CYP2J2 MOE Homology Modeling. A multiple sequence alignment was performed using the protein data bank to cross-reference the CYP2J2 sequence against 2A1, 2D6, 2E1, 2R1, 2A6, 2C8, 2B4, 2C5, 2C9, 2A13. For construction of the CYP2J2 homology model, crystal structure PDB coordinates were used for the isozymes’ with the closest homology to 2J2 and included CYP2B4 (backbone-42% primary sequence homology), 2C8 (B-region-38% primary sequence homology), 2R1 (FG region-40% primary sequence homology) and 2C9 ( 4-region-26% primary sequence homology). An open CYP2J2 conformation was initially used to dock AA, AEA, 2-AG following previously described procedures (Baudry et al., 2003). Two apparent CYP2J2 active site cavities, exhibited adjacent/overlapping areas near the heme in agreement with the highest propensity of ligand binding number (PLB) and were both used for docking of each respective substrate. Next, we conducted an energy minimization step for each respective ligand conformation and the induced fitting resulted in the CYP2J2-substrate model used to examine heme and ligand distances as well as the predicted interacting active site residues. Lastly, interaction energies were calculated for each substrate to examine the energetics and feasibility of substrate binding.
Spectral titrations of CYP2J2-ND. Spectral binding studies were performed by titrating AA (both arachidonic acid and sodium arachidonate), AEA, 2-AG, MS-PPOH and ebastine against CYP2J2- Nanodiscs (8 μM) in 100 mM phosphate buffer (pH 7.4). Importantly, substrate stocks were prepared with DMSO (5 mM and 50 mM) after the addition of ethanol and acetonitrile were determined to cause type I binding. CYP2J2-ND were incrementally titrated with increasing concentrations of each putative substrate (either 5 mM or 50 mM stocks) at 22oC and allowed to equilibrate for 5 minutes before measuring the spectra. The total dilution of the original CYP2J2 concentration was 1.5%. Absorbance spectra were collected from 800-200 nm using a Cary Bio 300 UV-Vis spectrophotometer (Agilent Technologies, Santa Clara CA). Raw absorbance spectra were processed to the corresponding subtraction spectra using a standard MATLAB subroutine (Mathworks Inc., MI). The absolute absorbance difference of the subtraction spectra (peak and trough) were plotted against the corresponding substrate concentration in Origin Lab (Origin Lab Inc., Northhampton, MA) and fitted using a single binding isotherm (Eq. 2).

(Equation 2) A = Amax [S] / (Ks + [S])
Where A is the absorbance difference at each difference spectra peak and trough, Amax is the amplitude corresponding to maximal spin shift, Ks is the spectral dissociation constant, S is the substrate concentration. None of the substrates investigated exhibited any homotropic cooperativity and hence the titration data was fitted to a single binding isotherm.
Cyanide binding and equilibrium constant (Ks) determination for CYP2J2-ND. The steady-state cyanide spectral equilibrium constants were determined by titrating cyanide concentrations (800 mM and 4 M stocks) against ~3 M of CYP2J2-ND in the absence and the presence of 40 μM of AA, AEA, 2-AG and MS-PPOH until saturation was reached (Das et al., 2014). The total addition of CYP2J2 concentration was 3% and did not affect Soret peak height. The raw data collected was converted into the difference spectra using a MATLAB subroutine (Mathworks Inc., MI). The A ( A 444-414nm) was plotted against the corresponding cyanide concentration and fitted to a single binding isotherm equation using Origin Pro 8.6 (San Clemente, CA) with equation 1 where A = A 444-414nm.
Transient cyanide binding kinetics to CYP2J2 active site measured using stopped-flow spectroscopy. The cyanide binding kinetics to the CYP2J2 active site in the absence and presence of varying concentrations of AA, AEA, 2-AG and MS-PPOH was determined using stopped-flow spectroscopy. Raw data was collected using an Applied Photophysics SX.18MV (Leatherhead, UK) spectrophotometer to monitor the rate of cyanide binding. A dual syringe system was prepared, with syringe I containing CYP2J2-ND in the absence or presence of varying concentrations (10, 20 and 40 μM) of substrate and syringe 2 containing 80 mM cyanide in the absence or presence of varying concentrations (10, 20 and 40 μM) of substrate. The relatively high cyanide concentration was selected based on steady state cyanide binding measurements and was necessary to ensure pseudo first-order binding kinetics (Denisov et al., 2007). The parameters for rapid mixing and data collection were as follows: 1.5 msec of dead time, external triggering mechanism, photodiode array detection and collection obtained on a logarithmic scale over 20 seconds. The raw data was converted to the difference spectra by subtracting each successive scan from the time = 0 scan using a MATLAB subroutine (Mathworks Inc., MI) that produced the typical 446 and 415 nm cyanide binding peaks. The difference of absorbance was plotted against time and analyzed with a double exponential equation using Origin Pro 8.6.

Where A1 and A2 are the amplitudes of the fast and slow phases of the reaction respectively and k1 and k2 are the rates of fast and slow phases of substrate binding to the protein, determined from fitting data obtained to equation 3.
Data Analysis. Data were analyzed and presented as means ± standard error. For all experiments, data were collected in triplicate (n = 3) or greater. Stopped flow data were analyzed for statistical significance using a student’s t-test where p < 0.05, p < 0.01 and p < 0.001 were denoted by *, **, and ***, respectively
Results
CYP2J2 mediated metabolism of AA and AEA determined using LC-ESI-MS. We first
investigated the metabolism of AA and AEA by CYP2J2. CYP2J2 was recombinantly expressed in E.coli, purified and assembled into nanodiscs as described in the materials and methods. Due to increased functional stability of the CYP2J2-CPR-nanodisc system, all the studies reported here were done with the proteins incorporated into nanodiscs. To examine the metabolism of AA and AEA, a LC-ESI-MS method was developed to qualitatively assess the product formation. Importantly, the use of a slow linear gradient was necessary for separation of the regioisomers and structural identification by mass and elution time. Product analysis of CYP2J2-ND-CPR incubations with AA produced four regioisomers (m/z = 319.5) corresponding to the mass and elution times of 14,15-, 11,12-, 8,9- and 5,6-EET (Fig. 2A). The additional peaks (m/z 319), 19-HETE, was confirmed at earlier elution times (11 min). Similarly, analysis of CYP2J2 and AEA incubations produced the same olefin oxygenation pattern as AA, with a regioisomer corresponding to the elution times of 14,15-EET-EA as confirmed using the authentic standard (Fig. 2B). The identity of the 11,12-, 8,9-, and 5,6-EET-EA were extrapolated based on literature describing the elution profile and peak mass confirmation. Similar to AA, the presence of two additional metabolites corresponding to 19- and 20-HETE-EAs were observed. These results show that both AA and AEA are substrates of CYP2J2 and form four epoxide products similar to previous reports with other P450s (Snider et al., 2010).

