A putative 8,7SI (sterol 8,7-isomerase) from Zea mays, termed Zm8,7SI, has been isolated from an EST (expressed sequence tag) library and subcloned into the yeast erg2 mutant lacking 8,7SI activity. Zm8,7SI restored endogenous ergosterol synthesis. An in vitro enzymatic assay in the corresponding yeast microsomal extract indicated that the preferred Δ8-sterol substrate possesses a single C4α methyl group, in contrast with 8,7SIs from animals and fungi, thus reflecting the diversity in the structure of their active site in relation to the distinct sterol biosynthetic pathways. In accordance with the proposed catalytic mechanism, a series of lipophilic ammonium-ion-containing derivatives possessing a variety of structures and biological properties, potently inhibited the Zm8,7SI in vitro. To evaluate the importance of a series of conserved acidic and tryptophan residues which could be involved in the Zm8,7SI catalytic mechanism, 20 mutants of Zm8,7SI were constructed as well as a number of corresponding mutants of the Saccharomyces cerevisiae 8,7SI. The mutated isomerases were assayed in vivo by sterol analysis and quantification of Δ5,7-sterols and directly in vitro by examination of the activities of the recombinant Zm8,7SI mutants. These studies have identified His74, Glu78, Asp107, Glu121, Trp66 and Trp193 that are required for Zm8,7SI activity and show that binding of the enzyme–substrate complex is impaired in the mutant T124I. They underline the functional homology between the plant and animal 8,7SIs on one hand, in contrast with the yeast 8,7SI on the other hand, in accordance with their molecular diversity and distinct mechanisms.
- plant sterol
- sterol 8,7-isomerase
Sterols are essential components of all eukaryotic cell membranes. The structural differences between sterols of animals, plants and fungi are linked to the biosynthetic pathways that differ significantly downstream of squalene epoxide [1–3]. Nevertheless, these pathways retain several enzymatic steps in common, for example those which together are necessary for the migration of the initial unsaturation in the B ring to the Δ5 position in the functional sterols. This includes isomerization of the C8 double bond to the C7 position catalysed by the 8,7SI (sterol 8,7-isomerase) (Figure 1). The sterol 8,7-isomerization is triggered by an α-protonation of the Δ8-double bond from an appropriately positioned general acid residue, giving a C8 carbocationic HEI (high-energy intermediate). It follows elimination of a proton at C7 accepted by a general base residue (Figure 1) . Loss of the 7β-H occurs in animals [5,6] and plants , whereas in fungi the enzyme proceeds with elimination of the 7α-H [7,8] (Figure 1).
8,7SI is one of the sterol biosynthesis enzymes that deserves special attention. In yeast, 8,7SI is a highly regulated step in sterol biogenesis . It is the primary target for a variety of compounds widely used in agriculture and medicine [2,10–14]. Azadecalines , morpholine and piperidine fungicides  inhibit the plant and yeast enzymes by mimicking the structure of the carbocationic HEI. The animal 8,7SI is inhibited by a variety of structurally distinct pharmacological compounds including receptor σ-ligand SR31747A , trifluoperazine [11,17] and tamoxifen . In addition, the σ-ligands, haloperidol, ifenprodil and verapamil were shown to inhibit the production of ergosterol in wild-type Saccharyomyces cerevisiae  and in the erg2 mutant complemented with Arabidopsis thaliana 8,7SI cDNA .
cDNAs of 8,7SI genes from fungi, animals and plants have been isolated and in some cases functionally characterized by complementation in the yeast erg2 mutant deficient in 8,7SI [16,18,19]. The mammalian 8,7SI was shown to correspond to the EBP (emopamil-binding protein), a previously isolated high-affinity receptor for the phenylalkylamine calcium-antagonist emopamil and other anti-ischaemic drugs . A number of syndromes caused by EBP gene mutations, such as X-linked dominant Conradi–Hünermann syndrome and CHILD (congenital hemidysplasia with ichthyosiform erythroderma and limb defect), or the ‘tattered’ mouse phenotype, were characterized by abnormally increased levels of cholest-8(9)-en-3β-ol due to inactivation of 8,7SI [21–23]. Human 8,7SI (EBP) and the yeast isoenzyme (Erg2p) are structurally unrelated and differ particularly in their substrate specificity [5,7]. Alanine-scanning mutagenesis was used to identify residues of human EBP required for in vivo sterol 8,7-isomerization . It is noteworthy that a mammalian protein that is structurally related to the yeast Erg2p has been described as a σ1 receptor but it exhibits no SI activity upon expression in yeast .
In plants, most of the reports on enzymes involved in the kinetic control of the post-squalene sterol pathway have been described in Zea mays . Moreover, Z. mays has been widely used as a monocot model to assay both in vivo and in vitro a variety of fungicides and herbicides inhibiting plant sterol biosynthesis . A plant 8,7SI cDNA has also been isolated by functional complementation of the corresponding yeast sterol mutant (erg2) by an Arabidopsis cDNA library and characterized by exposure to σ-ligands . The plant 8,7SI shows an homology with the animal enzyme and is not related to the yeast enzyme. The importance of a functional 8,7SI for normal plant growth and development has been shown using the Arabidopsis hydra mutants, deficient in 8,7SI .
In plants, little attention has been directed towards the in vitro enzymological characterization of 8,7SI, including features required for the specificity of sterol substrate. This is particularly due to the difficulties in obtaining the substrates and sufficient amounts of microsomal protein required, as the Vmax of the enzymes of post-squalene phytosterol synthesis are low . In addition, the integral membrane-bound nature of the 8,7SI renders a structural characterization through crystallization and/or NMR challenging.
Site-directed mutational analysis of plant and yeast 8,7SIs to establish the importance of select amino acids in the proteins and to identify amino acids essential for catalysis has not so far been performed.
