- Email: [email protected]
Induction of cytochrome P4501A (CYP1A) by clotrimazole, a non-planar aromatic compound. Computational studies on structural features of clotrimazole and related imidazole derivatives
Life Sciences 76 (2004) 699 – 714 www.elsevier.com/locate/lifescie
Induction of cytochrome P4501A (CYP1A) by clotrimazole, a non-planar aromatic compound. Computational studies on structural features of clotrimazole and related imidazole derivatives Jose´ Marı´a Navasa,*, Antonio Chanab, Bernardo Herrado´nb, Helmut Segnera,1 a
Department of Chemical Ecotoxicology, UFZ Center for Environmental Research, Permoserstr. 15, D-04318 Leipzig, Germany b Instituto de Quı´mica Orga´nica General, C.S.I.C., Juan de la Cierva 3, 28006 Madrid, Spain Received 17 February 2004; accepted 9 September 2004
Abstract The classical pathway for induction of cytochrome P4501A (CYP1A) by xenobiotics is ligand binding to the aryl hydrocarbon receptor (AhR). High-affinity AhR ligands are planar polyaromatic molecules such as the prototypic ligand, 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD). The present work investigated the ability of the imidazole derivative, clotrimazole [1-(2Vchlorotrityl)imidazole, CLO], to induce CYP1A in cultured rainbow trout (Oncorhynchus mykiss) hepatocytes at the catalytic activity (determined as 7-ethoxyresorufin-O-deethylase, EROD) and at the transcriptional level. CLO resulted in a significant increase of hepatocyte EROD activity and CYP1A mRNA at a concentration of 1.56 AM. Computational studies on the molecular structure of CLO show that CLO is unlikely to take a planar conformation. Further indications that CLO does not behave like a planar AhR ligand come from the experimental observation that co-incubation of trout hepatocytes with CLO and the AhR antagonist, a-naphthoflavone (a-NF), did not result in an inhibition of CLO induction of CYP1A mRNA, whereas a-NF was able to inhibit CYP1A induction by the prototpyic, planar AhR ligand, h-naphthoflavone. The experimental findings on CLO agree with previous results obtained for another non-planar imidazole derivative, 1benzylimidazole (BIM). Further, computational studies showed that the non-planar imidazoles, BIM and CLO, are highly similar with respect to some electrostatic properties, namely the dipole moment and the molecular
* Corrresponding author. Present address: INIA, Department of the Environment, Carretera de la Corun˜a, Km 7.5, 28040 Madrid, Spain. Tel.: +34 91 3 47 41 55; fax: +34 91 3 57 22 93. E-mail address: [email protected] (J.M. Navas). 1 Present address: Centre for Fish and Wildlife Health, Institute of Animal Pathology, University of Bern, L7nggass-Str. 122, CH-3012 Bern, Switzerland. 0024-3205/$ - see front matter D 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.lfs.2004.09.015
700
J.M. Navas et al. / Life Sciences 76 (2004) 699–714
electrostatic potential (MEP). Overall our experimental and computational studies suggest that transcriptional activation of CYP1A by the imidazole derivatives CLO and BIM is mediated by a mechanism different to that of prototypic CYP1A inducers such as the planar AhR-ligands. D 2004 Elsevier Inc. All rights reserved. Keywords: Aryl hydrocarbon receptor; EROD; Cytochrome P450 1A; Clotrimazole; Imidazole
Introduction The structural properties underlying the interaction of chemicals with biological molecules are an important topic of current research. The biological effects of numerous xenobiotics are mediated through binding to the arylhydrocarbon receptor (AhR) and the subsequent activation of AhRregulated genes such as the xenobiotic-metabolizing enzyme cytochrome P4501A (CYP1A) (Safe, 1990; Denison and Whitlock, 1995; Hahn, 1998). Ligands bind to the AhR in the cytoplasm of the target cell, thereafter the ligand-activated receptor undergoes a transformation process whereby it forms a heterodimer with the AhR nuclear translocator protein (ARNT), that is transported into the nucleus. The nuclear receptor complex then interacts with specific xenobiotic-responsive elements (XREs) located upstream of the CYP1A gene, leading to stimulated transcription of the gene, elevated CYP1A mRNA, and increased levels of CYP1A protein and its catalytic activity (Hahn, 1998; Rowlands and Gustafsson, 1997). Typical ligands for the AhR are hydrophobic aromatic compounds with at least two aromatic rings (either carbocyclic or heterocyclic) that are planar or can become planar, i.e. both aromatic rings are in the same plane with a dihedral angle of 08 between both rings (Hankinson, 1995; Safe, 1995). 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD, Fig. 1) is the most potent agonist of AhR found up to now, with a KD in the picomolar range (Rowlands and Gustafsson, 1997). High affinity ligands include certain polychlorinated biphenyls (PCBs, see Fig. 1), polyaromatic hydrocarbons (PAHs, i.e., h-naphthoflavone, h-NF), aromatic amines, indolocarbazoles and related compounds. A
Fig. 1. Chemical structures of a-naphthoflavone (a-NF), an AhR antagonist, and of some AhR activators: 2,3,7,8tetrachlorodibenzo-p-dioxin (TCDD), h-naphthoflavone (h-NF), 3,3V,4,4V,5-pentachlorobiphenyl (PCB no 126), 3,3V,4,4V,5,5V-hexachlorobiphenyl (PCB no 169), 1-benzylimidazole (BIM), and clotrimazole (CLO).
J.M. Navas et al. / Life Sciences 76 (2004) 699–714
701
comparative molecular field analysis (CoMFA) using a wider range of ligand types (dibenzo-pdioxins, dibenzofurans, biphenyls, naphthalenes and indolocarbazoles) has predicted a single ligandbinding pocket of 14 12 5 2, which would accommodate all the known ligands (Waller and McKinney, 1995). The importance of planarity for AhR binding and CYP1A induction was exemplified for polychlorinated biphenyls (PCBs), where congeners with para- or meta-substitution, which easily can take a co-planar configuration, are potent CYP1A inducers, but ortho-substituted congeners for which a co-planar configuration is energetically unfavorable, are weak or noninducers of CYP1A (Safe, 1990). However, this statement has been questioned by several authors who determined, by computational studies, that some PCBs acting as AhR ligands or CYP1A inducers were not planar molecules (Waller and McKinney, 1995; Kafafi et al., 1993; Chana et al., 2002). The objective of this work was to study the properties of the imidazole derivative clotrimazole (CLO, Fig. 1) as a CYP1A inducer. Imidazoles are a large group of compounds frequently used as antifungal agents, both clinically as well as in agriculture. Their antifungal activity is due to their capacity to inhibit cytochrome P450-mediated ergosterol synthesis (Henry and Sisler, 1984). These compounds can enter the aquatic environment and bioaccumulate in fish (Eglo et al., 1994; Castillo et al., 1997). Kobayashi et al. (1993) have reported that a variety of imidazole derivatives including CLO are inducers of CYP1A gene expression; and it has been hypothesized that this biological activity is due to the fact that these chemicals bind the AhR in a co-planar structure (i.e., the two aromatic rings are in the same plane, although the whole molecule is not planar). For a number of imidazole derivatives, however, the assumption that they can achieve a planar conformation is debatable from the point of view of structural organic chemistry, as we have previously demonstrated for 1-benzylimidazole (BIM, Fig. 1) (Navas et al., 2003). In order to further understand the mechanistic basis of CYP1A induction by imidazolic compounds, the present study explores AhR binding and CYP1A induction by CLO. Experimentally, we investigated the potency of CLO to induce CYP1A in cultured rainbow trout hepatocytes at both the enzyme catalytic and the transcriptional level, and we determined whether the AhR antagonist a-naphthoflavone (a-NF, Fig. 1) can block the CLO-related induction of CYP1A in the hepatocytes. Then, we performed computational studies to establish conformational preferences of the molecule, in particular its ability to take a planar configuration. Additionally, we have calculated electrostatic properties of CLO, more specifically the dipole moment and the molecular electrostatic potential (MEP). MEP is an indicator of the charge distribution of the molecule and has been employed to understand a variety of chemical and physical properties of molecules including binding or chemical reactivity, and thus has been used to rationalize the interaction between a biologically active molecule and its biomacromolecular target, e.g., the AhR (Chana et al., 2002; Murray and Sen, 1996).
Methods Animals Sexually immature male and female rainbow trout (250–350 g weight) from a local trout farm were maintained in 200 l steel tanks in the facilities of the Umweltforschungszentrum, (Centre for Environmental Research), Leipzig, Germany. Fish were held under artificial 12 hours
702
J.M. Navas et al. / Life Sciences 76 (2004) 699–714
Fig. 2. 7-Ethoxyresorufin-O-deethylase (EROD) activity induced by different AM concentrations of clotrimazole (CLO). Asterisks indicate significant differences with respect to the controls ( P b 0.05).
light/12 hours dark photoperiod in constant flow aerated water. Water temperature was maintained at 14–16 8C. Experimental design First, a dose-response curve was elaborated to establish the relationship between the concentration of CLO and the induction of EROD activity (Fig. 2). Data obtained after treatment of hepatocytes isolated from 3 different trout in three independent experiments were used at this point. Second, to compare the EROD induction of CLO, with that of a typical inducer like h-NF, four different experiments using hepatocytes from four trout were carried out. In these experiments, a single concentration of CLO (1.56 AM) and of h-NF (0.78 AM) were used (Fig. 3). Finally, to determine if the observed CYP1A induction was only provoked at the enzymatic level, or also at the transcriptional level, four independent
Fig. 3. 7-Ethoxyresorufin-O-deethylase (EROD) induction caused by 0.78 AM h-naphthoflavone (h-NF) and by 1.56 AM clotrimazole (CLO). Asterisks indicate significant differences with respect to the controls ( P b 0.05).
J.M. Navas et al. / Life Sciences 76 (2004) 699–714
703
Fig. 4. Induction of cytochrome P4501A (CYP1A) mRNA caused by 0.78 AM h-naphthoflavone (h-NF) and by 1.56 AM clotrimazole (CLO). 6.25 AM a-naphthoflavone (a-NF) did not provoke any reduction of the CYP1A mRNA levels detected after CLO treatment. Gels from one representative experiments of four are shown. On the right side appear data from the densitometric evaluation of the gels. Densitometric values for CYP1A were normalized with respect to densitometric values obtained with the housekeeping gene h-actin. The percentage of increase in this ratio obtained in the treatments with respect to the controls is represented.
