Homology modelling of human CYP2E1 based on the CYP2C5 crystal structure: investigation of enzyme–substrate and enzyme–inhibitor interactions

Homology modelling of human CYP2E1 based on the CYP2C5 crystal structure: investigation of enzyme–substrate and enzyme–inhibitor interactions

Toxicology in Vitro 17 (2003) 93–105 www.elsevier.com/locate/toxinvit Homology modelling of human CYP2E1 based on the CYP2C5 crystal structure: inves...

2MB Sizes 0 Downloads 16 Views

Toxicology in Vitro 17 (2003) 93–105 www.elsevier.com/locate/toxinvit

Homology modelling of human CYP2E1 based on the CYP2C5 crystal structure: investigation of enzyme–substrate and enzyme– inhibitor interactions D.F.V. Lewisa,*, B.G. Lakeb, M.G. Birdc, G.D. Loizoud, M. Dickinse, P.S. Goldfarba a

School of Biomedical and Life Sciences, University of Surrey, Guildford, Surrey GU2 7XH, UK b TNO BIBRA International Ltd, Woodmansterne Road, Carshalton, Surrey SM5 4DS, UK c Toxicology Division, ExxonMobil Biomedical Sciences Inc., Mettlers Road, CN 2350, East Millstone, NJ 08875-2350, USA d Health and Safety Laboratory, Broad Lane, Sheffield S3 7HQ, UK e GlaxoSmithKline Research and Development Ltd, Park Road, Ware, Hertfordshire SG12 0DP, UK Accepted 9 October 2002

Abstract The construction of a homology model of human cytochrome P450 2E1 (CYP2E1) is reported, based on the CYP2C5 crystallographic template. A relatively high degree of primary sequence homology (identity=59%), as expected for proteins of the same CYP family, ensured a straightforward generation of the 3-dimensional model due to relatively few deletions and insertions of amino acid residues with respect to the CYP2C5 crystal structure. Probing the CYP2E1 model with typical substrates of the enzyme showed a good agreement with experimental information in the form of positions of metabolism for substrates, and with sitedirected mutagenesis data on certain residues. Furthermore, quantitative relationships between substrate binding affinity and various structural parameters associated with the substrate molecules facilitated the formulation of a procedure for estimating relative binding energy and, consequently, Km or KD values towards the CYP2E1 enzyme. This method has been based on a consideration of the active site interactions between substrates and key amino acid residues lining the haem pocket, together with compound lipophilicity data from partition coefficients. # 2002 Elsevier Science Ltd. All rights reserved. Keywords: Cytochromes P450(CYP); Molecular modelling; Human CYP2E1 enzyme

1. Introduction The cytochromes P450 (CYP) constitute a superfamily of heme-thiolate enzymes, of which over 2700 individual members are currently known, and are present in most species for the Phase 1 metabolism of drugs and other xenobiotics, although endogenous roles have also been characterized (Gonzalez, 1992; Wrighton and Stevens, 1992; Lewis, 1996, 2001; Parkinson, 1996; Anzenbacher and Anzenbacherova, 2001; Guengerich, 2002). Enzymes of the CYP1, CYP2 and CYP3 families represent P450s most closely associated with the metabolism of foreign compounds in mammalian species, with the CYP2 family as a whole being involved in the Abbreviations: CYP2E1, cytochrome P450 2E1. * Corresponding author. Tel.: +44-1483-686477; fax: +44-1483300803. E-mail address: [email protected] (D.F.V. Lewis).

greater part of P450-mediated drug oxidations in man (Rendic and DiCarlo, 1997; Evans and Relling, 1999). Although the CYP2E1 enzyme only plays a relatively minor role (Rendic and DiCarlo, 1997) in human drug metabolism (approx. 4% involvement in total drug oxidations known to be mediated by P450s), it is apparent that a number of low molecular weight carcinogens and other toxicants undergo metabolic transformations via CYP2E-mediated pathways (Guengerich et al., 1991; Liu et al., 1993; Raucy et al., 1993; Terelius et al., 1993; Gonzalez and Gelboin, 1994; Kukielka and Cederbaum, 1994; Cai and Guengerich, 2001). Table 1 summarizes the metabolic properties of a number of typical human CYP2E1 substrates, including drugs, solvents and other industrial chemicals (Ronis et al., 1996; Rendic and DiCarlo, 1997). There is a relatively high degree of homology between human CYP2E1 and orthologous proteins from other mammalian species, such as the mouse and rat, although some species differences in

0887-2333/03/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved. PII: S0887-2333(02)00098-X

94

D.F.V. Lewis et al. / Toxicology in Vitro 17 (2003) 93–105

Table 1 Substrates of CYP2E1 and their metabolic properties Compound

Metabolic pathway

4-Nitrophenol Chlorzoxazone

2-Hydroxylation 6-Hydroxylation

Aniline Paracetamol Butadiene Ethanol Dimethylnitrosamine Benzene Halothane Lauric acid Tetrachloromethane

4-Hydroxylation N-Hydroxylation 1,2-Epoxidation 2-Hydroxylation N-Demethylation Epoxidation Dehalogenation o-1 Hydroxylation Dehalogenation

