PEDS Advance Access published online on December 6, 2007
Protein Engineering Design and Selection, doi:10.1093/protein/gzm063
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Helicoverpa zea CYP6B8 and CYP321A1: different molecular solutions to the problem of metabolizing plant toxins and insecticides
Departments of Cell and Developmental Biology, Biochemistry and Plant Biology, University of Illinois at Urbana-Champaign, 1201 W. Gregory Dr., 161 Edward R. Madigan Laboratory (ERML), Urbana, IL 61801, USA
1 To whom correspondence should be addressed. E-mail: maryschu{at}uiuc.edu
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and methodsResultsDiscussionAcknowledgementsReferences |
|---|
Under continual exposure to naturally occurring plant toxins and synthetic insecticides, insects have evolved cytochrome P450 monooxygenases (P450s) capable of metabolizing a wide range of structurally different compounds. Two such P450s, CYP6B8 and CYP321A1, expressed in Helicoverpa zea (a lepidopteran) in response to plant allelochemicals and plant signaling molecules metabolize these compounds with varying efficiencies. While sequence alignments of these proteins indicate highly divergent substrate recognition sites (SRSs), homology models developed for them indicate that the two active site cavities have essentially the same volume with distinct shapes dictated by side-chain differences in SRS1 and SRS5. CYP6B8 has a narrower active site cavity extending from substrate access channel pw2a with a very narrow access to the ferryl oxygen atom. This predicted shape suggests that bulkier molecules bind further from the ferryl oxygen at positions that are not as effectively metabolized. In contrast, CYP321A1 is predicted to have a more spacious cavity allowing larger molecules to access the heme-bound oxygen. The metabolic profiles for several plant toxins (xanthotoxin, angelicin) and insecticides (cypermethrin, aldrin and diazinon) correlate well with these predictive models. The absence of Thr in the I helix of CYP321A1 and hydroxyl groups on many of its substrates suggests that this insect P450 mediates oxygen activation by a mechanism different from that employed by CYP107A1 and CYP158A1, which are two bacterial P450s also lacking Thr in their I helix, and most other P450s that contain Thr in their I helix.
Keywords: cytochrome P450 monooxygenases (P450s)/detoxification of plant toxins and insecticides/molecular modeling
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionAcknowledgementsReferences |
|---|
In order to survive, insects must deal with both naturally occurring plant toxins and synthetic insecticides in their diets. From the inception of synthetic insecticide usage in the 1940s, insects have been exposed to several major classes of insecticides including DDT, organochlorines (DDT, cyclodienes), organophosphates, carbamates, pyrethroids and insect growth regulators. The continual use of a limited number of these compounds has resulted in the evolution of insecticide resistances in >500 insect species mediated variously by esterases, glutathione-S-transferases (GSTs) and coupled cytochrome P450 monooxygenase (P450): NADPH-dependent P450 reductase systems (Feyereisen, 1999
, 2005; Scott, 1999; Sheehan et al., 2001; Ranson et al., 2002; Li et al., 2007). Of these, detoxification by coupled P450: P450 reductase systems represents one of the most prominent routes responsible for the inactivation of synthetic and natural toxins.
Reflecting a wide diversity in their actual metabolic capabilities, P450 genes and proteins within individual insects are numerous with as few as 47 in Apis mellifera (honeybee), 85 in Drosophila melanogaster (fruit fly) and as many as 100 in Anopheles gambiae (mosquito) (http://drnelson.utmem.edu/CytochromeP450.html; http://p450.antibes.inra.fr/; Ranson et al., 2002; Claudianos et al., 2006). It is as yet unclear how many P450 loci exist in lepidopteran genomes since none of these have been completely sequenced. Sequence identities within the large P450 families within these organisms are limited, giving rise to a nomenclature system that numerically differentiates families of related P450s with >40% amino acid identity (CYP6, CYP321, etc.) and alphabetically differentiates subfamilies with >55% amino acid identity (CYP6A, CYP6B, etc.) (Nelson et al., 1996). With this level of amino acid divergence, individual P450s within a subfamily have potential to metabolize completely different substrates or very similar substrates making it difficult to predict substrate preferences based solely on primary sequence identities. Exemplifying one end of the catabolic spectrum are mouse CYP2A4 and CYP2A5 that differ substantially in their metabolic activities (CYP2A4, testosterone 15
-hydroxylase; CYP2A5, coumarin 7-hydroxylase) in spite of only 11 variations in their 494 amino acid proteins (Negishi et al., 1996). At the middle of the catabolic spectrum are Papilio polyxenes CYP6B1 and Helicoverpa zea CYP6B8 proteins that at 53% identity metabolize xanthotoxin, a naturally occurring toxic furanocoumarin, at 5-fold different efficiencies (22.2 µmol/µmol P450 per minute for CYP6B1, 3.7 µmol/µmol P450 per minute for CYP6B8) and cypermethrin, a pyrethroid insecticide, at completely different efficiencies (no metabolism for CYP6B1, 12.7 µmol/µmol P450 per minute for CYP6B8) (Li et al., 2004). And, at completely the other end of the spectrum are the previously mentioned H. zea CYP6B8 and CYP321A1 proteins that have evolved to metabolize xanthotoxin (CYP6B8 cited above, 4.2 µmol/µmol P450 per minute for CYP321A1) and cypermethrin (3.2 µmol/µmol P450 per minute for CYP321A1) despite only 32% identity between these two H. zea proteins (Sasabe et al., 2004).