experiments were performed (Fig. 4). The cells used in these four experiments were treated with 0.78 AM h-NF, or with 1.56 AM CLO, or received these chemicals together with a-NF 6.25 AM (a-NF is an AhR antagonist (Koley et al., 1997), so that the co-treatment of the cells with an AhR ligand and a-NF, should lead to an inhibition of the CYP1A induction caused by the ligand). In a previous work (Navas et al., 2003) we observed that another imidazole derivative, BIM, was able to induce CYP1A in rainbow trout cultured hepatocytes, although this compound was not planar, and that this induction was not inhibited by a-NF. Computational studies were performed on CLO and BIM to determine their most stable conformations as well as to assess the possibility that these compounds take co-planar structures; additionally some electrostatic properties were calculated with the aim to seek out structural similarities that could indicate similarities in the mechanism of action of both compounds. Xenobiotic treatments The chemicals h-NF, a-NF, and CLO were purchased from Sigma (St. Louis, MO, USA) at the highest purity available (95% for h-NF and a-NF, and 98% for CLO). Chemicals were dissolved in dimethyl sulfoxide (DMSO, Merck, Germany) and added to the culture media of hepatocytes to achieve the final desired concentration. Final organic solvent concentration in the assay was 0.1%. Hepatocyte isolation and culture Hepatocytes were isolated following a two-step perfusion technique as described by Braunbeck and Segner (2000). As culture medium, modified M199 medium (Sigma, St. Louis, MO, USA) supplemented with 2mM glutamine, 10 U/ml penicillin and 10 Ag/ml streptomycin was used. Medium was sterilized by a 0.22 Am filter. Cells were plated in 24-well FALCONR primary culture plates (Beckton Dickinson, Franklin Lakes, NJ, USA) precoated with Matrigel (BD Biosciences, Bedford, MA, USA) (0.1 mg protein ml1). Every well received 400 Al of the final suspension of cells (density: 2 105 cells cm2). The cells used for the extractions of mRNA in order to perform the RT-PCR analysis were plated on 60 mm diameter FALCONR tissue culture dishes and every plate received 5 ml of the
704
J.M. Navas et al. / Life Sciences 76 (2004) 699–714
cell suspension. During the first 24 hours of treatment the medium was supplemented with 5% fetal bovine serum (Sigma, St. Louis, MO, USA) to enhance the attachment of the cells. During exposure to the test compounds, serum free medium was used. The plates were maintained at 15 8C and 80% humidity. Twenty-four hours after plating, half of the medium was substituted with fresh medium containing the desired concentrations of CLO (0.19 AM to 3.12 AM, final concentrations in culture medium). The possible effect of a-NF on EROD activity was also observed by incubating the cells with a-NF concentrations ranging from 0.39 AM to 50 AM. Control wells received medium with solvent only. Treatments were applied in duplicate. At 48 and 72 hours after plating, half of the medium was replaced again and fresh medium containing the desired concentrations of CLO was added. 96 hours after plating (i.e., after 72 hours of treatment) the medium was removed, aliquoted, and stored at 80 8C until analysis of lactate dehydrogenase (LDH) activity. The cells were briefly washed with phosphate-buffered saline, pH 7.5, and the plates were frozen by using liquid nitrogen. They were maintained at -80 8C until analysis of EROD activity and protein contents. When the cells were co-treated with 6.25 AM a-NF and 1.6 AM CLO for RT-PCR, the following exposure regime was used: 24 hours after plating, the spent medium was removed and new medium with the desired concentrations of a-NF was added. After another 4 hours, half of the medium was removed and the same amount of fresh medium with the adequate concentrations of a-NF and CLO was added. In these experiments, cells were collected 48 hours after plating. The cells were briefly washed with phosphate-buffered saline, pH 7.5, and sampled by adding TRIZOL Reagent (Life Technologies, Gibco BRL, USA) to the plates to extract RNA (see below) that was stored at 80 8C. Cell viability assays Viability of the cells exposed to the test compounds was assessed by measuring the release of lactate dehydrogenase (LDH) into the extracellular medium according to the method of Denizeau and Marion (1990). Activity of LDH in the medium was determined following the oxidation of NADH at 340 nm in a spectrophotometer. Cell morphology was routinely observed using an inverted microscope. Protein content of the wells can also serve as an indicator of toxicity, and was routinely measured in parallel to EROD as described below. 7-Ethoxyresorufin-O-deethylase (EROD) activity and protein assays 7-Ethoxyresorufin-O-deethylase activity was used to estimate the catalytic activity of CYP1A. Measurements of EROD activity and total protein content of the wells were performed following the method of Kennedy et al. (1995). The EROD activity measured in the cells was normalized to protein quantity of the cells. RNA isolation and RT-PCR of CYP1A mRNA Total RNA was isolated from the hepatocytes using the TRIZOL Reagent (Life Technologies, Gibco BRL, USA). The quality of the isolated RNA was assessed by measuring the absorbance of the RNA solution and calculating the A260/A280 ratios. For the reverse transcriptase (RT)
J.M. Navas et al. / Life Sciences 76 (2004) 699–714
705
reaction, the first strand cDNA Synthesis Kit (Boehringer Mannheim, Germany) was used, and the PCR was performed with the Taq DNA polymerase kit from SIGMA (USA) as previously described (Navas and Segner, 2001). In brief, a 270 bp fragment of the CYP1A gene, and a 540 bp fragment of h-actin, used as housekeeping gene for the RT-PCR, were amplified using adequate primers (Navas and Segner, 2001). The primers used did not distinguish between CYP1A1 and CYP1A3 in rainbow trout. For the PCR, an initial incubation at 95 8C (10 min) was performed. Afterwards, reactions were run for 31 cycles with a 95 8C denaturing step (1 min), 56 8C annealing cycle (1 min) and a 72 8C extension cycle (1 min), plus final incubation at 72 8C for 10 min. In previous experiments, the optimal number of cycles was established in order to ensure that the PCR reaction is performed under linear conditions. In fact, the observed differences between treatments are also indicative that maximal levels of DNA (in which no differences can be detected) were not reached. Computational studies The calculations were performed in a Silicon Graphics O2 R5000 computer (Silicon Graphics Inc., Mountain View, CA), with Irix 6.5 operating system, and in a PC computer working with two 867 MHz processors and Windows XP operating system. All the computational simulations were carried out in the vacuum (1.0 as effective dielectric constant). The systematic conformational search on CLO was performed with the tool gridsearch as implemented in the Sybyl version 6.6 software (Sybyl molecular modeling system, available from Tripos Associated, Inc., St. Louis, MO), using the MMFF94 force field (Halgren, 1996). The conformational search was done by rotation in 608 of the four bonds connecting the central carbon to each aromatic ring. The quantum chemical calculations on CLO and BIM were realized using a hybrid Hartree-Fock-density functional methodology, employing the B3LYP functional (Becke, 1993) and 6-31G(d)* basis set as implemented in GAUSSIAN 98 suite of programs (Frisch et al., 1998). The calculations were carried out up to a high convergence, using the tight option in the GAUSSIAN 98 program. The starting geometries for the B3LYP/6-31G calculation were the one of the most stable conformer obtained in the conformational search, and they were optimized in any convergence cycle in the quantum chemical calculation. Once the geometry was optimized (Fig. 5), the molecular electrostatic potentials (MEPs) and dipole moments (ı`) of CLO and BIM were calculated following standard procedures implemented in the GAUSSIAN 98 program. The molecular surface electrostatic potentials (MSEPs) of the computed MEP were generated using the program gopenmol (Laaksonen, 1992; Bergman et al., 1997). Statistics For each set of experiments, three or four different trout were used (see Experimental Design section, in Material and Methods). Within each individual experiment, each treatment was done in duplicate (two wells of the plate received the same concentration of xenobiotic). Significant differences between treatments were determined by Repeated Measurements Analysis of Variance (RMANOVA, P b 0.05) using the Sigma Stat software from Jandel Scientific (San Rafael, CA). The program tests automatically the normality of the data using Kolmogorov Smirnoff test, and the equal variance by checking the variability about the group means. Means were contrasted using Dunnett test with control group. Cells receiving the solvent were used as control. In the dose-response curve (Fig. 2) only data in which no sign
706
J.M. Navas et al. / Life Sciences 76 (2004) 699–714
Fig. 5. Geometries resulting from B3LYP/6-31G(d) ab initio calculation on clotrimazole (CLO) and 1-benzylimidazole (BIM). CLO-I, CLO-II, CLO-III, and CLO-IV are the four most stable conformers of CLO. A planar conformer of CLO is depicted as CLO-PL. The most stable conformation of BIM is also represented. Energy corresponding to every conformer of CLO appears under the corresponding representation. Values of dipole moment (A) of every stable conformer of CLO (CLO-I, CLO-II, CLO-III, and CLO-IV) and of the stable conformation of BIM appear also under the figures.
of cytotoxicity was observed (Ctrl and CLO concentrations ranging from 0.19 AM to 1.56 AM) were compared.
Results Cell viability No visual alterations were observed in the cells due to CLO exposure. LDH content in media of CLO-treated cells was similar to that of controls, except for the concentration of 3.12 AM CLO, in which the LDH increase with time in medium was double than in controls (190% activity compared to control media). No alterations of protein content per well was observed for the various treatments and sampling times.
J.M. Navas et al. / Life Sciences 76 (2004) 699–714
707
7-Ethoxyresorufin-O-deethylase induction No effect of a-NF on EROD activity induction was detected when hepatocytes were treated with aNF concentrations ranging from 0.39 AM to 50 AM (Data not shown). CLO induced an increase of EROD activity statistically different ( P b 0.05) than control values only for the concentration of 1.56 AM (Fig. 2). EROD levels as obtained with CLO concentrations both higher and lower than 1.56 AM did not differ significantly from the EROD activity of the solvent control. The CLO concentration of 3.125 AM led to a numerically higher EROD activity than 1.56 AM, however, due to large variation among independent replicates, this increase was not significant. The large variation of EROD activity as observed for 3.125 AM CLO may have been caused by the onset of CLO cytotoxic action (as indicated from the LDH data) at this concentration level. In order to compare the CLO-related EROD induction with that of a prototype inducer, cells coming from 4 different isolations were treated with 0.78 AM h-NF, in parallel to CLO. For comparison, CLO at a concentration of 1.56 AM was tested. In these experiments, significant ( P b 0.05) EROD induction was observed for both h-NF and CLO with respect to the controls (Fig. 3). Effects of CLO on CYP1A mRNA expression In order to evaluate if CLO affects CYP1A only at the enzyme catalytic level or also at the transcriptional level, CYP1A mRNA levels were estimated by means of RT-PCR. In four independent cell incubation experiments, hepatocytes were treated either with 0.78 h-NF or with 1.56 AM CLO. In both cases an enhanced expression of CYP1A gene compared to the control cells was observed (Fig. 4). Effect of a-NF on CLO induced CYP1A mRNA The results of the EROD and CYP1A mRNA measurements indicate that CLO is a weak inducer of CYP1A. In order to examine whether the induction response involves ligand binding of CLO to the AhR, co-incubation experiments of CLO with the AhR antagonist a-NF were used. The effect of a-NF as an AhR antagonist was previously demonstrated in co-incubations experiments with h-NF (Navas et al., 2003). In that work, a-NF inhibited the CYP1A transcription induced by h-NF. If CLO acts through ligand binding to the AhR, the presence of a-NF should inhibit the CYP1A induction. The co-exposure of hepatocytes to 1.56 AM CLO and 6.25 AM a-NF did not reduce CYP1A mRNA levels compared to cells incubated with CLO in the absence of a-NF, i.e. the AhR antagonist, a-NF, was not able to inhibit the CLO-induced transcriptional activation of CYP1A (Fig. 4). Computational conformational study on clotrimazole (CLO) Since it has been proposed that all the AhR ligands should be planar molecules to fit into the AhR binding site (Safe, 1995), one objective in this work was to find the more stable conformers of CLO to assess the feasibility of a planar structure. For the size and functionality of our target molecule, the methods of the computational chemistry are very convenient in terms of reliability and time-cost. Conformational search algorithms are automated methods to generate many conformers and, then, to compare them based on their relative energies (Young, 2001). The choice of the different methodologies for conformational searches depends mainly on the size of the molecule and the
708
J.M. Navas et al. / Life Sciences 76 (2004) 699–714