CYP2J2 mediated metabolism of 2-AG determined using LC-ESI-MS. The in vitro incubations containing CYP2J2-ND-CPR and 2-AG produced the following products – two detectable 2-AG epoxide metabolites (m/z 393.3), arachidonic acid (m/z 303.3) and EETs (m/z 319.5) as shown in Figure 3. Authentic standards were used to confirm the 14,15-EET-G, EETs (14,15-, 11,12-, 8,9-, 5,6-EET) and AA metabolites. While the 11,12-EET-G was not commercially available, the results from EET-G saponification confirmed the production of the 11,12-regioisomer.
In contrast to AA and AEA, only two olefin sites of 2-AG were epoxidized to form 2-AG-EETs (m/z 393.3). Thus, the functional data suggests that the interaction of 2-AG within the CYP2J2 binding pocket is influenced by the presence of the bulky glycerol group of 2-AG (Fig. 1). Previous incubations with other P450s have failed to produce the EET-G regioisomers (Chen et al., 2008). Here we report that CYP2J2 is identified as an isoform of P450 that produces epoxide 2-AG metabolites. As a control we analyzed CYP2C8 (supersomes) with 2-AG following the same methodology and failed to observe any detectable EET-G formation (Supplemental Fig. 1). This suggests that CYP2J2 may be unique in the epoxidation of 2-AG as compared to other family 2 P450s.
Quantitative analysis of CYP2J2 mediated epoxidation of 2-AG and AEA using LC-ESI-MS/MS.The detection and quantification of fatty acids and their metabolites have been most commonly examined using radio flow HPLC due to the high sensitivity and ease of use (Wu et al., 1996). Alternatively, detection by mass spectrometry after derivatization, coupled with gas chromatography or liquid chromatography has also been widely used for EET analysis (Zhu et al., 1995). Here we successfully used HPLC coupled to a triple quadrupole detector utilizing one of the most sensitive commercially available ion trap (QTRAP 5500) which enabled a highly sensitive, accurate and reproducible method for detection of all the EET regioisomers in multiple reaction monitoring (MRM) mode. The limit of quantitation (LOQ) with a 10:1 signal to noise ratio for 5(6)-EET was 5.0 ng/mL, whereas 8,9-, 11,12-, 14,15-EET were calculated at 2.5 ng/mL. Prior to each run a six point standard curve was generated for each set of samples using authentic standards. Extraction efficiency was calculated for each of the EET regioisomers using the acidification and the three step ethyl acetate extraction. 14,15-, 11,12-, and 8,9-, and 5,6-EETs were extracted from the reaction mixture with a 95% or greater efficiency. Reaction linearity was demonstrated for each of the substrates at 15, 30 and 60 minutes with R2 values of 0.98 or greater. The advantages of a LC-MS/MS method are that it provides structural elucidation, quantification and efficiency with excellent sensitivity without the need of derivatization. This method should be widely applicable for complex biological samples with low levels of EETs. The identification of the metabolites of 2-AG and AEA were followed by quantification and measurement of the rate of formation using a LC- MS-MS method as described in the materials and method section (Fig. 4).

The rates of formation of the two EET-G regioisomers were determined by saponification of the EET-G to the corresponding EET regioisomers. The final value of EETs arising from EET-G was obtained by subtracting the free EETs as determined in the materials and methods. The kinetics of EET-G formation were as follows: 14,15-EET-G Vmax = 265.4 ± 12.3 pmol/min/nmol and Km = 32.6 ± 1.7 μM; 11,12- EET-G Vmax = 125.5 ± 5.2 pmol/min/nmol and Km = 13.6 ± 1.3 μM.
In a separate experiment, the rate of metabolism of arachidonic acid to form the four EET regio-isomers was measured in the presence of saturating concentrations of AA. Catalytic rates were calculated at 103 ± 4.6, 61.5 ± 5.6, 18.4 ± 1.7 and 8.9 ± 0.1 pmol/min/nmol of CYP2J2 for 14,15-, 11,12-, 8,9- and 5,6-EET, respectively.
Additionally, we examined the CYP2J2 mediated epoxidation of saturating amounts of AEA (60 μM) towards the production of 14,15-EET-EA which was calculated to be 391 ± 12 pmol/min/nmol of CYP2J2. Thus, the production of the 14,15-EET-EA appears to proceed at a significantly faster rate when compared to the production 14,15-EET from AA despite sharing a similar qualitative product profile and binding characteristics as discussed later.