From a variety of site-directed mutagenesis studies on closely related P450 proteins, several substrate recognition sites (designated SRS1–6) (Gotoh, 1992) have been identified in site-directed mutagenesis experiments as important for substrate metabolism. With 11 -helices (designatedA–K) and 4 β-pleated sheets (designated 1–4) conserved in most bacterial, insect, plant and vertebrate P450s (Graham and Peterson, 1999; Poulos and Johnson, 2005; Rupasinghe and Schuler, 2006), the loop region between the B and C helices positioned over the heme corresponds to SRS1, the F and G helices comprising part of the bacterial substrate access channel contain SRS2 and SRS3, the I helix extending over the heme pyrrole ring B contains SRS4, the N-terminus of β-sheet 1–4 and the β-turn at the end of β-sheet 4 protruding into the catalytic site contain SRS5 and SRS6, respectively.
Our initial comparisons between the specialist P. polyxenes CYP6B1 and generalist H. zea CYP6B8 proteins, which were conducted to molecularly define the evolutionary progression from allelochemical-metabolizing P450 to insecticide-metabolizing P450, indicated that replacements of three aromatic residues within SRS2, SRS5 and SRS6 of the constrained CYP6B1 active site with small hydrophobic residues significantly expanded the volume of the predicted CYP6B8 catalytic site (Li et al., 2004). Presumably as a result of these changes, the generalist CYP6B8 protein was capable of metabolizing a more diverse range of allelochemicals and insecticides than the specialist CYP6B1 protein. To define the molecular/evolutionary relationships among allelochemical- and insecticide-metabolizing P450s further, we have directly compared the significantly more divergent CYP6B8 and CYP321A1 proteins that exist within the same species and metabolize many of the same allelochemicals and insecticides (Li et al., 2004; Sasabe et al., 2004). Comparative activity profiling and molecular modeling of these two proteins presented here indicate that their clearance rates for many of these compounds differ substantially because of numerous differences in highly variable as well as conserved positions in their catalytic sites.
|
Materials and methods |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionAcknowledgementsReferences |
|---|
Sequence alignments
The sequence used for modeling CYP6B8v1 corresponded to Genbank accession AAF06 102.1 (Li et al., 2000). The sequence used for modeling CYP321A1 corresponded to Genbank accession AAM54724
[GenBank]
.1 (Sasabe et al., 2004) with Phe421 and Phe447 corrected to Ser421 and Ser447 because of errors in the original sequence reads.
Development of the CYP6B8v1 and CYP321A1 models proceeded by first aligning seven class II P450 crystal structures [human CYP3A4 (1TQN) (Yano et al., 2004), CYP2C8 (1PQ2) (Schoch et al., 2004), CYP2A6 (1Z1O) (Yano et al., 2005), CYP2C5 (1N6B) (Wester et al., 2003), CYP2C9 (1OG5) (Williams et al., 2003), rabbit CYP2B4 (1SUO) (Scott et al., 2004), Bacillus megaterium CYP102 (2HPD) (Ravichandran et al., 1993) and Thermus thermophilus CYP175A1 (1N97) (Yano et al., 2003)] using the ALIGN facility in MOE versions 2004 and 2005 (Chemical Computing Group Inc., Montreal, Canada) that uses the BLOSUM62 scoring matrix. The CYP6B8v1 and CYP321A1 sequences were then aligned to this fixed alignment and 10 models were generated for each target sequence using the CYP3A4 crystal structure as the primary template for most of the backbone and coordinates from the CYP2B4 crystal structure for the highly divergent B’ helix and B–C loop region as well as the F–G loop region as detailed in Baudry et al. (2006). Heme coordinates were copied to the best model from the CYP102 crystal structure with a covalent bond created between the heme’s iron atom and the sulfur of the conserved cysteine axial ligand. Further energy minimizations for each model were performed using the CHARMm22 force field (MacKerell et al., 1998) in the MOE distribution until the final energy gradient was <0.01 kcal/mol per Å. A distance-dependant dielectric constant was used in the calculations with a cutoff between 6.5 and 7.0 Å. The best model for each protein was further tested by Profiles 3D 3D–1D analysis provided in Insight II (MSI, San Diego, CA, USA) and Prosa II analysis (Center for Applied Molecular Engineering, University of Salzburg, Austria). No regions of low quality were identified in the CYP6B8v1 and CYP321A1 models while using Profiles 3D 3D-1D analysis and Prosa II analysis.
The binding site volumes for these models were defined using the Active Site Finder function in MOE. The sites located either on the surface of the models or underneath the heme were removed from consideration and the volumes were calculated by running the MOE site volume calculation SVL code on the remaining sites. The binding site volumes were predicted to be 2014 Å3 for the CYP321A1 model and 2044 Å3 for the CYP6B8v1 model. Putative substrate access channels, identified also using the Active Site Finder function in MOE, in both models correspond to pathway pw2a designated as in Wade et al. (2004).