Acetone Toluene Styrene Salicylic acid

Methyl hydroxylation Methyl hydroxylation Vinyl epoxidation 5-Hydroxylation

Km (mm) 21 40 15 1290 200 11200 2000 25 35 130 162.5*, 57 900* 14.8 10 280

References Tassaneeyakul et al., 1993 Peter et al., 1990; Gillam et al., 1994 Bourrie´ et al., 1996 Chen et al., 1998 Duescher and Elfarra, 1994 Bell and Guengerich, 1997 Guengerich et al., 1991 Nedelcheva et al., 1999 Spracklin et al., 1997 Clarke et al., 1994 Paustenbach et al., 1988, Zangar et al., 2000 Johansson et al., 1986 Tassaneeyakul et al., 1996 Carlson et al., 2000 Dupont et al., 1999

Km=Michaelis constant (mm) for substrate metabolism. All Km values have been determined for human CYP2E1 apart from *, which relates to data generated for CYP2E1-mediated metabolism in the rat. General references: Ronis et al., 1996; Rendic and DiCarlo, 1997.

CYP2E1-mediated metabolism exist such as that shown by butadiene (Melnick and Kohn, 1995; Lewis et al., 1997, 2000; Bird et al., 2001). In addition, it has been reported that a small number of allelic variants of human CYP2E1 exist, although these appear to exhibit only minor alteration in catalytic activity relative to the wild-type (Ingelman-Sundberg, 2001). Some of the physicochemical properties of CYP2E1 substrates are presented in Table 2, from which it can be appreciated that the majority are neutral, small molecular weight compounds with relatively low log P values (where P is the octanol/water partition coefficient). Apart from their relatively small molecular size being a common factor, CYP2E1 substrates exhibit structural diversity, although some contain a single aromatic ring and this is also a feature shown by certain inhibitors of the enzyme (Hargreaves et al., 1994; Rodrigues, 1999), with 3-

amino-1,2,4-triazole representing a selective inhibitor of human CYP2E1 (Koop, 1990). The availability of the crystallographic co-ordinates for the rabbit enzyme CYP2C5, a mammalian enzyme from within the same family, has provided an opportunity for developing homology models of CYP2 enzymes (Lewis, 2002) and there is an interest in human CYP2E1 from the point of view of evaluating the likely metabolic fate of small molecular weight environmental agents in Homo sapiens. CYP2C5 exhibits significantly high primary sequence identity with CYP2E1 (59%) than that shown by the previously used template CYP102 (25%) in our earlier studies (Lewis et al., 1997, 2000). Consequently, we have utilized the CYP2C5 crystal structure as a template for modelling human CYP2E1 such that the interactions of typical substrates and inhibitors can be investigated.

Table 2 Distances between sites of metabolism and haem iron for human CYP2E1 substrates Compound

Route of metabolism

References

Distance (A˚)

1. 4-Nitrophenol 2. Chlorzoxazone 3. Aniline 4. Paracetamol 5. Lauric acid 6. Butadiene 7. Halothane 8. Ethanol 9. Dimethylnitrosamine 10. Toluene

2-Hydroxylation 6-Hydroxylation 4-Hydroxylation N-Hydroxylation o-1 Hydroxylation 1,2-Epoxidation Dehalogenation 2-Hydroxylation N-Demethylation Methyl hydroxylation

Tassaneeyakul et al., 1993 Peter et al., 1990 Bourrie´ et al., 1996 Chen et al., 1998; Patten, et al., 1992 Clarke et al., 1994 Duescher and Elfarra, 1994 Spracklin et al., 1997; Madan and Parkinson, 1996 Bell and Guengerich, 1997 Guengerich et al., 1991 Tassaneeyakul et al., 1996

5.666 5.675 5.675 6.240 3.248 5.381 6.047 4.920 4.857 4.615

D.F.V. Lewis et al. / Toxicology in Vitro 17 (2003) 93–105

2. Materials and methods Fig. 1 shows an alignment between CYP2E1 enzymes with those of other CYP2 family P450s, including that of CYP2C5 for which the crystal structure is known (Williams et al., 2000). According to this alignment there is a 59% primary sequence identity between human CYP2E1 and the CYP2C5 crystallographic template (Lewis, 2002) as can be expected for proteins

95

within the same CYP family. The alignment shown in Fig. 1 was constructed using the GCG software package (Genetics Computer Group, Madison, WI, USA) and involved a small amount of manual editing to comply with retention of secondary structural elements, such as the a-helices (Lewis, 2002). Using the Sybyl Biopolymer package (Tripos Associates, St. Louis, MO, USA) a 3-dimensional model of human CYP2E1 was constructed from the CYP2C5 crystallographic template in

Fig. 1. An alignment between protein sequences of several enzymes from the CYP2 family including human CYP2E1 and rabbit CYP2C5 used as the template for homology modelling.

96

D.F.V. Lewis et al. / Toxicology in Vitro 17 (2003) 93–105

accordance with the alignment presented as Fig. 1. This involved replacement of amino acid residues, together with a small number of deletions and insertions as required by the alignment, but this only involved very few residues in each case as has been described previously (Lewis, 2002). The raw structure was first refined in order to eliminate any unfavourable steric interactions caused by residue changes, and then energy minimized using molecular mechanics (Tripos force field) to give rise to a low energy conformation which

was consistent with known protein geometries. The final model was then probed using a number of selective substrates and inhibitor molecules, which were docked interactively within the putative active site of the enzyme. This process was facilitated by the likely location of a bound substrate previously docked in the CYP2C5 crystal structure (Williams et al., 2000) and from consideration of optimal contacts with complementary amino acid residues. All molecular modelling procedures were carried out on a Silicon Graphics

Fig. 1. (continued)

D.F.V. Lewis et al. / Toxicology in Vitro 17 (2003) 93–105

Indigo2 IMPACT 10000 high-resolution graphics workstation, operating under UNIX.