CYP2J2 mediated 2-AG ester cleavage and production of free AA. The ESI-LC-MS product analysis of 2-AG incubations revealed significant amounts of free AA in solution (Fig. 3). Control experiments showed that the conversion of 2-AG to AA required the presence of both NADPH and CPR for catalytic turnover indicating that the reaction proceeded through an oxidative mechanism (controls not shown). As seen in Figure 4D, analysis of 2-AG incubations followed Michaelis-Menten kinetics, indicating that the reaction was dependent on substrate binding. The Km of the ester cleavage reaction was calculated at 3.25 μM with a Vmax of 2.2 ± 0.2 nmol/min/nmol of CYP2J2. Samples lacking NADPH and CYP2J2 were ran in parallel as controls. Arachidonic acid was not detected in the NADPH-free sample. Additionally, 2- AG was stable in buffer at 37oC with negligible hydrolysis. The ability of CYP to mediate ester cleavage reactions has long been reported and rates determined in this work show similar catalytic properties of other CYP isozymes (Guengerich, 1987; Peng et al., 1995). However, this is the first account of the P450 mediated cleavage of a fatty acid ester of an endogenous substrate. The same reactions were run with CYP2C8 (supersomes) which also produced AA in a NADPH dependent reaction (Supplemental Fig. 1). However, CYP2C8 epoxygenase did not demonstrate the ability to oxygenate the 2-AG olefin bonds.
Bovine and porcine heart microsomes metabolism of AA and 2-AG. To examine whether the novel endocannabinoid epoxygenation reactions would occur within a complex environment in the presence of other proteins, we prepared heart microsomes from the tissue of bovine and porcine left ventricular myocardium. The rate of metabolism of AA, AEA and 2-AG are summarized in Table 2. We chose bovine and porcine species because they both contain CYP2J2 with high sequence homology to humans (Supplemental Figure 3). CYP2J2 immunoblotting of the bovine and porcine microsomes were performed using a polyclonal rabbit IgG antibody for human CYP2J2. The microsomal preparations cross-reacted with the CYP2J2 antibody demonstrating the presence of CYP2J2 in bovine and porcine hearts (Supplemental Figure 4). First, the functionality was assessed by measuring AA metabolism of bovine and porcine microsomes which produced 1.8 ± 0.3 and 2.08 ± 0.02 pmol min-1 mg-1 microsomal protein of EETs (all regio-isomers), respectively. Next we quantified 2-AG metabolism by directly measuring the production of the 14,15-EET-G regioisomer by bovine and porcine microsomes that was calculated at 1.17 ± 0.57 and 1.1 ± 0.55 pmol min-1 mg-1 microsomal protein, respectively. The 14,15-EET-EA isomer was produced with a higher turnover in both bovine and porcine microsomes with values of 1.96 ± 0.05 and 2.25 ± 0.07 pmol min-1 mg-1 microsomal protein, respectively.

CYP2J2 mediated lipase activity. CYP2J2 mediated oxidative cleavage of 2-AG into arachidonic acid raised questions whether this enzyme could act as a lipase and use AA containing phospholipids as direct substrates. PIP2 (phosphatidylinositol 4,5-bisphospate) is present in the plasma membrane where it is cleaved by phospholipase C to produce AA for subsequent entry into the eicosanoid cascade (Rouzer and Marnett, 2011). Thus, we incubated PIP2 with CYP2J2-ND-CPR-NADPH system utilizing PIP2 concentrations below its critical micelle formation (CMC). In these experiments we did not to observe formation of AA or diacylglycerol (Huang et al., 2011).
Spectral binding studies of CYP2J2 with AA, AEA and 2-AG. In an effort to understand the binding interactions of AA and its derivatives (AEA and 2-AG) with the active site of CYP2J2, we performed spectroscopic binding studies. First, the interactions of AA, AEA, 2-AG, ebastine and MS-PPOH with the CYP2J2-ND active site were investigated using UV-Vis spectral binding titrations (Fig. 5). As controls, the spectral binding properties of the potent epoxygenase inhibitor N-methylsulfonyl-6-(2- propargyloxyphenyl) hexanamide (MS-PPOH) (Falck et al., 1997) and ebastine were investigated and both induced type-I shifts with apparent spectral binding constants (Ks) of 10.8 ± 1.5 μM and 5.6 ± 0.7, respectively.
The fatty acids presented with smaller changes in spin-state and minor shifts in Soret band. Specifically, the Soret band at 417 nm red-shifted by 0.5 nm with a slight decrease in amplitude for both AA and AEA. The difference spectra revealed prominent troughs for both ligands at 414 nm and broad peaks at 434 and 435 nm for AA and AEA, respectively (Figure 5A). Both substrates were fitted to single binding isotherm with apparent Ks of 11.3 ± 1.3 μM for AA and 13.1 ± 1.4 μM for AEA (Figure 5D). The similarities between AA and AEA spectral shifts and binding equilibrium suggests that the presence of the ethanolamide (AEA) motif does not significantly alter substrate distances and positioning in relation to the heme when compared with AA. Conversely, 2-AG spectral titrations resulted in a blue-shift while slightly decreasing the Soret. The apparent Ks of binding for 2-AG was calculated at 18 ± 2.7 μM. Taken together, these data demonstrate that the modification of the carboxylic acid of the arachidonic acid with the glycerol moiety affects binding in CYP2J2’s large yet narrow binding cavity (Lafite et al., 2007). Note that the observed Soret shifts are not as prominent as xenobiotic P450 substrates, yet are consistent with the few published spectra of other fatty acids binding to membrane bound CYPs (Loughran et al., 2001). Moreover, the lack of significant Soret perturbations corresponds to the relatively low turnover rates of the AA metabolizing epoxygenases such as CYP2J2 (Westphal C et al., 2011). Additionally, these experiments were further validated using steady state cyanide binding experiments in the presence of AA, AEA and 2-AG to probe the access and binding of cyanide to the heme (Supplemental Figure 2).