number of rotable bonds. On the basis of these premises, CLO was analyzed using a systematic grid search protocol. The four bonds connecting the central sp 3 -carbon and each aromatic ring of the CLO molecule were rotated in 608 steps, generating 64 = 1296 conformers that were analyzed in terms of symmetry grounds and classified by their relative energy content. The energy range for all the conformers was ca 330 kJ.mol1. The four most stable conformers of CLO, along with their relative energies, are shown as CLO-I, CLO-II, CLO-III, and CLO-IV in Fig. 5. These four conformers of CLO have similar energies (at a 2.0 kJ.mol1 range), indicating that the four conformers have roughly the same probability to be present in the conformational distribution of CLO. A visual inspection of CLO-I to CLO-IV structures, reveals that there is no single stable conformer of CLO that has two aromatic rings in the same plane (i.e. a co-planar conformer). This qualitative observation has been further confirmed by a detailed geometrical analysis showing that the angle between each pair of aromatic rings ranges from 518 to 648 (for two pairs) and from 878 to 908 (for the third pair) in conformers CLO-I to CLO-IV. It is worth to mention that a similar conformation has been found in the solid state of CLO (as determined by single crystal X-ray diffraction analysis) (Song and Shin, 1998). In order to assess the energetic content of any putative co-planar conformer, molecular mechanics calculations (MMFF94) were performed on two potential co-planar conformers (one of them is depicted as CLO-PL in Fig. 5). The results of the calculations indicate that the energy of any of the two conformers would be over 6000 kJ.mol-1 what means that these structures are extremely unstable. The source of the instability of the co-planar conformers is the unfavorable Van der Waals interactions between the substituents at the ortho-positions in the aromatic rings; for instance, the distance between H-2 in the imidazole ring and H-6 in the chlorinated ring in the conformer CLO-PL is only 0.63 2, what is much lower than twice the distance of the van der Waals radius of the hydrogen atom (1.06 2 for each H-atom) (Bondi, 1964). For comparison, the most stable conformation of BIM, another imidazole derivative able to induce CYP1A (Navas et al., 2003), was computed and is also portrayed in Fig. 5. Quantum chemical study on CLO structure The previous calculations indicate that a planar structure of CLO is highly unlikely due to energetic reasons. This observation points to the possibility that CLO does not bind to the AhR binding site as do planar AhR ligands such as TCDD. This conclusion from the computational part is corroborated by the experimental findings on the effect of a-NF (see above). Previously we observed that BIM, was also able to induce CYP1A in rainbow trout cultured hepatocytes at both enzymatic and transcriptional levels (Navas et al., 2003). Computational studies lead to the conclusion that a planar conformation of BIM is very unlikely. Given these functional resemblances between CLO and BIM, we computed here two electrostatic properties, namely the dipole moment and the MEP, with the aim to determine similarities between both molecules that could be indicative of similar mechanisms of action. The most stable BIM conformer (Fig. 5) and the four stable conformers of CLO (CLO-I; CLO-II, CLO-III, and CLO-IV in Fig. 5) were subjected to quantum chemical calculations as indicated in the Material and Methods section. The values of the dipole moments are indicated in Fig. 5. As expected for molecules with a high proportion of heteroatoms, low symmetry, and relatively large size, all the conformers are moderately polar with dipole moments of 4.49 D for the most stable conformer of BIM, and ranging from 3.78 D (CLO-I) to 5.58 D (CLO-III) for CLO. The dipole
J.M. Navas et al. / Life Sciences 76 (2004) 699–714
709
orientation for all the conformers of CLO is from the imidazole ring (negative end) towards the chlorine atoms (positive end); and the orientation of the dipole in BIM is from the phenyl ring to the imidazole ring. The molecular electrostatic potential at the point r [MEP, V(r)] is a representation of the electrostatic interaction energy between a molecule and a test charge of magnitude e (that is a proton) placed at that point, supposing that the molecule is not polarized by the test charge. The MEP is calculated by the Eq. (1) Z X ZA qðrVÞ drV; ð1Þ VðrÞ ¼ jr rVj jR rj A A where ZA is the charge on nucleus A, located at RA, that is considered to be a point charge, and the second term arises from the electron density, U(rV), of the molecule, that can be readily obtained computationally. The MEP is a property that indicates the interaction energy with a proton or an electrophilic chemical entity (such as some fragments of a biomacromolecular target): the more negative value of the electrostatic potential EP, the more stabilizing interaction. As indicated above, the MEPs were plotted on an isoelectronic density surface of 0.005 e.bohr3 (the molecular surface electrostatic potentials). The plots (Figs. 6 and 7) show regions ranging from positive (red) to negative (blue) electrostatic potentials; the values of the electrostatic potentials (in atomic units) are indicated in the scale at the left of the plot. An advantage of the use of the MSEPs is that allows a rapid visual comparison of results, readily identifying the different regions in a molecule.
Fig. 6. MSEPs for the four stable conformers (CLO-I, CLO-II, CLO-III, CLO-IV) of clotrimazole (CLO). Two views (a and b) of each conformer are depicted, showing the different electrostatic potential patterns in the two enantiotopic phenyl rings of CLO. The values of the EP ranges from +0.03 a.u. (deep red) to 0.03 a.u. (deep blue).
710
J.M. Navas et al. / Life Sciences 76 (2004) 699–714
Fig. 7. MSEP of the most stable conformer of 1-benzylimidazole (BIM).
Although the EP reflects both the electronic and nuclear charges, it can be as a representation of the electron density distribution on the molecules, particularly when the same atoms are compared. The regions with negative values of electrostatic potentials are electron-rich regions that would interact favorably with electrophilic and acidic targets. On analyzing the MSEPs of BIM (Fig. 7), and of the four conformers of CLO (Fig. 6), we observe quite similar trends. Thus, the region with the most negative electrostatic potential corresponds to the non-substituted nitrogen atom. Interestingly, the substituted nitrogen atom in the imidazole ring in all the compounds possesses a positive value of the electrostatic potential. The Fig. 6 shows two views for each conformer of CLO. Although the molecule of CLO is achiral, the two phenyl rings are enantiotopic (one is pro-R, the other is pro-S) (Eliel and Wilen, 1994), and they can be differentiated during the interaction with a chiral biomacromolecule (e. g., the aryl hydrocarbon receptor). It is worth to remark that the two enantiotopic phenyl rings of CLO show different EP patterns.
Discussion The present work shows that the imidazole derivative CLO is an inducer of CYP1A in rainbow trout hepatocytes, at both the enzymatic and transcriptional levels. Comparison with previous results on typical EROD inducers, such as h-NF and 3-methylcholanthrene (3MC), in rainbow trout hepatocytes allow to estimate the potency (measure of dose of inducer at which induction occurs) and efficacy (magnitude of induction) of CYP1A induction by CLO. The lowest concentrations of h-NF and 3MC for which a significant induction of EROD activity was observed were 0.024 AM and 0.39 AM, respectively (Navas and Segner, 2000). In the case of CLO this concentration was 1.56 AM. The maximal induction of EROD activity caused by h-NF and 3MC was about twice the EROD activity of that observed in controls: 35 pMol/mg/min in controls and around 75 to 80 pMol/mg/min in h-NF or 3MC treated cells (Navas and Segner, 2000). In the case of CLO the maximal induction observed was around 60 pMol/mg/min with respect to 36 pMol/mg/min detected in controls. The percent increase of CYP1A mRNA detected after treatment with CLO (31 F 7% with respect to controls) was similar to increases of CYP1A mRNA detected after treatment with h-NF (37 F 10% with respect to controls) (Navas et al., 2003). In the case of another imidazole derivative, BIM, the lowest concentration for which a significant EROD induction was observed, was also 1.56 AM (Navas et al., 2003) and the maximal values of EROD activity was around 80 pMol/mg/min for a BIM concentration of 3.12 AM. Taken together, these data indicate that CLO is an effective but weak inducer of CYP1A in cultured rainbow trout hepatocytes.
J.M. Navas et al. / Life Sciences 76 (2004) 699–714
711
The transcriptional activation of CYP1A by xenobiotics is typically mediated through ligand binding to the AhR (Hahn, 1998). The strength of the interaction between the ligand and the receptor is designated as affinity and is a property of both the ligand and the receptor molecule (Hestermann et al., 2000). Ligands with a high affinity for the AhR, for instance TCDD or h-NF, are hydrophobic polyannular aromatic (i.e. with two or more aromatic rings) compounds that are planar, or can take a planar structure (i. e., two aromatic rings in the same plane), and can be accommodated within a binding site of approximately 14 12 5 (2)3 (Waller and McKinney, 1995). In this work, computational studies were performed to estimate if the CLO molecule would be able to take a planar conformation. CLO shows a structure with four aromatic rings bonded to a tetrahedral carbon atom, causing a highly steric encumbrance on this atom. The computational modeling of CLO have yielded four stable conformers (CLO-I to CLO-IV, Fig. 5), and none of them have two aromatic rings in the same plane. Computational studies indicated that the energy content of a putative co-planar conformer is very high, resulting in an extremely unstable structure, due to interactions between the substituents at the ortho-positions in the aromatic rings. The stable conformers of CLO, CLO-I to CLOIV, are almost perfect propellers, a structural feature that has been found in organic compounds having aromatic fragments linked to fully saturated carbon (Tukada and Mochizuki, 2002), nitrogen (Yu et al., 2002), silicon (Lambert and Lin, 2001), or phosphorus (Whitnall et al., 1997). A common structural feature of this kind of compounds is that each aromatic ring is like the blade of the propeller (and, therefore, any blade is rotated with respect to the others), and this topological distribution permits to minimize destabilizing steric interactions. The conclusion from the computational analysis is that CLO is not a typical AhR ligand which can adopt a planar conformation. In line with this view is the experimental finding that the AhR antagonist a-NF was not able to inhibit CLO-mediated CYP1A induction, indicating that CLO is not behaving as a competitive AhR agonist. These data indicate that CLO is not a typical (i.e. planar) AhR ligand. Kafafi et al. (1993) have developed a model of AhR binding of aromatic xenobiotics in which receptor binding is not an exclusive function of planarity but is based on electron affinities, entropies and lipohilicities. Recently, Chana et al. (2002) have provided evidence that electrostatic properties, particularly MEP, are an important factor for the biological activity of PCBs, and that the possibility of a PCB congener to exist in a co-planar configuration is only a strengthening factor for bioactivity. The MEPs are not only a representation of the electronic surface of the molecules, but also they must be considered as valid three dimensional surfaces of the molecules, and it is through these surfaces that the molecules interact one with each other (Murray and Politzer, 1998). Therefore, we have calculated in the present work the MEP of CLO, and, additionally, of BIM, another imidazole compound which is able to induce CYP1A without being coplanar (Navas et al., 2003), in order to seek for molecular resemblances that could be indicative of similar mechanisms of action of both compounds. Interestingly, in both CLO and BIM the substituted nitrogen atom in the imidazole ring possesses a positive value of the electrostatic potential what, in principle, was unexpected on the basis of the higher electronegativity of nitrogen versus carbon as well as on the availability of one electron pair on the nitrogen atom. This effect can be rationalized assuming that this electron pair is engaged in the maintenance of the aromatic electron sextet in the imidazole ring. Qualitatively, the patterns of the electrostatic potential in BIM and in each conformer of CLO are quite similar: there is a peripheral electron rich region (that corresponds to the unsubstituted nitrogen atom) and, at least, one of the benzenic rings that has a nearly neutral or slightly positive value of the electrostatic potential. As a consequence both molecules will interact efficiently with acidic or electrophilic species that are
712
J.M. Navas et al. / Life Sciences 76 (2004) 699–714
present in biological targets. To conclude, our results have shown conformational and electrostatic similarities between BIM and CLO that are not shared by other CYP1A inducers. From a structural point of view CYP1A inducers could be classified in three groups: 1) Compounds having at least two aromatic rings in a plane, with a very low conformational mobility. This is the case of TCDD (Safe, 1990, 1995) or of h-NF, (Fig. 1) (Navas et al., 2003). Both compounds have been considered as typical AhR ligands and CYP1A inducers (Safe, 1990, 1995; Navas and Segner, 1998, 2000; Navas et al., 2003). 2) Molecules with conformational mobility, i. e., they can exist in a variety of conformers that are separated by small energy differences, so that the interconversion between them can be readily achieved. Some examples of this class are 3,3V,4,4V,5-pentachlorobiphenyl (PCB no 126, Fig. 1), 3,3V,4,4V,5,5V-hexachlorobiphenyl (PCB no 169, Fig. 1) (Chana et al., 2002). Although the coplanar structure of these compounds is not the most stable one, this conformation can be readily reached. These compounds have also been described as low affinity AhR ligands and CYP1A inducers in cultured rainbow trout hepatocytes (Anderson et al., 1996). 3) Compounds having at least two aromatic rings linked by a sp 3 -carbon in their structures. Due to the steric requirements of the tetrahedral carbon atom (i.e.: bond angles of ca 1098), the conformations having two aromatic rings in the same plane are energetically unfavorable, being unlikely that these molecules induce CYP1A after binding the AhR in a co-planar conformation. CLO and BIM would pertain to this specific class of CYP1A inducers. The specific characteristics of these compounds should be taken into account in the establishment of structure activity relationship (SAR) models to search for CYP1A inducers. It remains to be elucidated if CYP1A inducers such as BIM and CLO activate the AhR through direct interaction with the receptor molecule, or if the activation occurs indirectly, through kinase-dependent signaling pathways, as it has been suggested for benzimidazoles (Lesca et al., 1995; Backlund et al., 1997, 1999).