Dockings of prospective ligands were performed using LigandFit implemented in the program Cerius2 (Version 4.10, Accelrys, San Diego, USA). In short, LigandFit employs a cavity detection algorithm combined with a fast ligand conformational search engine, which allows the optimization of the bond structure of a ligand in a protein cavity, and calculations of several protein/ligand docking scores. For each potential substrate (xanthotoxin, angelicin, -naphthoflavone, -cypermethrin, aldrin and diazinon), 100 possible docking conformations were generated and ranked according to the Dock score provided in LigandFit. The binding conformation with the highest score (lowest interaction energy) and appropriate hydroxylation site closest to the heme was selected as the optimal conformation and subjected to further energy minimization using the MMFF94 force field in MOE, heme coordinates were fixed to prevent distortion of the heme plane originating from bonded parameters in the MOE’s implementation of the MMFF94 force field.
Heterologous expression and metabolism
Recombinant CYP6B8 and housefly P450 reductase viruses were co-expressed at an MOI ratio (multiplicity of infection ratio for each virus) of 1:1, and CYP321A1 and housefly P450 reductase viruses were co-expressed at an MOI ratio of 2:2 as detailed in Sasabe et al. (2004). These proportions of recombinant P450 and P450 reductase were determined to be optimal for the production of these P450s and P450 reductase with the final P450 activities obtained in the presence of P450 reductase significantly higher than in its absence. Metabolism assays for insecticides and plant allelochemicals were conducted as described for cleared cell lysates in Wen et al. (2003) and Sasabe et al. (2004) with some modifications. Briefly, 500 µl reaction mixture containing 50 pmol baculovirus-expressed CYP6B8 or CYP321A1 (100 nM final concentrations), 1 µl stock solution of a substrate in methanol and 50 µl 3 mM NADPH (final concentration of 0.3 mM) in cell lysate buffer (pH 7.8) were set up side by side in duplicates. The stock solutions were 5 mM (final concentration of 10 µM) for xanthotoxin, angelicin and -naphthoflavone, 10 mM (final concentration of 20 µM) for carbaryl and diazinon and 50 mM (final concentration of 100 µM) for aldrin and -cypermethrin. The reactions were initiated with the addition of NADPH. For xanthotoxin, angelicin and -naphthoflavone metabolism assays, the reactions were incubated at 30°C for 20 min, stopped by adding 125 µl 2 N HCl and internal standards were added. For carbaryl, diazinon, aldrin and -cypermethrin, the reactions were incubated at 30°C for 30 min and stopped by adding 250 µl acetonitrile containing the appropriate internal standard. To normalize for slight variations in metabolite extraction efficiency, flavone was used as an internal standard for xanthotoxin metabolism, parathion for diazinon metabolism and xanthotoxin for angelicin and -naphthoflavone metabolisms. For metabolism assays involving xanthotoxin, angelicin and -naphthoflavone, the reactions were extracted by adding 500 µl ethyl acetate, vortexing for 1 min and centrifuging at 2000 g for 10 min. For metabolism assays involving carbaryl, diazinon, aldrin and -cypermethrin, the terminated reactions were centrifuged at 16 000 g for 5 min and the supernatant was filtered through a 0.45-µm Alltech Micro-Spin filtration tube by centrifugation at 2500 g for 2 min. The proportions of the remaining substrates and internal standards were determined using normal phase HPLC (Waters Spherisorb 5 µm silica column, 150 x 4.6 mm) for xanthotoxin, angelicin and -naphthoflavone as described previously (Cohen et al., 1989). The proportion of the remaining diazinon was determined by reverse phase HPLC (XTerra 5 µm column, 150 x 4.6 mm) using acetonitrile:water (55:45) at a flow rate of 1 ml/min. The retention times for diazinon and parathion in this system were 9.8 and 12 min, respectively, when detected at 235 nm. P450 enzymatic activities are recorded as substrate disappearance by subtracting the remaining substrate (following correction for non-P450 activities by subtracting residuals in samples lacking NADPH) from zero-time substrate controls and expressed as nanomole substrate disappearance per minute per nanomole P450. For all compounds except aldrin and diazinon, the proportion of substrate metabolized in an assay was <20%, ensuring that rates are linear for the 30 min reaction period. The proportion of diazinon present in an assay that was metabolized was 20% for CYP321A1 and 30% for CYP6B8; the proportion of aldrin present in an assay that was metabolized was 20% for CYP321A1 and 50% for CYP6B8. Each experiment was replicated at least three times with cell lysates prepared from at least three independent batches of baculovirus-infected cells.
[14C]-Methoxy-labeled xanthotoxin was synthesized by heating 14.4 mg of xanthotoxol (Indofine Chemical Co.), 36.1 mg of potassium carbonate, 2.6 mg of [14C]-methyl iodide (1 mCi, Amersham Biosciences) and 0.36 ml of acetone at 90°C in a sealed ampule for 72 h. The resulting [14C]-xanthotoxin was purified by thin-layer chromatography (TLC) on analytical plates without indicators using benzene:ether (10:1) solvent system. The band containing the labeled xanthotoxin on the plate (as visualized under 254 nm UV light) was scraped off and eluted with methanol.