In the present study we report the cloning and in vivo functional expression in yeast, of a 8,7SI cDNA from Z. mays (Zm8,7SI), and the first in vitro enzymological characterization and inhibition analysis of a plant recombinant 8,7SI. Moreover, to provide a framework for future structural studies, we report the first site-directed mutagenesis analysis of plant and yeast 8,7SIs. We mutated a number of conserved acidic residues, or putative cation-stabilizing tryptophan residues, looking for amino acid residues that are essential for catalytic activity. The mutated isomerases were assayed both in vivo by sterol analysis and quantification of Δ5,7-sterols, and directly in vitro by examining the activities of the recombinant mutated Zm8,7SI in the corresponding yeast microsome extracts. This allowed particularly the identification of a number of amino acid residues essential for the plant 8,7SI activity. Taken together, the results revealed functional homology and diversity with the isomerases of animals and fungi.
Tamoxifen (Figure 2, compound 16) and trifluoperazine (Figure 2, compound 15) were purchased from Sigma. N-Benzyl-4α,10-dimethyl-8-aza-trans-decal-3β-ol (Figure 2, compound 17), N-[(1,5,9)-trimethyldecyl]-4α,10-dimethyl-8-aza-trans-decal-3β-ol (Figure 2, compound 18) and N-benzamido-4α,10-dimethyl-8-aza-trans-decal-3β-ol (Figure 2, compound 19) were synthesized as previously described [15,27]. AY9944 (Figure 2, compound 13) was a gift of Dr Dvornik (Ayerst Research Laboratory, Montreal, Canada). We thank the BASF Agrochemical Station (Limburgerhof, Germany) for providing tridemorph (Figure 2, compound 11) and fenpropimorph (Figure 2, compound 12). SR31747 (Figure 2, compound 14) was provided by Dr Loison (Sanofi-Aventis Recherche-Developpement, Labège, France).
Strains and plasmids
The 8,7SI-deficient strain, WA10-3-1D (Mat a, erg2-4 ::LEU2, ura3-52, leu2-3, leu2-112, his7-2, ade5) used in the present study has been described previously . The strain was grown aerobically at 30 °C on solid or liquid minimal medium [0.67% (w/v) yeast nitrogen base (DIFCO) and 2% (w/v) glucose] containing suitable supplements (50 mg/l each) and casamino acids (1 g/l) or complete medium [YPD: 1% (w/v) yeast extract, 2% (w/v) peptone and 2% (w/v) glucose]. The yeast strain Δerg2pEMR1235 containing the murine 8,7SI was provided by Dr Loison. The pVT102U  S. cerevisiae shuttle vector optimized for expressing recombinant proteins in yeast was used for cloning, sequencing and transformation of the erg2 strain. This plasmid contains an Escherichia coli origin of replication, a yeast 2μ origin of replication, an E. coli ampicillin-resistance gene and the yeast URA3 gene. It also contains an expression cassette including the ADH1 (alcohol dehydrogenase) promoter and terminator.
A BLAST search with the A. thaliana 8,7SI and the EBP from Homo sapiens revealed significant sequence similarities with the hypothetical protein of a Z. mays EST (expressed sequence tag) clone (GenBank® accession no BG840208) which was obtained from the Schnable laboratory of the Iowa State University of Science and Technology (Ames, IA, U.S.A.). Sequencing of this EST clone revealed that it contained a complete ORF (open reading frame) which was PCR-amplified with the primer pair P1 (Table 1) containing an XbaI site, and the reverse primer P2 containing an XhoI site. This 666-bp cDNA, termed Zm8,7SI (AY533175), was cloned between the XbaI and XhoI sites of the pVT102U shuttle vector and placed under the control of the constitutive ADH promoter to generate the plasmid pVT-Zm8,7SI. Both strands of the amplified cDNA were sequenced to ensure sequence fidelity.
Z. mays FLAG 8,7SI
In order to check the expression and accumulation of wild-type and mutated Zm8,7SIs in yeast microsomes, an N-terminal FLAG epitope (MDYKDDDDK) was fused to the 8,7SI protein. For this purpose a DNA molecule containing the corresponding nucleotide sequence at the 5′-end was synthesized by PCR using the following forward primer: P3 containing an XbaI site and the reverse primer P2 using the Zm8,7SI cDNA as a template. The ZmFLAG8,7SI was further cloned between the XbaI and XhoI sites of the pVT102U vector to generate pVT- ZmFLAG8,7SI.
Site-directed mutagenesis was performed using the Quikchange® site-directed mutagenesis kit (Stratagene) according to the manufacturer's instructions. pVT-Zm8,7SI or pVT-ZmFLAG8,7SI were used as a template and the synthetic oligonucleotide primers listed in Table 1. Putative, positive clones were picked, plasmids isolated and sequenced.
S. cerevisiae transformations were performed using the lithium acetate procedure as previously described . The transformed erg2 yeast strain was plated on minimal YNB medium containing adenine and histidine (50 μg·ml−1 each). Cells were grown aerobically at 30 °C.
Freeze-dried yeast cells (10–30 mg) were sonicated in the presence of methanol/KOH [6% (w/v), 2 ml] for 10 min and heated in the same medium at 70 °C under reflux conditions for 2 h. The mixture was diluted with 1 vol. of water and total sterols were extracted three times with 1 vol. of hexane. The extract was dried on Na2SO4 and evaporated to dryness. Sterols were analysed by GC. GC analysis was carried out with a Varian GC model 8300 equipped with a flame-ionization detector at 300 °C, the column injector at 250 °C and a fused capillary column [WCOT: 30 m×0.25 mm i.d. (internal diameter)] coated with DB1 (hydrogen flow-rate of 2 ml/min). The temperature programme used included a 30 °C/min increase from 60 to 240 °C, followed by a 2 °C/min increase from 240 to 280 °C. Relative retention times (tR) are given with respect to cholesterol (tR=1). Identification of individual sterols was performed using a GC-MS spectrometer (Agilent 5973N) equipped with an ‘on column’ injector and a capillary column (30 m×0.25 m i.d.) coated with DB5. Sterols were unequivocally identified by retention times and an electron-impact spectrum identical with that of authentic standards .