Conclusion In summary, our findings indicate that the imidazole derivative CLO is able to induce CYP1A in rainbow trout cultured hepatocytes without being a planar molecule, and that such induction is not inhibited by the AhR antagonist a-NF. These functional and structural features are shared with another imidazole derivative BIM (Navas et al., 2003) and computational studies have demonstrated that both compounds exhibit similar electrostatic properties. These data indicate that these molecules may activate the AhR and induce CYP1A by a mechanism different to classical ligand binding to the AhR. Further work must explore the capacity of different imidazole derivatives to induce CYP1A and examine the mechanism of such induction taking into account the variety of structures shown by this group of compounds.
Acknowledgments Jose´ Marı´a Navas holds a Ramo´n y Cajal contract from the Spanish Ministry of Science and Technology (MCYT). This work was financially supported by MCYT projects (REN2002-00639/
J.M. Navas et al. / Life Sciences 76 (2004) 699–714
713
GLO and BQU2001-2270), and by CAM (Comunidad Auto´noma de Madrid) project 07M/0032/ 2002.
References Anderson, M.J., Miller, M.R., Hinton, D.E., 1996. In vitro modulation of 17-h-estradiol-induced vitellogenin synthesis: Effects of cytochrome P4501A1 inducing compounds on rainbow trout (Oncorhynchus mykiss) liver cells. Aquatic Toxicology 34, 327 – 350. Backlund, M., Johansson, I., Mkrtchian, S., Ingelman-Sundberg, M., 1997. Signal transduction-mediated activation of the aryl hydrocarbon receptor in rat hepatoma H4IIE cells. Journal of Biological Chemistry 50, 31755 – 31763. Backlund, M., Weidolf, L., Ingelman-Sundberg, M., 1999. Structural and mechanistic aspects of transcriptional induction of cytochrome P450 1A1 by benzimidazole derivatives in rat hepatoma H4IIE cells. European Journal of Biochemistry 261, 66 – 71. Becke, A.D., 1993. Density-functional thermochemistry. III. The role of exact exchange. Journal of Chemical Physics 98, 5648 – 5652. Bergman, D.L., Laaksonen, L., Laaksonen, A., 1997. Visualization of solvation structures in liquid mixtures. Journal of Molecular Graphics 15, 301 – 306. Bondi, A., 1964. Van der Waals Volumes and Radii. Journal of Physical Chemistry 68, 441 – 451. Braunbeck, T., Segner, H., 2000. Isolation and cultivation of teleost hepatocytes. In: Berry, M.N., Edwards, A.M. (Eds.), The Hepatocyte Review. Kluwer, Amsterdam, pp. 49 – 71. Castillo, L.E., De la Cruz, E., Ruepert, C., 1997. Ecotoxicology and pesticides in tropical aquatic ecosystems in Central America. Environmental Toxicology and Chemistry 16, 41 – 51. Chana, A., Concejero, M.A., de Frutos, M., Gonza´lez, M.J., Herrado´n, B., 2002. Computacional studies on biphenyl derivatives. Analysis of the conformational mobility, molecular electrostatic potential, and dipole moment of chlorinated biphenyls: Searching for the rationalization of the selective toxicity of polychlorinated biphenyls. Chemical Research in Toxicology 15, 1514 – 1526. Denison, M.S., Whitlock, J.P., 1995. Xenobiotic-inducible transcription of cytochrome P450 genes. Journal of Biological Chemistry 270, 18175 – 18178. Denizeau, F., Marion, M., 1990. Toxicity of cadmium, copper, and mercury to isolated trout hepatocytes. Canadian Journal of Fisheries and Aquatic Sciences 47, 1038 – 1042. Eglo, O.M., Aspmi, R., Lode, O., 1994. Runoff and leaching experiments of dichlorprop, MCPA, propioconazole, dimethoate, and chlorsulfuron in outdoor lysimeters and field catchment areas. Norwegian Journal of Agricultural Sciences 13, 53 – 78. Eliel, E.L., Wilen, S.H., 1994. Stereochemistry of Organic Compounds. John Wiley & Sons, New York. Frisch, M.J., Trucks, G.W., Schlegel, H.B., Scuseria, G.E., Robb, M.A., Cheeseman, J.R., Zakrzewski, V.G., Montgomery Jr., J.A., Stratmann, R.E., Burant, J.C., Dapprich, S., Millam, J.M., Daniels, A.D., Kudin, K.N., Strain, M.C., Farkas, O., Tomasi, J., Barone, V., Cossi, M., Cammi, R., Mennucci, B., Pomelli, C., Adamo, C., Clifford, S., Ochterski, J., Petersson, G.A., Ayala, P.Y., Cui, Q., Morokuma, K., Salvador, P., Dannenberg, J.J., Malick, D.K., Rabuck, A.D., Raghavachari, K., Foresman, J.B., Cioslowski, J., Ortiz, J.V., Baboul, A.G., Stefanov, B.B., Liu, G., Liashenko, A., Piskorz, P., Komaromi, I., Gomperts, R., Martin, R.L., Fox, D.J., Keith, T., Al-Laham, M.A., Peng, C.Y., Nanayakkara, A., Challacombe, M., Gill, P.M.W., Johnson, B., Chen, W., Wong, M.W., Andres, J.L., Gonzalez, C., Head-Gordon, M., Replogle, E.S., Pople, J.A., 1998. Gaussian 98, Revision A.3. Gaussian Inc, Pittsburgh, PA. Hahn, M., 1998. The aryl hydrocarbon receptor. A comparative perspective. Comparative Biochemistry and Physiology 121 (C), 23 – 53. Halgren, T.A., 1996. Merck molecular force field I. Basis, form, scope, parameterization, and performance of MMFF94. Journal of Comparative Chemistry 17, 490 – 519. Hankinson, O., 1995. The aryl hydrocarbon receptor complex. Annual Reviews in Pharmacology and Toxicology 35, 307 – 340. Henry, M.J., Sisler, H.D., 1984. Effects of sterol biosynthesis-inhibiting (SBI) fungicides on cytochrome P-450 oxygenations in fungi. Pesticide Biochemistry and Physiology 22, 262 – 275. Hestermann, E.V., Stegeman, J.J., Hahn, M.E., 2000. Relative contributions of affinity and intrinsic efficacy to aryl hydrocarbon receptor ligand potency. Toxicology and Applied Pharmacology 168, 160 – 172.
714
J.M. Navas et al. / Life Sciences 76 (2004) 699–714
Kafafi, S.A., Afeefy, H.Y., Ali, A.H., Said, H.K., Kafafi, A.G., 1993. Binding of polychlorinated biphenyls to the aryl hydrocarbon receptor. Environmental Health Perspectives 101, 422 – 428. Kennedy, S.W., Jones, S.P., Bastien, L.J., 1995. Efficient analysis of cytochrome P4501A catalytic activity, porphyrins, and total proteins in chicken embryo hepatocyte cultures with a fluorescence plate reader. Analytical Biochemistry 226, 362 – 370. Kobayashi, Y., Matsuura, Y., Kotani, E., Fukuda, T., Aoyagi, T., Tobinaga, S., Yoshida, T., Kuroiwa, Y., 1993. Structural requirements for the induction of hepatic microsomal cytochrome P450 by imidazole- and pyridine-containing compounds in rats. The Journal of Biochemistry 114, 697 – 701. Koley, A.P., Buters, T.M., Robinson, R.C., Markowitz, A., Friedman, F.K., 1997. Differential mechanisms of cytochrome P450 inhibition and activation by a-naphthoflavone. Journal of Biological Chemistry 6, 3149 – 3152. Laaksonen, L., 1992. A graphics program for the analysis and display of molecular dynamics trajectories. Journal of Molecular Graphics 10, 33 – 34. Lambert, J.B., Lin, L., 2001. Propeller rotation of aryl groups in triarylsilanes and triarylstannanes. Magnetic Resonance in Chemistry 39, 714 – 718. Lesca, P., Peryt, B., Larrieu, G., Alvinerie, M., Galtier, P., Daujat, M., Maurel, P., Hoogenboom, L., 1995. Evidence for the ligand-independent activation of the Ah receptor. Biochemical and Biophysical Research Communications 209, 474 – 482. Murray, J.S., Sen, K., 1996. Molecular Electrostatic Potentials. Concepts and Applications. Elsevier, Amsterdam. Murray, J.S., Politzer, P., 1998. The molecular electrostatic potential: a tool for understanding and predicting molecular interactions. In: Sapse, A.M. (Ed.), Molecular Orbital Calculations for Biological Systems. Oxford University Press, New York, pp. 49 – 84. Navas, J.M., Segner, H., 1998. Antiestrogenic activity of anthropogenic and natural chemicals. Environmental Science and Pollution Research 5, 75 – 82. Navas, J.M., Segner, H., 2000. Antiestrogenicity of h-naphthoflavone and PAHs in cultured rainbow trout hepatocytes: evidence for a role of the arylhydrocarbon receptor. Aquatic Toxicology 51, 79 – 92. Navas, J.M., Segner, H., 2001. Estrogen-mediated suppression of cytochrome P4501A (CYP1A) expression in rainbow trout hepatocytes: role of estrogen receptor. Chemico-Biological Interactions 138, 285 – 298. Navas, J.M., Chana, A., Herrado´n, B., Segner, H., 2003. Induction of CYP1A by the N-imidazole derivative, 1benzylimidazole. Environmental Toxicology and Chemistry 22, 830 – 836. Rowlands, J.C., Gustafsson, J.A., 1997. Aryl hydrocarbon receptor-mediated signal transduction. Critical Reviews in Toxicology 27, 109 – 134. Safe, S., 1990. Polychlorinated biphenyls (PCBs), dibenzo-p-dioxins (PCDDs), dibenzofurans (PCDFs) and related compounds: environmental and mechanistic considerations which support the development of toxic equivalency factors (TEFs). Critical Reviews in Toxicology 21, 51 – 88. Safe, S., 1995. Modulation of gene expression and endocrine response pathways by 2,3,7,8-tetrachlorodibenzo-p-dioxin and related compounds. Pharmacology and Therapeutics 67, 247 – 281. Song, H., Shin, H.-S., 1998. The antifungal drug clotrimazole. Acta Crystallographica 54, 1675 – 1677. Tukada, H., Mochizuki, K., 2002. Structures and short contacts of F. . .O(nitro) and H. . .O(nitro) in 2,2-di(4-nitrophenyl)hexafluoropropane and 2,2-di(4-nitrophenyl)propane. Journal of Molecular Structure 643, 85 – 91. Waller, C.L., McKinney, J.D., 1995. Three-dimensional quantitative structure activity relationships of dioxins and dioxin-like compounds-model validation and Ah receptor characterization. Chemical Research in Toxicology 8, 847 – 857. Whitnall, M.R., Hii, K.K., Thornton-Pett, M., Kee, T.P., 1997. Conformational diastereoisomerism in tris-(2-alkylimino) triphenylphosphines. Journal of Organometallic Chemistry 529, 35 – 50. Young, D.C., 2001. Computational Chemistry: A Practical Guide for Applying Techniques to Real-World Problems. John Wiley & Sons, New York. Yu, H., Huang, Y., Zhang, W., Matsuura, T., Meng, J., 2002. Crystal structures of two new compounds containing 2,5-diaryl1,3,4-oxadiazole and triarylamine units. Journal of Molecular Structure 642, 53 – 62.