Metabolism of [14C]-methoxy-labeled xanthotoxin by baculovirus-expressed CYP6B8 and CYP321A1 was carried out in 500 µl reactions set up in 1 drum glass vials (
4 ml, Fisher Scientific) containing final concentrations of 0.1 µM CYP6B8 or 0.02 µM CYP321A1 protein co-expressed with house fly P450 reductase (P450: P450 reductase MOI ratio is 1:1 for CYP6B8 and 2:2 for CYP321A1), 0.3 mM NADPH, 10 µM cold xanthotoxin and trace amount of [14C]-xanthotoxin (1 µl delivered in methanol). Reactions were initiated with NADPH, incubated at 30°C for 1 h in a shaking water bath and terminated with the addition of 125 µl 2 N HCl. Control reactions included reactions that were quenched at the beginning of the reactions with 125 µl 2 N HCl (zero time controls) and reactions that were incubated in the absence of NADPH (no NADPH control). A total of 32 µl of each HCl-quenched reaction was spotted directly on a Silica Gel 60 F254 precoated TLC plate (20 x 20 cm plate dimensions, 250 µm thickness) (EMD Chemicals, Darmstadt, Germany). The plate was developed using ethyl acetate:methanol:acetic acid (75:25:1) (Ivie et al., 1983). 14C radioactivity was detected by autoradiography and phosphorimager (Molecular Dynamics, Sunnyvale, CA, USA) analysis.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionAcknowledgementsReferences |
|---|
Molecular models
Molecular models of the substrate-free CYP6B8v1 and CYP321A1 proteins were built as detailed in Baudry et al. (2006) by generating a fixed alignment of eight class II P450 crystal structures (human CYP3A4, CYP2C8, CYP2C5, CYP2C9, CYP2A6, rabbit CYP2B4, B. megaterium CYP102, T. thermophilus CYP175A1) using the ALIGN facility in MOE versions 2004 and 2005 (Chemical Computing Group Inc., Montreal, Canada) that employs the BLOSUM62 scoring matrix. The individual CYP6B8 and CYP321A1 sequences (Fig. 1) were then aligned to this multisequence alignment and the best sequence identity matches for the full-length sequences were found to be to CYP3A4 for both insect sequences (Table I). No major insertions or deletions occur in the six SRS regions found in these proteins. As described in Baudry et al. (2006), the three most highly divergent regions harboring SRS1 (designated as the B region), SRS2 and SRS3 (designated as the FG region) and SRS6 (designated as the β4 region) were aligned separately to these eight class II sequences (Table I). In these shorter sequence alignments, the B region of CYP6B8v1 displayed the highest identity to the CYP2B4 sequence whereas B region of CYP321A1 showed similar levels of identity to the CYP2B4, CYP2C8 and CYP2C9 sequences. The FG regions of both insect sequences showed highest identity to the CYP2B4 sequence and the β4 regions showed the highest identity to the CYP3A4 sequence. On the basis of these sequence similarities, the CYP3A4 structure was used as the main template for both sequences with replacements of the B and FG regions with those from the CYP2B4 structure.
|
|
Comparisons between CYP321A1 and CYP6B8 of the C
-backbones in the loop sequences within the catalytic site (SRS1, SRS5, SRS6), the I helix at the ‘back’ of the catalytic site (SRS4) and the loop sequences near the top of the catalytic site (SRS2 and SRS3) indicate only marginal differences in the SRS5 backbone, which contains more Pro residues in CYP321A1, and the SRS1 backbone, which contains an additional amino acid in CYP321A1.
Closer analysis of the predicted side chains shows substantial variation in nearly all SRS regions. Even in the most conserved of these SRS regions, the side-chain positions vary substantially with only the following number of amino acids being identical: SRS1 (9/30), SRS2 (3/8), SRS3 (2/8), SRS4 (9/19), SRS5 (2/10) and SRS6 (4/13). Even with the higher degree of conservation in SRS4, it is noteworthy that Thr309 in CYP6B8 in the middle of the I helix is replaced with Pro307 in CYP321A1 at a position that somewhat distorts the I helix without breaking it. Other noteworthy changes in the I helix are the replacement of adjacent Val304–Ala305 residues in CYP6B8 with Thr302–Ala303 residues in CYP321A1. The comparison with bacterial P450s indicates that this pair of amino acid variations in the CYP321A1 sequence corresponds exactly to the amino acids present in bacterial CYP108 (P450terp). In a comparison of the CYP101 (P450cam) and CYP108 crystal structures, Hasemann et al. (1995) identified Ala267 in CYP108 (equivalent to Ala303 in CYP321A1) as hydrogen bonding with conserved Thr271 in the next helical turn of CYP108 and identified Thr266 in CYP108 (equivalent to Thr302 in CYP321A1) as hydrogen bonding to waters that have intercalated themselves into the heme-proximal side of this helix.
Analysis of their binding site cavities using the Active Site Finder function in MOE predicts that their volumes are almost the same with CYP6B8v1 having a site volume of 2044 Å3 and CYP321A1 with 2014 Å3. Closer inspection of these predicts that substrates access both catalytic sites via channel pw2a (Wade et al., 2004) and that each of these sites can be subdivided into two with a smaller subsite located close to the ferryl oxygen linked by a narrow channel to a larger subsite extending to the substrate access channel (Fig. 2). In CYP6B8, the heme proximal subsite is predicted to have a volume of 180 Å3 and more distal subsite is predicted to have a volume of 1864 Å3. In CYP321A1, the smaller subsite is predicted to be much larger with a volume of 325 Å3 and the larger subsite is predicted to occupy 1689 Å3. Additionally, the CYP321A1 region linking the subsites appears broader than the linking region in the CYP6B8 model. The amino acid side chains responsible for constricting the CYP6B8 catalytic site shown in Fig. 2 occur in SRS1 and SRS5.