Melting points are uncorrected. Proton magnetic resonance was monitored in a [2H]chloroform solution with a Brucker 400 or 500 MHz spectrometer. Chemical shifts (δ) (p.p.m.) were determined relative to tetramethylsilane. Coupling constants (J) were in Hertz.
Δ8-Sitostenol [Figure 2, compound 1; 24(R)-24-ethyl-5α-stigmasta-8-en-3β-ol] was extracted from bramble (Rubus fruticosus) suspension cultures treated with AY 9944 as previously described : GC purity>98% (tR=1.233, DB1), MS m/z [rel. int. (percentage relative abundance of molecular and prominent fragment ions)] M+=414(100), 399(42), 381(8), 315(13), 273(20), 255(19), 229(20), 213(20); 1H-NMR (CDCl3; acetate derivative): δ: 0 ;605 (3 H, s, H18), 0.812 (3 H, d, J 6.6 Hz, H26), 0.835 (3 H, d, J 6.6 Hz, H27), 0.843 (3 H, t, J 6 Hz, H29), 0.962 (3 H, s, H19), 4.698 (1 H, m, H3α).
Zymosterol (Figure 2, compound 2) was isolated from the yeast strain erg6, erg2 and crystallized from methanol: melting point 107–109 °C; GC purity>98% (tR=1.057, DB1); MS m/z (rel. int.) M+=384(88), 369(100), 351(25), 271(39), 229(40); 1H-NMR (CDCl3): δ: 0.610 (3 H, s, H18), 0.944 (3 H, d, J 5.2 Hz, H21), 0.950 (3 H, s, H19), 1.601 (3H, s, H27), 1.681 (3 H, d, J 0.8 Hz, H26), 3.614 (1 H, m, H3α), 5.093 (1 H, tt, J 7.1 Hz, J 1.4 Hz, H24).
4α-Methyl-fecosterol (Figure 2, compound 3), 4,4-dimethyl-zymosterol (Figure 2, compound 4) and 4,4-dimethyl-fecosterol (Figure 2, compound 5) were isolated from the wild-type yeast strain FL530 grown in the presence of APB (6-amino-2-n-pentylthiobenzothiazole) or from the mutant erg-25-25c, as previously described . 4α-Methyl-fecosterol (Figure 2, compound 3): GC purity>98% (tR=1.198, DB1; tR=1.175, DB5); MS m/z (rel. int.) M+=412(100), 397(68), 379(27), 285(46), 227(31). 4,4-Dimethyl-zymosterol (Figure 2, compound 4): GC purity>98% (tR=1.241, DB1; tR=1.210, DB5); MS m/z (rel. int.) M+=412(100), 397(70), 379(39), 299(23), 259(35), 241(46). 4,4-Dimethyl-fecosterol (Figure 2, compound 5): GC purity>98% (tR=1.313, DB1; tR=1.267, DB5); MS m/z (rel. int.) M+=426(100), 411(45), 393(26), 342(3), 327(8), 299(26), 259(21), 241(22).
Obtusifoliol (Figure 2, compound 6) was synthezised as previously described : melting point 139–141 °C from methanol; GC purity>98% (tR=1.165, DB5); MS m/z (rel. int.) M+=426(30), 411(100), 393(15), 327(13), 259(8), 245(25), 215(10).
Cholest-8(14)-en-3β-ol (Figure 2, compound 7) was synthezised as previously described : melting point=129–131 °C from methanol; GC purity>98% (tR=1.010, DB1); MS m/z (rel. int.) M+=386(100), 371(33), 353(12), 273(13), 255(10). 1H-NMR (CDCl3): δ: 0.829 (3 H, s, H18), 0.872 (3 H, d, J 6.6 Hz, H26 or H27), 0.875 (3 H, d, J 6.6 Hz, H26 or H27), 0.906 (3 H, s, H19), 0.916 (3 H, d, J 6.4 Hz, H21), 3.803 (1 H, tt, J 11 Hz, J 4.7 Hz, H3α), 5.159 (1 H, s, ω 6.5, H15).
Standard assay for recombinant maize 8,7SI in yeast microsomes
Yeast microsomes were prepared as previously described . Microsomes (0.4 ml, 1.0 mg of protein) were incubated in the presence of exogenous Δ8-sitostenol (Figure 2, compound 1) (20–150 μM) emulsified with Tween 80 (final concentration 1.5 g/l). Incubations were continued aerobically at 30 °C with gentle stirring for 90 min. During this period the progression of the reaction was linear. The reaction was stopped by adding 1 ml of 6% (w/v) KOH-ethanol. Sterols were extracted three times with a total volume of 15 ml of n-hexane and, after drying with Na2SO4, the extract was concentrated to dryness. The extracts were further analysed by TLC on silica gel, using dichloromethane as the eluant (developed twice). The 4-desmethylsterols (Rf=0.30) were separated from 4α-methylsterols (Rf=0.40) and from 4,4-dimethylsterols (Rf=0.45). After elution from the silica gel, an aliquot of the 4-desmethyl fraction was analysed by GLC. The residual substrate Δ8-sitostenol (Figure 2, compound 1) (tR=1.233) and the product formed, Δ7-sitostenol (Figure 2, compound 8) (tR=1.265) were readily separated from each other and from the bulk of endogenous 4-desmethylsterols, including ergosterol (tR=1.083) (Supplementary Figure S1 at http://www.BiochemJ.org/bj/414/bj4140247add.htm). The Δ7 metabolite (Figure 2, compound 8) produced by the reaction was unequivocally identifed by its retention time on GC and by an electron-impact mass spectrum identical with that of an authentic standard (Table 2). The conversion ratio was calculated from the areas of the peaks of Δ7-sitostenol (Figure 2, compound 8) and Δ8-sitostenol (Figure 2, compound 1) and corrected with the values obtained in the corresponding assay using boiled microsomes. The rate of substrate isomerization was calculated from the conversion ratio of compound 1 into compound 8, and the concentration of substrate used in the reaction. Apparent maximum velocity (Vmax) and Km values were determined by fitting the data to the Michaelis–Menten equation using the nonlinear regression program DNRP-EASY derived by Duggleby  from DNRP53.
