 76% to a 64% low spin state, in comparison to 41% and 36% low- to high-spin-state changes for the S279F and S279Y CaCYP51 proteins, respectively. Previous studies have shown that the spin state change caused by lanosterol binding to wild-type CYP51 enzymes is relatively small (6, 24), with spin state changes usually not exceeding 10%. Such small changes in spin state are due to the enzyme-substrate complex existing in an equilibrium of low- and high-spin conformations (4, 40). Lepesheva et al. (22) attributed the small spin state change observed for CYP51 with substrate to be due to the axial ligated water molecule also having direct contact with several I-helix residues via a network of hydrogen bonds which strengthens the water-heme coordination.
Fluconazole bound tightly to all three CaCYP51 proteins, producing type II binding spectra (Fig. 4A, B, and C) with a peak at 429 nm and a trough at 412 nm. Type II binding spectra are caused by the triazole ring N-4 nitrogen coordinating as the sixth ligand with the heme iron (15). Fluconazole saturation curves (Fig. 4D) in the absence of lanosterol were best fitted using a rearrangement of the Morrison equation (26), indicating that the observed binding was “tight.” Tight binding is observed when the Kd for a ligand is similar to or lower than the concentration of the enzyme present (8). The S279F and S279Y CaCYP51 proteins had 4.3- and 5.5-fold lower affinities for fluconazole than the wild-type CaCYP51 protein (Table 2) in the absence of lanosterol, indicating that the S279F and S279Y point mutations conferred an increase in resistance to fluconazole. Voriconazole and itraconazole also bound tightly to all three CaCYP51 proteins (Table 2), producing similar type II binding spectra to those observed with fluconazole (data not shown). The S279F and S279Y CaCYP51 proteins had 3.6- and 3.7-fold lower affinities for voriconazole and 3.5- and 4.2-fold lower affinities for itraconazole (Table 2) than the wild-type CaCYP51 protein, mirroring the results observed with fluconazole. This was in contrast to a recent study by Park et al. (34) which found that F105L, D116E, Y132H, and R467K point mutations in CaCYP51 did not significantly alter the affinity of fluconazole binding to purified CaCYP51 protein.
Incubation of the CaCYP51 proteins with 80 μM lanosterol prior to commencing titration with fluconazole resulted in dramatic changes in the observed fluconazole binding with the S279F and S279Y CaCYP51 proteins (Fig. 5 and Table 2), with both S279 mutant proteins no longer binding fluconazole tightly, resulting in 304- and 247-fold reductions in the apparent affinity for fluconazole. In contrast, wild-type CaCYP51 still bound fluconazole tightly, with a similar affinity to that observed in the absence of lanosterol (Table 2). Binding of fluconazole to the three CaCYP51 proteins in the presence of 0.3% (vol/vol) Tween 80 resulted in no significant changes in Kd values, with fluconazole binding remaining tight, confirming that the Tween 80 present in the stock lanosterol solution did not affect fluconazole binding. Most of the observed increase in ΔAmax for fluconazole binding in the presence of lanosterol (Table 2) can be attributed to the increased proportion of CaCYP51 molecules occupying the high spin state (Soret peak at 393 nm) prior to fluconazole binding. The subsequent azole binding spectra coupled with the high- to low-spin-state change of CaCYP51 led to increased absorbance maxima at 426 to 428 nm and to the appearance of double-dip absorbance minima at 385 and 409 nm, best shown in Fig. 5A. However, the large increase in apparent Kd for fluconazole in the presence of lanosterol (Table 2) with the two S279 mutations suggests a substantially decreased affinity for fluconazole and/or increased affinity for lanosterol under these conditions compared to that predicted by the increased substrate affinity and decreased fluconazole affinity determined in isolation.