Induction of cytochrome P4501A (CYP1A) by clotrimazole, a non-planar aromatic compound. Computational studies on structural features of clotrimazole and related imidazole derivatives Jose´ Marı´a Navasa,*, Antonio Chanab, Bernardo Herrado´nb, Helmut Segnera,1 a
Department of Chemical Ecotoxicology, UFZ Center for Environmental Research, Permoserstr. 15, D-04318 Leipzig, Germany b Instituto de Quı´mica Orga´nica General, C.S.I.C., Juan de la Cierva 3, 28006 Madrid, Spain Received 17 February 2004; accepted 9 September 2004
Abstract The classical pathway for induction of cytochrome P4501A (CYP1A) by xenobiotics is ligand binding to the aryl hydrocarbon receptor (AhR). High-affinity AhR ligands are planar polyaromatic molecules such as the prototypic ligand, 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD). The present work investigated the ability of the imidazole derivative, clotrimazole [1-(2Vchlorotrityl)imidazole, CLO], to induce CYP1A in cultured rainbow trout (Oncorhynchus mykiss) hepatocytes at the catalytic activity (determined as 7-ethoxyresorufin-O-deethylase, EROD) and at the transcriptional level. CLO resulted in a significant increase of hepatocyte EROD activity and CYP1A mRNA at a concentration of 1.56 AM. Computational studies on the molecular structure of CLO show that CLO is unlikely to take a planar conformation. Further indications that CLO does not behave like a planar AhR ligand come from the experimental observation that co-incubation of trout hepatocytes with CLO and the AhR antagonist, a-naphthoflavone (a-NF), did not result in an inhibition of CLO induction of CYP1A mRNA, whereas a-NF was able to inhibit CYP1A induction by the prototpyic, planar AhR ligand, h-naphthoflavone. The experimental findings on CLO agree with previous results obtained for another non-planar imidazole derivative, 1benzylimidazole (BIM). Further, computational studies showed that the non-planar imidazoles, BIM and CLO, are highly similar with respect to some electrostatic properties, namely the dipole moment and the molecular
* Corrresponding author. Present address: INIA, Department of the Environment, Carretera de la Corun˜a, Km 7.5, 28040 Madrid, Spain. Tel.: +34 91 3 47 41 55; fax: +34 91 3 57 22 93. E-mail address: [email protected] (J.M. Navas). 1 Present address: Centre for Fish and Wildlife Health, Institute of Animal Pathology, University of Bern, L7nggass-Str. 122, CH-3012 Bern, Switzerland. 0024-3205/$ - see front matter D 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.lfs.2004.09.015
700
J.M. Navas et al. / Life Sciences 76 (2004) 699–714
electrostatic potential (MEP). Overall our experimental and computational studies suggest that transcriptional activation of CYP1A by the imidazole derivatives CLO and BIM is mediated by a mechanism different to that of prototypic CYP1A inducers such as the planar AhR-ligands. D 2004 Elsevier Inc. All rights reserved. Keywords: Aryl hydrocarbon receptor; EROD; Cytochrome P450 1A; Clotrimazole; Imidazole
Introduction The structural properties underlying the interaction of chemicals with biological molecules are an important topic of current research. The biological effects of numerous xenobiotics are mediated through binding to the arylhydrocarbon receptor (AhR) and the subsequent activation of AhRregulated genes such as the xenobiotic-metabolizing enzyme cytochrome P4501A (CYP1A) (Safe, 1990; Denison and Whitlock, 1995; Hahn, 1998). Ligands bind to the AhR in the cytoplasm of the target cell, thereafter the ligand-activated receptor undergoes a transformation process whereby it forms a heterodimer with the AhR nuclear translocator protein (ARNT), that is transported into the nucleus. The nuclear receptor complex then interacts with specific xenobiotic-responsive elements (XREs) located upstream of the CYP1A gene, leading to stimulated transcription of the gene, elevated CYP1A mRNA, and increased levels of CYP1A protein and its catalytic activity (Hahn, 1998; Rowlands and Gustafsson, 1997). Typical ligands for the AhR are hydrophobic aromatic compounds with at least two aromatic rings (either carbocyclic or heterocyclic) that are planar or can become planar, i.e. both aromatic rings are in the same plane with a dihedral angle of 08 between both rings (Hankinson, 1995; Safe, 1995). 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD, Fig. 1) is the most potent agonist of AhR found up to now, with a KD in the picomolar range (Rowlands and Gustafsson, 1997). High affinity ligands include certain polychlorinated biphenyls (PCBs, see Fig. 1), polyaromatic hydrocarbons (PAHs, i.e., h-naphthoflavone, h-NF), aromatic amines, indolocarbazoles and related compounds. A
Fig. 1. Chemical structures of a-naphthoflavone (a-NF), an AhR antagonist, and of some AhR activators: 2,3,7,8tetrachlorodibenzo-p-dioxin (TCDD), h-naphthoflavone (h-NF), 3,3V,4,4V,5-pentachlorobiphenyl (PCB no 126), 3,3V,4,4V,5,5V-hexachlorobiphenyl (PCB no 169), 1-benzylimidazole (BIM), and clotrimazole (CLO).
J.M. Navas et al. / Life Sciences 76 (2004) 699–714
701
comparative molecular field analysis (CoMFA) using a wider range of ligand types (dibenzo-pdioxins, dibenzofurans, biphenyls, naphthalenes and indolocarbazoles) has predicted a single ligandbinding pocket of 14 12 5 2, which would accommodate all the known ligands (Waller and McKinney, 1995). The importance of planarity for AhR binding and CYP1A induction was exemplified for polychlorinated biphenyls (PCBs), where congeners with para- or meta-substitution, which easily can take a co-planar configuration, are potent CYP1A inducers, but ortho-substituted congeners for which a co-planar configuration is energetically unfavorable, are weak or noninducers of CYP1A (Safe, 1990). However, this statement has been questioned by several authors who determined, by computational studies, that some PCBs acting as AhR ligands or CYP1A inducers were not planar molecules (Waller and McKinney, 1995; Kafafi et al., 1993; Chana et al., 2002). The objective of this work was to study the properties of the imidazole derivative clotrimazole (CLO, Fig. 1) as a CYP1A inducer. Imidazoles are a large group of compounds frequently used as antifungal agents, both clinically as well as in agriculture. Their antifungal activity is due to their capacity to inhibit cytochrome P450-mediated ergosterol synthesis (Henry and Sisler, 1984). These compounds can enter the aquatic environment and bioaccumulate in fish (Eglo et al., 1994; Castillo et al., 1997). Kobayashi et al. (1993) have reported that a variety of imidazole derivatives including CLO are inducers of CYP1A gene expression; and it has been hypothesized that this biological activity is due to the fact that these chemicals bind the AhR in a co-planar structure (i.e., the two aromatic rings are in the same plane, although the whole molecule is not planar). For a number of imidazole derivatives, however, the assumption that they can achieve a planar conformation is debatable from the point of view of structural organic chemistry, as we have previously demonstrated for 1-benzylimidazole (BIM, Fig. 1) (Navas et al., 2003). In order to further understand the mechanistic basis of CYP1A induction by imidazolic compounds, the present study explores AhR binding and CYP1A induction by CLO. Experimentally, we investigated the potency of CLO to induce CYP1A in cultured rainbow trout hepatocytes at both the enzyme catalytic and the transcriptional level, and we determined whether the AhR antagonist a-naphthoflavone (a-NF, Fig. 1) can block the CLO-related induction of CYP1A in the hepatocytes. Then, we performed computational studies to establish conformational preferences of the molecule, in particular its ability to take a planar configuration. Additionally, we have calculated electrostatic properties of CLO, more specifically the dipole moment and the molecular electrostatic potential (MEP). MEP is an indicator of the charge distribution of the molecule and has been employed to understand a variety of chemical and physical properties of molecules including binding or chemical reactivity, and thus has been used to rationalize the interaction between a biologically active molecule and its biomacromolecular target, e.g., the AhR (Chana et al., 2002; Murray and Sen, 1996).
Methods Animals Sexually immature male and female rainbow trout (250–350 g weight) from a local trout farm were maintained in 200 l steel tanks in the facilities of the Umweltforschungszentrum, (Centre for Environmental Research), Leipzig, Germany. Fish were held under artificial 12 hours
702
J.M. Navas et al. / Life Sciences 76 (2004) 699–714
Fig. 2. 7-Ethoxyresorufin-O-deethylase (EROD) activity induced by different AM concentrations of clotrimazole (CLO). Asterisks indicate significant differences with respect to the controls ( P b 0.05).
light/12 hours dark photoperiod in constant flow aerated water. Water temperature was maintained at 14–16 8C. Experimental design First, a dose-response curve was elaborated to establish the relationship between the concentration of CLO and the induction of EROD activity (Fig. 2). Data obtained after treatment of hepatocytes isolated from 3 different trout in three independent experiments were used at this point. Second, to compare the EROD induction of CLO, with that of a typical inducer like h-NF, four different experiments using hepatocytes from four trout were carried out. In these experiments, a single concentration of CLO (1.56 AM) and of h-NF (0.78 AM) were used (Fig. 3). Finally, to determine if the observed CYP1A induction was only provoked at the enzymatic level, or also at the transcriptional level, four independent
Fig. 3. 7-Ethoxyresorufin-O-deethylase (EROD) induction caused by 0.78 AM h-naphthoflavone (h-NF) and by 1.56 AM clotrimazole (CLO). Asterisks indicate significant differences with respect to the controls ( P b 0.05).
J.M. Navas et al. / Life Sciences 76 (2004) 699–714
703
Fig. 4. Induction of cytochrome P4501A (CYP1A) mRNA caused by 0.78 AM h-naphthoflavone (h-NF) and by 1.56 AM clotrimazole (CLO). 6.25 AM a-naphthoflavone (a-NF) did not provoke any reduction of the CYP1A mRNA levels detected after CLO treatment. Gels from one representative experiments of four are shown. On the right side appear data from the densitometric evaluation of the gels. Densitometric values for CYP1A were normalized with respect to densitometric values obtained with the housekeeping gene h-actin. The percentage of increase in this ratio obtained in the treatments with respect to the controls is represented.