In the case of inhibition assays, microsomes were incubated for 90 min at 30 °C in the presence of compound 1 (100 μM) and a range of concentration of inhibitors (0.01–50 μM) from which the dose–response curves were obtained allowing the corresponding IC50 values to be determined. The deviation between two determinations did not exceed 15%.
Previous studies revealed that inhibition of post-squalene enzymes of sterol biosynthesis by ammonium analogues of carbocationic HEIs follows a non-competitive kinetic inhibition pattern as a consequence of a slow rate of dissociation of the enzyme–inhibitor complex [10,36]. Assuming that inhibition of 8,7SI by tertiary ammonium derivatives follows such a non-competitive kinetic inhibition pattern, the IC50 values measured should be independent of the substrate concentration as well as of the Δ8-sterol substrate used.
Membrane proteins were quantified using the Bio-Rad protein assay according to the method of Bradford .
Western blot analysis of microsomes (40 μg of protein) or of 100000 g supernatants from ZmFLAG8,7SIs-transformed yeast were achieved after separation of proteins on SDS/PAGE (14% gel). After electrophoretic transfer on to PVDF Immobilon P membrane (Millipore), ZmFLAG8,7SI was immunoblotted with affinity purified murine monoclonal anti-FLAG M2 antibodies from Sigma (1:6000 dilution) according to the manufacturer's instructions. A goat anti-mouse IgG-AP (alkaline phosphatase) conjugate (Bio-Rad) was used as a secondary antibody (1:10000 dilution). The membrane was then treated with the chemiluminescent AP substrate and the blot was further used to expose an instant film for detection.
Incubation of substrate analogues and identification of enzyme-generated products
The apparent Km and Vmax of analogues (1), (2) and (3) were determined by incubating them for 60–90 min at 30 °C under standard assay conditions in a yeast microsomal preparation of recombinant Zm8,7SI. The concentration of substrate was 40–150 μM. In the case of compounds (3) and (6), the 4α-methylsterol fraction was analysed by GC-MS. In the case of Δ8-sterols (4) and (5), the 4,4-dimethylsterol fraction was analysed by GC-MS. The 8,7SI products were unequivocally identified by their retention times and an electron-impact spectrum identical with that of authentic standards (Table 2). The extracts of incubations of Δ8-sterols (4), (5), (6) and (7) and of the corresponding controls performed with inactivated microsomes revealed the absence of isomerization product and complete recovery of these compounds, which was confirmed by ion monitoring that corresponded to the mass of the substrate and expected Δ7-sterol product.
Multiple amino-acid sequences were aligned with ClustalW algorithm.
Characterization of 8,7SI from Z. mays
Cloning of 8,7SI in Z. mays and sequence analysis
We identified a maize EST clone encoding a protein presenting 54% identity with the previously identified 8,7SI from A. thaliana  and 36% identity with the H. sapiens EBP protein . However, the predicted protein showed only 15% identity with the ERG2 protein from S. cerevisiae  (Figure 3). Analysis of the Zm8,7SI protein sequence indicated the presence of four transmembrane spanning domains similar in order, spacing and length to those found in murine 8,7SI  but clearly distinct from the three hydrophobic domain model observed for the yeast 8,7SI. In addition, Zm8,7SI possesses a C-terminal ER (endoplasmic reticulum)-retrieval signal KKXK.
Zm8,7SI can complement the erg2 strain deficient in 8,7SI
To further characterize the function of the cloned full-length Zm8,7SI, we performed a yeast complementation assay in the ERG2-deficient strain lacking the 8,7SI activity necessary to synthesize ergosterol. The Zm8,7SI ORF was cloned into the pVT102U shuttle vector under the control of the constitutive ADH promotor. Several pVT transformants were picked from the selection plate and propagated in liquid medium. After sterol extraction, the sterol profiles were analysed by GC and GC-MS. The strains erg2-pVT-Zm8,7SI and erg2-pVT-ZmFLAG8,7SI accumulated ergosterol (44–57%) as the major sterol, several Δ7-sterols (8–10%) and residual amounts (26–37%) of Δ8-sterols (Table 3). In comparison, the erg2-pVT-Void (without insert) control strain produced exclusively Δ8-sterols (Table 3). These results demonstrate that Zm8,7SI and ZmFLAG8,7SI can efficiently complement the erg2 strain and restore endogenous ergosterol synthesis.
Zm8,7SI has 8,7SI activity in vitro
An enzymatic assay was performed to test whether the recombinant putative Zm8,7SI protein in the transformed erg2 strain indeed possesses 8,7SI activity. Purification of plant membrane proteins for functional analysis after expression in yeast is still a relatively unexplored field with little documentation in the literature. Additionally, in the case of an enzyme which is membrane-bound, interactions with other components of the membrane may be necessary for optimum enzymatic activity. Thus 8,7SI activity was assayed in the microsome extracts prepared from erg2-pVT-Void, erg2-pVT-Zm8,7SI and erg2-pVT-ZmFLAG8,7SI by using the standard assay conditions for recombinant 8,7SI described in the Experimental section. The results from these studies revealed that microsome extracts obtained from erg2-pVT-Zm8,7SI and erg2-pVT-ZmFLAG8,7SI were able to isomerize the 8,7SI substrate, Δ8-sitostenol (Figure 2, compound 1), to produce a single Δ7-sterol metabolite, Δ7-sitostenol (Figure 2, compound 8) which was unequivocally identified by GC-MS analysis (Table 2 and Supplementary Figure S1). 8,7SI activity was undetectable in reactions with microsomal extract of erg2-pVT-Void. The observed catalytic competence of Zm8,7SI revealed that it indeed encodes a membrane-bound maize 8,7SI.