|
Metabolism of [14C]-methoxy labeled xanthotoxin
Previous studies (Li et al., 2004; Sasabe et al., 2004) have defined some of the catabolic activities for these proteins in relation to CYP6B1, the highly specialized P450 inP. polyxenes, using disappearance of nonradioactive substrate as a measure of catalytic activity. With both proteins capable of metabolizing xanthotoxin, their products derived from [14C]-methoxy-labeled xanthotoxin were compared by TLC analysis of in vitro microsomal assays conducted with CYP6B8 or CYP321A1 protein co-expressed in Sf9 insect cells with house fly P450 reductase. Comparison of these products with the CYP6B1 products generated in the presence of NADPH indicates that all three of these P450s generate the same major product presumably from NADPH-dependent epoxidation of the furan ring of xanthotoxin (Fig. 3). The fact that all 14C-labeled xanthotoxin is converted into this product with little apparent loss in radioactivity indicates that these P450s do not O-demethylate xanthotoxin as is the case for some mammalian P450s (Ivie, 1987). The comparison between these sets of products indicates that the CYP6B1 protein metabolizes xanthotoxin at very high efficiency and that the CYP6B8 and CYP321A1 proteins metabolize xanthotoxin at similar low efficiencies.
|
Comparative metabolism assays
Side-by-side activity comparisons of the CYP321A1 and CYP6B8v1 proteins co-expressed with P450 reductase at MOI ratios determined to be non-limiting for this electron transfer partner indicate that both proteins metabolize the linear furanocoumarin xanthotoxin at 1.28 µmol/µmol P450 per minute for CYP321A1 and 1.88 µmol/µmol P450 per minute for CYP6B8 (Table II). In comparison, CYP321A1 metabolizes the angular furanocoumarin angelicin at a rate (0.88 µmol/µmol P450 per minute) 31% lower than that for xanthotoxin and CYP6B8 metabolizes angelicin at a rate (0.45 µmol/µmol P450 per minute) 76% lower than that for xanthotoxin. Among the other activities detailed in Table II, CYP321A1 metabolizes the pyrethroid insecticide -cypermethrin at a rate of 1.60 µmol/µmol P450 per minute and CYP6B8 metabolizes it at a rate of 2.44 µmol/µmol P450 per minute. CYP321A1 metabolizes the cyclodiene insecticide aldrin at a rate of 5.28 µmol/µmol P450 per minute and CYP6B8 metabolizes it (to dieldrin) at a rate of 14.33 µmol/µmol P450 per minute. CYP321A1 metabolizes the organophosphate insecticide diazinon at a rate of 1.35 µmol/µmol P450 per minute and CYP6B8 metabolizes it at a rate of 2.33 µmol/µmol P450 per minute.
|
Substrate-binding simulations
Seven substrates defined for these proteins given in Table II were docked within the identified sites using LigandFit provided in Cerius2 program. For each substrate, 100 possible conformations were generated and ranked according to Dock score provided in LigandFit. The binding conformation with the highest score, lowest interaction energy and appropriate reactive site closest to the heme was selected as the optimal conformation and subjected to energy minimization using the MMFF94 force field in MOE. Interaction energies (E_int) between the minimized protein and the ligand were calculated as the difference between the total potential energy of the minimized complex and the sum of the individual protein and ligand components of the complex.
Comparison of the distances to the reaction center on the substrate molecule from the ferryl oxygen atom indicate that the bulkier, rigid molecules bind further from the ferryl oxygen in CYP6B8 whereas the smaller flexible substrates (aldrin, -cypermethrin) bind within similar distances in both P450s (Table II). Examination of the fully minimized binding modes for xanthotoxin, a representative linear furanocoumarin, predicts that it binds to the CYP321A1 catalytic site with its furan ring positioned at a distance of 3.27 Å from the oxygen of the iron-oxo intermediate (Fig. 4A). Similarly, xanthotoxin is predicted to bind in the CYP6B8v1 catalytic site with only its furan ring positioned at a distance of 4.28 Å from the oxygen of the iron-oxo intermediate; the predicted interaction energy for this docking mode (–27.0 kcal/mol) is approximately the same as that for CYP321A1 (–30.0 kcal/mol) (Table II). The predicted binding mode for angelicin, the angular furanocoumarin metabolized by CYP321A1 (0.88 µmol/µmol P450 per minute) more efficiently than the linear furanocoumarin xanthotoxin (0.45 µmol/µmol P450 per minute) (Table II), positions the furan ring at approximately the same distance (3.32 Å) from the oxygen of the iron-oxo intermediate in the CYP321A1 model (Fig. 4B). In CYP6B8v1, which metabolizes angelicin less efficiently than xanthotoxin, the binding mode for angelicin positions the furan ring at a much greater distance (5.32 Å) and different orientation than in the CYP321A1 model with a higher predicted interaction energy (–25.0 kcal/mol for CYP6B8 versus –37.0 kcal/mol for CYP321A1). -Napthoflavone, another substrate differentially metabolized by CYP321A1 (0.94 µmol/µmol P450 per min) and CYP6B8v1 (0.23 µmol/µmol P450 per minute), is predicted to bind in completely different orientations in these two proteins (Fig. 4C) with substantially shorter distances to the 7- and 8-carbons for CYP321A1 than CYP6B8 (Table II). The optimal binding mode for aflatoxin B1 is predicted to place its reactive methyl group at 3.23 Å from the ferryl oxygen in CYP321A1, a position that is significantly closer than the 7.63 Å distance predicted for its binding in CYP6B8 (Fig. 4D). The predicted binding mode for diazinon, an organophosphate, in the CYP321A1 catalytic site places it at a significantly shorter distance (3.40 Å) from the heme oxygen than in the CYP6B8v1 catalytic site (5.35 Å) (Fig. 4E) but with a higher energy for the diazinon-docked CYP6B8v1 model. The predicted binding modes and interaction energies for the smaller aldrin molecule, a cyclodiene, are similar for both models (Fig. 4F). The predicted binding modes for -cypermethrin, a representative pyrethroid, also differ for CYP321A1 and CYP6B8v1 (Fig. 4G), but in both cases have the 4’-position on the benzene ring in similar proximity to the heme oxygen with similar interaction energies.