experiments were performed (Fig. 4). The cells used in these four experiments were treated with 0.78 AM h-NF, or with 1.56 AM CLO, or received these chemicals together with a-NF 6.25 AM (a-NF is an AhR antagonist (Koley et al., 1997), so that the co-treatment of the cells with an AhR ligand and a-NF, should lead to an inhibition of the CYP1A induction caused by the ligand). In a previous work (Navas et al., 2003) we observed that another imidazole derivative, BIM, was able to induce CYP1A in rainbow trout cultured hepatocytes, although this compound was not planar, and that this induction was not inhibited by a-NF. Computational studies were performed on CLO and BIM to determine their most stable conformations as well as to assess the possibility that these compounds take co-planar structures; additionally some electrostatic properties were calculated with the aim to seek out structural similarities that could indicate similarities in the mechanism of action of both compounds. Xenobiotic treatments The chemicals h-NF, a-NF, and CLO were purchased from Sigma (St. Louis, MO, USA) at the highest purity available (95% for h-NF and a-NF, and 98% for CLO). Chemicals were dissolved in dimethyl sulfoxide (DMSO, Merck, Germany) and added to the culture media of hepatocytes to achieve the final desired concentration. Final organic solvent concentration in the assay was 0.1%. Hepatocyte isolation and culture Hepatocytes were isolated following a two-step perfusion technique as described by Braunbeck and Segner (2000). As culture medium, modified M199 medium (Sigma, St. Louis, MO, USA) supplemented with 2mM glutamine, 10 U/ml penicillin and 10 Ag/ml streptomycin was used. Medium was sterilized by a 0.22 Am filter. Cells were plated in 24-well FALCONR primary culture plates (Beckton Dickinson, Franklin Lakes, NJ, USA) precoated with Matrigel (BD Biosciences, Bedford, MA, USA) (0.1 mg protein ml1). Every well received 400 Al of the final suspension of cells (density: 2 105 cells cm2). The cells used for the extractions of mRNA in order to perform the RT-PCR analysis were plated on 60 mm diameter FALCONR tissue culture dishes and every plate received 5 ml of the
704
J.M. Navas et al. / Life Sciences 76 (2004) 699–714
cell suspension. During the first 24 hours of treatment the medium was supplemented with 5% fetal bovine serum (Sigma, St. Louis, MO, USA) to enhance the attachment of the cells. During exposure to the test compounds, serum free medium was used. The plates were maintained at 15 8C and 80% humidity. Twenty-four hours after plating, half of the medium was substituted with fresh medium containing the desired concentrations of CLO (0.19 AM to 3.12 AM, final concentrations in culture medium). The possible effect of a-NF on EROD activity was also observed by incubating the cells with a-NF concentrations ranging from 0.39 AM to 50 AM. Control wells received medium with solvent only. Treatments were applied in duplicate. At 48 and 72 hours after plating, half of the medium was replaced again and fresh medium containing the desired concentrations of CLO was added. 96 hours after plating (i.e., after 72 hours of treatment) the medium was removed, aliquoted, and stored at 80 8C until analysis of lactate dehydrogenase (LDH) activity. The cells were briefly washed with phosphate-buffered saline, pH 7.5, and the plates were frozen by using liquid nitrogen. They were maintained at -80 8C until analysis of EROD activity and protein contents. When the cells were co-treated with 6.25 AM a-NF and 1.6 AM CLO for RT-PCR, the following exposure regime was used: 24 hours after plating, the spent medium was removed and new medium with the desired concentrations of a-NF was added. After another 4 hours, half of the medium was removed and the same amount of fresh medium with the adequate concentrations of a-NF and CLO was added. In these experiments, cells were collected 48 hours after plating. The cells were briefly washed with phosphate-buffered saline, pH 7.5, and sampled by adding TRIZOL Reagent (Life Technologies, Gibco BRL, USA) to the plates to extract RNA (see below) that was stored at 80 8C. Cell viability assays Viability of the cells exposed to the test compounds was assessed by measuring the release of lactate dehydrogenase (LDH) into the extracellular medium according to the method of Denizeau and Marion (1990). Activity of LDH in the medium was determined following the oxidation of NADH at 340 nm in a spectrophotometer. Cell morphology was routinely observed using an inverted microscope. Protein content of the wells can also serve as an indicator of toxicity, and was routinely measured in parallel to EROD as described below. 7-Ethoxyresorufin-O-deethylase (EROD) activity and protein assays 7-Ethoxyresorufin-O-deethylase activity was used to estimate the catalytic activity of CYP1A. Measurements of EROD activity and total protein content of the wells were performed following the method of Kennedy et al. (1995). The EROD activity measured in the cells was normalized to protein quantity of the cells. RNA isolation and RT-PCR of CYP1A mRNA Total RNA was isolated from the hepatocytes using the TRIZOL Reagent (Life Technologies, Gibco BRL, USA). The quality of the isolated RNA was assessed by measuring the absorbance of the RNA solution and calculating the A260/A280 ratios. For the reverse transcriptase (RT)
J.M. Navas et al. / Life Sciences 76 (2004) 699–714
705
reaction, the first strand cDNA Synthesis Kit (Boehringer Mannheim, Germany) was used, and the PCR was performed with the Taq DNA polymerase kit from SIGMA (USA) as previously described (Navas and Segner, 2001). In brief, a 270 bp fragment of the CYP1A gene, and a 540 bp fragment of h-actin, used as housekeeping gene for the RT-PCR, were amplified using adequate primers (Navas and Segner, 2001). The primers used did not distinguish between CYP1A1 and CYP1A3 in rainbow trout. For the PCR, an initial incubation at 95 8C (10 min) was performed. Afterwards, reactions were run for 31 cycles with a 95 8C denaturing step (1 min), 56 8C annealing cycle (1 min) and a 72 8C extension cycle (1 min), plus final incubation at 72 8C for 10 min. In previous experiments, the optimal number of cycles was established in order to ensure that the PCR reaction is performed under linear conditions. In fact, the observed differences between treatments are also indicative that maximal levels of DNA (in which no differences can be detected) were not reached. Computational studies The calculations were performed in a Silicon Graphics O2 R5000 computer (Silicon Graphics Inc., Mountain View, CA), with Irix 6.5 operating system, and in a PC computer working with two 867 MHz processors and Windows XP operating system. All the computational simulations were carried out in the vacuum (1.0 as effective dielectric constant). The systematic conformational search on CLO was performed with the tool gridsearch as implemented in the Sybyl version 6.6 software (Sybyl molecular modeling system, available from Tripos Associated, Inc., St. Louis, MO), using the MMFF94 force field (Halgren, 1996). The conformational search was done by rotation in 608 of the four bonds connecting the central carbon to each aromatic ring. The quantum chemical calculations on CLO and BIM were realized using a hybrid Hartree-Fock-density functional methodology, employing the B3LYP functional (Becke, 1993) and 6-31G(d)* basis set as implemented in GAUSSIAN 98 suite of programs (Frisch et al., 1998). The calculations were carried out up to a high convergence, using the tight option in the GAUSSIAN 98 program. The starting geometries for the B3LYP/6-31G calculation were the one of the most stable conformer obtained in the conformational search, and they were optimized in any convergence cycle in the quantum chemical calculation. Once the geometry was optimized (Fig. 5), the molecular electrostatic potentials (MEPs) and dipole moments (ı`) of CLO and BIM were calculated following standard procedures implemented in the GAUSSIAN 98 program. The molecular surface electrostatic potentials (MSEPs) of the computed MEP were generated using the program gopenmol (Laaksonen, 1992; Bergman et al., 1997). Statistics For each set of experiments, three or four different trout were used (see Experimental Design section, in Material and Methods). Within each individual experiment, each treatment was done in duplicate (two wells of the plate received the same concentration of xenobiotic). Significant differences between treatments were determined by Repeated Measurements Analysis of Variance (RMANOVA, P b 0.05) using the Sigma Stat software from Jandel Scientific (San Rafael, CA). The program tests automatically the normality of the data using Kolmogorov Smirnoff test, and the equal variance by checking the variability about the group means. Means were contrasted using Dunnett test with control group. Cells receiving the solvent were used as control. In the dose-response curve (Fig. 2) only data in which no sign
706
J.M. Navas et al. / Life Sciences 76 (2004) 699–714
Fig. 5. Geometries resulting from B3LYP/6-31G(d) ab initio calculation on clotrimazole (CLO) and 1-benzylimidazole (BIM). CLO-I, CLO-II, CLO-III, and CLO-IV are the four most stable conformers of CLO. A planar conformer of CLO is depicted as CLO-PL. The most stable conformation of BIM is also represented. Energy corresponding to every conformer of CLO appears under the corresponding representation. Values of dipole moment (A) of every stable conformer of CLO (CLO-I, CLO-II, CLO-III, and CLO-IV) and of the stable conformation of BIM appear also under the figures.
of cytotoxicity was observed (Ctrl and CLO concentrations ranging from 0.19 AM to 1.56 AM) were compared.
Results Cell viability No visual alterations were observed in the cells due to CLO exposure. LDH content in media of CLO-treated cells was similar to that of controls, except for the concentration of 3.12 AM CLO, in which the LDH increase with time in medium was double than in controls (190% activity compared to control media). No alterations of protein content per well was observed for the various treatments and sampling times.
J.M. Navas et al. / Life Sciences 76 (2004) 699–714
707
7-Ethoxyresorufin-O-deethylase induction No effect of a-NF on EROD activity induction was detected when hepatocytes were treated with aNF concentrations ranging from 0.39 AM to 50 AM (Data not shown). CLO induced an increase of EROD activity statistically different ( P b 0.05) than control values only for the concentration of 1.56 AM (Fig. 2). EROD levels as obtained with CLO concentrations both higher and lower than 1.56 AM did not differ significantly from the EROD activity of the solvent control. The CLO concentration of 3.125 AM led to a numerically higher EROD activity than 1.56 AM, however, due to large variation among independent replicates, this increase was not significant. The large variation of EROD activity as observed for 3.125 AM CLO may have been caused by the onset of CLO cytotoxic action (as indicated from the LDH data) at this concentration level. In order to compare the CLO-related EROD induction with that of a prototype inducer, cells coming from 4 different isolations were treated with 0.78 AM h-NF, in parallel to CLO. For comparison, CLO at a concentration of 1.56 AM was tested. In these experiments, significant ( P b 0.05) EROD induction was observed for both h-NF and CLO with respect to the controls (Fig. 3). Effects of CLO on CYP1A mRNA expression In order to evaluate if CLO affects CYP1A only at the enzyme catalytic level or also at the transcriptional level, CYP1A mRNA levels were estimated by means of RT-PCR. In four independent cell incubation experiments, hepatocytes were treated either with 0.78 h-NF or with 1.56 AM CLO. In both cases an enhanced expression of CYP1A gene compared to the control cells was observed (Fig. 4). Effect of a-NF on CLO induced CYP1A mRNA The results of the EROD and CYP1A mRNA measurements indicate that CLO is a weak inducer of CYP1A. In order to examine whether the induction response involves ligand binding of CLO to the AhR, co-incubation experiments of CLO with the AhR antagonist a-NF were used. The effect of a-NF as an AhR antagonist was previously demonstrated in co-incubations experiments with h-NF (Navas et al., 2003). In that work, a-NF inhibited the CYP1A transcription induced by h-NF. If CLO acts through ligand binding to the AhR, the presence of a-NF should inhibit the CYP1A induction. The co-exposure of hepatocytes to 1.56 AM CLO and 6.25 AM a-NF did not reduce CYP1A mRNA levels compared to cells incubated with CLO in the absence of a-NF, i.e. the AhR antagonist, a-NF, was not able to inhibit the CLO-induced transcriptional activation of CYP1A (Fig. 4). Computational conformational study on clotrimazole (CLO) Since it has been proposed that all the AhR ligands should be planar molecules to fit into the AhR binding site (Safe, 1995), one objective in this work was to find the more stable conformers of CLO to assess the feasibility of a planar structure. For the size and functionality of our target molecule, the methods of the computational chemistry are very convenient in terms of reliability and time-cost. Conformational search algorithms are automated methods to generate many conformers and, then, to compare them based on their relative energies (Young, 2001). The choice of the different methodologies for conformational searches depends mainly on the size of the molecule and the
708
J.M. Navas et al. / Life Sciences 76 (2004) 699–714