Substrate specificity of Zm8,7SI
A series of Δ8-sterols with distinct nucleus or side-chain structures were assayed with the recombinant Zm8,7SI (Table 4). The apparent kinetic parameters of Zm8,7SI with substrates Δ8-sitostenol (Figure 2, compound 1), zymosterol (Figure 2, compound 2) and 4α-methylfecosterol (Figure 2, compound 3) were determined by varying their concentrations under our standard assay conditions. The velocity/substrate concentration curves obey simple Michaelis–Menten kinetics with respect to compounds (1), (2) and (3) (Supplementary Figure S2 at http://www.BiochemJ.org/bj/414/bj4140247add.htm) and the kinetics data obtained are summarized in Table 4. The results indicate that Δ8-sterol (compound 1), and compound (2) without the methyl group at C4β and C14α were productive substrates and that compound (3), possessing a single C4α-methyl substituent, had the highest apparent relative specificity constant (Vmax/Km) in the series and thus appeared to be the preferred substrate of the isomerase. In contrast, the presence of an additional C4β-methyl group in substrates (4) and (5) abolished activity as did the addition of a C14-methyl group as in substrate (6). Although the enzyme can catalyse the isomerization of various Δ8-sterols, the Δ8(14)-monoene analogue (compound 7) is not isomerized, indicating the strong regioselectivity of the isomerase for the substrate double bond localization. Finally, the terminal part of the side chain was not essential to activity. As shown in Table 4, absence of a substituent at C24 (compound 2), or addition of one or two carbon atom substituents at C24, compounds (3) and (1), were not structural deterrents for the 8,7SI activity.
Inhibition of 8,7SI by lipophilic tertiary ammonium derivatives
We first synthesized in our laboratory a series of rationally designed carbocationic HEI analogues, including compounds (17) and (18), possessing a nitrogen atom with a steady positive charge in place of carbenium-C, which could simulate the C8-HEI [10,15]. With this mechanism in mind, inhibition of the recombinant Zm8,7SI activity by a series of lipophilic tertiary ammonium derivatives susceptible to interact with its active site, and possessing a variety of structures and biological properties, have been examined. This included the agronomical N-substituted morpholine fungicides tridemorph (Figure 2, compound 11) and fenpropimorph (Figure 2, compound 12) [13,14], the fungicide AY-9944 (Figure 2, compound 13) , the human receptor σ ligand SR31747A (Figure 2, compound 14) , the high-affinity human EBP ligand trifluoperazine (Figure 2, compound 15) , the oestrogen receptor modulator tamoxifen (Figure 2, compound 16) , and the rationally designed 8-aza-decalins (Figure 2, compounds 17, 18 and 19). Table 5 provides a summary of the dose–response curves obtained for in vitro inhibition of the recombinant plant 8,7SI by these compounds, allowing the corresponding IC50 values to be determined. In addition, we similarly determined IC50 values for a number of these compounds for the yeast and murine 8,7SIs. The results of the present study and those found in the literature for in vitro inhibition of yeast and animal 8,7SIs are shown in Table 5.
By examining first at pH 7.5 the cationic azadecalin (compound 17) and the corresponding neutral amide analogue (compound 19), yielding IC50 values of 0.85 μM and more than 100 μM respectively, it was confirmed that the HEI analogues function with a charged ammonium group. Six ammonium derivatives were potent inhibitors of Zm8,7SI with IC50 values in the range 0.1–2.0 μM and IC50/Km (compound 1 or 3)=(2.10×10−4)–(2.10×10−2). Two ammonium derivatives (compounds 15 and 16) with a higher molecular mass and less structural flexibility were more poorly accommodated, yielding IC50 values of 8.5–10 μM. The lower inhibition by these two ammonium derivatives indicates that the binding site has a limited tolerance in terms of size and flexibility of the lipophilic domain of the HEI analogues.
Our in vitro assay revealed the potent inhibition of the plant 8,7SI by the σ-ligand SR31747 (compound 14) (IC50=0.1 μM) which is comparable with the inhibition of the recombinant yeast (IC50=0.6 μM)  and mammalian 8,7SI (IC50=0.40 μM) by this derivative (Table 5). These results indicate that the diverse 8,7SIs contain a high-affinity binding site for the σ-ligand SR31747. In addition, the observed affinity of compound (14) is in the same order as that measured for the Zm8,7SI with HEI analogues (compounds 17 and 18) (IC50=0.85 and 0.10 μM respectively), and compound (18) is a good inhibitor of the mammalian (IC50=10 μM)  and yeast 8,7SI (IC50=0.2 μM). Moreover, we observed a similar affinity of trifluoperazine (compound 15) for the plant 8,7SI (IC50=10 μM) and for the human EBP (IC50=7 μM) , whereas it is a poor inhibitor of yeast 8,7SI (IC50>500 μM).
Requirement of acidic amino acid residues for Zm8,7SI activity
Multiple sequence alignment of 8,7SI across seven species (Z. mays, Oryza sativa, A. thaliana, H. sapiens, Mus musculus, Rattus norvegicus and S. cerevisiae; Figure 3) revealed that across all seven species, 8,7SIs show identity for a variety of amino acids. In addition, the plant and animal 8,7SIs show additional identities which are not shared with the yeast enzyme.
Based on the catalytic mechanism of 8,7SI, good candidates for both the proton donor and the proton acceptor would be acidic amino acids as previously suggested . To identify amino acid residues that might be critically important for the proton delivery and (or) abstraction in the plant sterol isomerase, we mutated a number of acidic residues conserved in the 8,7SIs between plants, animals and yeast, or conserved only between plants and animals, so that they could no longer serve as proton-donating or -accepting residues. In addition, we also mutated a conserved histidine residue which could function as a proton-donating or -accepting residue during the isomerization reaction. Moreover, previous studies from our laboratory emphasized the importance of electrostatic interactions during the binding of carbocationic HEI analogue inhibitors to the catalytic site . The nature of the amino acid residue of the enzyme interacting with these inhibitors is not known, but it could be a delocalized carboxylate anion such as the two aspartate and glutamate residues which have been shown to be obligatory for ammonium derivative binding of the σ1 receptor .