Fluconazole displacement experiments (Fig. 6B) indicated that fluconazole could not be displaced from 5 μM wild-type CaCYP51 protein (tight 1:1 binding) by 80 μM lanosterol, with no type I lanosterol binding spectra produced at fluconazole concentrations greater than the P450 concentration. In contrast, fluconazole was displaced from the two S279 mutant CaCYP51 proteins by 80 μM lanosterol, with weak type I lanosterol binding spectra obtained even in the presence of 12 μM fluconazole. These results indicate that the 250- to 300-fold higher apparent Kd values obtained for fluconazole in the presence of 80 μM lanosterol with the two S279 mutant CaCYP51 proteins were due to the measured Kd values actually being composites of the Kd values of fluconazole and lanosterol and the equilibrium constants between the three enzyme species, i.e., CaCYP51, CaCYP51-lanosterol, and CaCYP51-fluconazole. Therefore, fluconazole binding in the presence of lanosterol and displacement of fluconazole by lanosterol can be used to screen for resistance prior to performing CYP51 reconstitution assay IC50 determinations.
S279F and S279Y CaCYP51 proteins gave 2.4- and 2.8-fold higher IC50s (Fig. 6A) for fluconazole (3.28 and 3.87 μM, respectively) than the wild-type CaCYP51 protein (1.39 μM), confirming the reduced sensitivity to fluconazole of the two S279 mutant CaCYP51 proteins. A representative gas chromatogram of the sterol metabolites, along with mass fragmentation patterns of substrate and product, is available in Fig. S1 in the supplemental material. The IC50 of the wild-type CaCYP51 protein was as expected for a tightly binding inhibitor that could not be displaced by substrate, i.e., half the enzyme concentration present. However, the CYP51 reconstitution assay IC50s for the two S279 mutant CaCYP51 proteins were higher than the equivalent ΔA50 point for fluconazole displacement (Fig. 6B). This was due to CYP51 catalysis proceeding by two routes in the S279F and S279Y CaCYP51 assays. First, lanosterol displaced fluconazole from CaCYP51, leading to catalysis and release of product (slower route), and second, lanosterol bound directly to free CaCYP51 immediately after product release, leading to further catalysis (faster route), with the two routes combined producing significant catalytic turnover at a fluconazole concentration twice that of the P450 concentration (Fig. 6A).
The ΔAmax/Ks values for lanosterol binding with the two S279 mutant CaCYP51 proteins were 6 (S279F)- and 5 (S279Y)-fold greater than that for the wild-type protein, suggesting that increased catalytic turnover was possible. However, there were no significant differences in the catalytic turnover numbers of the three CaCYP51 enzymes in the CYP51 reconstitution assays, indicating that the two S279 mutations did not appear to alter the catalytic efficiency of the CaCYP51 enzyme in the absence of fluconazole. While Ks values are generally indicative of Km values during catalysis, ΔAmax values cannot be compared directly with kcat (Vmax) values for CYP51 enzymes due to the spin state equilibrium that exists for the CYP51-lanosterol complex (22). The increase in ΔAmax values observed with the two S279 mutant proteins could be due to the S279F and S279Y proteins having CYP51-lanosterol spin state equilibrium constants that favor the high-spin conformation relative to the wild-type protein. Previous experiments (27) have shown a 5-fold higher IC50 for fluconazole as measured by sterol biosynthesis and a 1.5-fold higher IC50 for fluconazole as measured by cytochrome P450 CO displacement using the subcellular fraction isolated from azole-resistant C. albicans strain NCPF3363 (containing S279F, Y132H, and G465S CaCYP51 mutations) instead of the subcellular fraction isolated from azole-sensitive strain ATCC 22516.
Structural modeling of the wild-type, S279F, and S279Y CaCYP51 proteins revealed notable differences in secondary and tertiary structure brought about by the point mutations at position 279. In the wild-type protein, S279 is located outside the protein, on an external β turn (Fig. 7A). With the S279F and S279Y substitutions, F279 and Y279 become incorporated within a short section of α helix preceding the long I helix, resulting in a large movement of position (Fig. 7B and C). In the wild-type protein, the subsequent section of β turn returns toward the center of the protein, and I282, D294, and Q295 directly border the heme cavity, with the last two located at the end of the long I helix. These three residues are removed from the heme cavity in the S279F and S279Y mutants. Other residues that are removed from the heme cavity include K128 and F145, located at positions that undergo mutations (K128T and F145L) that affect fluconazole sensitivity (27). Conversely, Y132, involved in the Y132H mutation (27), does not border the cavity in the wild-type protein but is introduced into the cavity in the S279 mutants. The rearrangement of sections surrounding the heme cavity leads to changes in the predicted interaction of the distal oxygen atom of S507 with fluconazole (Fig. 8A). Structural modeling of fluconazole binding gave a distance of 3.5 Å between S507 and fluconazole for the wild-type CaCYP51 protein. This distance increased dramatically with the S279 mutations, to 8.7 Å in the case of the S279F protein (Fig. 8B) and 8.8 Å for the S279Y protein (Fig. 8C). The distances between fluconazole and the other residues predicted to be within the interaction range, namely, G307, T311, and L376, were found to be similar for the wild-type and S279 mutant CaCYP51 proteins. The loss of interaction with S507 is consistent with the observed reduction in inhibition by fluconazole, while the maintenance of interactions with the other residues is in keeping with fluconazole continuing to exert some inhibition. This suggests that heme-coordinated binding of fluconazole occurs in the S279F and S279Y CaCYP51 proteins but that binding is impaired in comparison with that of the wild-type CaCYP51 protein by involving fewer interactions with heme cavity residues, resulting in the lower binding affinities observed. It is likely that the interaction with S507 is particularly important, as it is the only residue that interacts with the azole ring of fluconazole most distal from the heme group, serving to assist in the correct orientation of the triazole molecule for coordination of the proximal azole ring by the heme group. Loss of the interaction with S507 in the S279F and S279Y CaCYP51 mutants is likely to result in less efficient alignment of fluconazole in the heme cavity and in impaired binding at the heme group, supporting the 4.3- and 5.5-fold lower affinities for fluconazole observed with the S279F and S279Y proteins than that observed with wild-type CaCYP51.
S279 was previously observed to be altered in the CYP51 protein of azole-resistant C. albicans strain NCPF3363 (27), but the azole resistance mechanism was not fully resolved at the protein level. Strain NCPF3363 CaCYP51 (27) contained the S279F mutation in combination with Y132H and G465S mutations, with the conferred azole resistance being attributed primarily to the Y132H mutation (43). However, this was contradicted by in vitro studies using recombinant CaCYP51 proteins containing the Y132H point mutation (5, 34) where the mutation did not significantly alter the affinity of fluconazole binding. Our aim was to assess the impact of S279 point mutations in isolation on the azole resistance of CaCYP51. The S279F and S279Y CaCYP51 mutations resulted in a 2-fold increased affinity for the substrate lanosterol and in 4.3- to 5.5-fold lower affinities for fluconazole than those of wild-type CaCYP51. Affinities for voriconazole and itraconazole were also reduced 3- to 4-fold for the two S279 mutants compared with the wild-type CaCYP51 protein. The two S279 point mutations also conferred increased resistance to inhibition of sterol 14-α demethylase activity by fluconazole, with 2.4- to 2.8-fold increases in IC50 compared with the wild-type CaCYP51 protein. This suggested that the residue change from serine (β-hydroxyalanine) to the bulkier phenylalanine or tyrosine (p-hydroxyphenylalanine) residue, rather than the removal of the hydroxyl functional group (as in the S279F mutant), altered the three-dimensional structure of CaCYP51 so that fluconazole bound less tightly to the heme prosthetic group, especially when the CaCYP51 proteins were saturated with lanosterol. Molecular modeling of fluconazole binding suggested that the lack of interaction of the oxygen atom of S507 with the second triazole ring of fluconazole in the two S279 mutant CaCYP51 proteins was responsible for the lower observed affinities for fluconazole, as the interaction of S507 with fluconazole in the wild-type CaCYP51 protein ensures the optimal orientation for fluconazole binding to the heme. This study demonstrates how an S279 mutation in the amino acid sequence of CaCYP51 can confer enhanced resistance of the CYP51 target to inhibition by azole antifungal agents and highlights the need for continuous development of new azole antifungal compounds to combat the increasing emergence of clinical azole-resistant C. albicans strains.
We are grateful to the European Union for support by the FP6 EURESFUN project under Genomics for Biotechnology and Health and to the Biotechnology and Biological Science Research Council (BBSRC) of the United Kingdom for supporting this work.
We are grateful to the Engineering and Physical Sciences Research Council National Mass Spectrometry Service Centre at Swansea University for assistance.
Published ahead of print 17 January 2012
Supplemental material for this article may be found at http://aac.asm.org/.