number of rotable bonds. On the basis of these premises, CLO was analyzed using a systematic grid search protocol. The four bonds connecting the central sp 3 -carbon and each aromatic ring of the CLO molecule were rotated in 608 steps, generating 64 = 1296 conformers that were analyzed in terms of symmetry grounds and classified by their relative energy content. The energy range for all the conformers was ca 330 kJ.mol1. The four most stable conformers of CLO, along with their relative energies, are shown as CLO-I, CLO-II, CLO-III, and CLO-IV in Fig. 5. These four conformers of CLO have similar energies (at a 2.0 kJ.mol1 range), indicating that the four conformers have roughly the same probability to be present in the conformational distribution of CLO. A visual inspection of CLO-I to CLO-IV structures, reveals that there is no single stable conformer of CLO that has two aromatic rings in the same plane (i.e. a co-planar conformer). This qualitative observation has been further confirmed by a detailed geometrical analysis showing that the angle between each pair of aromatic rings ranges from 518 to 648 (for two pairs) and from 878 to 908 (for the third pair) in conformers CLO-I to CLO-IV. It is worth to mention that a similar conformation has been found in the solid state of CLO (as determined by single crystal X-ray diffraction analysis) (Song and Shin, 1998). In order to assess the energetic content of any putative co-planar conformer, molecular mechanics calculations (MMFF94) were performed on two potential co-planar conformers (one of them is depicted as CLO-PL in Fig. 5). The results of the calculations indicate that the energy of any of the two conformers would be over 6000 kJ.mol-1 what means that these structures are extremely unstable. The source of the instability of the co-planar conformers is the unfavorable Van der Waals interactions between the substituents at the ortho-positions in the aromatic rings; for instance, the distance between H-2 in the imidazole ring and H-6 in the chlorinated ring in the conformer CLO-PL is only 0.63 2, what is much lower than twice the distance of the van der Waals radius of the hydrogen atom (1.06 2 for each H-atom) (Bondi, 1964). For comparison, the most stable conformation of BIM, another imidazole derivative able to induce CYP1A (Navas et al., 2003), was computed and is also portrayed in Fig. 5. Quantum chemical study on CLO structure The previous calculations indicate that a planar structure of CLO is highly unlikely due to energetic reasons. This observation points to the possibility that CLO does not bind to the AhR binding site as do planar AhR ligands such as TCDD. This conclusion from the computational part is corroborated by the experimental findings on the effect of a-NF (see above). Previously we observed that BIM, was also able to induce CYP1A in rainbow trout cultured hepatocytes at both enzymatic and transcriptional levels (Navas et al., 2003). Computational studies lead to the conclusion that a planar conformation of BIM is very unlikely. Given these functional resemblances between CLO and BIM, we computed here two electrostatic properties, namely the dipole moment and the MEP, with the aim to determine similarities between both molecules that could be indicative of similar mechanisms of action. The most stable BIM conformer (Fig. 5) and the four stable conformers of CLO (CLO-I; CLO-II, CLO-III, and CLO-IV in Fig. 5) were subjected to quantum chemical calculations as indicated in the Material and Methods section. The values of the dipole moments are indicated in Fig. 5. As expected for molecules with a high proportion of heteroatoms, low symmetry, and relatively large size, all the conformers are moderately polar with dipole moments of 4.49 D for the most stable conformer of BIM, and ranging from 3.78 D (CLO-I) to 5.58 D (CLO-III) for CLO. The dipole
J.M. Navas et al. / Life Sciences 76 (2004) 699–714
709
orientation for all the conformers of CLO is from the imidazole ring (negative end) towards the chlorine atoms (positive end); and the orientation of the dipole in BIM is from the phenyl ring to the imidazole ring. The molecular electrostatic potential at the point r [MEP, V(r)] is a representation of the electrostatic interaction energy between a molecule and a test charge of magnitude e (that is a proton) placed at that point, supposing that the molecule is not polarized by the test charge. The MEP is calculated by the Eq. (1) Z X ZA qðrVÞ drV; ð1Þ VðrÞ ¼ jr rVj jR rj A A where ZA is the charge on nucleus A, located at RA, that is considered to be a point charge, and the second term arises from the electron density, U(rV), of the molecule, that can be readily obtained computationally. The MEP is a property that indicates the interaction energy with a proton or an electrophilic chemical entity (such as some fragments of a biomacromolecular target): the more negative value of the electrostatic potential EP, the more stabilizing interaction. As indicated above, the MEPs were plotted on an isoelectronic density surface of 0.005 e.bohr3 (the molecular surface electrostatic potentials). The plots (Figs. 6 and 7) show regions ranging from positive (red) to negative (blue) electrostatic potentials; the values of the electrostatic potentials (in atomic units) are indicated in the scale at the left of the plot. An advantage of the use of the MSEPs is that allows a rapid visual comparison of results, readily identifying the different regions in a molecule.
Fig. 6. MSEPs for the four stable conformers (CLO-I, CLO-II, CLO-III, CLO-IV) of clotrimazole (CLO). Two views (a and b) of each conformer are depicted, showing the different electrostatic potential patterns in the two enantiotopic phenyl rings of CLO. The values of the EP ranges from +0.03 a.u. (deep red) to 0.03 a.u. (deep blue).
710
J.M. Navas et al. / Life Sciences 76 (2004) 699–714
Fig. 7. MSEP of the most stable conformer of 1-benzylimidazole (BIM).
Although the EP reflects both the electronic and nuclear charges, it can be as a representation of the electron density distribution on the molecules, particularly when the same atoms are compared. The regions with negative values of electrostatic potentials are electron-rich regions that would interact favorably with electrophilic and acidic targets. On analyzing the MSEPs of BIM (Fig. 7), and of the four conformers of CLO (Fig. 6), we observe quite similar trends. Thus, the region with the most negative electrostatic potential corresponds to the non-substituted nitrogen atom. Interestingly, the substituted nitrogen atom in the imidazole ring in all the compounds possesses a positive value of the electrostatic potential. The Fig. 6 shows two views for each conformer of CLO. Although the molecule of CLO is achiral, the two phenyl rings are enantiotopic (one is pro-R, the other is pro-S) (Eliel and Wilen, 1994), and they can be differentiated during the interaction with a chiral biomacromolecule (e. g., the aryl hydrocarbon receptor). It is worth to remark that the two enantiotopic phenyl rings of CLO show different EP patterns.
Discussion The present work shows that the imidazole derivative CLO is an inducer of CYP1A in rainbow trout hepatocytes, at both the enzymatic and transcriptional levels. Comparison with previous results on typical EROD inducers, such as h-NF and 3-methylcholanthrene (3MC), in rainbow trout hepatocytes allow to estimate the potency (measure of dose of inducer at which induction occurs) and efficacy (magnitude of induction) of CYP1A induction by CLO. The lowest concentrations of h-NF and 3MC for which a significant induction of EROD activity was observed were 0.024 AM and 0.39 AM, respectively (Navas and Segner, 2000). In the case of CLO this concentration was 1.56 AM. The maximal induction of EROD activity caused by h-NF and 3MC was about twice the EROD activity of that observed in controls: 35 pMol/mg/min in controls and around 75 to 80 pMol/mg/min in h-NF or 3MC treated cells (Navas and Segner, 2000). In the case of CLO the maximal induction observed was around 60 pMol/mg/min with respect to 36 pMol/mg/min detected in controls. The percent increase of CYP1A mRNA detected after treatment with CLO (31 F 7% with respect to controls) was similar to increases of CYP1A mRNA detected after treatment with h-NF (37 F 10% with respect to controls) (Navas et al., 2003). In the case of another imidazole derivative, BIM, the lowest concentration for which a significant EROD induction was observed, was also 1.56 AM (Navas et al., 2003) and the maximal values of EROD activity was around 80 pMol/mg/min for a BIM concentration of 3.12 AM. Taken together, these data indicate that CLO is an effective but weak inducer of CYP1A in cultured rainbow trout hepatocytes.
J.M. Navas et al. / Life Sciences 76 (2004) 699–714
711
The transcriptional activation of CYP1A by xenobiotics is typically mediated through ligand binding to the AhR (Hahn, 1998). The strength of the interaction between the ligand and the receptor is designated as affinity and is a property of both the ligand and the receptor molecule (Hestermann et al., 2000). Ligands with a high affinity for the AhR, for instance TCDD or h-NF, are hydrophobic polyannular aromatic (i.e. with two or more aromatic rings) compounds that are planar, or can take a planar structure (i. e., two aromatic rings in the same plane), and can be accommodated within a binding site of approximately 14 12 5 (2)3 (Waller and McKinney, 1995). In this work, computational studies were performed to estimate if the CLO molecule would be able to take a planar conformation. CLO shows a structure with four aromatic rings bonded to a tetrahedral carbon atom, causing a highly steric encumbrance on this atom. The computational modeling of CLO have yielded four stable conformers (CLO-I to CLO-IV, Fig. 5), and none of them have two aromatic rings in the same plane. Computational studies indicated that the energy content of a putative co-planar conformer is very high, resulting in an extremely unstable structure, due to interactions between the substituents at the ortho-positions in the aromatic rings. The stable conformers of CLO, CLO-I to CLOIV, are almost perfect propellers, a structural feature that has been found in organic compounds having aromatic fragments linked to fully saturated carbon (Tukada and Mochizuki, 2002), nitrogen (Yu et al., 2002), silicon (Lambert and Lin, 2001), or phosphorus (Whitnall et al., 1997). A common structural feature of this kind of compounds is that each aromatic ring is like the blade of the propeller (and, therefore, any blade is rotated with respect to the others), and this topological distribution permits to minimize destabilizing steric interactions. The conclusion from the computational analysis is that CLO is not a typical AhR ligand which can adopt a planar conformation. In line with this view is the experimental finding that the AhR antagonist a-NF was not able to inhibit CLO-mediated CYP1A induction, indicating that CLO is not behaving as a competitive AhR agonist. These data indicate that CLO is not a typical (i.e. planar) AhR ligand. Kafafi et al. (1993) have developed a model of AhR binding of aromatic xenobiotics in which receptor binding is not an exclusive function of planarity but is based on electron affinities, entropies and lipohilicities. Recently, Chana et al. (2002) have provided evidence that electrostatic properties, particularly MEP, are an important factor for the biological activity of PCBs, and that the possibility of a PCB congener to exist in a co-planar configuration is only a strengthening factor for bioactivity. The MEPs are not only a representation of the electronic surface of the molecules, but also they must be considered as valid three dimensional surfaces of the molecules, and it is through these surfaces that the molecules interact one with each other (Murray and Politzer, 1998). Therefore, we have calculated in the present work the MEP of CLO, and, additionally, of BIM, another imidazole compound which is able to induce CYP1A without being coplanar (Navas et al., 2003), in order to seek for molecular resemblances that could be indicative of similar mechanisms of action of both compounds. Interestingly, in both CLO and BIM the substituted nitrogen atom in the imidazole ring possesses a positive value of the electrostatic potential what, in principle, was unexpected on the basis of the higher electronegativity of nitrogen versus carbon as well as on the availability of one electron pair on the nitrogen atom. This effect can be rationalized assuming that this electron pair is engaged in the maintenance of the aromatic electron sextet in the imidazole ring. Qualitatively, the patterns of the electrostatic potential in BIM and in each conformer of CLO are quite similar: there is a peripheral electron rich region (that corresponds to the unsubstituted nitrogen atom) and, at least, one of the benzenic rings that has a nearly neutral or slightly positive value of the electrostatic potential. As a consequence both molecules will interact efficiently with acidic or electrophilic species that are
712
J.M. Navas et al. / Life Sciences 76 (2004) 699–714
present in biological targets. To conclude, our results have shown conformational and electrostatic similarities between BIM and CLO that are not shared by other CYP1A inducers. From a structural point of view CYP1A inducers could be classified in three groups: 1) Compounds having at least two aromatic rings in a plane, with a very low conformational mobility. This is the case of TCDD (Safe, 1990, 1995) or of h-NF, (Fig. 1) (Navas et al., 2003). Both compounds have been considered as typical AhR ligands and CYP1A inducers (Safe, 1990, 1995; Navas and Segner, 1998, 2000; Navas et al., 2003). 2) Molecules with conformational mobility, i. e., they can exist in a variety of conformers that are separated by small energy differences, so that the interconversion between them can be readily achieved. Some examples of this class are 3,3V,4,4V,5-pentachlorobiphenyl (PCB no 126, Fig. 1), 3,3V,4,4V,5,5V-hexachlorobiphenyl (PCB no 169, Fig. 1) (Chana et al., 2002). Although the coplanar structure of these compounds is not the most stable one, this conformation can be readily reached. These compounds have also been described as low affinity AhR ligands and CYP1A inducers in cultured rainbow trout hepatocytes (Anderson et al., 1996). 3) Compounds having at least two aromatic rings linked by a sp 3 -carbon in their structures. Due to the steric requirements of the tetrahedral carbon atom (i.e.: bond angles of ca 1098), the conformations having two aromatic rings in the same plane are energetically unfavorable, being unlikely that these molecules induce CYP1A after binding the AhR in a co-planar conformation. CLO and BIM would pertain to this specific class of CYP1A inducers. The specific characteristics of these compounds should be taken into account in the establishment of structure activity relationship (SAR) models to search for CYP1A inducers. It remains to be elucidated if CYP1A inducers such as BIM and CLO activate the AhR through direct interaction with the receptor molecule, or if the activation occurs indirectly, through kinase-dependent signaling pathways, as it has been suggested for benzimidazoles (Lesca et al., 1995; Backlund et al., 1997, 1999).