In order to have minimal effects on secondary structure, we chose to mutate the different amino acids by sterically conservative hydrophobic and electrically neutral residues. Thus we neutralized residues (i) His74, Glu78 and Glu121 which are conserved in animals, plants and yeast and (ii) Glu102, Asp107, Asp114 and Asp170 which are conserved only in animals and plants. The mutants proteins (H74L, E78V, E102V, D107V, D114V, E121V and D170V) were expressed in erg2 null mutants and the mutated isomerases were assayed in vivo by sterol analysis and quantification of Δ5,7-sterols. In addition, the activities of the recombinant plant 8,7SIs were examined directly in vitro in the corresponding yeast microsomal preparations.
The effects of these mutations on in vivo and in vitro activities of Zm8,7SI are shown in Table 6. In four out of the seven mutants, H74L, E78V, D107V and E121V, replacement of the proton-delivering residue with a leucine or valine residue totally eliminated the activity both in vivo, since neither Δ5,7- nor Δ7-sterols were detected, and in vitro, since no 8,7SI activity could be detected (Table 6), indicating that these residues were essential for the enzyme activity. The other mutants, D114V and D170V, exhibited similar sterol profiles as the wild-type and an in vitro activity that was approx. 60% of the activity of the wild-type enzyme. Therefore these residues are not essential for the catalysis but contribute to the activity through conformational or other effects.
For the mutants that failed to complement it was conceivable that conservation of functionality and charge at these positions might be sufficient for 8,7SI activity. We explored this possibility in one case, Glu78, and changed this residue to a number of other residues such that it could either no longer serve as a general acid or its pK would be significantly perturbed. We constructed five additional mutants: mutant E78D, E78H, E78R, E78K and E78Q. None of these mutants complemented erg2, suggesting that both functionality and length of the side chain of residue 78 are critical for 8,7SI activity.
Involvement of critical tryptophan residues in Zm8,7SI activity
Electron-rich aromatic amino acids have been suggested in general terms as particular stabilizers of intermediate carbocations . The crystal structure of the hopene cyclase indeed revealed that several tryptophan residues and phenylalanine residues were well positioned to stabilize different cations of the cyclization cascade . Thus we mutated (i) Trp183 which is mostly conserved in animals, plants and yeast, (ii) Trp66, Trp100 and Trp193 which are conserved only in animals and plants and (iii) Trp67 and Trp205, which are not conserved, to leucine residues.
In two out of the six mutants, W66L and W193L, replacement of the tryptophan residue with a leucine residue totally eliminated the activity both in vivo and in vitro (Table 6) indicating that these residues were essential for the enzyme activity. The mutants, W67L, W100L, W183L and W205L, exhibited similar sterol profiles as the wild-type and an in vitro activity that was 60–100% of the activity of the wild-type enzyme. Therefore these residues are not essential for the catalysis but contribute to the activity through conformational or other effects.
An hydroxy function at residue 124 is needed for maximal binding of the enzyme–substrate complex
We looked for conserved residues that could be involved in the binding of the 3β-hydroxy group of the sterol substrate at the end of the binding pocket. Some of the aforementioned carboxy residues could form an hydrogen bond with the 3β-hydroxy group of the sterol substrate. For example, the oestrogen receptor donates a hydrogen bond to a glutamate residue . Moreover, in Mycobacterium tuberculosis sterol 14α-demethylase, an aspartate residue is hypothesized to form a hydrogen bond with the 3β-hydroxy group . However, in Candida albicans sterol 14α-demethylase, a threonine residue was predicted to form a hydrogen bond with the 3-hydroxy group of the sterol substrate and helped to locate it in the active site . Thus we mutated the hydroxy-containing residue Thr124 conserved in all organisms (Figure 3).
Although replacement of Thr124 by a functionally conservative serine residue in mutant T124S led to a sterol profile similar to the wild-type, its replacement by an hydrophobic isoleucine residue in mutant T124I led to a strong decrease in Δ5,7-sterol content (6%), absence of Δ7-sterols and high accumulation of Δ8-sterols (Table 6). Because this mutation led in vivo to a sterol profile clearly dictinct from both those of wild-type-like or inactive mutants, we characterized more precisely its enzyme kinetic properties in vitro. The fact that the Km for mutant T124I, is increased 25-fold (3.03 mM) indicates that Thr124 contributes to the stabilization of the enzyme–substrate complex in the ground state. In contrast, Vmax/Km was decreased approx. only 3-fold, indicating that Thr124 has a minor role in stabilizing the transition state of a rate-controlling step of the 8,7SI reaction, which is much less destabilized in T124I. The result is a substantial decrease in the activation energy to reach the transition state thus leading to a significant improvement in the isomerase maximum rate  (101 nmol·mg·h−1) in mutant T124I.
Functional diversity between the plant and yeast 8,7SIs
To know whether the sterol 8,7-isomerization reaction in plants and yeast would involve identical essential amino acid residues, in Sc8,7SI we mutated conserved residues corresponding to amino acid residues found to be essential or important for the Zm8,7SI activity. Thus yeast mutants H69L, E73V, E116V and T119I, corresponding respectively to His74, Glu78, Glu121 and Thr124 in Zm8,7SI, were constructed and analysed as described above. Mutants H69L, E73V and E116V exhibited similar sterol profiles as the wild-type yeast 8,7SI (Table 6). Therefore, in contrast with Zm8,7SI and animal 8,7SI, acidic amino acid residues at these positions, as well as at position 102 corresponding to Asp107 in Zm8,7SI, are not essential for the yeast 8,7SI catalytic activity (Table 7). In addition, Table 7 also shows that two essential Zm8,7SI tryptophan residues (Trp66 and Trp193) were also not conserved in Sc8,7SI.