Kinetics of cyanide binding to CYP2J2-ND in the presence of substrates. The cyanide binding kinetics of CYP2J2 was investigated in the absence or presence of hydrophobic substrates using stopped- flow spectroscopy. For each substrate tested, rapid mixing and formation of the heme-cyanide complex red-shifted the Soret (Fig. 6A). As shown in Fig. 6B, the difference spectra (A444 – 414 nm) were analyzed as a function of time and yielded an exponential biphasic curve that were comprised of a fast and slow phase (Bidwai et al., 2013). The “fast phase” corresponds to the rate of cyanide binding to the ferric heme active site. Specifically, it can be used as a direct measure of substrate egress from the immediate vicinity of the heme active site to accommodate formation of the cyanide complex. In the fast phase the presence of increasing concentrations ebastine significantly decreased cyanide binding rates in a concentration- dependent manner. Conversely, the fatty acid ligands only slightly decreased cyanide binding (Fig. 6C). The ratio of the fast-and slow-phase amplitudes were compared to determine how ligand concentrations affect the presence of the ratios. As seen in Figure 6D-6G, saturating concentrations of ebastine significantly increases the ratio of the slow-phase (17.4% to 66.4%) whereas the arachidonate containing ligands only slightly increased the slow-phase ~10%. To rule out non-specific fatty acid binding we examined the rates of cyanide binding in the presence of saturating concentrations of palmitate. As shown in Fig. 6H the presence of palmitate does not noticeably alter the ratio of the two phases. While not dramatic, the reported results provide important characterization of the nature of fatty acid binding.

CYP2J2 Homology Modeling with AA, AEA and 2-AG. Molecular Operating Environment (MOE, Chemical Computing Group) software was used to identify substrate-binding characteristics within the active site of a CYP2J2 homology model. A hybrid structure model was constructed from the crystal PDB coordinates of similar CYP isozymes (class dependent sequence alignment strategy) as described in the materials and methods section (Baudry et al., 2006). This approach takes advantage of the conserved sequences that maintain the secondary and tertiary P450 folds to generate the core structures surrounding the buried catalytic sites. Given the fatty acids relatively low turnover there are likely many ‘unproductive’ conformations within the active site. As shown in Fig. 7A, AA was docked in the most preferred configuration for epoxidation of the 14,15-olefin when closest to the heme (5.1 Å), followed by the 11,12-, 8,9-, and 5,6-olefins with calculated heme distances of 8.6, 11.3 and 12.9 Å, respectively. As shown in Fig. 7B, the distances of the AEA 14,15-, 11,12-, 8,9- and 5,6-olefins were calculated at 6, 8.5, 12.1 and 13.5 Å, respectively. Interestingly, 2-AG (Fig. 7C) yielded significantly different active site interactions albeit with similar heme distances of the 14,15-, 11,12-, 8,9- and 5,6-olefin estimated at 5.9, 7.6, 9 and 11.5 Å, respectively. The interaction potential energies predict feasibility of binding and were similar for each polyunsaturated fatty acid with calculations of -68.3 kcal/mol, -69.4 kcal/mol, and -73.4 kcal/mol for AA, 2-AG and AEA, respectively. Thus, the olefin distances of the MOE model were in agreement experimental differences observed in AA, AEA and 2-AG metabolism. Additionally, the model predicted significantly different binding characteristics based on the presence of the carboxylate/ethanolamide or glycerol moiety. For example, a patch of hydrophobic residues appeared to mediate AA and AEA positioning within the CYP2J2 active site with Pro 112, Ile 487 and Ile 376 common for both substrates. Interactions of polar residues were fewer with only Gln 228 interacting with the polar carboxylate/ethanolamide groups. Conversely, positioning of the 2-AG utilized both hydrophobic residues (Ile127, Val 380 and Ile 487) as well as interactions of the hydrophilic glycerol moiety with polar CYP2J2 residues (Asn 231, Ser 224, Gln 228 and Ser 64). The relatively fewer polar interactions of AA and AEA might enable sampling of different conformations of these substrates in the active site leading to the epoxidation of the 8,9- and 5,6-olefin groups. Conversely, the glycerol group appears tethered and cannot sample orientations for epoxidation of the 8,9- and 5,6-olefins. In absence of a high resolution CYP2J2 crystal structure, this model provides a useful tool to predict the interactions of AA, AEA and 2-AG within the active that provide CYP2J2’s regioselectivity.

Discussion
Extrahepatic CYP2J2 epoxygenase metabolizes AA to form EETs that are essential for regulation of blood pressure and cardiovascular function (Spector, 2009). As previously mentioned, AA derived endocannabinoids play an important role in the cardiovascular system. Therefore we explored the metabolism of functionalized arachidonate containing lipids such as the endocannabionids with the cytochrome P450 most highly expressed in the human heart. The two most well characterized, arachidonic acid derived endocannabinoids are AEA and 2-AG whose dysregulation have been implicated in a wide range of pathophysiological states (Pacher et al., 2006; Kogan and Mechoulam, 2007; Pacher and Steffens, 2009). Herein, we demonstrate CYP2J2 mediated metabolism of AEA and 2-AG using recombinant human CYP2J2 and heart microsomes derived from bovine and porcine tissues. The results provided here show the endocannanbinoids 2-AG and AEA are both substrates of human CYP2J2 and that the predominant products, 14,15-EET-EA and 14,15-EET-G, are produced in heart microsomal systems of two different mammalian species, lending credibility to the proposed pathway in tissues.
The qualitative product analysis of CYP2J2-AEA incubations by LC-MS revealed four epoxide regio- isomers at the 5,6-, 8,9-, 11,12- and 14,15-olefin positions. The quantitative analysis of 14,15-EET-EA exhibited 4-fold more product when compared to the rate of formation of 14,15-EET regioisomer from AA. Previously, the oxygenation of AEA into a single hydroxy-product, 20-HETE ethanolamide, as well as four epoxide regio-isomers 5,6-, 8,9-, 11,12- and 14,15-EET-EA were demonstrated by several hepatic P450s including CYP2D6, CYP3A4 and CYP4F2 (Snider et al., 2010). However, this is the first report of the AEA metabolism by a CYP epoxygenase that is highly expressed in the myocardium and endocardium. Given the increasing reports of an endocannabinoid mediated regulation of blood pressure and the cardiovascular system, the role of the oxygenated AEA metabolites may provide an interesting avenue for future studies. For instance, AEA is a partial agonist of both CB1 and CB2 receptors whereas the 5,6-EET-EA metabolite selectively binds CB2 with much greater affinity (Snider et al., 2009). Additionally there is mounting evidence of CB2-cardioprotection versus CB1-cardiotoxicity and the ability to selectively target these pathways may provide a useful therapeutic target for the treatment of heart disease (Mukhopadhyay et al., 2008; Pacher and Steffens, 2009). Moreover, less is known about the physiological effects of the other EET-EA metabolites and whether they also exhibit differential receptor binding profiles and in vivo effects.