Conclusion In summary, our findings indicate that the imidazole derivative CLO is able to induce CYP1A in rainbow trout cultured hepatocytes without being a planar molecule, and that such induction is not inhibited by the AhR antagonist a-NF. These functional and structural features are shared with another imidazole derivative BIM (Navas et al., 2003) and computational studies have demonstrated that both compounds exhibit similar electrostatic properties. These data indicate that these molecules may activate the AhR and induce CYP1A by a mechanism different to classical ligand binding to the AhR. Further work must explore the capacity of different imidazole derivatives to induce CYP1A and examine the mechanism of such induction taking into account the variety of structures shown by this group of compounds.
Acknowledgments Jose´ Marı´a Navas holds a Ramo´n y Cajal contract from the Spanish Ministry of Science and Technology (MCYT). This work was financially supported by MCYT projects (REN2002-00639/
J.M. Navas et al. / Life Sciences 76 (2004) 699–714
713
GLO and BQU2001-2270), and by CAM (Comunidad Auto´noma de Madrid) project 07M/0032/ 2002.
References Anderson, M.J., Miller, M.R., Hinton, D.E., 1996. In vitro modulation of 17-h-estradiol-induced vitellogenin synthesis: Effects of cytochrome P4501A1 inducing compounds on rainbow trout (Oncorhynchus mykiss) liver cells. Aquatic Toxicology 34, 327 – 350. Backlund, M., Johansson, I., Mkrtchian, S., Ingelman-Sundberg, M., 1997. Signal transduction-mediated activation of the aryl hydrocarbon receptor in rat hepatoma H4IIE cells. Journal of Biological Chemistry 50, 31755 – 31763. Backlund, M., Weidolf, L., Ingelman-Sundberg, M., 1999. Structural and mechanistic aspects of transcriptional induction of cytochrome P450 1A1 by benzimidazole derivatives in rat hepatoma H4IIE cells. European Journal of Biochemistry 261, 66 – 71. Becke, A.D., 1993. Density-functional thermochemistry. III. The role of exact exchange. Journal of Chemical Physics 98, 5648 – 5652. Bergman, D.L., Laaksonen, L., Laaksonen, A., 1997. Visualization of solvation structures in liquid mixtures. Journal of Molecular Graphics 15, 301 – 306. Bondi, A., 1964. Van der Waals Volumes and Radii. Journal of Physical Chemistry 68, 441 – 451. Braunbeck, T., Segner, H., 2000. Isolation and cultivation of teleost hepatocytes. In: Berry, M.N., Edwards, A.M. (Eds.), The Hepatocyte Review. Kluwer, Amsterdam, pp. 49 – 71. Castillo, L.E., De la Cruz, E., Ruepert, C., 1997. Ecotoxicology and pesticides in tropical aquatic ecosystems in Central America. Environmental Toxicology and Chemistry 16, 41 – 51. Chana, A., Concejero, M.A., de Frutos, M., Gonza´lez, M.J., Herrado´n, B., 2002. Computacional studies on biphenyl derivatives. Analysis of the conformational mobility, molecular electrostatic potential, and dipole moment of chlorinated biphenyls: Searching for the rationalization of the selective toxicity of polychlorinated biphenyls. Chemical Research in Toxicology 15, 1514 – 1526. Denison, M.S., Whitlock, J.P., 1995. Xenobiotic-inducible transcription of cytochrome P450 genes. Journal of Biological Chemistry 270, 18175 – 18178. Denizeau, F., Marion, M., 1990. Toxicity of cadmium, copper, and mercury to isolated trout hepatocytes. Canadian Journal of Fisheries and Aquatic Sciences 47, 1038 – 1042. Eglo, O.M., Aspmi, R., Lode, O., 1994. Runoff and leaching experiments of dichlorprop, MCPA, propioconazole, dimethoate, and chlorsulfuron in outdoor lysimeters and field catchment areas. Norwegian Journal of Agricultural Sciences 13, 53 – 78. Eliel, E.L., Wilen, S.H., 1994. Stereochemistry of Organic Compounds. John Wiley & Sons, New York. Frisch, M.J., Trucks, G.W., Schlegel, H.B., Scuseria, G.E., Robb, M.A., Cheeseman, J.R., Zakrzewski, V.G., Montgomery Jr., J.A., Stratmann, R.E., Burant, J.C., Dapprich, S., Millam, J.M., Daniels, A.D., Kudin, K.N., Strain, M.C., Farkas, O., Tomasi, J., Barone, V., Cossi, M., Cammi, R., Mennucci, B., Pomelli, C., Adamo, C., Clifford, S., Ochterski, J., Petersson, G.A., Ayala, P.Y., Cui, Q., Morokuma, K., Salvador, P., Dannenberg, J.J., Malick, D.K., Rabuck, A.D., Raghavachari, K., Foresman, J.B., Cioslowski, J., Ortiz, J.V., Baboul, A.G., Stefanov, B.B., Liu, G., Liashenko, A., Piskorz, P., Komaromi, I., Gomperts, R., Martin, R.L., Fox, D.J., Keith, T., Al-Laham, M.A., Peng, C.Y., Nanayakkara, A., Challacombe, M., Gill, P.M.W., Johnson, B., Chen, W., Wong, M.W., Andres, J.L., Gonzalez, C., Head-Gordon, M., Replogle, E.S., Pople, J.A., 1998. Gaussian 98, Revision A.3. Gaussian Inc, Pittsburgh, PA. Hahn, M., 1998. The aryl hydrocarbon receptor. A comparative perspective. Comparative Biochemistry and Physiology 121 (C), 23 – 53. Halgren, T.A., 1996. Merck molecular force field I. Basis, form, scope, parameterization, and performance of MMFF94. Journal of Comparative Chemistry 17, 490 – 519. Hankinson, O., 1995. The aryl hydrocarbon receptor complex. Annual Reviews in Pharmacology and Toxicology 35, 307 – 340. Henry, M.J., Sisler, H.D., 1984. Effects of sterol biosynthesis-inhibiting (SBI) fungicides on cytochrome P-450 oxygenations in fungi. Pesticide Biochemistry and Physiology 22, 262 – 275. Hestermann, E.V., Stegeman, J.J., Hahn, M.E., 2000. Relative contributions of affinity and intrinsic efficacy to aryl hydrocarbon receptor ligand potency. Toxicology and Applied Pharmacology 168, 160 – 172.
714
J.M. Navas et al. / Life Sciences 76 (2004) 699–714
Kafafi, S.A., Afeefy, H.Y., Ali, A.H., Said, H.K., Kafafi, A.G., 1993. Binding of polychlorinated biphenyls to the aryl hydrocarbon receptor. Environmental Health Perspectives 101, 422 – 428. Kennedy, S.W., Jones, S.P., Bastien, L.J., 1995. Efficient analysis of cytochrome P4501A catalytic activity, porphyrins, and total proteins in chicken embryo hepatocyte cultures with a fluorescence plate reader. Analytical Biochemistry 226, 362 – 370. Kobayashi, Y., Matsuura, Y., Kotani, E., Fukuda, T., Aoyagi, T., Tobinaga, S., Yoshida, T., Kuroiwa, Y., 1993. Structural requirements for the induction of hepatic microsomal cytochrome P450 by imidazole- and pyridine-containing compounds in rats. The Journal of Biochemistry 114, 697 – 701. Koley, A.P., Buters, T.M., Robinson, R.C., Markowitz, A., Friedman, F.K., 1997. Differential mechanisms of cytochrome P450 inhibition and activation by a-naphthoflavone. Journal of Biological Chemistry 6, 3149 – 3152. Laaksonen, L., 1992. A graphics program for the analysis and display of molecular dynamics trajectories. Journal of Molecular Graphics 10, 33 – 34. Lambert, J.B., Lin, L., 2001. Propeller rotation of aryl groups in triarylsilanes and triarylstannanes. Magnetic Resonance in Chemistry 39, 714 – 718. Lesca, P., Peryt, B., Larrieu, G., Alvinerie, M., Galtier, P., Daujat, M., Maurel, P., Hoogenboom, L., 1995. Evidence for the ligand-independent activation of the Ah receptor. Biochemical and Biophysical Research Communications 209, 474 – 482. Murray, J.S., Sen, K., 1996. Molecular Electrostatic Potentials. Concepts and Applications. Elsevier, Amsterdam. Murray, J.S., Politzer, P., 1998. The molecular electrostatic potential: a tool for understanding and predicting molecular interactions. In: Sapse, A.M. (Ed.), Molecular Orbital Calculations for Biological Systems. Oxford University Press, New York, pp. 49 – 84. Navas, J.M., Segner, H., 1998. Antiestrogenic activity of anthropogenic and natural chemicals. Environmental Science and Pollution Research 5, 75 – 82. Navas, J.M., Segner, H., 2000. Antiestrogenicity of h-naphthoflavone and PAHs in cultured rainbow trout hepatocytes: evidence for a role of the arylhydrocarbon receptor. Aquatic Toxicology 51, 79 – 92. Navas, J.M., Segner, H., 2001. Estrogen-mediated suppression of cytochrome P4501A (CYP1A) expression in rainbow trout hepatocytes: role of estrogen receptor. Chemico-Biological Interactions 138, 285 – 298. Navas, J.M., Chana, A., Herrado´n, B., Segner, H., 2003. Induction of CYP1A by the N-imidazole derivative, 1benzylimidazole. Environmental Toxicology and Chemistry 22, 830 – 836. Rowlands, J.C., Gustafsson, J.A., 1997. Aryl hydrocarbon receptor-mediated signal transduction. Critical Reviews in Toxicology 27, 109 – 134. Safe, S., 1990. Polychlorinated biphenyls (PCBs), dibenzo-p-dioxins (PCDDs), dibenzofurans (PCDFs) and related compounds: environmental and mechanistic considerations which support the development of toxic equivalency factors (TEFs). Critical Reviews in Toxicology 21, 51 – 88. Safe, S., 1995. Modulation of gene expression and endocrine response pathways by 2,3,7,8-tetrachlorodibenzo-p-dioxin and related compounds. Pharmacology and Therapeutics 67, 247 – 281. Song, H., Shin, H.-S., 1998. The antifungal drug clotrimazole. Acta Crystallographica 54, 1675 – 1677. Tukada, H., Mochizuki, K., 2002. Structures and short contacts of F. . .O(nitro) and H. . .O(nitro) in 2,2-di(4-nitrophenyl)hexafluoropropane and 2,2-di(4-nitrophenyl)propane. Journal of Molecular Structure 643, 85 – 91. Waller, C.L., McKinney, J.D., 1995. Three-dimensional quantitative structure activity relationships of dioxins and dioxin-like compounds-model validation and Ah receptor characterization. Chemical Research in Toxicology 8, 847 – 857. Whitnall, M.R., Hii, K.K., Thornton-Pett, M., Kee, T.P., 1997. Conformational diastereoisomerism in tris-(2-alkylimino) triphenylphosphines. Journal of Organometallic Chemistry 529, 35 – 50. Young, D.C., 2001. Computational Chemistry: A Practical Guide for Applying Techniques to Real-World Problems. John Wiley & Sons, New York. Yu, H., Huang, Y., Zhang, W., Matsuura, T., Meng, J., 2002. Crystal structures of two new compounds containing 2,5-diaryl1,3,4-oxadiazole and triarylamine units. Journal of Molecular Structure 642, 53 – 62.