3. Results and discussion 3.1. Molecular modelling of substrates The three-dimensional structure of human CYP2E1 minimized smoothly over 100 iterative cycles of mole-

97

cular mechanics to give an optimized geometry of 1156.9 kcal/mol, which contained no regions of disallowed protein conformation. The final structure was investigated using typical marker substrates and selective inhibitors, as detailed below. Fig. 2 shows the relatively selective substrate chlorzoxazone positioned within the putative active site of CYP2E1, where favourable contacts with complementary amino acid residues orientate the substrate for 6-hydroxylation, which is the experimentally

Fig. 1. (continued)

98

D.F.V. Lewis et al. / Toxicology in Vitro 17 (2003) 93–105

observed site of metabolism (Peter et al., 1990; Lucas et al., 1999; Shimada et al., 1999; Easterbrook et al., 2001; Loizou and Cocker, 2001). In particular, phenylalanine205 and phenylalanine-293 in the alignment (see Fig. 1) are able to form p–p stacking interactions with the benzoxazole ring of the substrate; whereas, asparagine202 and threonine-298 enter into hydrogen bonding with the chlorzoxazone hydroxyl and chloro groups, respectively. These favourable contacts co-operatively

assist in positioning the substrate such that the 6hydrogen is directly over the haem iron at a distance of 5.7A˚, as shown in Fig. 2. Other substrates of human CYP2E1 also fit the putative active site in positions that are consistent with their reported metabolism. For example, 4-nitrophenol undergoes CYP2E1-mediated 2-hydroylation (Koop et al., 1989; Tassaneeyakul et al., 1993), and it is possible to show that the 4-nitrophenol molecule can be super-

Fig. 1. (continued)

D.F.V. Lewis et al. / Toxicology in Vitro 17 (2003) 93–105

imposed with chlorzoxazone such that their respective sites of hydroxylation overlay, together with making similar contacts with nearby amino acid residues. In this respect, phenylalanine-205 and phenylalanine-293 ‘sandwich’ the benzene ring of 4-nitrophenol, and hydrogen bonds to asparagine-202 and threonine-298 serve to anchor the 4-nitrophenol molecule via the nitro and hydroxyl groups, respectively, such that the 2-hydrogen lies 5.7A˚ from the haem iron. A template of substrates

99

that includes: aniline, paracetamol, butadiene, toluene, dimethylnitrosamine, ethanol, halothane and lauric acid, can be built up from the initial location of the superimposed structures of chlorzoxazone and 4-nitrophenol, and this is displayed in Fig. 3. In each case, the distances between known sites of metabolism on the substrate molecules and haem iron lie within the 3–6 A˚ range, as shown in Table 2. It is possible that a small number of solvating water molecules may be present

Fig. 1. (continued)

100

D.F.V. Lewis et al. / Toxicology in Vitro 17 (2003) 93–105

within the haem environment, following the binding of small molecular weight polar substrates (e.g. ethanol) in the CYP2E1 active site, and Fig. 3 shows that a bound water molecule can mediate hydrogen-bonded interactions between certain types of substrate and the sidechain of threonine-301.

Fig. 2. Chlorzoxazone is shown positioned within the putative active site of human CYP2E1, where a combination of hydrogen bonded and p–p stacking interactions orientate the substrate for 6-hydroxylation. Hydrogen bonds are displayed as dashed lines.

3.2. Inhibitor interactions Furthermore, typical inhibitors of CYP2E1 are also able to fit within the putative active site of the enzyme in a way which is consistent with their likely mode of interaction, namely, haem ligation. For example, 3-amino1,2,4-triazole (Koop, 1990) and pyridine (Hargreaves et al., 1994) both ligate the haem iron, in agreement with their Type II binding spectra, and these show distances of 3.6 and 5.6 A˚, respectively, between their heterocyclic nitrogen atoms and the iron atom of the haem moiety. Both of these inhibitors are also able to make favourable contacts with active site amino acid residues such as phenylalanine-205 and threonine-298, as mentioned previously with respect to substrate interactions. However, 3-amino-1,2,4-triazole also hydrogen bonds with the side-chain of asparagine-362, which corresponds to a site-directed mutagenesis position in CYP2C3v (Richardson and Johnson, 1994). Other inhibitors include diallylsulfide and diallyldisulfide (Tessier et al., 1999), and it is possible that S-oxidation is associated with their inhibitory activity. These small molecular

Fig. 3. A molecular template of ten CYP2E1 substrates including:- chlorzoxazone, 4-nitrophenol, aniline, paracetamol, butadiene, toluene, dimethylnitrosamine, ethanol, halothane and lauric acid, superimposed within the putative active site of human CYP2E1. The distance between each site of metabolism and the haem iron is listed in Table 2. Hydrogen bonds are displayed as dashed lines and residues are numbered according to the alignment with CYP2C5 shown in Fig. 1.