Previously, the epoxide metabolites of 2-AG, 14,15- and 11,12-EET-G were found endogenously in rat kidney and spleen tissues. These metabolites bound both CB1 and CB2 with greater affinity than 2-AG (Chen et al., 2008). Incubations with family 2 CYPs – CYP2C8, CYP2C11, and CYP2C3 failed to produce these metabolites (Chen et al., 2008). Therefore it was concluded that these epoxide metabolites do not arise from epoxidation of 2-AG through the CYP pathway. Our modeling studies show that CYP2J2 contains a relatively large active site that can accommodate 2-AG as compared to other family 2 CYPs. Moreover, we show it is possible to directly oxygenate 2-AG by CYP2J2 to produce relatively high levels of both 14,15- and 11,12-EET-G. Additionally our results show that CYP2J2 produces only two epoxide metabolites – 14,15-EET-G and 11,12-EET-G similar to the 2-AG epoxide metabolites identified in vivo. Thus, these results provide an alternative pathway for the biosynthesis of 2-AG epoxides through the direct enzymatic oxygenation by CYP2J2.
Due to the demonstrated beneficial functions of the endocannabinoids, there are several drugs in the market that target the enzymes that attenuate the in vivo levels of these substrates (Cravatt and Lichtman, 2003). Previous inhibition of enzymes that inactivate AEA and 2-AG have shown that the inhibition is incomplete (Blankman et al., 2007) indicating that there may be alternative pathways of attenuation of these endocannabinoids. We demonstrate here that 2-AG is converted to AA by CYP2J2 and CYP2C8 using an oxidative ester cleavage mechanism. Subsequent control experiments confirmed that the cleavage proceeded via an oxidative NADPH-dependent CYP2J2-CPR mechanism. To our knowledge, these results provide the first evidence of a P450 mediated endocannabinoid attenuation pathway for production of free arachidonic acid for subsequent metabolism in the classical eicosanoid pathways (COX, LOX and EPOX). Thus far, there have been no reports of an oxidative ester cleavage mechanism of endogenous 2-AG. However, other P450 isozymes have long been recognized as important sources of carbonyl products through non-hydrolytic ester cleavage reactions (Guengerich, 1987; Peng et al., 1995).

The potential of CYP2J2 to produce its own substrate through the cleavage of arachidonic acid from PIP2 was also explored. Several different conditions were examined and it was concluded the PIP2 is not a CYP2J2 substrate. Considering the tissue distribution of cytochrome P450s, the role of a CYP mediated attenuation of 2-AG may constitute further examination in vivo.
The binding interactions of arachidonic acid and its derivatives with the active site of CYP2J2 revealed relatively low spin state changes. Thus, the uses of the nanoscale lipid bilayers of nanodiscs were essential for the characterization of AA, AEA and 2-AG binding in spectroscopic studies with minimal spectral scattering. In general, there are very few examples of human CYP titrations in the literature, likely due to issues with solubility, scattering and low spin state changes. Interestingly, the few that have been published typically present with a similar spectra to what is presented in this work (Loughran et al., 2001). The ebastine, MS-PPOH and 2-AG Soret shifts present with classical type I binding modes (Fig. 5). The binding of AA and AEA was less clear and difficult to conclusively assign the classical type I and type II shifts due to the low signal to noise ratio. However, all the binding shifts were concentration dependent and reach saturation and are present with clear peaks/troughs lending credibility to the calculated binding constants. While the spin-state changes of the arachidonic acid and the derivatives are minor when compared with ebastine they do correspond to the relatively low catalytic turnover of these substrates. Importantly, the relatively slow turnover rates of CYP2J2 resemble the functions of the steroid synthesizing P450’s that regulate complex homeostatic processes rather than the very fast detoxification functions of liver CYPs.
The results for the CYP2J2-ND substrate-free and substrate-bound stopped-flow cyanide binding experiments are summarized in Fig. 6 and Supplemental Table 1. For cyanide binding studies we employed ebastine and showed that the rates of the fast phase were significantly decreased with increasing ebastine concentrations similar to other P450-substrate systems (Denisov et al., 2007). Conversely, AA, AEA and 2-AG only showed modest inhibition of cyanide binding as shown in Fig. 6C and indicated that they either more easily egress from the active site or are bound as to not block cyanide binding. Similarly, the ratio of the fast- to slow-phase was significantly decreased in the presence of increasing concentrations of ebastine and only modestly decreased in the presence of increasing concentrations of AA, AEA and 2-AG (Fig. 6E – 6G). This effect was consistent and repeatable and not due to non-specific fatty acid binding as evidenced by the lack of change induced by palmitate (Fig. 6H).

The enzymatic turnover of CYP2J2 is several orders of magnitude lower than xenobiotic metabolizing hepatic P450s and it might be tempting to disregard the potential biological significance of the CYP2J2 pathway based solely on these low enzymatic turnover rates. However, the extrahepatic epoxygenases, including CYP2J2, function to regulate complex homeostatic functions within the body and exhibit significantly lower turnover of the endogenous fatty acids (Westphal C et al., 2011) in a highly regulated manner. Thus, the formation of these metabolites likely represents a far more elegant mechanism for control of the many delicate homeostatic processes within organs such as the heart and kidneys. Therefore, similar to the steroid synthesizing CYPs, the enzymatic expression levels rather than the catalytic turnover are likely the prominent mechanism for metabolic control in each respective organ.
In summary, we have demonstrated the ability of CYP2J2 to metabolize a diverse range of AA- containing substrates into a number of unique metabolites that have previously been shown to contain important physiological effects in vivo. Additionally, we provide important evidence that CYP2J2 can directly attenuate 2-AG levels through a NADPH-dependent oxidative cleavage mechanism, thereby producing free arachidonic acid for subsequent metabolism within the classical eicosanoid cascade. We further show that the formation of the metabolites of 2-AG and AEA are in agreement with the spectral binding studies and modeling.