|
Contact residues predicted to exist within 4 Å of each substrate are shown in Table III. Despite the fact that the six molecules docked in the CYP6B8v1 and CYP321A1 catalytic sites are significantly different in their dimensions, most contact the SRS1, SRS4 and SRS5 that form the bulk of the catalytic site directly above the heme plane. Depending on the size and binding mode for some of these compounds, additional residues in SRS2, SRS3 and SRS6 or downstream from SRS2 are predicted to contact substrates in CYP6B8 and/or CYP321A1. As an example, the large cypermethrin molecule is predicted to contact six residues in or adjacent to SRS2 in CYP321A1 and three residues downstream of SRS2 in CYP6B8v1. The significantly smaller aldrin molecule contacts no residues near SRS2 in CYP321A1 and three residues in or downstream of SRS2 in CYP6B8v1. The planar xanthotoxin contacts three residues downstream of SRS2 in CYP6B8v1 and none in CYP321A1.
|
Contact residues within SRS4 include several of those previously discussed as providing different degrees of kinking in the I helix. Specifically, all substrates except xanthotoxin, angelicin and
-napthoflavone contact Thr302 and Pro307 residues in CYP321A1. In CYP6B8v1, all substrates contact Thr309 (which aligns with Pro307 in CYP321A1) and all, except angelicin and -cypermethrin, contact Val304 and Ala305 (which align with Thr302 and Ala303 in CYP321A1). Other than these side-chain contacts in the I helix, very few contact residues are absolutely conserved between CYP6B8 and CYP321A1. Several that are conserved but do not overlay in the models are Arg479/Ile482/Gly483 in CYP321A1 and Arg483/Ile486/Gly487 in CYP6B8 (in SRS6). The very small number of side chains overlaying one another indicates in a manner more compelling than the primary sequence alignments that H. zea has evolved two different structural solutions to the problem of metabolizing these compounds.
Using these models, we have evaluated the positions of naturally occurring sequence variations in the CYP321A1 and CYP6B8 proteins using amino acid replacement functions in MOE with subsequent rounds of energy-minimization and substrate docking. To date, three CYP6B8 variants in addition to CYP6B8v1 have been cloned from cDNA libraries (CYP6B1v1) and PCR amplifications of genomic DNA (CYP6B8v2, v3 and v4) from laboratory colonies and wild-type populations of H. zea (Li et al., 2000, 2002). The amino acid replacements in each sequence are as follows: CYP6B8v2 (Leu226Ser, Ser322Thr, Gln375Arg and Lys377Arg), CYP6B8v3 (Asp283Gly, Met367Val and Asp388Gly) and CYP6B8v4 (Thr379Ser) with each designated according to the amino acid in CYP6B8v1 followed by the replacement in the variant. The positions of these variations in the CYP6B8v1 backbone shown in Fig. 5 indicate that, of these, the Gln375Arg in the CYP6B8v2 variant has potential for altering substrate specificity since this residue is predicted to contact four of the docked substrates; the closely positioned Lys377Arg in CYP6B8v2 and Thr379Ser in CYP6B8v4 also have potential to affect the configuration of SRS5 although they are not directly predicted to contact any of the substrates.
|
The sole CYP321A1v2 variant that has been cloned from PCR amplification of genomic DNA differs from CYP321A1v1 at two positions (Asp197Glu and Thr289Met). The first of these variations maps between the E and F helices to a region near the exterior of the CYP321A1 model and the second of these maps to the N-terminal end of the I helix. Neither would be expected to alter substrate reactivity.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionAcknowledgementsReferences |
|---|
These models and metabolic data have several implications in the evolution of insecticide resistance. Among these is the fact that two P450 proteins significantly diverged at the primary sequence level retain the ability to metabolize a range of naturally occurring plant allelochemicals as well as synthetic insecticides using catalytic sites that are comparable in their overall size. Side-by-side metabolic analyses indicate that CYP321A1 metabolizes larger and more rigid substrates (angelicin,
-naphthoflavone and aflatoxin B1) several-fold faster than CYP6B8v1 and more flexible substrates (diazinon) somewhat more slowly than CYP6B8 (Table II). The smaller and more flexible aldrin molecule is metabolized by CYP6B8v1 at a much higher rate than it is metabolized by CYP321A1. The flexible cypermethrin molecule is also metabolized faster by CYP6B8v1 than CYP321A1.