101

D.F.V. Lewis et al. / Toxicology in Vitro 17 (2003) 93–105 Table 3 CYP2E1 substrates physicochemical properties Compound

Log P

pKa

Log D7.4

Mr

Km (mm)

Gbind (kcal/mol)

1. 4-Nitrophenol 2. Paracetamol 3. Dimethylnitrosamine 4. Aniline 5. Benzene 6. Ethanol 7. Chlorzoxazone 8. Butadiene 9. Tetrachloromethane 10. Halothane 11. Lauric acid 12. Toluene

1.91 0.25 1.35 0.90 2.13 0.32 2.36c 1.99 2.83 2.3 4.60 2.73

Neutral Neutral Neutral 4.7b Neutral Neutral 8.3b Neutral Neutral Neutral 5.3a Neutral

1.91 0.25 1.35 .09 2.13 0.32 1.41c 1.99 2.83 2.30 2.50 2.73

139.12 151.2 74.10 93.14 78.11 46.1 169.57 54.09 153.84 197.39 200.31 92.14

21 1290 2000 15 25 11,200 40 300 57 500 130 14.8

6.6353 4.0985 3.8284 6.8426 6.5279 2.7671 6.2539 5.2469 6.0201 4.6824 5.5122 6.8508

Range of log P values=1.35–4.60. Average log P value=1.69.Gbind=RTlnKm where Km is the Michaelis constant (mm) for substrate metabolism. a=acidic. b=basic. c=calculated value (Pallas System, CompuDrug Ltd, Budapest, Hungary). log P=logarithm of the octanol/water partition coefficient. pKa=negative logarithm of the acid/base dissociation constant. log D7.4=logarithm of the ionization-corrected log P value (logarithm of the distribution coefficient at pH 7.4). Mr=relative molecular mass. Reference: Lewis et al., 2002.

weight compounds are also able to fit the putative active site of CYP2E1, giving additional support for a relatively small haem cavity in this enzyme (Mackman et al., 1996). Consequently, there is good agreement between the modelling results and experimental findings on CYP2E1 substrate and inhibitor selectivity, together with evidence from site-specific mutation studies in CYP2 family enzymes. For example, threonine-303 in CYP2E1 (corresponding to threonine-298 in the current model) has been shown to be important for fatty acid regioselectivity of metabolism (Fukuda et al., 1993) and this residue may play a role in hydrogen bond interactions with substrates.

3.3. Structure–activity relationships and estimation of binding affinity Table 3 indicates particular characteristics of selected CYP2E1 substrates and this information can be used to generate QSARs with substrate binding affinity to CYP2E1, as shown in Table 4. However, active site interactions provide useful structural descriptors for assessing relative binding affinity, and Table 5 presents similar information for CYP2E1 inhibitors where lipophilic character (in the form of compound log P values) appears to be important for inhibitory potency with congeneric series of primary alcohols and carboxylic

Table 4 Dataset for QSARs in CYP2E1 substrates Compound

NHB

Npp

HBa

Mr

log P

Gbbind

1. 4-Nitrophenol 2. Ethanol 3. Chlorzoxazone 4. Butadiene 5. Halothane 6. Dimethylnitrosamine 7. Tetrachloromethane 8. Aniline 9. Lauric acid 10. Benzene

4 1 2 0 1 1 0 2 2 0

1 0 1 0.5 0 0 0 1 0 1

4 2 4 0 3 2 0 2 2 0

139.12 46.1 169.57 54.09 197.39 74.10 153.84 93.14 200.31 78.11

1.91 0.32 2.36c 1.00 2.30 1.35 2.83 0.90 4.60 2.13

6.6353 2.7671 6.2539 5.2469 4.6824 3.8284 6.0201 6.8426 5.5122 6.5279

a=total number of hydrogen bond acceptors and donors. NHB=number of active site hydrogen bonds. b=RTlnKm, where Km is the Michaelis constant (mm) for substrate metabolism. Npp=number of active site p–p stacking interactions. c=calculated value (Pallas Software, CompuDrug Ltd, Budapest, Hungary). Mr=relative molecular mass. (continued on next page)

102

D.F.V. Lewis et al. / Toxicology in Vitro 17 (2003) 93–105

Table 4 (continued) Relationship

n

s

R

F

1. Gbind ¼ 2:19N  0:33logP  3:73 ( 0.39) ( 0.11)

9

0.5473

0.94

22.95

2. DGbind ¼ 0:41HB  2:59N  3:53 logMr  2:13 ( 0.10) ( 0.27) ( 0.53)

9

0.3269

0.98

42.17

3.DGbind ¼ 0:58HB  2:33N  3:42 logMr  0:32NHB þ 1:89 ( 0.09) ( 0.21) ( 0.36) ( 0.12)

9

0.2214

0.99

70.69

4. DGbind ¼ 0:33HB  2:39N  3:48 logMr þ 2:12 ( 0.09) ( 0.24) ( 0.56)

10

0.3478

0.98

40.86

5. DGbind ¼ 0:56HB  2:18N  3:37 logMr þ 0:36NHB þ 1:86 ( 0.09) ( 0.17) ( 0.37) ( 0.12)

10

0.2287

0.99

73.08

n=Number of points; s=standard error; R=correlation coefficient; F=variance ratio.