Acknowledgments
We thank Dr. Ilia Denisov and Dr. Matthew Edin for helpful discussions. We thank Mr. William Arnold for editing the paper. We greatly appreciate the contributions of Dr. Zhong Li at the Metabolomics Lab of Roy J. Carver Biotechnology Center, University of Illinois at Urbana-Champaign. We want to thank Furong Sun and Dr. Kevin Tucker for their continued support at the School of Chemical Sciences mass spectrometry facility, University of Illinois at Urbana-Champaign. We thank Prof. Stephen Sligar for providing the gene encoding MSP1D1 and MSP1E3D1 and Prof. Robert Gennis for the use of stopped flow equipment, and Dr. Hanlin Ouyang for technical assistance. We thank Prof. Mary Schuler, Brendon Colón, and Eryk Radziszewski for training and assistance with the MOE modeling software. We thank Dr. Fan and Holly Pondenis for assistance with CYP2J2 immunoblotting. We want to thanks Dr. Bunick for providing us with the homogenizer.

Downloaded from jpet.aspetjournals.org at ASPET Journals on October 22, 2014

62

895612# TEPJ

JPET Fast Forward. Published on October 2, 2014 as DOI: 10.1124/jpet.114.216598
This article has not been copyedited and formatted. The final version may differ from this version.
Authorship Contributions
Participated in research design: McDougle, Kambalyal, Meling, Das Conducted experiments: McDougle, Kambalyal
Performed data analysis: McDougle, Meling, Kambalyal, Das
Wrote or contributed to the writing of the manuscript: McDougle, Das

 

References

JPET Fast Forward. Published on October 2, 2014 as DOI: 10.1124/jpet.114.216598
This article has not been copyedited and formatted. The final version may differ from this version.

Alghasham A, Ali A, Ismail H, Dowaidar M and Settin AA (2012) CYP2J2 -50 G/T and ADRB2 G46A gene polymorphisms in Saudi subjects with hypertension. Genet Test Mol Biomarkers 16:1027-1031.
Baudry J, Li W, Pan L, Berenbaum MR and Schuler MA (2003) Molecular docking of substrates and inhibitors in the catalytic site of CYP6B1, an insect cytochrome p450 monooxygenase. Protein Eng 16:577-587.
Baudry J, Rupasinghe S and Schuler MA (2006) Class-dependent sequence alignment strategy improves the structural and functional modeling of P450s. Protein Eng Des Sel 19:345- 353.
Bayburt TH and Sligar SG (2010) Membrane protein assembly into Nanodiscs. FEBS Lett 584:1721-1727.
Bidwai AK, Meyen C, Kilheeney H, Wroblewski D, Vitello LB and Erman JE (2013) Apolar distal pocket mutants of yeast cytochrome c peroxidase: Hydrogen peroxide reactivity and cyanide binding of the TriAla, TriVal, and TriLeu variants. Bba-Proteins Proteom 1834:137-148.
Blankman JL, Simon GM and Cravatt BF (2007) A comprehensive profile of brain enzymes that hydrolyze the endocannabinoid 2-arachidonoylglycerol. Chem Biol 14:1347-1356.
Capdevila JH, Dishman E, Karara A and Falck JR (1991) Cytochrome P450 arachidonic acid epoxygenase: stereochemical characterization of epoxyeicosatrienoic acids. Methods Enzymol 206:441-453.

Chen JK, Chen J, Imig JD, Wei S, Hachey DL, Guthi JS, Falck JR, Capdevila JH and Harris RC (2008) Identification of novel endogenous cytochrome p450 arachidonate metabolites with high affinity for cannabinoid receptors. J Biol Chem 283:24514-24524.
Cravatt BF and Lichtman AH (2003) Fatty acid amide hydrolase: an emerging therapeutic target in the endocannabinoid system. Curr Opin Chem Biol 7:469-475.
Das A, Varma SS, Mularczyk C and Meling DD (2014) Functional investigations of thromboxane synthase (CYP5A1) in lipid bilayers of nanodiscs. Chembiochem 15:892- 899.
Delozier TC, Kissling GE, Coulter SJ, Dai D, Foley JF, Bradbury JA, Murphy E, Steenbergen C, Zeldin DC and Goldstein JA (2007) Detection of human CYP2C8, CYP2C9, and CYP2J2 in cardiovascular tissues. Drug Metab Dispos 35:682-688.
Denisov IG, Grinkova YV, McLean MA and Sligar SG (2007) The one-electron autoxidation of human cytochrome P450 3A4. The Journal of Biological Chemistry 282:26865-26873.
Denisov IG and Sligar SG (2011) Cytochromes P450 in nanodiscs. Biochim Biophys Acta 1814:223-229.
Falck JR, Belosludtsev YY, Reddy KK, Reddy KM, Shortt MF, Chauhan K, Capdevila JH and Wei SZ (1997) Eicosanoid biosynthesis: Differential inhibition of cytochrome P450 epoxygenase and omega-hydroxylase. Bioorg Med Chem Lett 7:3053-3056.
Grinkova YV, Denisov IG and Sligar SG (2010) Functional reconstitution of monomeric CYP3A4 with multiple cytochrome P450 reductase molecules in Nanodiscs. Biochem Biophys Res Commun 398:194-198.
Guengerich FP (1987) Oxidative cleavage of carboxylic esters by cytochrome P-450. J Biol Chem 262:8459-8462.