Homology modeling of these substrate-free P450s suggests that their catalytic sites consist of two cavities connected by a narrow linker region, with the smaller subcavity located closer to the ferryl oxygen atom. In the CYP321A1 model, this small subcavity appears much larger (325 Å3) and the connecting linker region appears broader than in the CYP6B8 model. As is consistent with their respective activities, theoretical docking experiments indicate that the CYP321A1 model can harbor bulkier and more rigid molecules with reactive positions closer to the ferryl oxygen, whereas the CYP6B8 model docks these same molecules in the more distal large subcavity with their reactive centers beyond reactive distance from the ferryl oxygen. Among the compounds docked here, the most extreme example of this difference is the largest AFB1 molecule that is predicted to bind in the CYP321A1 cavity with its reactive methyl group at 3.23 Å from the ferryl oxygen, in contrast to its predicted distance (7.63 Å) in the CYP6B8v1 cavity. Consistent with this longer distance, AFB1 is not metabolized by CYP6B8v1 (G. Niu et al., submitted). Closer inspection of the AFB1-docked CYP6B8 model suggests that, although the methyl group is able to bind in the small subsite of CYP6B8, the bulky and rigid ring system in the rest of this molecule is sequestered in the larger subcavity in an orientation that prevents the methyl group from accessing the smaller subcavity. Two of the other large molecules docked in this site, -cypermethrin and diazonin, are much more flexible and capable of positioning the reactive benzene ring on -cypermethrin and reactive sulfur group on diazinon in the smaller subcavity of CYP6B8 within reactive distance of the heme oxygen.
Differences between these predictive models have suggested that variations in two regions (SRS1 and SRS5) constrain the catalytic site of CYP6B8v1 compared with that of CYP321A1. The replacement of Thr119 in CYP321A1 with the bulky aromatic Phe118 in CYP6B8v1 as well as Pro368 and Thr373 in CYP321A1 with the larger Ile370 and Gln375 in CYP6B8v1 limits the range of proximal positions that each of the larger substrates can assume in the CYP6B8v1 catalytic site. As a result, many of the substrates compared in side-by-side assays are hydroxylated or epoxidated faster, and, sometimes exclusively, by CYP321A1.
From a biochemical perspective, the differences in the activities of these two proteins are especially interesting in light of the fact that CYP321A1 does not contain the highly conserved threonine that participates in oxygen activation in most other P450s (Atkins and Sligar, 1988; Nagano and Poulos, 2005). This conserved threonine is also lacking in at least one other well-known P450, CYP107A1 (P450eryF), with its function in oxygen activation originally proposed to be replaced by a water molecule in the active site (Cupp-Vickery et al., 1996). More recent structural determinations on 6-deoxyerythronilide B-bound CYP107A1 have suggested that a hydroxyl group on this substrate directly interacts with dioxygen to facilitate the proton transfer (Nagano et al., 2005). Structural determinations on CYP158A1, another bacterial P450 known to lack this conserved threonine, suggest that it mediates oxygen activation using water molecules complexed with two hydroxyls on its flaviolin substrate and an ordered hydrogen-bonding network of waters extending from the catalytic site to the surface of this protein (Zhao et al., 2005). The fact that many of the substrates for CYP321A1 (e.g. xanthotoxin, -cypermethrin and diazinon) lack hydroxyl groups suggests that, in this insect P450, oxygen activation is mediated by yet another mechanism.
Comparison of their gene organizations has already indicated that CYP321A1 and CYP6B8 have evolved independently of one another: one intron exists in the CYP6B8 gene (Li et al., 2002) and none in the CYP321A1 gene (not shown). Despite their different evolutionary histories and different activities toward toxins, both proteins remain capable of metabolizing broad ranges of compounds that Helicoverpa has evolved to live with. Redundancy in its detoxification systems derived from the presence of two distinct P450s provides this insect with a competitive advantage over others that contain fewer P450s and/or lower levels of P450s metabolizing more limited sets of toxic compounds. Comparisons between these catalytic sites indicate that convergent evolution to handle many of the same compounds has been achieved both by variation in the side chains of residues predicted to contact these substrates as well as by variation in at least one residue that is presumed to be involved in catalytic activation of the dioxygen needed for substrate modification. Determination of the crystal structure of one or more of these P450s will allow testing of this hypothesis.
|
Footnotes |
|---|
Edited by Anthony Wilkinson
|
Acknowledgements |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionAcknowledgementsReferences |
|---|
The authors gratefully thank Dr. Wenfu Mao for synthesizing 14C xanthotoxin for use in these metabolic assays, Ms. Lauren Chin for cloning the CYP321A1v2 variant and Dr. May Berenbaum for scientific input. This project was funded by National Institutes of Health grant R01 GM071826.
|
References |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionAcknowledgementsReferences |
|---|
Atkins W.M., Sligar S.G. J. Biol. Chem. (1988) 263:18842–18849.
Baudry J., Rupasinghe S., Schuler M.A. Protein Eng. Des. Sel. (2006) 19:345–353.
Claudianos C., Ranson H., Johnson R.M., Biswas S., Schuler M.A., Berenbaum M.R., Feyereisen R., Oakeshott J.G. Insect Mol. Biol. (2006) 15:615–636.[CrossRef][Web of Science][Medline]
Cohen P., Klumpp S., Schelling D.L. FEBS Lett. (1989) 250:596–600.[CrossRef][Web of Science][Medline]
Cupp-Vickery J.R., Han O., Hutchinson C.R., Poulos T.L. Nat. Struc. Biol. (1996) 3:632–637.[CrossRef][Web of Science][Medline]
Feyereisen R. Annu. Rev. Entomol. (1999) 44:507–533.[CrossRef][Web of Science][Medline]
Feyereisen R. Comprehensive Molecular Insect Science (2005) vol. 4. Oxford: Elsevier. 1–77.