Table 5 Dataset for QSARs in CYP2E1 inhibitors 1. Primary alcohols

Log P

Ki (mm)

pKi

Methanol Ethanol Propanol Butanol Pentanol Hexanol Heptanol Octanol Nonanol Decanol

0.74 0.30 0.25 0.84 1.51 2.03 2.62 3.07 4.02 4.57

610 102 20 13 9 6 3 6 6 53

2.7853 2.0086 1.3010 1.1139 0.952 0.7782 0.4771 0.7782 0.7782 1.7243

2. Carboxylic acids

Log P

pKa

Log D7.4

Ki (mm)

pKi

Hexanoic acid Heptanoic acid Octanoic acid Nonanoic acid Decanoic acid Undecanoic acid Dodecanoic acid Tridecanoic acid Tetradecanoic acid Pentadecanoic acid Hexadecanoic acid

1.92 2.37 3.05 3.39 4.09 4.41 4.60 5.43 5.94 6.45 7.17

4.87 4.90 4.89 4.91 4.90 4.91 5.30 4.91 4.91 4.91 4.50

0.61 0.12 0.54 0.95 1.59 1.95 2.50 2.95 3.46 3.97 4.27

2610 328 162 113 66 23 22 45 120 296 340

3.4166 2.5159 2.2095 2.0531 1.8195 1.3617 1.3424 1.6532 2.0792 2.4713 2.5315

pKi=Log Ki where Ki is the constant for inhibition of CYP2E1 activity. pKa=Log Ka where Ka is the acid/base dissociation constant. Log D7.4=Logarithm of the distribution coefficient at pH 7.4 (logarithm of the ionization-corrected log P value). Log P=Logarithm of the octanol–water partition coefficient. Relationship

n

s

R

1. pKi ¼ 0:983logP  0:203logP 2  1:789 ( 0.107) ( 0.024)

10

0.2177

0.963

2 2. pKi ¼ 1:047logD7:4  0:244logD7:4  2:659 (0.096) ( 0.024)

11

0.1689

0.968

References to physicochemical and Ki data, respectively: Sangster, 1989; Wang et al., 1995.

F 89.4 119

103

D.F.V. Lewis et al. / Toxicology in Vitro 17 (2003) 93–105

(distal to the haem) is sterically hindered by the presence of these two relatively large amino acid residues, together with a substantial number of hydrophobic sidechains which line the haem pocket. Previous modelling studies (Wang et al., 1995; Lewis et al., 1997, 2000; Tan et al., 1997) tend to support this viewpoint, as does evidence from active site topology studies involving porphyrin adducts (Mackman et al., 1996). Furthermore, the significant relationship between lipophilicity and binding affinity for CYP2E1 substrates and inhibitors (see Tables 5 and 6, respectively) provides additional support to the important role of hydrophobic interactions within the CYP2E1 active site. However, it is likely that there are additional factors involved in the binding of some substrates, including hydrogen bonding and p–p stacking interactions with complementary aromatic residues in the CYP2E1 active site.

acids. This is further emphasized by a relationship between log P and binding affinity for six human P450 substrates (see Table 6), whereas Table 7 shows that it is possible to obtain satisfactory estimates of CYP2E1 binding affinities for six typical substrates, based on an analysis of the various contributions to Gbind. This analysis employed a method which has been reported previously for CYP2 family substrates (Lewis, 2002). The relatively restricted substrate binding site of CYP2E1, where two phenylalanine residues (phenylalanine-205 and phenylalanine-293) occupy positions directly above the haem moiety, could explain the observation (Guengerich and Johnson, 1997) that this enzyme is predominantly in the high-spin state even without the presence of bound substrate. It is possible that access for water molecules to the sixth ligand site Table 6 Baseline lipophilicity relationship for CYP2E1 substrates Compound

Log P

Gpart

Km (mm)

Gbind

1. Toluene 2. Styrene 3. o-Xylene 4. m-Xylene 5. Chlorzoxazone 6. Butadiene

2.73 3.05 3.12 3.2 2.36c 1.99

3.8724 4.3263 4.4256 4.5391 3.3476 2.8227

14.8 10 16.4 11.8 40 20

6.8508 7.0923 6.7876 6.9904 6.2539 5.2469

c=calculated vale (Pallas System, CompuDrug Ltd, Budapest, Hungary). Gpart=RTlnP where P is the octanol–water partition coefficient. Gbind=RTlnKm where Km is the Michaelis constant (mm) for substrate metabolism. Relationship

n

s

R

F

Gbind=0.942Gpart2.876

6

0.2968

0.92

23.5

References to Km data: Tassanneyakul et al., 1996; Carlson et al., 2000; Peter et al., 1990; Duescher and Elfarra, 1994.

Table 7 Comparison between experimental and calculated binding energies for CYP2E1 substrates Compound

Log P

Gpart

Ghb

Gpp

Grot

DGcalc bind

a DGexpt bind

1. 4-Nitrophenol 2. Chlorzoxazone 3. Tetrachloromethane 4. Paracetamol 5. Aniline 6. Halothane

1.91 2.36c 2.83 0.25 0.9 2.30

2.7093 3.3476 4.0143 0.3546 1.2766 3.2625

4.0 2.0 2.0 4.0 4.0 2.0

0.9 0.9 0 0.9 0.9 0

0.6 0 0 0.6 0 0.6

7.0093 6.2476 6.0143 4.6546 6.1766 4.6625

6.6353 6.2539 6.0201 4.0985 6.8426 4.6824

Gpart=RTlnP where P is the octanol–water partition coefficient. Ghb=Hydrogen bond interaction energy (2 kcal/mol per hydrogen bond). Gpp=p–p stacking energy (0.9 kcalmol1 per p–p stacking interaction). Grot=rotatable bond energy (0.6 kcal/mol per rotatable bond restricted on binding). DGcalc bind =Gpart+Ghb+Gpp+Grot. DGexpt bind =RTlnKm where Km is the Michaelis constant (mm). c=Calculated value (Pallas Software, CompuDrug Ltd, Budapest, Hungary). a The agreement between calculated and experimental Gbind values is 0.93 with a slope of unity.

104

D.F.V. Lewis et al. / Toxicology in Vitro 17 (2003) 93–105

4. Conclusions The homology model of human CYP2E1 appears to show consistency with available experimental information in the form of substrate selectivity and position of metabolism, together with the findings of site-directed mutagenesis studies within CYP2 family enzymes. Furthermore, it is possible to utilize modelling information to derive quantitative relationships between substrate binding affinity (related to Km values via the relationship G=RTlnKm) and structural features on the substrates themselves, in addition to providing a means of estimating CYP2E1 substrate binding affinity from a summation of various contributions to the overall binding interaction. In this respect, it is apparent that hydrogen bonding and p–p stacking play important roles, together with desolvation of the haem environment as estimated from compound lipophilicity (log P values). Further work is currently in progress for evaluating the rates of metabolism and binding affinities for CYP2E1 substrates such as alkylbenzenes.

Acknowledgements The financial support of GlaxoSmithKline Research & Development Limited, Merck Sharp & Dohme Ltd and the University of Surrey Foundation Fund is gratefully acknowledged by DFVL.

References Anzenbacher, P., Anzenbacherova, E., 2001. Cytochromes P450 and metabolism of xenobiotics. Cellular and Molecular Life Sciences 58, 737–747. Bell, L.C., Guengerich, F.P., 1997. Oxidation kinetics of ethanol by human cytochrome P450 2E1. Journal of Biological Chemistry 272, 29643–29651. Bird, M.G., Lewis, D.F.V., Whitman, F.T., Lewis, J., Przygoda, R.T., Witz, G., 2001. Application of process chemistry and SAR Modelling to the evaluation of health findings of lower olefins. ChemicoBiological Interactions 135/136, 571–584. Bourrie´, M., Meunier, V., Berger, Y., Fabre, G., 1996. Cytochrome P450 isoform inhibitors as a tool for the investigation of metabolic reactions catalyzed by human liver microsomes. Journal of Pharmacology and Experimental Therapeutics 277, 321–332. Cai, H., Guengerich, F.P., 2001. Reaction of trichloroethylene and trichloroethylene oxide with cytochrome P450 enzymes: inactivation and sites of modification. Chemical Research in Toxicology 14, 451– 458. Carlson, G.P., Mantick, N.A., Powley, M.W., 2000. Metabolism of styrene by human liver and lung. Journal of Toxicology and Environmental Health 59, 591–595. Chen, W., Koenigs, L.L., Thompson, S.J., Peter, R.M., Rettie, A.E., Trager, W.F., Nelson, S.D., 1998. Oxidation of acetamidophen to its toxic quinone imine and nontoxic catechol by baculovirusexpressed and purified human cytochromes P450 2E1 and 2A6. Chemical Research in Toxicology 11, 295–301. Clarke, S.E., Baldwin, S.J., Bloomer, J.C., Ayrton, A.D., Sozio, R.S.,

Chenery, R.J., 1994. Lauric acid as a model substrate for the simultaneous determination of cytochrome P450 2E1 and 4A in hepatic microsomes. Chemical Research in Toxicology 7, 836–842. Duescher, R.J., Elfarra, A.A., 1994. Human liver microsomes are efficient catalysts of 1,3-butadiene oxidation: evidence for major roles by cytochromes P450 2A6 and 2E1. Archives of Biochemistry and Biophysics 311, 342–349. Dupont, I., Berthou, F., Bodenez, P., Bardou, L., Guirriec Stephan, N., Dreano, Y., Lucas, D., 1999. Involvement of cytochromes P450 2E1 and 3A4 in the 5-hydroxylation of salicylate in humans. Drug Metabolism and Disposition 27, 322–326. Easterbrook, J., Fackett, D., Li, A.P., 2001. A comparison of aroclor 1254-induced and uninduced rat liver microsomes to human liver microsomes in phenacetin O-deethylation, coumarin 7-hydroxylation, tolbutamide 4-hydroxylation, S-mephenytoin 40 -hydroxylation, chlorzoxazone 6-hydroxylation and testosterone 6bhydroxylation. Chemico-Biological Interactions 134, 243–249. Evans, W.E., Relling, M.V., 1999. Pharmacogenomics: translating functional genomics into rational therapeutics. Science 286, 489– 491. Fukuda, T., Imai, Y., Komori, M., Nakamura, M., Kusunose, E., Satouchi, K., Kusunose, M., 1993. Replacement of Thr-303 of P4502E1 with serine modifies the regioselectivity of its fatty acid hydroxylase activity. Journal of Biochemistry 113, 7–12. Gillam, E.M.J., Guo, Z., Guengerich, F.P., 1994. Expression of modified human cytochrome P4502E1 in Escherichia coli, purification, and spectral and catalytic properties. Archives of Biochemistry and Biophysics 312, 59–66. Gonzalez, F.J., 1992. Human cytochromes P450: problems and prospects. Trends in Pharmaceutical Science 13, 346–352. Gonzalez, F.J., Gelboin, H.V., 1994. Role of human cytochromes P450 in the metabolic activation of chemical carcinogens and toxins. Drug Metabolism Reviews 26, 165–183. Guengerich, F.P., 2002. Cytochrome P450.. In: Ioannides, C. (Ed.), Enzyme Systems that Metabolize Drugs and other Xenobiotics. Wiley, New York, pp. 33–65. Chapter 2. Guengerich, F.P., Johnson, W.W., 1997. Kinetics of ferric cytochrome P450 reduction by NADPH-cytochrome P450 reductase: rapid reduction in the absence of substrate and variations among cytochrome P450 systems. Biochemistry 36, 14741–14750. Guengerich, F.P., Kim, D.H., Iwasaki, M., 1991. Role of human cytochrome P450IIE1 in the oxidation of several low molecular weight cancer suspects. Chemical Research in Toxicology 4, 168– 179. Hargreaves, M.B., Jones, B.C., Smith, D.A., Gescher, A., 1994. Inhibition of p-nitrophenol hydroxylase in rat liver microsomes by small aromatic and heterocyclic molecules. Drug Metabolism and Disposition 22, 610–806. Ingelman-Sundberg, M., 2001. Implications of polymorphic cytochrome P450-dependent drug metabolism for drug development. Drug Metabolism and Disposition 29, 570–573. Johansson, I., Eliasson, E., Norsten, C., Ingelman-Sundberg, M., 1986. Hydroxylation of acetone by ethanol- and acetone-inducible cytochrome P450 in liver microsomes and reconstituted membranes. FEBS Letters 196, 59–64. Koop, D.R., 1990. Inhibition of ethanol-inducible cytochrome P450IIE1 by 3-amino-1,2,4-triazole. Chemical Research in Toxicology 3, 377–383. Koop, D.R., Laethem, C.L., Tierney, D.J., 1989. The utility of pnitrophenol hydroxylation in P450IIE1 analysis. Drug Metabolism Reviews 20, 541–551. Kukielka, E., Cederbaum, A.I., 1994. DNA strand cleavage as a sensitive assay for the production of hydroxyl radicals by microsomes: role of cytochrome P4502E1 in the increased activity after ethanol treatment. Biochemical Journal 302, 773–779. Lewis, D.F.V., 1996. Cytochromes P450: Structure, Function and Mechanism. Taylor & Francis, London.

D.F.V. Lewis et al. / Toxicology in Vitro 17 (2003) 93–105 Lewis, D.F.V., 2001. Guide to Cytochromes P450 Structure and Function. Taylor & Francis, London. Lewis, D.F.V., 2002. Homology modelling of huma CYP2 family enzymes based on the CYP2C5 crystal structure. Xenobiotica 32, 305–323. Lewis, D.F.V., Bird, M.G., Parke, D.V., 1997. Molecular modelling of CYP2E1 enzymes for mouse, rat and man: an explanation for species differences in butadiene metabolism and carcinogenicity, and rationalization of CYP2E substrate specificity. Toxicology 118, 93– 113. Lewis, D.F.V., Bird, M.G., Dickins, M., Lake, B.G., Eddershaw, P.J., Tarbit, M.H., Goldfarb, P.S., 2000. Molecular modelling of human CYP2E1 by homology with the CYP102 haemoprotein domain: investigations of the interactions of substrates and inhibitors within the putative active site of the human CYP2E1 isoform. Xenobiotica 30, 1–25. Lewis, D.F.V., Modi, S., Dickins, M., 2002. Quantitative structureactivity relationships in substrates of human hepatic microsomal cytochromes P450. Drug Metabolism Reviews 34, 69–82. Liu, P.T., Ioannides, C., Shavida, J., Symons, A.M., Parke, D.V., 1993. Effects of ether anaesthesia and fasting on various cytochromes P450 of rat liver and kidney. Biochemical Pharmacology 45, 871–877. Loizou, G.D., Cocker, J., 2001. The effects of alcohol and diallyl sulphide on CYP2E1 activity in humans: a phenotyping study using chlorzoxazone. Human and Experimental Toxicology 20, 321–327. Lucas, D., Ferrara, R., Gonzalez, E., Bodenez, P., Albores, A., Manno, M., Berthou, F., 1999. Chlorzoxazone, a selective probe for phenotyping CYP2E1 in humans. Pharmacogenetics 9, 377–388. Mackman, R., Guo, Z., Guengerich, F.P., Ortiz de Montellano, P.R., 1996. Active site topology of human cytochrome P450 2E1. Chemical Research in Toxicology 9, 223–226. Madan, A., Parkinson, A., 1996. Characterization of the NADPHdependent covalent binding of [14C]halothane to human liver microsomes. Drug Metabolism and Disposition 24, 1307–1313. Melnick, R.L., Kohn, M.L., 1995. Mechanistic data indicate that 1,3butadiene is a human carcinogen. Carcinogenesis 16, 157–163. Nedelcheva, V., Gut, I., Soucek, P., Tichavska, B., Tynkova, L., Mraz, J., Guengerich, F.P., Ingelman-Sundberg, M., 1999. Metabolism of benzene in human liver microsomes: individual variations in relation to CYP2E1 expression. Archives of Toxicology 73, 33– 40. Parkinson, A., 1996. An overview of current cytochrome P450 technology for assessing the safety and efficacy of new materials. Toxicologic Pathology 24, 45–57. Patten, C.J., Ishizaki, H., Aoyama, T., Lee, M., Ning, S.M., Huang, W., Gonzalez, F.J., Yang, C.S., 1992. Catalytic properties of the human cytochrome P450 2E1 produced by cDNA expression in mammalian cells. Archives of Biochemistry and Biophysics 299, 163–171. Paustenbach, D.J., Clewell, I.H.J., Gargas, M.L., Andersen, M.E., 1988. A physicologically based pharmacokinetics model for inhaled carbon tetrachloride. Toxicology and Applied Pharmacology 96, 191–211. Peter, R., Bocker, R., Beaune, P.H., Iwasaki, M., Guengerich, F.P., Yang, C.S., 1990. Hydroxylation of chlorzoxazone as a specific probe for human liver cytochrome P450IIE1. Chemical Research in Toxicology 3, 566–573. Raucy, J.L., Kramer, K.C., Lasker, J.M., 1993. Bioactivation of halogenated hydrocarbons by cytochrome P4502E1. Critical Reviews in Toxicology 23, 1–20.

105

Rendic, S., DiCarlo, F.J., 1997. Human cytochrome P450 enzymes: a status report summarizing their reactions, substrates, inducers and inhibitors. Drug Metabolism Reviews 29, 413–580. Richardson, T.H., Johnson, E.F., 1994. Alterations of the regiospecificity of progesterone metabolism of two key amino acid residues in rabbit cytochrome P450 2C3v. Journal of Biological Chemistry 269, 23927–23943. Rodrigues, A.D., 1999. Integrated cytochrome P450 reaction phenotyping: attempting to bridge the gap between cDNA-expressed cytochromes P450 and native human liver microsomes. Biochemical Pharmacology 57, 465–480. Ronis, M.J.J., Lindros, K.O., Ingelman-Sundberg, M., 1996. The CYP2E subfamily. In: Ioannides, C. (Ed.), Cytochromes P450Metabolic and Toxicological Aspects. CRC Press, Boca Raton, FL, pp. 211–239. Chapter 9. Sangster, J., 1989. Octanol-water partition coefficients of simple organic compounds. Journal of Physical and Chemical Reference Data 18, 1111–1229. Shimada, T., Tsumura, F., Yamazaki, H., 1999. Prediction of human liver microsomal oxidations of 7-ethoxycoumarin and chlorzoxazone with kinetic parameters of recombinant cytochrome P450 enzymes. Drug Metabolism and Disposition 27, 1274–1280. Spracklin, D.K., Hankins, D.C., Fisher, J.M., Thummel, K.E., Karasch, E.D., 1997. Cytochrome P450 2E1 is the principal catalyst of human oxidative halothane metabolism in vitro. Journal of Pharmacology and Experimental Therapeutics 281, 400–411. Tan, Y., White, S.P., Paranawithana, S.R., Yang, C.S., 1997. A hypothetical model for the active site of human cytochrome P4502E1. Xenobiotica 27, 287–299. Tassaneeyakul, W., Birkett, D.J., Edwards, J.W., Veronese, M.E., Tassaneeyakul, W., Tukey, R.H., Miners, J.O., 1996. Human cytochrome P450 isoform specificity in the regioselective metabolism of toluene and o-, m- and p-xylene. Journal of Pharmacology and Experimental Therapeutics 273, 101–108. Tassaneeyakul, W., Veronese, M.E., Birkett, D.J., Gonzalez, F.J., Miners, J.O., 1993. Validation of 4-nitrophenol as an in vitro substrate probe for human liver CYP2E1 using cDNA expression and microsomal kinetic techniques. Biochemical Pharmacology 46, 1975–1981. Terelius, Y., Lindros, K.O., Albano, E., Ingelman-Sunberg, M., 1993. Isoenzyme-specificity of cytochrome P450 mediated hepatotoxicity. Frontiers in Biotransformation 8, 186–232. Tessier, C., Guenot, L., Suschetet, M., Siess, M.-H., 1999. Metabolism of dialkyl disulfide by human liver microsomal cytochromes P450 and flavin-containing monooxygenases. Drug Metabolism and Disposition 27, 835–841. Wang, M.-H., Wade, D., Chen, L., White, S., Yang, C.S., 1995. Probing the active sites of rat and human cytochrome P450 2E1 with alochols and carboxylic acids. Archives of Biochemistry and Biophysics 317, 299–304. Williams, P.A., Cosme, J., Sridhar, V., Johnson, E.F., McRee, D.E., 2000. Mammalian cytochrome P450 monooxygenase: structural adaptations for membrane binding and functional diversity. Molecular Cell 5, 121–131. Wrighton, S.A., Stevens, J.C., 1992. The human hepatic cytochromes P450 involved in drug metabolism. Critical Reviews in Toxicology 22, 1–21. Zangar, R.C., Benson, J.M., Burnett, V.L., Springer, D.L., 2000. Cytochrome P4502E1 is the primary enzyme responsible for lowdose carbon tetrachloride metabolism in human liver microsomes. Chemico-Biological Interactions 125, 233–243.