Huang W, Jiang D, Wang X, Wang K, Sims CE, Allbritton NL and Zhang Q (2011) Kinetic analysis of PI3K reactions with fluorescent PIP2 derivatives. Anal Bioanal Chem 401:1881-1888.
Imig JD and Hammock BD (2009) Soluble epoxide hydrolase as a therapeutic target for cardiovascular diseases. Nat Rev Drug Discov 8:794-805.
Kogan NM and Mechoulam R (2007) Cannabinoids in health and disease. Dialogues Clin Neurosci 9:413-430.
Lafite P, Andre F, Zeldin DC, Dansette PM and Mansuy D (2007) Unusual regioselectivity and active site topology of human cytochrome P450 2J2. Biochemistry 46:10237-10247.
Loughran PA, Roman LJ, Miller RT and Masters BS (2001) The kinetic and spectral characterization of the E. coli-expressed mammalian CYP4A7: cytochrome b5 effects vary with substrate. Arch Biochem Biophys 385:311-321.
McDougle DR, Palaria A, Magnetta E, Meling DD and Das A (2013) Functional studies of N- terminally modified CYP2J2 epoxygenase in model lipid bilayers. Protein Sci 22:964- 979.
Mukhopadhyay P, Mohanraj R, Batkai S and Pacher P (2008) CB1 cannabinoid receptor inhibition: promising approach for heart failure? Congest Heart Fail 14:330-334.
Nath A, Grinkova YV, Sligar SG and Atkins WM (2007) Ligand binding to cytochrome P450 3A4 in phospholipid bilayer nanodiscs: the effect of model membranes. J Biol Chem 282:28309-28320.
Orlando BJ, McDougle DR, Lucido MJ, Eng ET, Graham LA, Schneider C, Stokes DL, Das A and Malkowski MG (2014) Cyclooxygenase-2 catalysis and inhibition in lipid bilayer nanodiscs. Arch Biochem Biophys 546:33-40.

Pacher P, Batkai S and Kunos G (2006) The endocannabinoid system as an emerging target of pharmacotherapy. Pharmacol Rev 58:389-462.
Pacher P and Steffens S (2009) The emerging role of the endocannabinoid system in cardiovascular disease. Semin Immunopathol 31:63-77.
Panigrahy D, Edin ML, Lee CR, Huang S, Bielenberg DR, Butterfield CE, Barnes CM, Mammoto A, Mammoto T, Luria A, Benny O, Chaponis DM, Dudley AC, Greene ER, Vergilio JA, Pietramaggiori G, Scherer-Pietramaggiori SS, Short SM, Seth M, Lih FB, Tomer KB, Yang J, Schwendener RA, Hammock BD, Falck JR, Manthati VL, Ingber DE, Kaipainen A, D’Amore PA, Kieran MW and Zeldin DC (2012) Epoxyeicosanoids stimulate multiorgan metastasis and tumor dormancy escape in mice. J Clin Invest 122:178-191.
Peng HM, Raner GM, Vaz AD and Coon MJ (1995) Oxidative cleavage of esters and amides to carbonyl products by cytochrome P450. Arch Biochem Biophys 318:333-339.
Pretorius PJ, Pohl WG, Smithen CS and Inesi G (1969) Structural and functional characterization of dog heart microsomes. Circ Res 25:487-499.
Rouzer CA and Marnett LJ (2011) Endocannabinoid Oxygenation by Cyclooxygenases, Lipoxygenases, and Cytochromes P450: Cross-Talk between the Eicosanoid and Endocannabinoid Signaling Pathways. Chem Rev 111:5899-5921.
Snider NT, Kornilov AM, Kent UM and Hollenberg PF (2007) Anandamide metabolism by human liver and kidney microsomal cytochrome p450 enzymes to form hydroxyeicosatetraenoic and epoxyeicosatrienoic acid ethanolamides. J Pharmacol Exp Ther 321:590-597.

Snider NT, Nast JA, Tesmer LA and Hollenberg PF (2009) A cytochrome P450-derived epoxygenated metabolite of anandamide is a potent cannabinoid receptor 2-selective agonist. Mol Pharmacol 75:965-972.
Snider NT, Sikora MJ, Sridar C, Feuerstein TJ, Rae JM and Hollenberg PF (2008) The Endocannabinoid Anandamide Is a Substrate for the Human Polymorphic Cytochrome P450 2D6. Journal of Pharmacology and Experimental Therapeutics 327:538-545.
Snider NT, Walker VJ and Hollenberg PF (2010) Oxidation of the endogenous cannabinoid arachidonoyl ethanolamide by the cytochrome P450 monooxygenases: physiological and pharmacological implications. Pharmacol Rev 62:136-154.
Spector AA (2009) Arachidonic acid cytochrome P450 epoxygenase pathway. J Lipid Res 50 Suppl:S52-56.
Sridar C, Snider NT and Hollenberg PF (2011) Anandamide Oxidation by Wild-Type and Polymorphically Expressed CYP2B6 and CYP2D6. Drug Metabolism and Disposition 39:782-788.
Stark K, Dostalek M and Guengerich FP (2008) Expression and purification of orphan cytochrome P450 4X1 and oxidation of anandamide. Febs J 275:3706-3717.
Westphal C, Konkel A and WH. S (2011) CYP-eicosanoids–a new link between omega-3 fatty acids and cardiac disease? Prostaglandins Other Lipid Mediat 96:99-108.
Wu S, Moomaw CR, Tomer KB, Falck JR and Zeldin DC (1996) Molecular cloning and expression of CYP2J2, a human cytochrome P450 arachidonic acid epoxygenase highly expressed in heart. J Biol Chem 271:3460-3468.
Yang J, Dong H and Hammock BD (2011) Profiling the regulatory lipids: another systemic way to unveil the biological mystery. Curr Opin Lipidol 22:197-203.

Zelasko S, Palaria A and Das A (2013) Optimizations to achieve high-level expression of cytochrome P450 proteins using Escherichia coli expression systems. Protein Expres Purif 92:77-87.
Zeldin DC, Foley J, Goldsworthy SM, Cook ME, Boyle JE, Ma J, Moomaw CR, Tomer KB, Steenbergen C and Wu S (1997) CYP2J subfamily cytochrome P450s in the gastrointestinal tract: expression, localization, and potential functional significance. Mol Pharmacol 51:931-943.
Zhu Y, Schieber EB, McGiff JC and Balazy M (1995) Identification of arachidonate P-450 metabolites in human platelet phospholipids. Hypertension 25:854-859.