Gotoh O. J. Biol. Chem. (1992) 267:83–90.
Graham S.E., Peterson J.A. Arch. Biochem. Biophys. (1999) 369:24–29.[CrossRef][Web of Science][Medline]
Hasemann C.A., Kurumbail R.G., Boddupalli S.S., Peterson J.A., Deisenhofer J. Structure (1995) 3:41–62.[Medline]
Ivie G.W. Amer. Chem. Soc. Symp. Series (1987) 339. Washington, D.C: American Chemical Society. 217–230.
Ivie G.W., Bull D.L., Beier R.C., Pryor N.W., Oertli E.H. Science (1983) 221:374–376.
Li X., Berenbaum M.R., Schuler M.A. Insect Biochem. Mol. Biol. (2000) 30:75–84.[CrossRef][Web of Science][Medline]
Li X., Berenbaum M.R., Schuler M.A. Insect Biochem. Mol. Biol. (2002) 32:311–320.[CrossRef][Web of Science][Medline]
Li X., Baudry J., Berenbaum M.R., Schuler M.A. Proc. Natl. Acad. Sci. USA (2004) 101:2939–2944.
Li X., Schuler M.A., Berenbaum M.R. Ann. Rev. Entomol. (2007) 52:231–253.[CrossRef][Web of Science][Medline]
MacKerell A.D. Jr, et al. J. Phys. Chem. B (1998) 102:3586–3616.
Nagano S., Poulos T.L. J. Biol. Chem. (2005) 280:31659–31663.
Nagano S., Cupp-Vickery J.R., Poulos T.L. J. Biol. Chem. (2005) 280:22102–22107.
Negishi M., Uno T., Honkakoski P., Sueyoshi T., Darden T.A., Pedersen L.P. Biochimie (1996) 78:685–694.[Medline]
Nelson D.R., et al. Pharmacogenetics (1996) 6:1–41.[Web of Science][Medline]
Poulos T.L., Johnson E.F. Cytochrome P450: Structure, Mechanism, and Biochemistry (2005) 3e. New York: Kluwer Academic/Plenum Publishers.
Ranson H., Claudianos C., Ortelli F., Abgrall C., Hemingway J., Sharakhova M.V., Unger M.F., Collins F.H., Feyereisen R. Science (2002) 298:179–181.
Ravichandran K.G., Boddupalli S.S., Hasemann C.A., Peterson J.A., Deisenhofer J. Science (1993) 261:731–736.
Rupasinghe S., Schuler M.A. Phytochem. Rev. (2006) 5:473–505.[CrossRef]
Sasabe M., Wen Z., Berenbaum M.R., Schuler M.A. Gene (2004) 338:163–175.[CrossRef][Web of Science][Medline]
Schoch G.A., Yano J.K., Wester M.R., Griffin K.J., Stout C.D., Johnson E.F. J. Biol. Chem. (2004) 279:9497–9503.
Scott J.G. Insect Biochem. Mol. Biol. (1999) 29:757–777.[CrossRef][Web of Science][Medline]
Scott E.E., White M.A., He Y.A., Johnson E.F., Stout C.D., Halpert J.R. J. Biol. Chem. (2004) 279:27294–27301.
Sheehan D., Meade G., Foley V.M., Dowd C.A. Biochem. J. (2001) 360:1–16.[CrossRef][Web of Science][Medline]
Wade R.C., Winn P.J., Schlichting I., Sudarko. J. Inorg. Biochem. (2004) 98:1175–1182.[CrossRef][Web of Science][Medline]
Wen Z., Pan L., Berenbaum M.R., Schuler M.A. Insect Biochem. Mol. Biol. (2003) 33:937–947.[CrossRef][Web of Science][Medline]
Wester M.R., Johnson E.F., Marques-Soares C., Dansette P.M., Mansuy D., Stout C.D. Biochemistry (2003) 42:6370–6379.[CrossRef][Medline]
Williams P.A., Cosme J., Ward A., Angove H.C., Vinkovi
D.M., Jhoti H. Nature (2003) 424:464–468.[CrossRef][Medline]
Yano J.K., Blasco F., Li H., Schmid R.D., Henne A., Poulos T.L. J. Biol. Chem. (2003) 278:608–616.
Yano J.K., Wester M.R., Schoch G.A., Griffin K.J., Stout C.D., Johnson E.F. J. Biol. Chem. (2004) 279:38091–38094.
Yano J.K., Hsu M.-H., Griffin K.J., Stout C.D., Johnson E.F. Nat. Struct. Mol. Biol. (2005) 12:822–823.[CrossRef][Web of Science][Medline]
Zhao B., Guengerich F.P., Voehler M., Waterman M.R. J. Biol. Chem. (2005) 280:42188–42197.
Received April 2, 2007; revised October 9, 2007; accepted October 10, 2007.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  T.-L. Chiu, Z. Wen, S. G. Rupasinghe, and M. A. Schuler Comparative molecular modeling of Anopheles gambiae CYP6Z1, a mosquito P450 capable of metabolizing DDT PNAS, July 1, 2008; 105(26): 8855 - 8860. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||