Expression of mutated Zm8,7SIs
Because it is an integral membrane-bound protein, Zm8,7SI has not yet been purified allowing production of antibodies to examine its expression levels in the microsome extracts used herein. However, two mutational studies of membrane-bound enzymes of the post-squalene sterol synthesis, performed in similar yeast expression systems as the present study (erg2 and erg6 mutants), indicated no major differences in the expression level of the different mutants and wild-type enzymes [24,50]. To express our different 8,7SI mutants in the yeast erg2 strain we used a multicopy plasmid containing a strong promoter (yeast ADH1 promoter). These conditions should favour the similar expression of these mutants. To check protein accumulation from wild-type and mutated Zm8,7-SI cDNAs, for a number of them we constructed the corresponding 8,7SI proteins fused to an N-terminal FLAG epitope (ZmFLAG8,7SI). Because of the amount of work needed to develop FLAG constructs corresponding to all mutations, we developed FLAG constructs only for a sub-set of essential (D107V) and non-essential (D114V, D170V) acidic residues, as well as for essential (W193L) and non-essential (W205L) tryptophan residues in addition to the wild-type enzyme. The microsome extracts from the corresponding transformed yeast cells were subjected to SDS/PAGE followed by Western blot analysis using commercial anti-FLAG serum (Figure 4). There is evidence that all mutated and wild-type FLAG8,7SI proteins did accumulate. However, there appear to be some differences in the level of accumulation which could be due to the yeast culture conditions which were not optimized for heterologous protein production.
Considering these results together, it seems unlikely that the lack of activity in the mutated 8,7SIs would result from lack of expression or accumulation in the yeast membranes.
The predicted amino acid sequence of the present Zm8,7SI confirms that the plant 8,7SIs, including the Arabidopsis isomerase , are much more closely structurally related to the animal 8,7SI protein  than to the ERG2 protein from S. cerevisiae .
In vitro kinetics with a variety of potential substrates revealed that 4α-methylsterol (Figure 2, compound 3) has the highest specificity constant (Vmax/Km) and thus is the preferred substrate of the recombinant maize 8,7SI. The results are in agreement with previous in vivo results obtained in plant cells treated with the 8,7SI inhibitor AY9944, indicating an accumulation of 4α-methyl-5α-ergosta-8,24(241)-dien-3β-ol (Figure 2, compound 3) and 4α-methyl-5α-stigmasta-8,24(241)-dien-3β-ol in addition to a variety of C4-demethylated-Δ8-sterols [1,31]. In contrast, substrate-specificity studies with membrane-bound rat liver isomerase indicated that the isomerization occurs primarily after complete nuclear demethylation in the formation of cholesterol from lanosterol . In addition, fecosterol (24-methylene-Δ8-cholestenol) was shown to be an isomerase substrate using yeast extracts . Such unique substrate specificity of the plant enzyme is unusual when data for the human and fungal enzymes are compared. The distinct substrate specificities observed for the plant as opposed to human/fungal 8,7SIs presumably reflects the early divergence of the sterol biosynthetic pathway in the plant kingdom and underlines the molecular diversity of 8,7SI substrates in plants, yeast and animals.
The potent inhibition of maize 8,7SI by the σ-ligand SR31747A is in good agreement with the previously reported inhibition by other σ-ligands of the production of ergosterol in an erg2 mutant complemented with an Arabidopsis 8,7SI cDNA . Incidentally, drugs such as the σ-ligand SR31747A which inhibit the activity of the σ receptor and yeast 8,7SI , cause defects in a variety of cellular processes including immunosuppressive effects such as inhibition of graft rejection or lymphocyte proliferation . As shown particularly for Zm8,7SI, the striking ability of 8,7SIs from different sources to bind a variety of structurally distinct lipophilic tertiary ammonium derivatives with various biological properties is remarkable. It is worth noting that SR31747 (Figure 2, compound 14) is structurally not related to the sterol substrate nor to the rationally designed HEI analogues (compounds 17 and 18), although the protonated form of its tertiary amine function presumably interacts with the active-site domain that stabilizes the putative HEI. The results underline the primary importance of electrostatic interactions in the binding of such tertiary ammonium derivatives to the isomerase. Accordingly, human 8,7SI (human EBP), has been shown to be able to bind a variety of structurally distinct drugs, suggesting an intimate pharmacological relationship among EBP, Erg2p and σ1-receptor . As previously suggested for human 8,7SI , the propensity to bind structurally distinct compounds could also be related to the presence of a sterol-binding site in the protein. Indeed, plants contain proteins able to bind non-specifically a variety of lipid derivatives including fatty acids and sterols , and the multidrug-resistance protein involved in extrusion of xenobiotics takes part in cholesterol biosynthesis .
In the present study, our in vivo and in vitro mutational analysis of recombinant 8,7SIs clearly show that acidic residues Glu78, Asp107, Glu121 and His74 are essential for Zm8,7SI enzymatic activity, whereas no essential acidic residue was found in the homologous positions in Sc8,7SI. In contrast, the corresponding EBP amino acid residues Glu81, Asp108, Glu122 and His77 have been shown to be also critical for the in vivo 8,7SI activity of the human EBP, albeit the mutations were not tested directly, and in vitro for the enzymatic activity . In the plant and animal 8,7SIs, Asp107 lies within a sequence box that is completely conserved in a cytoplasmic domain. In contrast His74, Glu78 and Glu121 are located into two putative transmembrane segments. His74, Glu78 and Glu121 could include the two obligatory distinct residues that deliver and receive a proton during the trans addition–elimination of hydrogen atoms during the plant and animal 8,7SIs catalytic process. Incidentally, it has been demonstrated that the initial protonation of the carbon–carbon bond in the squalene hopene cyclase involves a carboxylic acid .
For the mutants that failed to complement, we could show in the case of residue Glu78 that the functionality as well as the length of the side-chain residue are critical for 8,7SI activity. The aforementioned detailed catalytic mechanism of the 8,7SI suggests a precise organization within the Michaelis complex to preferentially add a proton on a C8(9) double bond rather than on a C14 double bond, and to rigidly control the stereochemistry of the proton abstraction at C7. It would be conceivable to disturb the precise molecular interactions of the enzyme–substrate complex by changes in the nature or the distance of the catalytical amino-acid residues to produce an inactive enzyme.
The results of the present study clearly show that tryptophan residues Trp66 and Trp193 play crucial roles in Zm8,7SI, either for the correct folding of the protein, or for the catalytic process. In this latter case they could possibly be involved in the stabilization of the C8 carbocationic intermediate. In accordance with this hypothesis, Trp66 and Trp193 are located in two putative transmembrane segments. Substitution of the homologous residues in the human 8,7SI by an alanine residue was also shown to reduce dramatically the amount of Δ5,7-sterols in the corresponding yeast transformants compared with the wild-type isomerase . However, no aromatic residue was found in the homologous positions in the yeast isomerase (Table 7), underlying the functional diversity of the different 8,7SIs.
Suppression of the hydroxy function at residue 124 apparently decreased more substantially the stability of the enzyme–substrate complex in the ground state than in the transition-state. The data could reflect the formation of a hydrogen bond between Thr124 and the 3-hydroxy group of the sterol, helping to dock the sterol in the active site in the ground-state as previously proposed in the case of the C. albicans sterol 14α-demethylase . This residue could have a similar function in the different 8,7SIs since it is important or critical in all of them (Table 7). Incidentally, a more important role for the interaction of the 3β-hydroxy group with a specific amino acid for initial substrate binding than for transition-state stabilization would be in accordance with previous studies showing that a number of structural features that are critical for binding of the sterol substrates in the ground state (such as the 3β-hydroxy group) play a minor role for binding of transition-state analogues .
The complementing mutations obtained herein did not alter the sensitivity of the transformed yeast strain erg2 to the inhibitors azadecaline (Figure 2, compound 18) and SR31747 (Figure 2, compound 14), and they showed a similar sensitivity to nystatin as the wild-type, as observed by spotting the corresponding cells on medium with or without these inhibitors. In contrast, the non-complementing mutants affected in essential amino acids were all insensitive to nystatin and, in addition, exhibited a significant increase in sensitivity to compounds 18 and 14 (EC50 decreased 5- to 10-fold; results not shown). It has been previously shown that ERG2 is not an essential gene, in contrast with ERG24 coding for the sterol C14 reductase, and that overexpression of ERG2 does not lead to fenpropimorph resistance, while overexpression of ERG24 does . The increased sensitivity of non-complementing mutants to compounds 18 and 14 could thus reflect a higher concentration of these inhibitors available for blockade of endogenous C14 reductase in the yeast cells because of the absence of binding to the mutated Zm8,7SI. These results would be consistent with the hypothesis that mutations found to be critical for Zm8,7SI activity also lower the binding of these lipophilic tertiary ammonium derivatives to the enzyme. This hypothesis of an intimate structural relationship between the catalytic site and the binding domain of the inhibitor is in accordance with their function as putative HEI analogues and would need a combination of binding assay studies of a radiolabelled inhibitor with the present site-directed mutagenesis study.
Biochemical analysis of A. thaliana hyd1 mutant strongly affected in embryonic development [26,55], revealed deficiency in 8,7SI activity. The hyd1-E508 allele was found to encode a D102N substitution in 8,7SI that corresponds to the D107V mutation in Z. mays, that we found to also be essential for the maize isomerase. All of the essential amino acids found in the present work are conserved between Z. mays and other plant 8,7SIs including A. thaliana 8,7SI. Thus we think that the conclusions obtained with the present maize mutants could be extrapolated with confidence, particularly to the Arabidopsis isomerase, and the data used for appropriate Arabidopsis plant mutants for further characterization.
This first series of mutants of plant sterol 8,7SI in the present study has allowed the identification of six essential amino acid residues for plant 8,7SI activity by in vivo complementation of the yeast erg2 mutant and corroboration of the results by direct in vitro measurement of the 8,7SI activity in the corresponding microsome extracts. The absence of any structural information for this membrane-associated enzyme limits our ability to verify the hypotheses about the specific role played by the essential amino acid residues identified herein. However, it is likely that they are located at the substrate-binding domain of the active-site of the enzyme.
Remarkably, the critical residues identified in the plant 8,7SI are conserved and also crucial in the animal 8,7SI, whereas they are either not essential or not conserved in the yeast 8,7SI. These results underline the functional homology between the plant and animal 8,7SIs, and their high divergence with the yeast 8,7SI, in accordance with their structural diversity and the distinct stereochemistry of their mechanisms. Moreover, enzymes involved in post-squalene sterol biosynthesis in higher plants generally share amino acid identity ranging from 28% to 38% with their corresponding S. cerevisiae counterparts . In contrast, with less than 15% identity, clear distinct substrate specificities and non-conserved essential amino acid residues, it appears that the 8,7-isomerization step is performed by completely different enzymes in higher plants and animals on one hand and in S. cerevisiae and probably most fungi on the other hand. This finding, however is not unprecedented in the pathway since no orthologues of sterol C24(241) reductase isomerase from A. thaliana have been reported in the S. cerevisiae genome [1,56]. In addition to the present study, further structural information about the active site of the fungal, plant and mammalian 8,7SI could provide a basis for rational design of more efficacious and specific antifungal agents in addition to a better insight into the molecular mechanism of 8,7SI.
We are indebted to Dr M. Bard (Department for Biology, Indiana University, Indianapolis, IN, U.S.A.) for providing the yeast mutant erg2. We thank Dr Geneviève Genot for her help in Western blot analysis. We are grateful to Dr G. Loison (Sanofi-Aventis Recherche-Developpement, Labège, France) for providing SR31747 and the yeast strain Δerg2pEMR1235. We acknowledge Dr Marc Fischer for help in editing the manuscript prior to submission.
Abbreviations: ADH1, alcohol dehydrogenase 1; EBP, emopamil-binding protein; EST, expressed sequence tag; HEI, high-energy intermediate; i.d., internal diameter; ORF, open reading frame; 8,7SI, sterol 8,7-isomerase; Sc8,7SI, Saccharomyces cerevisiae 8,7SI; Zm8,7SI, Zea mays 8,7SI
- © The Authors Journal compilation © 2008 Biochemical Society