Footnotes
This work was funded in part through the American Heart Predoctoral Fellowship [14PRE20130015] (D.R.M). The Roy J. Carver Biotechnology Center’s 5500 QTrap MS was funded by NIH National Center for Research Resources [S10RR024516]. The School of Chemical Sciences Q-TOF Ultima mass spectrometer was purchased in part with a grant from the National Science Foundation, Division of Biological Infrastructure [DBI-0100085].

Legends for Figures
FIGURE 1. Chemical structures of substrates that bind to CYP2J2-nanodiscs. Endogenous CYP2J2 ligands include (A) arachidonic acid, (B) anandamide and (C) 2-arachidonoyl glycerol. (D) The xenobiotic, ebastine, is a cardiotoxic 2J2 substrate. (E) MS-PPOH (N-(methylsulfonyl)-2-(2- propynyloxy)-benzenehexanamide) is a non-specific CYP inhibitor. (F) Schematic of CYP2J2 and CPR incorporated into the membrane bilayers of nanodiscs. The membrane scaffold protein (MSP) is represented in cyan surrounding a bilayer of phospholipids colored white with gold phosphate head groups. CYP2J2 is red and its obligate redox partner cytochrome P450 reductase is colored blue.
FIGURE 2. Metabolism of AA and AEA by human CYP2J2-ND-CPR identified qualitatively using LC- ESI-MS. CYP2J2-ND incubations with substrates were analyzed as described in the materials and methods. The different regio-isomers of (A) AA (m/z = 319.5 of 14,15-EET at 22.5 min) and (B) AEA (m/z = 363.4 of 14,15-EET-EA at 14 min) produced six different products including four epoxide regio- isomers at the 5,6-, 8,9-, 11,12- and 14,15-olefin positions and two HETEs.
FIGURE 3. Metabolism of 2-AG by CYP2J2-ND-CPR identified qualitatively using LC-ESI-MS. Metabolism of 2-AG by incubations with CYP2J2-ND-CPR revealed the following metabolic profile as represented in the chromatogram that consists of an overlay of two EET-G regio-isomers, 14,15-EET-G and 11,12-EET-G (m/z 393), three detectable EET regio-isomers, 14,15-EET, 11,12-EET and 8,9-EET (m/z = 319) and arachidonic acid (m/z = 303).
FIGURE 4. Kinetics of CYP2J2 mediated metabolism of 2-AG. (A) Schematic of the proposed routes of 2-AG metabolism by CYP2J2 that produces AA, EET-G and EET. The kinetic rates of the three products were quantified as described in the materials and methods. (B) Kinetics of the formation of the four regio- isomers, 14,15-, 11,12-, 8,9- and 5,6-EET were detected by LC-MS-MS. These correspond to the free EETs in solution that were generated either from epoxidation of AA or oxidative ester cleavage of EET-G. (C) Kinetics of the formation of the two EET-G regioisomers, 14,15-EET-G and 11,12-EET-G analyzed as described in the materials and methods section. (D) The kinetics of CYP2J2-ND mediated 2- AG ester cleavage and arachidonic acid production were determined by monitoring AA formation. The data were fitted to Michaelis-Menten kinetics in Origin Pro 8.6 for calculation of Vmax and Km. The kinetic rates of the epoxidation reactions and side reactions are summarized in Table 1.

FIGURE 5. Equilibrium binding using absorption spectra for interactions of substrate with CYP2J2- Nanodiscs. (A) AEA titration Ks was determined to be 13.1 ± 1.4 μM (inset: difference spectrum of AEA bound CYP2J2-Nanodiscs). (B) 2-AG titration Ks was determined to be 18 ± 2.7 μM (inset: difference spectrum of 2-AG bound CYP2J2-Nanodiscs) (C) MS-PPOH titration Ks was determined to be 10.8 ± 1.5 μM (inset: difference spectrum of MS-PPOH bound CYP2J2-Nanodiscs showing a type I binding mode) (D) Summary of the max (nm) of peak/trough, Ks and max for binding of arachidonic acid, anandamide, 2-arachidonoyl glycerol, ebastine and MS-PPOH to CYP2J2-Nanodiscs.
FIGURE 6. Stopped-flow cyanide binding spectroscopy with CYP2J2-ND and AA, 2-AG, AEA and ebastine. (A) Stopped-flow spectroscopy was used to monitor kinetics of cyanide binding to CYP2J2-ND by recording the formation of the cyanide-heme complex at 444 nm as a function of time in the presence of saturating [CN-]. (B) The difference spectra ( Abs 444-414 nm) was plotted against time and fitted to a double exponential equation for determination of the rate of the (C) fast-phase for each substrate concentration tested and determination of statistical significance relative to the substrate free condition when p < 0.05, p < 0.01 and p < 0.001 which are denoted by *, **, and ***, respectively. The ratio of the amplitudes of the fast-phase (black) to slow-phase (gray) were dramatically altered in the presence of (D) ebastine whereas the presence of the polyunsaturated fatty acids such as (E) AEA (F) 2-AG and (G) AA were modestly changed. (H) The saturated fatty acid palmitate did not alter cyanide binding characteristics.

FIGURE 7. Substrate bound CYP2J2 Homology model with arachidonic acid, anandamide and 2- arachidonoyl glycerol. Proposed docking of the polyunsaturated fatty acids into the active site of CYP2J2. (Left) the substrate (A) AA (B) AEA and (C) 2-AG are represented as stick figures (green) surrounded by space-filling dots. The CYP2J2 ribbon structure is colored blue and the heme-oxygen motif is in red. The interaction potential energies predict feasibility of binding and were similar for each polyunsaturated fatty acid with calculations of -69.4 kcal/mol, and -73.4 kcal/mol for 2-AG and AEA, respectively. (Right) predicted ligand-residue interactions within the CYP2J2 active site for (A) AA (B) AEA and (C) 2-AG.

Images to follow: