Structure-Function Correlations of Two Highly Conserved Motifs in Saccharomyces cerevisiae Squalene Epoxidase

Abstract

Saccharomyces cerevisiae squalene epoxidase contains two highly conserved motifs, 1 and 2, of unknown function. Amino acid substitutions in both regions reduce enzyme activity and/or alter allylamine sensitivity. In the homology model, these motifs flank the flavin adenine dinucleotide cofactor and form part of the interface between cofactor and substrate binding domains.

Squalene epoxidase (SE) is an essential flavin adenine dinucleotide (FAD)-dependent monooxygenase in sterol biosynthesis. The fungal enzyme is the target for the allylamines terbinafine and naftifine which inhibit SE in a noncompetitive manner (11). The homology model of SE (Erg1p) of Saccharomyces cerevisiae (9) shows a two-domain structure typical for this class of flavin-dependent enzymes (12): the FAD binding domain (Fig. 1A, lower part, pink) with the characteristic nucleotide binding αβ-fold (1, 9, 13), and a second domain typically referred to asthe substrate binding domain (Fig. 1A, upper part, blue). The active site is at the interface of these domains.

FIG. 1.

Modeled structure of Erg1p from S. cerevisiae based on the crystal structure of p-hydroxybenzoate hydroxylase. (A) Schematic representation of the three-dimensional model. The cofactor binding domain is shown in pink, the substrate binding domain in blue. The FAD cofactor is depicted as a yellow stick model. The two conserved regions are highlighted in green (motif-1) and cyan (motif-2). (B) Close-up view of motif-1. Amino acid side chains for which mutations are described in the text are displayed as CPK models (C, orange; N, blue; O, red). (C) Close-up view of motif-2. This figure was prepared using PyMol software (http://www.pymol.org).

We previously identified two interesting regions of SE, based on sequence conservation; the regions were denoted motif-1 and motif-2 (Fig. 1A, green and cyan) (10). To investigate the role of the motifs, scanning alanine mutagenesis was performed by two-step PCR. Subsequently, the respective erg1 alleles were cloned into the low-copy-number vector pRS315, and the recombinant plasmids were transformed into S. cerevisiae KLN1 (9). This mutant fails to express SE and consequently shows an aerobically lethal phenotype (5).

The amino acid sequence motif-1, ₅₅DRIVGELMQPGG₆₆, forms a loop at the interface between the FAD and the substrate binding domains (Fig. 1A, green). Alanine scanning of the eight conserved residues (in bold) gave rise to functional SEs. The most-pronounced alterations regarding sensitivity to antifungal compounds, enzyme activity, or protein levels were observed for SE with substitutions at residues E₆₀ and G₆₆.

The E₆₀A variant poorly complemented growth of KLN1, drastically increased sensitivity to terbinafine and naftifine (>50-fold compared to that in the wild type) and ketoconazole (>5-fold), and conferred temperature-sensitive growth (Fig. 2C). Strongly reduced enzymatic activity was accompanied by increased Erg1p levels (Fig. 2D and E), as already demonstrated previously for other hypersensitive Erg1p variants (2, 9). E₆₀ lies in a mechanistically sensitive region at the tip of the loop structure described above, near the cofactor and the putative active site (Fig. 1B). Therefore, we also generated an E₆₀Q variant. This showed all the effects of the E₆₀A variant, especially a reduced enzymatic activity which was accompanied by high Erg1p levels (Fig. 2C, D, and E). This indicates a potential role for the negatively charged E₆₀ in the catalytic mechanism. In addition, the structural model indicates a possible electrostatic interaction with R₂₆₉, an amino acid which is known to be specifically involved in allylamine sensitivity (9). This interaction could at least in part be complemented by hydrogen bonds to the glutamine in E₆₀Q, accounting for the slight difference in drug sensitivity compared to E₆₀A (Fig. 2C).

FIG. 2.

Properties of S. cerevisiae KLN1 expressing motif-1 Erg1p variants E₆₀A and G₆₆A in comparison to wild-type NS1. (A) Sensitivity to terbinafine in liquid medium. Optical density at 600 nm (OD₆₀₀) was measured after 24 h at 30°C in yeast extract-peptone-dextrose medium supplemented with the indicated concentrations of terbinafine. (B and C) Spot assays. Serial dilutions of overnight cultures were spotted on yeast-peptone-dextrose agar plates containing the indicated inhibitor concentrations. Plates were incubated for 3 days at 30°C or 37°C. In each experiment, at least five concentrations of the respective inhibitor were tested; only one concentration is shown. (D) In vitro SE activity in cell extracts: sterol intermediate formation. Three milligrams of total protein in crude cell extracts was incubated with ¹⁴C-mevalonic acid and the cofactor-substrate mix. Lipid extracts were subjected to thin-layer chromatography analysis. Values of the lipid species (S, squalene; SE, squalene epoxide; L, lanosterol) represent the percentages of total radioactivity per lane. (E) Western blot analysis of Erg1p. Aliquots of whole-cell extracts were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Western blot analysis was performed using anti-Erg1p antibodies and anti-glyceraldehyde-3-phosphate dehydrogenase (GAPDH) antibodies as a loading control, as described previously (8).

The G₆₆A variant showed only slightly altered enzyme activity and protein levels (Fig. 2D and E) and an approximately 10-fold increase in sensitivity to terbinafine (Fig. 2A). However, this was not accompanied by a general sensitivity to other sterol biosynthesis inhibitors such as ketoconazole (Fig. 2B).

The conserved motif-2 (10), ₃₃₈NMRHPLTGGGMTV₃₅₀, is located close to the previously described FADII binding site (₃₃₄GD₃₃₅) (9) and the potential substrate binding residues identified in rat SE (7). This motif also forms a loop near the cofactor at the interface between the two domains of SE and is located opposite to motif-1 (Fig. 1A, cyan). Alanine scanning of all amino acids of motif-2 resulted in functional Erg1p variants. Only four amino acid substitutions (R₃₄₀A, G₃₄₅A, G₃₄₆A, and M₃₄₈A) lead to significantly altered phenotypes.

The variant R₃₄₀A exhibited very low enzymatic activity (Fig. 3C). The steady-state protein levels were the highest of all mutants tested at present (Fig. 3D). This can in part be explained by the high stability of this SE variant, which shows no degradation within 24 h at 30°C compared to the 60-min half-life of wild-type Erg1p (9) (Fig. 3E). The homology model indicates that R₃₄₀ may form a salt bridge with Glu_432. This would tend to destabilize the protein, as typically observed for internal salt bridges (3). Therefore, a loss of the salt bridge might in part explain the increased protein stability. Low enzymatic activity of the R₃₄₀A variant should lead to high sensitivity toward allylamines and ketoconazole, as already shown in other mutants, such as G₃₀S (Fig. 3A) or L₃₇P (8, 9). However, sensitivity of the R₃₄₀A variant toward allylamines was increased only twofold (Fig. 3A). A likely explanation for these data is that a structural alteration in SE induced by this amino acid substitution interferes with allylamine binding.

FIG. 3.

Properties of S. cerevisiae KLN1 strains expressing Erg1p variants with alterations in motif-2 (R₃₄₀A, G₃₄₅A, G₃₄₆A, and M₃₄₈A) in comparison to the wild type (NS1) and the terbinafine-hypersensitive G₃₀S and R₂₆₉G variants. For details see the Fig. 2 legend. (A) Spot assay of the KLN1 strains expressing motif-2 variants in the presence of terbinafine, naftifine, and ketoconazole. (B) Growth of strain KLN1[G₃₄₆A] in the presence of terbinafine after 24 h at 30°C. (C) In vitro SE activity in cell extracts: sterol intermediate formation. S, squalene; SE, squalene epoxide; L, lanosterol. (D) Western blot analysis of Erg1p variants. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) served as a loading control. (E) Stability of the Erg1p variant R₃₄₀A. KLN1[R₃₄₀A] was grown at 30°C in yeast extract-peptone-dextrose medium to early log phase. After addition of cycloheximide, 4 optical density at 600 nm (OD₆₀₀) equivalents were removed at the indicated time points. Aliquots of whole-cell extracts were subjected to Western blot analysis using anti-Erg1p antibodies and anti-GAPDH antibodies as a loading control.

Substitution of G₃₄₅ by alanine resulted in increased allylamine sensitivity without cross-sensitivity to ketoconazole, decreased enzyme activity, and induced Erg1p levels (Fig. 3A, C, and D). The G₃₄₆A variant exhibited wild-type enzyme activity (Fig. 3C), steady-state protein levels (data not shown), and naftifine and ketoconazole sensitivity (Fig. 3A), but it was less sensitive toward terbinafine (Fig. 3B). To the best of our knowledge, this is the first mutation which confers terbinafine resistance without affecting an amino acid in the proposed terbinafine binding site on the surface of Erg1p (4, 6, 9).

The M₃₄₈A variant was more sensitive toward terbinafine and naftifine and slightly more sensitive toward ketoconazole. Furthermore, enzyme activity was reduced and protein levels were induced (Fig. 3A, C, and D). In the model, the main chain amide group of M₃₄₈ interacts with N1 of FAD, indicating that structural perturbations in this particular region are likely to affect enzyme activity.

Variants G₆₆A (motif-1), G₃₄₅A, and M₃₄₈A (motif-2) confer increased allylamine sensitivity, similar to that observed with R₂₆₉G (Fig. 3) (9). None of these amino acid residues lie in the presumed terbinafine binding site on the surface of Erg1p. This precludes direct interactions with the inhibitor. In the same way as R₂₆₉G, we believe that these mutations facilitate conformational changes upon terbinafine binding which are required for noncompetitive inhibition. In order to prove this at the molecular level, the interaction of Erg1p and the inhibitor have to be measured directly. However, none of the enzyme variants is available in a purified form yet.

Together with our previous studies (8, 9), we have completed the characterization of all highly conserved motifs in fungal SEs. In the structural model, motif-1 and -2 (10) flank the FAD cofactor and form part of the interface between the cofactor and substrate binding domains (Fig. 1). The present data clearly show that specific mutations within these motifs do affect enzyme activity and sensitivity toward allylamines, thus underlining the structural and functional importance of the interface region.